From: <wim...@ch...> - 2001-05-23 12:19:48
|
[06:59] Nick change: aj_uni -> ajmitch [07:00] iGN_ (ig...@lo...) left irc: Ping timeout for iGN_[login1.simplemente.net] [07:01] omega_away (om...@bo...) joined #gstreamer. [07:02] iGN_ (ig...@lo...) joined #gstreamer. [07:07] steveb (st...@no...) joined #gstreamer. [07:10] Nick change: ajmitch -> ajbusy [07:12] Nick change: taaz -> taazzzz [07:18] taazzzz (dlehn@66.37.66.32) left irc: Ping timeout for taazzzz[66.37.66.32] [07:20] taazzzz (dlehn@66.37.66.32) joined #gstreamer. [07:48] Nick change: ajbusy -> ajmitch [08:02] chillywilly (bau...@d4...) left irc: brb [08:06] omega_away (om...@bo...) left irc: Ping timeout for omega_away[boisdslgw5poolb76.bois.uswest.net] [08:24] Nick change: ajmitch -> ajbusy [10:13] iGN__ (ig...@lo...) joined #gstreamer. [10:13] big_T (th...@fy...) left irc: Ping timeout for big_T[fysgr763.sn.umu.se] [10:13] iGN_ (ig...@lo...) left irc: Ping timeout for iGN_[login1.simplemente.net] [10:13] ajbusy (aj...@p4...) left irc: Ping timeout for ajbusy[p47-max4.dun.ihug.co.nz] [10:13] ajbusy (aj...@p4...) joined #gstreamer. [10:14] omega_away (om...@bo...) joined #gstreamer. [10:15] big_T (th...@fy...) joined #gstreamer. [11:17] thomas (th...@21...) joined #gstreamer. [11:17] <thomas> hi [12:04] omega_away (om...@bo...) left irc: Ping timeout for omega_away[boisdslgw5poolb76.bois.uswest.net] [12:26] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee4c.a2000.nl] [12:26] steveb (st...@no...) joined #gstreamer. [12:38] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee4c.a2000.nl] [12:41] arik (ar...@sd...) joined #gstreamer. [12:41] <arik> anyone awake who uses gnus? [12:41] <thomas> what's gnus ? [12:42] <arik> a mail client for emacs [12:42] <thomas> no, sorry [12:42] <arik> heh [12:42] <arik> well [12:42] <arik> do you know how to setup sendmail? ;-P [12:43] <thomas> arik: use qmail [12:43] <thomas> arik: but you can try, I've done sendmail a few times as well [12:43] <thomas> though it's hard [12:44] <thomas> arik: what's the problem ? [12:44] <arik> well [12:44] <arik> gnus just passes off the mail to the local spool i think [12:44] <arik> i just want to set that up to actually send mail somewhere [12:44] <arik> but i know [12:44] <arik> a total of _nothing_ about this [12:45] <thomas> arik: do you know what mailbox format gnus handles ? [12:45] <thomas> arik: i've never used gnus [12:45] <thomas> arik: your best bet is to see if you can send mail using the "mail" command first [12:45] <arik> ok [12:45] <thomas> most of the times sendmail works out of the box [12:46] <arik> just tested [12:46] <arik> i don't think it worked [12:46] <thomas> is sendmail running ? [12:46] <arik> wait [12:46] <arik> testing again [12:46] <arik> i didn't actually test before [12:47] <thomas> btw: is nautilus cvs-only or is there a released version of it ? [12:47] <arik> there are released versions [12:47] <arik> it is part of 1.4 [12:47] <arik> sendmail is running [12:47] <thomas> ok, try mailing to yourself first on the box [12:48] <thomas> you have to make sure you end the mail with a dot. [12:48] <arik> um ok [12:48] <thomas> I just tested it too, that should work. [12:48] <thomas> then read it with pine for example [12:48] <thomas> arik: what distro are you using ? [12:48] <arik> mail ad...@an... -s fdfdf test. [12:48] <arik> rh7 [12:48] <arik> that is what i am doing [12:48] <thomas> arik: do it interactively [12:49] <thomas> do mail adevens@localhost [12:49] <thomas> fill in the blanks [12:49] <thomas> end the body with a line with only "." on it [12:49] <arik> ok [12:49] <arik> how do i actually send it? [12:49] <arik> oh [12:49] <arik> i see [12:49] <arik> done [12:50] <arik> ok [12:50] <arik> that worked [12:50] <thomas> ok... let's try the next step then [12:50] <arik> but that wasn't actually outgoing [12:50] <thomas> do the same, but replace localhost with the full name [12:50] <arik> ok [12:50] <thomas> you have root right ? [12:50] <thomas> open a second window and tail /var/log/maillog [12:50] <arik> ok [12:50] <arik> done [12:51] <thomas> now, when you send a mail outbound, you should get a few lines into your maillog file [12:51] steveb (st...@no...) joined #gstreamer. [12:51] <arik> ok [12:51] <arik> but i don't think it worked [12:51] <thomas> arik: how so ? [12:51] <arik> i don't get the mail from fetchmail ;-) [12:52] <thomas> but did you get an entry in the maillog ? [12:52] <arik> i think so, i'm not sure [12:52] <arik> yes [12:52] <arik> i do get an entry [12:52] <thomas> what's it say ? [12:53] <arik> Deferred: Connection timed out with college.antioch-college.edu. [12:53] <arik> don't know why it tried that [12:53] <thomas> and that's the same machine as your localhost or not ? [12:53] <arik> no [12:53] <arik> not at all [12:54] <thomas> ok, what's your localhost ? [12:54] <thomas> and what machine is your mail account on ? [12:54] <thomas> and what type is it ? [12:54] <arik> heh [12:54] <arik> my localhost is this machine [12:54] <arik> or am i not understanding ;- [12:54] <arik> ) [12:54] <arik> my mail account is on a machine at my school [12:54] <thomas> arik: i mean, what's it's FQDN [12:54] <arik> fqdn? [12:54] <thomas> fully qualified domain name [12:54] <arik> man i am a novice with mail [12:54] <arik> oh [12:55] <thomas> what's it's name on the network [12:55] <arik> it's just a dynamic ip on a ppp connection [12:55] <thomas> and what is the name of the mailserver [12:55] <arik> mail.antioch-college.edu [12:55] <arik> but i dpn't have that setup [12:55] <arik> at all [12:55] <arik> i mean i don't have it setup to use that [12:55] <thomas> arik: uhm, now you've lost me [12:55] <arik> heh [12:55] <thomas> how do you want to receive mail from a server then ? [12:55] <arik> i have that setup [12:55] <arik> i don't have smtp setup [12:56] <thomas> ok, the connection with college.antioch-college.edu was made ... [12:56] <thomas> ... because that's the same server as mail.antioch-college.edu [12:56] <arik> ok [12:56] <thomas> so what connection type are you using ? SMTP ? [12:56] <arik> i guess so [12:56] <arik> i don't know, i didn't setup sendmail atall [12:56] <arik> i have no idea [12:56] <arik> sorry i am so confused [12:57] <arik> i feel like a total novice again ;-) weird [12:57] <thomas> arik: no problem, I'm confused as well ;) [12:57] <arik> hehe [12:57] <arik> i have never setup sendmail at all [12:57] <arik> i setup fetchmail to get mail from my account [12:58] <thomas> arik: the thing that confuses me is that sending a mail contacts your mailserver of an account you established [12:58] <thomas> normally, sendmail figures out the mx record of the recipient [12:58] <thomas> and sends it itself [12:58] <arik> hmm [12:58] <arik> i don't know [12:59] <thomas> what does fetchmail do ? I never used that as well... [12:59] <arik> it gets mail from a server and puts it in /var/mail/user [12:59] <arik> so that mail clients can view it [13:00] <thomas> arik: ok, try to send a mail to another account that is not at antioch-college... [13:00] <thomas> ... see what line sendmail prints for that. [13:00] <arik> yeah i am trying that as we speak ;-) [13:00] <arik> May 23 04:00:47 localhost sendmail[16414]: f4NB0l216414: from=arik, size=39, class=0, nrcpts=1, msgid=<200105231100.f4NB0l216414@localhost.localdomain>, relay=arik@localhost [13:00] <arik> so far [13:00] <arik> that's what is says [13:01] <thomas> nothing more than that ? [13:01] <arik> nop [13:01] <arik> e [13:01] <arik> oh wait [13:01] <arik> Connection timed out with relay1-no.simplemente.net. [13:01] <thomas> ok, can you ping those hosts ? [13:02] <arik> yes [13:02] <thomas> then try this : "telnet mail.antioch-college.edu 25" [13:02] <thomas> this way you establish a manual smtp connection [13:03] <thomas> wait for a line starting with 220 [13:03] <thomas> if that doesn't work, you have a network connectivity problem [13:03] <arik> ok [13:04] <thomas> do you get a reply ? [13:04] <arik> not yet [13:04] <arik> i can ping it though [13:05] <thomas> hmmm... it seems like your port is blocked. [13:05] <thomas> have nmap installed ? [13:05] <thomas> if you don't get a relatively fast reply from the telnet command, something is wrong. [13:06] <arik> don't know ;-) [13:06] <thomas> You did add "25" at the end of the line, right ? [13:06] <arik> yes [13:06] <thomas> because when I do it, I get an immediate reply [13:06] <thomas> ok, can you give me your ip address ? I'll scan it [13:06] <thomas> arik: other stuff like browsing, telnetting works, right ? [13:06] <arik> 168.191.230.141 [13:06] <arik> yes [13:06] <arik> works fine [13:08] <thomas> ok, I can make an ESMTP connection to your machine... [13:08] <arik> hm [13:10] <arik> where should nmap be? [13:11] <thomas> I think www.fyodor.org [13:11] <thomas> but i'll check first [13:11] <arik> no i mean on my system ;-P [13:11] <thomas> not ;) it's a port scanner [13:11] <arik> og [13:11] <arik> er oh [13:11] <thomas> you have to get it, it's probably illegal in some countries i suppose [13:11] <arik> ok [13:11] <arik> i wonder why this isn't working [13:12] <thomas> arik: if the telnet to the port directly isn't working, then that's the problem [13:12] <arik> right [13:12] <thomas> that's also what the line in your maillog indicates [13:12] <arik> but i wonder why [13:12] <arik> right [13:12] <thomas> arik: me too [13:12] <thomas> ;) [13:12] <arik> so how do i fix it ;-P [13:12] <thomas> arik: but a portscan can help in diagnosing [13:12] <arik> ok [13:13] <thomas> arik: since I can access it from where I am [13:13] <thomas> arik: how concerned are they about security at your uni ? [13:13] <arik> not at all i don't think [13:13] <thomas> arik: btw nmap is at www.insecure.org [13:13] <arik> ok [13:13] <thomas> arik: portscanning can be detected with the right tools, if they're conscious about that [13:13] <thomas> sometimes it's explicitly forbidden [13:13] <arik> i doubt it [13:13] <arik> it's possible i suppose [13:14] <thomas> well it's your call of course <grin> [13:14] <arik> heh [13:15] <thomas> when you have it installed (get an rpm), just type "nmap mail......" [13:15] <thomas> that will scan all of their ports and list what is open. [13:15] <arik> ok [13:15] <thomas> I'll do the same and then we can compare the results [13:15] <arik> ok [13:15] <arik> i hope i don't get in trouble [13:15] <arik> oh well [13:16] <thomas> well, did you ever hear about a student getting in trouble ? [13:16] <thomas> I'll try first if you want [13:16] <arik> no [13:16] <arik> i really think it will be fine [13:16] <thomas> let me know if you get a mail on your localhost [13:16] <arik> i got your maikl [13:16] <thomas> directly, that is [13:16] <thomas> ok [13:17] <arik> yes [13:17] <arik> got it [13:17] <thomas> your esmtp is fine then, I did it manually [13:17] <arik> hmm [13:18] <thomas> arik: start with namp -p 20-30 mail.antioch-college.edu [13:18] <arik> ok [13:18] <thomas> that only scans a few ports [13:18] <arik> done [13:19] <thomas> ok, paste the output [13:19] <arik> Port State Service [13:19] <arik> 21/tcp open ftp [13:19] <arik> 22/tcp open ssh [13:19] <arik> 23/tcp open telnet [13:19] <arik> 25/tcp filtered smtp [13:19] <thomas> ok, you see the "filtered" ? [13:19] <arik> yeah [13:19] <thomas> that indicates they're deliberately blocking based on IP [13:19] <thomas> when I do the same, I have the port as "open" like all the others [13:19] <arik> weird [13:19] <arik> why would they block my ip? [13:20] <thomas> don't know. it's nothing personal I guess ;) [13:20] <arik> hah [13:20] <thomas> try a normal telnet to mail.... [13:20] <thomas> no port specified [13:20] <thomas> you should get a quick response, right ? [13:20] <arik> yes [13:20] <thomas> well, the response on port 25 should be just as quick... [13:20] <thomas> ... since it's filtered, telnet hangs [13:20] <arik> ok [13:21] <thomas> ... at this point, the best you can try is nmap the same ports on that other server you used [13:21] <thomas> that way you can check if it's the server or your isp that's filtering [13:21] <arik> ok [13:21] <thomas> I'd suppose it's your isp since I have the port listed as open [13:21] <thomas> and they don't know me [13:21] <thomas> ;) [13:22] <arik> actually [13:22] <arik> the other one is gnome.org [13:22] <arik> that i could get in trouble for ;-) [13:22] <thomas> well, just start with a telnet to port 25 then [13:22] <arik> that didn't work [13:22] <arik> i tried it [13:23] <thomas> btw: when you limit the portscan range, you can't do much harm [13:23] <arik> ij [13:23] <arik> er ok [13:23] <thomas> btw: I can connect to the port just fine [13:23] <arik> ok [13:23] <arik> so this is my isp? [13:23] <thomas> maybe nmap -p 25 gnome.org also works [13:23] <arik> fuck [13:23] <arik> filtered [13:23] <arik> for me [13:23] <thomas> ok, sounds like your ISP is to blame [13:24] <arik> fucking a [13:24] <thomas> is it DSL ? [13:24] <arik> no [13:24] <thomas> what type ? [13:24] <arik> modem [13:26] <thomas> what's your isp ? [13:26] <arik> earthlink [13:26] <thomas> hang on, i'll check if it's blacklisted [13:26] <arik> ok [13:29] <thomas> arik: It's probably something else [13:29] <arik> hmm [13:29] <thomas> arik: it could be that your ISP only allows SMTP relays through their own server [13:30] <arik> maybe [13:30] <thomas> normally they don't do that for modems, since the speed is slow [13:30] <thomas> arik: have you ever sent *any* mail through that isp ever ? [13:30] <arik> no [13:30] <thomas> arik: because if that's the case, you'll first need to know the smtp relay for earthlink... [13:30] <thomas> ... and add that to sendmail.cf [13:30] <thomas> then it should be fixed [13:30] <arik> hmm [13:31] <thomas> ok, try this then : [13:31] <thomas> telnet mail.earthlink.net 25 [13:31] <arik> ok [13:31] <thomas> does that work ? [13:31] <arik> yes [13:31] <thomas> arik: ok, then that's probably your relay [13:32] <arik> ok [13:32] <thomas> do you also get this : 220-hawk.mail.pas.earthlink.net ESMTP Sendmail EL-8_9_3_3/8.9.3; Wed, 23 May 2001 04:32:38 -0700 (PDT) [13:32] <thomas> 220-NO UCE. Earthlink Network does not authorize the use of its computers [13:32] <thomas> 220 or network equipment to deliver unsolicited electronic mail. [13:32] <thomas> Q [13:32] <thomas> ? [13:32] <arik> yes [13:32] <thomas> without the Q that is [13:32] <thomas> ok... [13:32] <thomas> edit /etc/sendmail.cf [13:32] <thomas> find the line containing "relay" [13:32] <thomas> it looks like this : [13:32] <thomas> # "Smart" relay host (may be null) [13:32] <thomas> DS [13:32] <thomas> [13:32] <arik> um [13:32] <arik> i don't have a /etc/sendmail.cf [13:32] <thomas> huh ? [13:32] <arik> wait [13:32] <arik> yeah i do [13:32] <arik> dumb dumb [13:33] <thomas> no problem ;) [13:33] <arik> fount it [13:33] <arik> what now? [13:34] <thomas> ok, change the line "DS" to "DSmail.earthlink.net" [13:34] <thomas> no space inbetween DS and mail.... [13:34] <thomas> save that [13:34] <arik> ok [13:34] <thomas> then run /etc/rc.d/init.d/sendmail restart [13:34] <thomas> then try sending a mail again [13:34] <thomas> and paste the sendmail output [13:34] <thomas> (hmm... this shouldn't go in the gstreamer logs) [13:35] <arik> hehe [13:36] <arik> May 23 04:38:36 localhost sendmail[16910]: f4NBcaR16910: from=root, size=52, class=0, nrcpts=1, msgid=<200105231138.f4NBcaR16910@localhost.localdomain>, relay=root@localhost [13:36] <thomas> did sendmail restart ? [13:36] <arik> yes [13:37] <arik> May 23 04:38:04 localhost sendmail[16890]: alias database /etc/aliases rebuilt by arik [13:37] <arik> May 23 04:38:04 localhost sendmail[16890]: /etc/aliases: 38 aliases, longest 10 bytes, 377 bytes total [13:37] <arik> May 23 04:38:04 localhost sendmail[16905]: starting daemon (8.11.0): SMTP+queueing@01:00:00 [13:38] <thomas> arik: where did you send the mail to ? [13:38] <arik> ad...@an... [13:38] <arik> Connection timed out with xwing.aoltw.net. [13:38] <arik> wtf? [13:39] <thomas> ok, that's america online i suppose [13:39] <arik> right [13:39] <arik> why is it connecting to that? [13:39] <thomas> that's weird [13:39] <arik> yes [13:39] <thomas> arik: don't know [13:39] <arik> very strange [13:39] <thomas> arik: try "nslookup mail.earthlink.net" [13:39] <thomas> what do you get ? [13:40] <arik> Server: dns1-sf.snfc21.pacbell.net [13:40] <arik> Address: 206.13.28.12 [13:40] <arik> Non-authoritative answer: [13:40] <arik> Name: mail.earthlink.net [13:40] <arik> Addresses: 207.217.120.202, 207.217.121.203, 207.217.120.204, 207.217.121.205 [13:42] <thomas> arik: try sending the mail to me, th...@ur... [13:42] <arik> ij [13:42] <arik> er ok [13:42] <arik> you are in belgium? [13:43] <thomas> I have some message on deja claiming that AOL disallows delivery to locals [13:43] <thomas> don't see why [13:43] <thomas> arik: yes [13:43] <arik> i was in belgium for new years [13:43] <arik> in brussels [13:43] <arik> sent [13:43] <thomas> ok, what does sendmail say now ? [13:43] <arik> May 23 04:45:49 localhost sendmail[16920]: f4NBjmf16920: from=root, size=48, class=0, nrcpts=1, msgid=<200105231145.f4NBjmf16920@localhost.localdomain>, relay=root@localhost [13:43] <thomas> arik: I'm in brussels now [13:43] <arik> ;-) [13:43] <arik> i like brussels [13:43] <thomas> I'm from Gent, Gent is better ;) [13:43] <arik> hehe [13:44] <arik> flemish? [13:44] <thomas> ja ;) [13:44] <arik> hah [13:44] <arik> ich sprache deutsch [13:44] <thomas> heh [13:44] <arik> but not flemish ;-P [13:44] <arik> did you get the mail? [13:45] <thomas> well, it should give a second line before I can get it ;) [13:45] <arik> heh [13:45] <thomas> the first says it's going to send it [13:45] <arik> nope [13:45] <arik> another timeout [13:45] <thomas> but I think the relay line should be different [13:45] <thomas> what server did it try to connect to ? [13:46] <arik> ywind [13:46] <arik> er ywing.aoltw.net [13:46] <arik> that is so weird [13:46] <arik> i don't have aol [13:46] <arik> although i do work for them ;-) [13:46] <arik> oh! [13:46] <arik> i bet i know what's up [13:46] <arik> mayve [13:46] <arik> er maybe [13:46] <arik> no [13:46] <arik> that wouldn't matter [13:46] <arik> nm [13:47] <thomas> ok, now i'm worried [13:47] <arik> heh [13:48] <arik> what on earth have i done to have this happen? [13:48] <thomas> let's try this: change mail..... in your sendmail.cf to an IP address: [13:48] <arik> ok [13:48] <thomas> DS207.217.121.202 [13:49] <arik> done [13:49] <thomas> now restart sendmail again [13:49] <arik> done [13:50] <thomas> now send a mail [13:50] <arik> done [13:50] <thomas> what does sendmail say now ? [13:50] <arik> May 23 04:53:17 localhost sendmail[16982]: f4NBrH416982: from=root, size=48, class=0, nrcpts=1, msgid=<200105231153.f4NBrH416982@localhost.localdomain>, relay=root@localhost [13:51] <arik> same thing [13:51] <arik> xwing [13:52] <arik> so weird [13:53] <thomas> ok, we'll probably need to add some masquerading as well. [13:53] <thomas> ok, first make a backup of sendmail.cf [13:53] <arik> done [13:54] <thomas> now find the line saying # who I masquerade as (null for no masquerading) [13:54] <arik> done [13:55] <thomas> on the next line, change DM to DMarik...@.... [13:55] <thomas> the address you want to send as [13:55] <thomas> ... wait. your isp is earthlink... [13:55] <arik> yes [13:55] <thomas> ... but you're trying to send mail as a uni user, right ? [13:56] <arik> yes [13:59] <thomas> well, let's just try this first. [13:59] <arik> ok [13:59] <thomas> save it, restart sendmail, send mail, see what sendmail says [13:59] <arik> what do i put after ...? [14:00] <thomas> what you have as an address there [14:00] <arik> ij [14:00] <arik> er ok [14:01] <arik> May 23 05:03:45 localhost sendmail[17031]: f4NC3jq17031: from=root, size=48, class=0, nrcpts=1, msgid=<200105231203.f4NC3jq17031@localhost.localdomain>, relay=root@localhost [14:03] <arik> same fscking timeout [14:03] <arik> anyway i'm gonna sleep [14:03] <arik> i'll try again late [14:03] <arik> r [14:03] <arik> thanks alot ;-) [14:04] <thomas> arik: didn't help much did it ;( [14:04] <arik> heh [14:04] <arik> s'ok [14:04] <arik> i learned some stuff [14:04] <thomas> anyway, one thing you should try if you have the time is to put [14:04] <thomas> mailer: after the DS, before the host [14:04] <arik> ok [14:04] <thomas> or smtp: [14:04] <arik> ok [14:04] <thomas> or esmtp: [14:04] <arik> which one? [14:04] <thomas> depends on your sendmail [14:04] <thomas> check all three of them in row [14:04] <thomas> so change sendmail.cf, restart, send mail, watch for output [14:05] <thomas> you should get a different "relay=..." line to get it working [14:05] <thomas> but do it when you're awake ;) [14:05] <arik> heh [14:05] <arik> i have time for that [14:05] <thomas> cu [14:05] <thomas> (wtay-bot: clear the logs please ;)) [14:05] arik (ar...@sd...) left #gstreamer. [14:13] omega_away (om...@bo...) joined #gstreamer. [14:13] <thomas> hi omega |
From: <wim...@ch...> - 2001-05-24 04:30:08
|
[06:59] Nick change: aj_uni -> ajmitch [07:00] iGN_ (ig...@lo...) left irc: Ping timeout for iGN_[login1.simplemente.net] [07:01] omega_away (om...@bo...) joined #gstreamer. [07:02] iGN_ (ig...@lo...) joined #gstreamer. [07:07] steveb (st...@no...) joined #gstreamer. [07:10] Nick change: ajmitch -> ajbusy [07:12] Nick change: taaz -> taazzzz [07:18] taazzzz (dlehn@66.37.66.32) left irc: Ping timeout for taazzzz[66.37.66.32] [07:20] taazzzz (dlehn@66.37.66.32) joined #gstreamer. [07:48] Nick change: ajbusy -> ajmitch [08:02] chillywilly (bau...@d4...) left irc: brb [08:06] omega_away (om...@bo...) left irc: Ping timeout for omega_away[boisdslgw5poolb76.bois.uswest.net] [08:24] Nick change: ajmitch -> ajbusy [10:13] iGN__ (ig...@lo...) joined #gstreamer. [10:13] big_T (th...@fy...) left irc: Ping timeout for big_T[fysgr763.sn.umu.se] [10:13] iGN_ (ig...@lo...) left irc: Ping timeout for iGN_[login1.simplemente.net] [10:13] ajbusy (aj...@p4...) left irc: Ping timeout for ajbusy[p47-max4.dun.ihug.co.nz] [10:13] ajbusy (aj...@p4...) joined #gstreamer. [10:14] omega_away (om...@bo...) joined #gstreamer. [10:15] big_T (th...@fy...) joined #gstreamer. [11:17] thomas (th...@21...) joined #gstreamer. [11:17] <thomas> hi [12:04] omega_away (om...@bo...) left irc: Ping timeout for omega_away[boisdslgw5poolb76.bois.uswest.net] [12:26] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee4c.a2000.nl] [12:26] steveb (st...@no...) joined #gstreamer. [12:38] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee4c.a2000.nl] [12:41] arik (ar...@sd...) joined #gstreamer. [12:41] <arik> anyone awake who uses gnus? [12:41] <thomas> what's gnus ? [12:42] <arik> a mail client for emacs [12:42] <thomas> no, sorry [12:42] <arik> heh [12:42] <arik> well [12:42] <arik> do you know how to setup sendmail? ;-P [12:43] <thomas> arik: use qmail [12:43] <thomas> arik: but you can try, I've done sendmail a few times as well [12:43] <thomas> though it's hard [12:44] <thomas> arik: what's the problem ? [12:44] <arik> well [12:44] <arik> gnus just passes off the mail to the local spool i think [12:44] <arik> i just want to set that up to actually send mail somewhere [12:44] <arik> but i know [12:44] <arik> a total of _nothing_ about this [12:45] <thomas> arik: do you know what mailbox format gnus handles ? [12:45] <thomas> arik: i've never used gnus [12:45] <thomas> arik: your best bet is to see if you can send mail using the "mail" command first [12:45] <arik> ok [12:45] <thomas> most of the times sendmail works out of the box [12:46] <arik> just tested [12:46] <arik> i don't think it worked [12:46] <thomas> is sendmail running ? [12:46] <arik> wait [12:46] <arik> testing again [12:46] <arik> i didn't actually test before [12:47] <thomas> btw: is nautilus cvs-only or is there a released version of it ? [12:47] <arik> there are released versions [12:47] <arik> it is part of 1.4 [12:47] <arik> sendmail is running [12:47] <thomas> ok, try mailing to yourself first on the box [12:48] <thomas> you have to make sure you end the mail with a dot. [12:48] <arik> um ok [12:48] <thomas> I just tested it too, that should work. [12:48] <thomas> then read it with pine for example [12:48] <thomas> arik: what distro are you using ? [12:48] <arik> mail ad...@an... -s fdfdf test. [12:48] <arik> rh7 [12:48] <arik> that is what i am doing [12:48] <thomas> arik: do it interactively [12:49] <thomas> do mail adevens@localhost [12:49] <thomas> fill in the blanks [12:49] <thomas> end the body with a line with only "." on it [12:49] <arik> ok [12:49] <arik> how do i actually send it? [12:49] <arik> oh [12:49] <arik> i see [12:49] <arik> done [12:50] <arik> ok [12:50] <arik> that worked [12:50] <thomas> ok... let's try the next step then [12:50] <arik> but that wasn't actually outgoing [12:50] <thomas> do the same, but replace localhost with the full name [12:50] <arik> ok [12:50] <thomas> you have root right ? [12:50] <thomas> open a second window and tail /var/log/maillog [12:50] <arik> ok [12:50] <arik> done [12:51] <thomas> now, when you send a mail outbound, you should get a few lines into your maillog file [12:51] steveb (st...@no...) joined #gstreamer. [12:51] <arik> ok [12:51] <arik> but i don't think it worked [12:51] <thomas> arik: how so ? [12:51] <arik> i don't get the mail from fetchmail ;-) [12:52] <thomas> but did you get an entry in the maillog ? [12:52] <arik> i think so, i'm not sure [12:52] <arik> yes [12:52] <arik> i do get an entry [12:52] <thomas> what's it say ? [12:53] <arik> Deferred: Connection timed out with college.antioch-college.edu. [12:53] <arik> don't know why it tried that [12:53] <thomas> and that's the same machine as your localhost or not ? [12:53] <arik> no [12:53] <arik> not at all [12:54] <thomas> ok, what's your localhost ? [12:54] <thomas> and what machine is your mail account on ? [12:54] <thomas> and what type is it ? [12:54] <arik> heh [12:54] <arik> my localhost is this machine [12:54] <arik> or am i not understanding ;- [12:54] <arik> ) [12:54] <arik> my mail account is on a machine at my school [12:54] <thomas> arik: i mean, what's it's FQDN [12:54] <arik> fqdn? [12:54] <thomas> fully qualified domain name [12:54] <arik> man i am a novice with mail [12:54] <arik> oh [12:55] <thomas> what's it's name on the network [12:55] <arik> it's just a dynamic ip on a ppp connection [12:55] <thomas> and what is the name of the mailserver [12:55] <arik> mail.antioch-college.edu [12:55] <arik> but i dpn't have that setup [12:55] <arik> at all [12:55] <arik> i mean i don't have it setup to use that [12:55] <thomas> arik: uhm, now you've lost me [12:55] <arik> heh [12:55] <thomas> how do you want to receive mail from a server then ? [12:55] <arik> i have that setup [12:55] <arik> i don't have smtp setup [12:56] <thomas> ok, the connection with college.antioch-college.edu was made ... [12:56] <thomas> ... because that's the same server as mail.antioch-college.edu [12:56] <arik> ok [12:56] <thomas> so what connection type are you using ? SMTP ? [12:56] <arik> i guess so [12:56] <arik> i don't know, i didn't setup sendmail atall [12:56] <arik> i have no idea [12:56] <arik> sorry i am so confused [12:57] <arik> i feel like a total novice again ;-) weird [12:57] <thomas> arik: no problem, I'm confused as well ;) [12:57] <arik> hehe [12:57] <arik> i have never setup sendmail at all [12:57] <arik> i setup fetchmail to get mail from my account [12:58] <thomas> arik: the thing that confuses me is that sending a mail contacts your mailserver of an account you established [12:58] <thomas> normally, sendmail figures out the mx record of the recipient [12:58] <thomas> and sends it itself [12:58] <arik> hmm [12:58] <arik> i don't know [12:59] <thomas> what does fetchmail do ? I never used that as well... [12:59] <arik> it gets mail from a server and puts it in /var/mail/user [12:59] <arik> so that mail clients can view it [13:00] <thomas> arik: ok, try to send a mail to another account that is not at antioch-college... [13:00] <thomas> ... see what line sendmail prints for that. [13:00] <arik> yeah i am trying that as we speak ;-) [13:00] <arik> May 23 04:00:47 localhost sendmail[16414]: f4NB0l216414: from=arik, size=39, class=0, nrcpts=1, msgid=<200105231100.f4NB0l216414@localhost.localdomain>, relay=arik@localhost [13:00] <arik> so far [13:00] <arik> that's what is says [13:01] <thomas> nothing more than that ? [13:01] <arik> nop [13:01] <arik> e [13:01] <arik> oh wait [13:01] <arik> Connection timed out with relay1-no.simplemente.net. [13:01] <thomas> ok, can you ping those hosts ? [13:02] <arik> yes [13:02] <thomas> then try this : "telnet mail.antioch-college.edu 25" [13:02] <thomas> this way you establish a manual smtp connection [13:03] <thomas> wait for a line starting with 220 [13:03] <thomas> if that doesn't work, you have a network connectivity problem [13:03] <arik> ok [13:04] <thomas> do you get a reply ? [13:04] <arik> not yet [13:04] <arik> i can ping it though [13:05] <thomas> hmmm... it seems like your port is blocked. [13:05] <thomas> have nmap installed ? [13:05] <thomas> if you don't get a relatively fast reply from the telnet command, something is wrong. [13:06] <arik> don't know ;-) [13:06] <thomas> You did add "25" at the end of the line, right ? [13:06] <arik> yes [13:06] <thomas> because when I do it, I get an immediate reply [13:06] <thomas> ok, can you give me your ip address ? I'll scan it [13:06] <thomas> arik: other stuff like browsing, telnetting works, right ? [13:06] <arik> 168.191.230.141 [13:06] <arik> yes [13:06] <arik> works fine [13:08] <thomas> ok, I can make an ESMTP connection to your machine... [13:08] <arik> hm [13:10] <arik> where should nmap be? [13:11] <thomas> I think www.fyodor.org [13:11] <thomas> but i'll check first [13:11] <arik> no i mean on my system ;-P [13:11] <thomas> not ;) it's a port scanner [13:11] <arik> og [13:11] <arik> er oh [13:11] <thomas> you have to get it, it's probably illegal in some countries i suppose [13:11] <arik> ok [13:11] <arik> i wonder why this isn't working [13:12] <thomas> arik: if the telnet to the port directly isn't working, then that's the problem [13:12] <arik> right [13:12] <thomas> that's also what the line in your maillog indicates [13:12] <arik> but i wonder why [13:12] <arik> right [13:12] <thomas> arik: me too [13:12] <thomas> ;) [13:12] <arik> so how do i fix it ;-P [13:12] <thomas> arik: but a portscan can help in diagnosing [13:12] <arik> ok [13:13] <thomas> arik: since I can access it from where I am [13:13] <thomas> arik: how concerned are they about security at your uni ? [13:13] <arik> not at all i don't think [13:13] <thomas> arik: btw nmap is at www.insecure.org [13:13] <arik> ok [13:13] <thomas> arik: portscanning can be detected with the right tools, if they're conscious about that [13:13] <thomas> sometimes it's explicitly forbidden [13:13] <arik> i doubt it [13:13] <arik> it's possible i suppose [13:14] <thomas> well it's your call of course <grin> [13:14] <arik> heh [13:15] <thomas> when you have it installed (get an rpm), just type "nmap mail......" [13:15] <thomas> that will scan all of their ports and list what is open. [13:15] <arik> ok [13:15] <thomas> I'll do the same and then we can compare the results [13:15] <arik> ok [13:15] <arik> i hope i don't get in trouble [13:15] <arik> oh well [13:16] <thomas> well, did you ever hear about a student getting in trouble ? [13:16] <thomas> I'll try first if you want [13:16] <arik> no [13:16] <arik> i really think it will be fine [13:16] <thomas> let me know if you get a mail on your localhost [13:16] <arik> i got your maikl [13:16] <thomas> directly, that is [13:16] <thomas> ok [13:17] <arik> yes [13:17] <arik> got it [13:17] <thomas> your esmtp is fine then, I did it manually [13:17] <arik> hmm [13:18] <thomas> arik: start with namp -p 20-30 mail.antioch-college.edu [13:18] <arik> ok [13:18] <thomas> that only scans a few ports [13:18] <arik> done [13:19] <thomas> ok, paste the output [13:19] <arik> Port State Service [13:19] <arik> 21/tcp open ftp [13:19] <arik> 22/tcp open ssh [13:19] <arik> 23/tcp open telnet [13:19] <arik> 25/tcp filtered smtp [13:19] <thomas> ok, you see the "filtered" ? [13:19] <arik> yeah [13:19] <thomas> that indicates they're deliberately blocking based on IP [13:19] <thomas> when I do the same, I have the port as "open" like all the others [13:19] <arik> weird [13:19] <arik> why would they block my ip? [13:20] <thomas> don't know. it's nothing personal I guess ;) [13:20] <arik> hah [13:20] <thomas> try a normal telnet to mail.... [13:20] <thomas> no port specified [13:20] <thomas> you should get a quick response, right ? [13:20] <arik> yes [13:20] <thomas> well, the response on port 25 should be just as quick... [13:20] <thomas> ... since it's filtered, telnet hangs [13:20] <arik> ok [13:21] <thomas> ... at this point, the best you can try is nmap the same ports on that other server you used [13:21] <thomas> that way you can check if it's the server or your isp that's filtering [13:21] <arik> ok [13:21] <thomas> I'd suppose it's your isp since I have the port listed as open [13:21] <thomas> and they don't know me [13:21] <thomas> ;) [13:22] <arik> actually [13:22] <arik> the other one is gnome.org [13:22] <arik> that i could get in trouble for ;-) [13:22] <thomas> well, just start with a telnet to port 25 then [13:22] <arik> that didn't work [13:22] <arik> i tried it [13:23] <thomas> btw: when you limit the portscan range, you can't do much harm [13:23] <arik> ij [13:23] <arik> er ok [13:23] <thomas> btw: I can connect to the port just fine [13:23] <arik> ok [13:23] <arik> so this is my isp? [13:23] <thomas> maybe nmap -p 25 gnome.org also works [13:23] <arik> fuck [13:23] <arik> filtered [13:23] <arik> for me [13:23] <thomas> ok, sounds like your ISP is to blame [13:24] <arik> fucking a [13:24] <thomas> is it DSL ? [13:24] <arik> no [13:24] <thomas> what type ? [13:24] <arik> modem [13:26] <thomas> what's your isp ? [13:26] <arik> earthlink [13:26] <thomas> hang on, i'll check if it's blacklisted [13:26] <arik> ok [13:29] <thomas> arik: It's probably something else [13:29] <arik> hmm [13:29] <thomas> arik: it could be that your ISP only allows SMTP relays through their own server [13:30] <arik> maybe [13:30] <thomas> normally they don't do that for modems, since the speed is slow [13:30] <thomas> arik: have you ever sent *any* mail through that isp ever ? [13:30] <arik> no [13:30] <thomas> arik: because if that's the case, you'll first need to know the smtp relay for earthlink... [13:30] <thomas> ... and add that to sendmail.cf [13:30] <thomas> then it should be fixed [13:30] <arik> hmm [13:31] <thomas> ok, try this then : [13:31] <thomas> telnet mail.earthlink.net 25 [13:31] <arik> ok [13:31] <thomas> does that work ? [13:31] <arik> yes [13:31] <thomas> arik: ok, then that's probably your relay [13:32] <arik> ok [13:32] <thomas> do you also get this : 220-hawk.mail.pas.earthlink.net ESMTP Sendmail EL-8_9_3_3/8.9.3; Wed, 23 May 2001 04:32:38 -0700 (PDT) [13:32] <thomas> 220-NO UCE. Earthlink Network does not authorize the use of its computers [13:32] <thomas> 220 or network equipment to deliver unsolicited electronic mail. [13:32] <thomas> Q [13:32] <thomas> ? [13:32] <arik> yes [13:32] <thomas> without the Q that is [13:32] <thomas> ok... [13:32] <thomas> edit /etc/sendmail.cf [13:32] <thomas> find the line containing "relay" [13:32] <thomas> it looks like this : [13:32] <thomas> # "Smart" relay host (may be null) [13:32] <thomas> DS [13:32] <thomas> [13:32] <arik> um [13:32] <arik> i don't have a /etc/sendmail.cf [13:32] <thomas> huh ? [13:32] <arik> wait [13:32] <arik> yeah i do [13:32] <arik> dumb dumb [13:33] <thomas> no problem ;) [13:33] <arik> fount it [13:33] <arik> what now? [13:34] <thomas> ok, change the line "DS" to "DSmail.earthlink.net" [13:34] <thomas> no space inbetween DS and mail.... [13:34] <thomas> save that [13:34] <arik> ok [13:34] <thomas> then run /etc/rc.d/init.d/sendmail restart [13:34] <thomas> then try sending a mail again [13:34] <thomas> and paste the sendmail output [13:34] <thomas> (hmm... this shouldn't go in the gstreamer logs) [13:35] <arik> hehe [13:36] <arik> May 23 04:38:36 localhost sendmail[16910]: f4NBcaR16910: from=root, size=52, class=0, nrcpts=1, msgid=<200105231138.f4NBcaR16910@localhost.localdomain>, relay=root@localhost [13:36] <thomas> did sendmail restart ? [13:36] <arik> yes [13:37] <arik> May 23 04:38:04 localhost sendmail[16890]: alias database /etc/aliases rebuilt by arik [13:37] <arik> May 23 04:38:04 localhost sendmail[16890]: /etc/aliases: 38 aliases, longest 10 bytes, 377 bytes total [13:37] <arik> May 23 04:38:04 localhost sendmail[16905]: starting daemon (8.11.0): SMTP+queueing@01:00:00 [13:38] <thomas> arik: where did you send the mail to ? [13:38] <arik> ad...@an... [13:38] <arik> Connection timed out with xwing.aoltw.net. [13:38] <arik> wtf? [13:39] <thomas> ok, that's america online i suppose [13:39] <arik> right [13:39] <arik> why is it connecting to that? [13:39] <thomas> that's weird [13:39] <arik> yes [13:39] <thomas> arik: don't know [13:39] <arik> very strange [13:39] <thomas> arik: try "nslookup mail.earthlink.net" [13:39] <thomas> what do you get ? [13:40] <arik> Server: dns1-sf.snfc21.pacbell.net [13:40] <arik> Address: 206.13.28.12 [13:40] <arik> Non-authoritative answer: [13:40] <arik> Name: mail.earthlink.net [13:40] <arik> Addresses: 207.217.120.202, 207.217.121.203, 207.217.120.204, 207.217.121.205 [13:42] <thomas> arik: try sending the mail to me, th...@ur... [13:42] <arik> ij [13:42] <arik> er ok [13:42] <arik> you are in belgium? [13:43] <thomas> I have some message on deja claiming that AOL disallows delivery to locals [13:43] <thomas> don't see why [13:43] <thomas> arik: yes [13:43] <arik> i was in belgium for new years [13:43] <arik> in brussels [13:43] <arik> sent [13:43] <thomas> ok, what does sendmail say now ? [13:43] <arik> May 23 04:45:49 localhost sendmail[16920]: f4NBjmf16920: from=root, size=48, class=0, nrcpts=1, msgid=<200105231145.f4NBjmf16920@localhost.localdomain>, relay=root@localhost [13:43] <thomas> arik: I'm in brussels now [13:43] <arik> ;-) [13:43] <arik> i like brussels [13:43] <thomas> I'm from Gent, Gent is better ;) [13:43] <arik> hehe [13:44] <arik> flemish? [13:44] <thomas> ja ;) [13:44] <arik> hah [13:44] <arik> ich sprache deutsch [13:44] <thomas> heh [13:44] <arik> but not flemish ;-P [13:44] <arik> did you get the mail? [13:45] <thomas> well, it should give a second line before I can get it ;) [13:45] <arik> heh [13:45] <thomas> the first says it's going to send it [13:45] <arik> nope [13:45] <arik> another timeout [13:45] <thomas> but I think the relay line should be different [13:45] <thomas> what server did it try to connect to ? [13:46] <arik> ywind [13:46] <arik> er ywing.aoltw.net [13:46] <arik> that is so weird [13:46] <arik> i don't have aol [13:46] <arik> although i do work for them ;-) [13:46] <arik> oh! [13:46] <arik> i bet i know what's up [13:46] <arik> mayve [13:46] <arik> er maybe [13:46] <arik> no [13:46] <arik> that wouldn't matter [13:46] <arik> nm [13:47] <thomas> ok, now i'm worried [13:47] <arik> heh [13:48] <arik> what on earth have i done to have this happen? [13:48] <thomas> let's try this: change mail..... in your sendmail.cf to an IP address: [13:48] <arik> ok [13:48] <thomas> DS207.217.121.202 [13:49] <arik> done [13:49] <thomas> now restart sendmail again [13:49] <arik> done [13:50] <thomas> now send a mail [13:50] <arik> done [13:50] <thomas> what does sendmail say now ? [13:50] <arik> May 23 04:53:17 localhost sendmail[16982]: f4NBrH416982: from=root, size=48, class=0, nrcpts=1, msgid=<200105231153.f4NBrH416982@localhost.localdomain>, relay=root@localhost [13:51] <arik> same thing [13:51] <arik> xwing [13:52] <arik> so weird [13:53] <thomas> ok, we'll probably need to add some masquerading as well. [13:53] <thomas> ok, first make a backup of sendmail.cf [13:53] <arik> done [13:54] <thomas> now find the line saying # who I masquerade as (null for no masquerading) [13:54] <arik> done [13:55] <thomas> on the next line, change DM to DMarik...@.... [13:55] <thomas> the address you want to send as [13:55] <thomas> ... wait. your isp is earthlink... [13:55] <arik> yes [13:55] <thomas> ... but you're trying to send mail as a uni user, right ? [13:56] <arik> yes [13:59] <thomas> well, let's just try this first. [13:59] <arik> ok [13:59] <thomas> save it, restart sendmail, send mail, see what sendmail says [13:59] <arik> what do i put after ...? [14:00] <thomas> what you have as an address there [14:00] <arik> ij [14:00] <arik> er ok [14:01] <arik> May 23 05:03:45 localhost sendmail[17031]: f4NC3jq17031: from=root, size=48, class=0, nrcpts=1, msgid=<200105231203.f4NC3jq17031@localhost.localdomain>, relay=root@localhost [14:03] <arik> same fscking timeout [14:03] <arik> anyway i'm gonna sleep [14:03] <arik> i'll try again late [14:03] <arik> r [14:03] <arik> thanks alot ;-) [14:04] <thomas> arik: didn't help much did it ;( [14:04] <arik> heh [14:04] <arik> s'ok [14:04] <arik> i learned some stuff [14:04] <thomas> anyway, one thing you should try if you have the time is to put [14:04] <thomas> mailer: after the DS, before the host [14:04] <arik> ok [14:04] <thomas> or smtp: [14:04] <arik> ok [14:04] <thomas> or esmtp: [14:04] <arik> which one? [14:04] <thomas> depends on your sendmail [14:04] <thomas> check all three of them in row [14:04] <thomas> so change sendmail.cf, restart, send mail, watch for output [14:05] <thomas> you should get a different "relay=..." line to get it working [14:05] <thomas> but do it when you're awake ;) [14:05] <arik> heh [14:05] <arik> i have time for that [14:05] <thomas> cu [14:05] <thomas> (wtay-bot: clear the logs please ;)) [14:05] arik (ar...@sd...) left #gstreamer. [14:13] omega_away (om...@bo...) joined #gstreamer. [14:13] <thomas> hi omega [14:19] omega_away (om...@bo...) left irc: Ping timeout for omega_away[boisdslgw5poolb76.bois.uswest.net] [14:20] Nick change: wtay-sleeping -> wtay [14:20] <wtay> yo [14:21] <wtay> sorry, I sent out the dailt IRC log again 'cause I thought something was wrong, but there wasn't. I just didn't notice the mails in my mailbox :) [14:21] <wtay> s/dailt/daily [14:26] <thomas> hi wtay [14:27] <thomas> i'm installing ximian gnome, is that a good or bad idea ? ;) [14:28] <wtay> good idea :) [14:28] <thomas> yeah well it's going to upgrade half of my packages ;) [14:28] <thomas> oh well, let's get something to eat then first [14:28] <thomas> wtay: are you around today ? [14:29] <wtay> thomas: a bit, I'm trying to get my scanner working [14:29] <thomas> wtay: I'd like to work on the plugin writer manual [14:29] <wtay> It used to work with older kernels but in 2.4.4 it oopses and hangs the system [14:29] <thomas> wtay: ok, see you in a bit then [14:30] <wtay> right [14:47] hadess (ha...@pc...) joined #gstreamer. [14:48] <hadess> hey dudes [14:49] <wtay> yo hadess [14:50] <hadess> hooo, you're on leave, right :) [14:51] <wtay> yeah, fixing a scanner problem :( [14:51] <wtay> 2.4.4 oopses [14:51] <wtay> trying with 2.4.0 now [14:51] <hadess> wave at michele: "Woohoo" [14:52] <hadess> wtay: you know what ? to rip a cd you can use gnomevfssrc ;) [14:55] <wtay> hadess: cool :) [14:56] <hadess> cdda:///dev/cdroms/cdrom0/ in nautilus shows me all the tracks on my CD [14:56] <wtay> neat [14:57] <hadess> what scanner is it you have ? [15:00] <wtay> Microtek Scanmaker X6 [15:00] <wtay> SCSI [15:00] <wtay> looks like 2.4.4 broke something in scsc.c, applying AC patches now... [15:00] <hadess> huh. and it locks up hard, bad :/ [15:00] <wtay> yeah, scsi oops [15:12] <thomas> hmmm, this redcarpet installer thingy isn't moving anymore [15:18] Nick change: hadess -> hds-afk [15:29] dobey (do...@dr...) joined #gstreamer. [15:31] <thomas> hi dobey... i'm installing ximian ;) [15:32] <dobey> heh [15:32] <thomas> tough redcarpet stopped downloading stuff twice [15:32] <dobey> ;-P [15:32] <dobey> what server are you grabbing from? [15:32] <thomas> I first took the one in france twice... [15:32] <thomas> ... now I'm on ximian [15:32] <thomas> good thing it saves it's downloads [15:32] <thomas> nice installer btw [15:33] <dobey> thanks [15:33] <thomas> did you do it ? [15:33] <dobey> no [15:33] <dobey> the RC people did [15:34] <thomas> it is *big* [15:34] sienap (sy...@ip...) joined #gstreamer. [15:34] <thomas> hey sienap [15:34] <sienap> wieeeeeeeeeeh i've got my ximian monkey :) [15:34] <sienap> Thomas! [15:34] <dobey> ls [15:34] <dobey> eek [15:34] <sienap> ehehe [15:34] <sienap> :) [15:34] <sienap> a /window new shell would be nice in BitchX :) [15:36] <thomas> wtay: still around ? [15:37] <thomas> [15:37] <thomas> dobey: hmm, installer isn't downloading again... [15:37] <thomas> dobey: ... is this something other people have as well ? [15:38] <dobey> thomas: how fast is your connection? [15:38] <thomas> ADSL [15:38] <thomas> 60 K/sec on average [15:38] <thomas> that's not it anyway [15:38] <thomas> I'm streaming 128 kbit/sec audio all the way as well so [15:38] <wtay> thomas: nope, still strugling with the scanner [15:38] <thomas> wtay: good luck ;) [15:38] <thomas> wtay: I'm trying to find where the sgml source for the plugin stuff is [15:38] <thomas> is it ok to edit that ? or is that not done ? [15:39] <dobey> thomas: hrmm, i've no idea [15:39] <thomas> oh well... I'll restart it again then [15:39] <dobey> thomas: the servers may be having trouble with large amounts of users [15:39] <thomas> dobey: I'll try a different mirror then [15:44] <thomas> wtay: if you have the time, let me know what kind of tool to use for sgml ;) [15:44] <thomas> or does everyone do it by hand ? sgml source looks ugly [15:45] <wtay> thomas: yup, I do it with vi [15:45] Nick change: hds-afk -> hadess [15:45] <wtay> pfff fsck again and again... :( [15:46] <thomas> wtay: why are all of the tags closed on a new line, is that required ? [15:46] <dobey> wtay: are you handling bonobo-media dev right now? [15:46] <dobey> allo hadess [15:47] <hadess> hey dobey [15:50] Sopwith (so...@na...) joined #gstreamer. [15:50] <hadess> hey Sopwith [15:51] <Sopwith> hi [15:53] <hadess> right... [15:58] <hadess> Sopwith: btw, there's a glib helper for strtol and strtod [15:59] <Sopwith> are you a markobot? [15:59] <dobey> hahahha [16:00] <hadess> Sopwith: you told me about it the other day when i wanted to have numbers in my xml file [16:00] <Sopwith> oh I don't remember, sorry. [16:00] <Sopwith> glad you have it working though :) [16:01] <sienap> opswith! [16:01] <sienap> sopwith maybe :) [16:09] <dobey> hadess: be glad you aren't standing next to me right now ;-P [16:10] <hadess> dobey: why so ? [16:10] <dobey> hadess: because you are french, and i feel like kicking a french person in the pants because of mandrake [16:11] <hadess> hehe [16:18] Action: wtay hates fsck (especially on a 12GB HD) :( [16:19] <dobey> use xfs then [16:20] <wtay> dobey: perhaps, but that doesn't make my current fsck go away... [16:20] <dobey> C-c it [16:22] <hadess> wtay: but that makes fsck faster [16:24] Action: thomas is restarting to test ximian... [16:24] thomas (th...@21...) left irc: [x]chat [16:27] omega_away (om...@bo...) joined #gstreamer. [16:30] <hadess> hey omega_away [16:34] omega_away (om...@bo...) left irc: Ping timeout for omega_away[boisdslgw5poolb76.bois.uswest.net] [16:35] thomas (th...@21...) joined #gstreamer. [16:36] <thomas> dobey: still here ? [16:36] <dobey> yeah [16:36] <thomas> ximian looks nice ... [16:36] <thomas> ... but how do I get my menu bar to be integrated with the bottom panel again ? [16:36] <dobey> how do you mean? [16:36] <thomas> well, before I had all my launch buttons, plus the gnome menu button, on the bottom panel [16:37] <thomas> now I have a menu and launch bar on top [16:37] <thomas> and my window switcher stuff at the bottom [16:37] <thomas> I can't find it in preferences [16:37] <dobey> oh, did you change stuff in doorman? [16:37] <dobey> thomas: run doorman-reset [16:37] <dobey> (i think) [16:38] <thomas> hmm... restart gnome as well ? [16:38] <thomas> btw where's evolution ? [16:39] <dobey> yes [16:39] <dobey> how do you mean? the menu item? or the package? [16:39] <thomas> evolution the exchange replacement ;) [16:40] <dobey> uh [16:40] <dobey> well [16:40] <dobey> that doesn't tell me much ;-) [16:40] <thomas> evolution is supposed to be the best linux groupware solution ... [16:40] <thomas> ... but is still unfinished... [16:41] <thomas> ... I need a way to replace outlook so I can switch to linux for my work system ;) [16:42] <dobey> so you want the package? [16:42] <thomas> dobey: well, supposedly it should be in ximian gnome anyway [16:42] <thomas> but since you don't know it, it probably isn't [16:42] <dobey> uhm [16:43] <dobey> no, it's "beta" software [16:43] <dobey> it is in a separate channel in RC [16:43] <dobey> probably "Evolution" [16:43] <thomas> ah ok, I'll check that then [16:43] <thomas> ok, restarting... [16:43] thomas (th...@21...) left irc: [x]chat [16:50] sienap (sy...@ip...) left irc: sienap_ has no reason [16:52] thomas (th...@21...) joined #gstreamer. [16:56] <hadess> thomas: evo is not an exchange replacement [16:57] <thomas> hadess: as in "not yet" or "never will be" ? [16:57] <hadess> thomas: never will [16:57] <hadess> thomas: it's a groupware, not a server [16:57] <thomas> hadess: so it can't talk to exchange servers ? [16:58] <hadess> thomas: tell me, what is exchange... [16:58] <thomas> exchange is the microsoft groupware server ... [16:58] <thomas> ... integrating nicely with the outlook client ... [16:58] <thomas> to deliver mail (not a problem) ... [16:58] <hadess> ok, and what is evolution ? [16:58] <dobey> uhm [16:59] <thomas> ... but also integrated calendaring and address books [16:59] <dobey> evolution does not yet work with exchange [16:59] <thomas> hadess: I was led to believe that evolution was sometime going to be an outlook-replacement [16:59] <dobey> sometime, yes [16:59] <thomas> I don't like outlook that much, but it's standard policy over here so... [16:59] <hadess> thomas: so why do you say it's an exchange replacement ? [17:00] <thomas> hadess: did I ? I meant to say outlook replacement but exchange-capable. [17:00] <dobey> it is now if you don't use proprietary mail server crap [17:00] <hadess> <thomas> evolution the exchange replacement ;) [17:00] <dobey> heh [17:00] <thomas> hadess: sorry ;) [17:00] <thomas> I don't care if our server here runs windows & exchange [17:00] <thomas> actually I do, but I have enough work on my hands ... [17:00] <thomas> ... so I won't convince management here [17:00] <dobey> heh [17:00] <thomas> but I *do* care about what my desktop runs [17:01] <dobey> wine outlook [17:01] <dobey> heh [17:01] <thomas> Outlook is the only thing that's stopping me [17:01] <thomas> dobey: yeah, maybe [17:01] <thomas> hadess: you actually lead me to believe you didn't know what exchange was, bastard ;) [17:02] <hadess> thomas: i did exchange installs long ago [17:05] <thomas> hmm... dobey, stupid question : how do you get to the title bar menu allowing you to make a window sticky, when there's no titlebar (like in gkrellm) ? [17:06] <dobey> thomas: you configure matched windows in the sawfish config [17:19] thomas (th...@21...) left irc: [x]chat [17:19] thomas (th...@21...) joined #gstreamer. [17:35] taazzzz (dlehn@66.37.66.32) left irc: Ping timeout for taazzzz[66.37.66.32] [17:35] taazzzz (dlehn@66.37.66.32) joined #gstreamer. [17:39] greg_ (gr...@ho...) joined #gstreamer. [17:52] matth (ma...@bo...) joined #gstreamer. [17:52] Nick change: matth -> matth_ [17:53] <matth_> hadess: did you ever figure out the bonobo stuff? [17:55] taazzzz (dlehn@66.37.66.32) left irc: Ping timeout for taazzzz[66.37.66.32] [17:55] taazzzz (dlehn@66.37.66.32) joined #gstreamer. [18:04] <hadess> no not [18:05] <matth_> hadess: i'm familiar with what Orbit ORB can/can't do w.r.t pthreads, but not familiar with BONOBO... [18:07] <hadess> bonobo uses orbit... [18:07] <hadess> i trashed bonobo for now [18:07] <matth_> i assume you were using bonobo to remote gstreamer... were you remoting gst directly, or remoting a higher level api (like gstplay)? [18:07] <matth_> ah [18:08] <hadess> no, i was using gst in the container, and bonobo controllers [18:09] <matth_> so the interface remoted was some bonobo control interface? [18:09] <hadess> yep [18:09] <matth_> hmm... [18:10] <hadess> i should try it again, now that i have gdb 5.0 [18:11] Parapraxis (Sp...@us...) joined #gstreamer. [18:11] <hadess> hey Parapraxis [18:12] <Parapraxis> hey [18:12] <Parapraxis> hadess: how much more work do you have on yer gst frontend? [18:12] matth_ (ma...@bo...) left irc: Ping timeout for matth_[boisdslgw5poolb76.bois.uswest.net] [18:14] <hadess> Parapraxis: a lot, why ? [18:14] Action: Parapraxis is lookin fer some help on pn ;) [18:15] <Parapraxis> I'll just go put an add on sf [18:15] <hadess> i already maintain 5 programs... [18:15] <hadess> plus i work on a couple of others [18:15] <Parapraxis> lol, nm then ;) [18:18] <dobey> heh [18:18] <dobey> pn? [18:21] <hadess> dobey: paranormal.sf.net [18:21] <dobey> is it that visiplugin? [18:22] <Parapraxis> yup [18:22] <hadess> going to do some dishwashing, bbiab [18:22] <Parapraxis> sounds like fun [18:23] <dobey> hrmm [18:23] <Parapraxis> dobey: you interested in helping? ;) [18:23] <dobey> not really [18:23] <Parapraxis> hehe ;P [18:23] <dobey> i have WAY too much to do [18:23] <Parapraxis> dobey: what are you working on? [18:24] <dobey> Parapraxis: oh, like at least 10 things [18:24] <dobey> plus my day job [18:24] <Parapraxis> yikes [18:25] <dobey> yes [18:25] <dobey> this is why i "need" people to help me [18:28] matth_ (ma...@bo...) joined #gstreamer. [18:34] matth_ (ma...@bo...) left irc: Ping timeout for matth_[boisdslgw5poolb76.bois.uswest.net] [18:43] omega_away (om...@bo...) joined #gstreamer. [18:44] <hadess> dudes, you have some serious connectivity problems... [18:44] <dobey> heh [18:53] omega_away (om...@bo...) left irc: Ping timeout for omega_away[boisdslgw5poolb76.bois.uswest.net] [18:54] <wtay> pfff, finally... it works... [18:54] <hadess> what was the problem / [18:55] <wtay> DMA buffer size for the scsi driver [18:55] <dobey> hey wtay [18:55] <wtay> yo dobey [18:55] <hadess> wtf, this is dodgy [18:55] <wtay> hadess: yeah, that's why it took so long to find [18:56] <wtay> and lots of fscks and lockups.. :( [18:56] <wtay> compile the kernel a few time and libsane etc... not funny [18:58] <hadess> i'm adding multiple playlists support to rhythmbox... [18:58] <wtay> dobey: are you aware of the fact that the current .debs of ximian gnome have a malfunctioning caplet for sawmill? [18:59] <dobey> wtay: i don't know anything about the debs, nor care, what is the problem? [18:59] <wtay> dobey: when I want to configure sawmill in the control panel, clicking on a node (appearance) consumer 100% cpu and nothing happens [19:00] <wtay> s/consumer/consumes [19:00] <dobey> wtay: that isn't necessarily a problem with our packages, i've seen that happen several times even when installing from source [19:00] <wtay> dobey: maYam is very pissed off by that <g> [19:00] <wtay> dobey: any clue what would be the problem? [19:01] <hadess> wtay: use Window Maker like me ;) [19:01] <dobey> mv .sawfish .sawfish-bak, logout, login, and see if it works then [19:01] Nick change: taazzzz -> taaz [19:01] <Parapraxis> wtay: when gst moves to GObject, are all GTypes going to be registered as dynamic types? [19:01] <Parapraxis> hadess: wm rocks ;) [19:01] <taaz> wtay: you could just remove all ximian debs and go with standard stuff [19:01] <wtay> Parapraxis: GTypes? [19:01] <wtay> taaz: works better? [19:01] <dobey> taaz: that won't fix a problem that isn't cause by our packages [19:02] <wtay> likely a problem with sawmill... [19:02] <taaz> better? i have no idea... i've not used the ximian packages [19:02] <dobey> haah [19:02] <hadess> Parapraxis: i know ;) [19:02] <wtay> ok [19:02] <Parapraxis> GType's, yeah [19:02] <dobey> wtay: try what i said [19:02] <Parapraxis> the runtime type system of glib [19:02] <wtay> dobey: trying [19:03] <taaz> Parapraxis: i think the first stage will be s/gtk_/g_/ all over. then later look at new glib features like GTypes and so on [19:04] <Parapraxis> taaz: gtk_type_register -> g_type_register_static or g_type_register_dynamic [19:04] <hadess> Parapraxis: you should know next friday [19:04] <wtay> brb [19:04] <hadess> Parapraxis: pioneer ppl are doing the port [19:04] <Parapraxis> hadess: pioneer? [19:05] <hadess> Parapraxis: like in pioneer [19:05] <Parapraxis> ...? [19:05] <dobey> like in [19:06] <dobey> laserdisc [19:06] Action: Parapraxis is really confused [19:06] <hadess> hifi stereos, and stuff [19:06] Action: taaz has an old Pioneer amp [19:06] <Parapraxis> what does that have to do w/ this? [19:06] <hadess> <hadess> Parapraxis: pioneer ppl are doing the port [19:07] <taaz> read the irc log from yesterday, search for Pioneer [19:07] <Parapraxis> ok [19:07] <wtay> dobey: no change... [19:08] <dobey> wtay: hrmm [19:08] Nick change: taaz -> taaz-away [19:08] <dobey> wtay: check bugzilla to see if anyone has reported it then [19:08] <wtay> dobey: btw, is it sawfish or sawmill currently <g> [19:08] <dobey> wtay: i have no idea what else would cause that [19:08] <dobey> wtay: sawfish [19:09] <dobey> wtay: sawmill got renamed about a year ago [19:09] ajbusy (aj...@p4...) left irc: Ping timeout for ajbusy[p47-max4.dun.ihug.co.nz] [19:12] <wtay> dobey: yeah, it's a know problem, no solution posted yet... [19:12] <dobey> hrmm, what's the url for the roadmap page again? [19:13] <wtay> http://www.linuxrising.com/files/gstreroadmap2.html [19:14] <hadess> wtay: will you help me fix gnomevfssrc when incsched is merged ? [19:16] <wtay> hadess: sure [19:16] <hadess> ... need ... to .. finish ... playlist code... [19:17] CHW-cleaning_office (pa...@te...) joined #gstreamer. [19:17] <hadess> and the only thing i want to do is nothing [19:18] ajbusy (aj...@p4...) joined #gstreamer. [19:19] Nick change: CHW-cleaning_office -> ChiefHighwater [19:19] <dobey> hey aj [19:21] <hadess> wtay: there's a guy writing a ripper like grip with gst that mailed me [19:22] <dobey> cool [19:22] <hadess> wtay: he will mail in some changes to cdparanoia [19:24] <Parapraxis> does gst have a 32 bit color format? [19:30] <wtay> Parapraxis: fourcc RGB and depth 32/width 32 + masks etc... [19:31] <wtay> Parapraxis: flx decoder has 24 bpp, 32 is similar [19:34] Zeenix (s_l...@ho...) joined #gstreamer. [19:35] Action: wtay has a headache now :( [19:35] Nick change: wtay -> wtay-away [19:35] <Zeenix> hi there [19:35] <Zeenix> need an advice ? [19:35] <Zeenix> i need an advice [19:36] <wtay-away> yes? [19:37] <Zeenix> i want to develope an internet voice chat app. [19:37] <Zeenix> i want to decide whether to use threads or processes [19:38] <dobey> uhm [19:38] <wtay-away> Zeenix: why? [19:39] <wtay-away> processes are a pain in the ass unless you use corba [19:39] <wtay-away> or bonobo [19:39] <Zeenix> what ? [19:40] <Zeenix> then why do many good apps on linux uses processes [19:40] <wtay-away> Zeenix: define processes for me [19:42] <Zeenix> if you dont know about this, how could you answer my question [19:42] <hadess> Zeenix: are you stupid or what ? [19:42] <dobey> uhm [19:42] <wtay-away> Zeenix: I see a process as a fork/execvp setup [19:42] <dobey> everything is a pain in the ass [19:42] <dobey> just use the app that bjack wrote when he releases it ;-) [19:42] <hadess> threads are easier... [19:42] <wtay-away> If you use processes, then you need some kind of communication mechanism between then [19:42] <hadess> dobey: exactly =) [19:43] <wtay-away> 'cause they don't share memory like a thread does [19:44] <wtay-away> you can write your own communication protocol (like gstoss, a gstreamer plugin does) or you can use corba/bonobo [19:44] <wtay-away> or you can use threads and then everything is simple, sorta [19:45] <Zeenix> but if you a single process with many treads, the process still gets the same time slice from the cpu, which is then divided b/w the threads [19:45] <Zeenix> but with more process you get more time slices [19:46] <wtay-away> no [19:47] <Zeenix> plz explain [19:47] <dobey> ok [19:47] <wtay-away> in linux, threads are implemented in the same way a process is [19:48] <wtay-away> so each thread shares the same amount of CPU time [19:48] <Zeenix> thats very nice [19:48] <hadess> the only difference in linux is that threads share the same memory as the main process [19:48] <Zeenix> are you really sure about it [19:48] <wtay-away> yup, a different arg to the clone()call [19:48] <wtay-away> fairly sure, yes [19:48] <Zeenix> sharing memory is another plus another plus [19:48] <wtay-away> indeed [19:49] <wtay-away> threads are much easier to coordinate then processes [19:49] <Zeenix> I am nothing without this irc server [19:49] <Zeenix> i know that [19:49] <Zeenix> i'll have to learn those [19:49] <Zeenix> i have done it in Java [19:49] <wtay-away> although reuse of components is not possible with threads [19:50] <wtay-away> but it is with corba/bonobo, which are processes [19:50] <Zeenix> what do you mean by that [19:51] <wtay-away> each component runs in its own process, they talk to eachother using corba [19:51] <wtay-away> if one component crashes, the whole app doesn't crash with it [19:51] <wtay-away> you can see that in evolution, where parts of the app crash :) [19:52] <wtay-away> with threads, if one thread crashes, the whole app crashes [19:52] <Zeenix> i wont like to concider this problem in this app. [19:52] <Zeenix> As i need to do it fast. [19:53] <wtay-away> Zeenix: what kind of components will be in your app that will run concurrently? [19:53] wayne (wa...@ra...) left irc: Client Exiting [19:53] <Zeenix> 1. it'll be divided into a client & a server [19:54] <wtay-away> ah [19:54] Nick change: wtay-away -> wtay [19:54] <Zeenix> 2. then the client'll contain a thread for takin sound from the mic. [19:55] <wtay> ok [19:55] <Zeenix> 3. client contain a thread for compressing data( MP3 i think ) [19:55] <wtay> right [19:55] <hadess> mp3 is a very idea for streaming voice [19:55] <Zeenix> 4. another thread in client will be responsible for sending data [19:56] <wtay> Zeenix: why use threads for that? [19:56] <Zeenix> 5. Concider the oposite side of stuff for the server [19:56] <wtay> read -> encode -> send [19:56] <wtay> no threads needed IMO [19:56] <wtay> well, maybe one for the GUI [19:57] <wtay> how are you going to communicta this mp3 to the server? [19:57] <Zeenix> sockets [19:59] <Zeenix> read -> encode -> send will be a synchronous process, that'll make data lose [20:00] <Zeenix> i have done some mathematical stuff for that [20:00] <Zeenix> reading, encoding & sending wont be done at the same speed [20:00] <wtay> true, a little thread and a queue will somwhat imporve this [20:01] <Zeenix> you are right [20:01] <wtay> read->encode->queue->send [20:01] <Zeenix> i think a producer consumer model as i stated will be a better idea [20:02] <Zeenix> not doing everything in a single thread [20:02] <wtay> ok [20:02] <wtay> read->queue->encode->queue->send [20:02] <Zeenix> any way, i havent yet started to learn gstreamer, should i learn gtk+ first [20:03] <wtay> yeah, the object model is helpfull [20:03] <Zeenix> i could not understand, they say its object oriented, i could not see any object orientation in any gtk+ code in c [20:05] <wtay> Zeenix: look at the classes, virtual methods, inheritence etc. [20:06] <Zeenix> classes in C [20:06] <Zeenix> they are in C++ [20:07] <wtay> classes == struct [20:08] <wtay> remember how the early C++ compilers precompiled to C first... [20:09] <Zeenix> i am not that old [20:09] <wtay> heh [20:09] <Zeenix> but when i see a gtk+ code , all i find are simple function calls like ordinary c [20:10] <wtay> Zeenix: yup, and the first arg is the object to perform this method on [20:10] <wtay> so pad1->connect(pad2) becomes pad_connect (pad1, pad2) [20:11] <wtay> same thing, other syntax.. [20:12] <Zeenix> i thought they'll be using struct as objects [20:13] <Zeenix> if i put i pointer to function as a member of struct, then i think i'll be using struct just like an object [20:13] <Parapraxis> Zeenix: the object oriented part is that there is inheritance [20:14] <wtay> and virtual functions, which are pointers in the class struct [20:14] <Parapraxis> for example, a GtkButton can be cast to a GtkContainer, and treated just as if it were one (and in fact it is) [20:14] <Parapraxis> (it contains a label usually) [20:16] <Sopwith> hmm, no uraeus today, I was going to spew the $0.02 on AF [20:16] <Zeenix> class struct ? [20:17] omega_rr (om...@bo...) joined #gstreamer. [20:17] <Zeenix> wtay: what do you mean ? [20:17] <Zeenix> which class struct ? [20:17] <Parapraxis> hey omega_rr [20:17] <omega_rr> yo [20:17] <wtay> yo [20:18] <wtay> Zeenix: look at the definition of an object and its class [20:19] <steveb> yo [20:19] <Parapraxis> wtay: I'm not sure that's such a good idea ;) [20:19] <Zeenix> wtay: ? [20:20] <Zeenix> wtay: there is no class in c [20:20] <Parapraxis> Zeenix: you should go look up what the C++ vtable is, then think about how it could be implimented in C [20:21] <Zeenix> whats the vtable [20:21] <Parapraxis> Zeenix: that's what you need to look up ;) [20:22] steveb (st...@no...) left irc: [x]chat [20:22] <Parapraxis> Zeenix: basically, it's the table of all functions that could possibly be overriden by child classes, but it's stored with the instance data of the object [20:23] <wtay> Parapraxis: in gtk it's stored in the Class struct [20:23] <Zeenix> is Class implemented by a struct [20:23] steveb (st...@no...) joined #gstreamer. [20:23] <wtay> yes [20:23] <Parapraxis> wtay: I know, but in C++ it's stored w/ the instance data (i believe), in gtk they just have a pointer to the class at the beginning [20:23] <wtay> yo [20:23] <steveb> yo [20:24] <Parapraxis> yo [20:24] <wtay> Parapraxis: not sure that is correct... [20:24] <Parapraxis> wtay: neither am I, actually [20:24] <wtay> doesn't realy make sense to override the method for each instance when it's just a virtaul function [20:24] <steveb> omega_rr: so are pioneer going to do the port openly in a branch of CVS? [20:24] <Parapraxis> in fact, it probably isn't, since that would mess up the offsets of the data members [20:25] <Zeenix> ok ok i'll learn gtk+ first & then talk to you about it [20:25] <wtay> http://developer.gnome.org/doc/GGAD/cha-objects.html [20:26] <Zeenix> wtay: for me ? [20:26] <omega_rr> steveb: that's the idea, but we haven't heard back from the developers yet [20:26] <wtay> Zeenix: yup, good stuff to learn gtk [20:27] <omega_rr> http://ars.userfriendly.org/cartoons/?id=20010523 [20:27] <wtay> hehe [20:30] <omega_rr> Parapraxis: I have a potential solution for the problem you and dichro have with unpredictable sink rates [20:30] <Parapraxis> do tell [20:31] <omega_rr> use a thread for audio decode and a thread for the viz code [20:31] <omega_rr> put a 'leaky' queue between them, which is leaking on the downstream side [20:32] <omega_rr> as audio goes into the queue, the queue fills [20:32] <omega_rr> when the queue is full, new buffers on the upstream side push old buffers off the end of the downstream side, destroying them [20:33] <omega_rr> when you pull, you only get the latest buffer, offset by the depth of the queue [20:33] <omega_rr> that allows you to completely separate the clocks of the audio and viz threads [20:33] <Parapraxis> omega_rr: hehe, that's what wtay said the 1st day ;) [20:33] <omega_rr> I implemented the leaky queue last night [20:33] <Parapraxis> cool [20:33] Action: wtay creates a leaky queue [20:33] hadess (ha...@pc...) left irc: mooooh! [20:33] <omega_rr> same model works for dichro's webcam [20:33] Nick change: taaz-away -> taaz [20:34] <taaz> what do you mean leaky? [20:34] Zeenix (s_l...@ho...) left irc: [20:34] <omega_rr> taaz: if the queue fills up, new buffers are either dropped, or they force the other end to drop to make room [20:34] <wtay> taaz: buffers are not queued when the queue is filled [20:34] <dobey> hmm [20:35] <omega_rr> wtay: it's the downstream leaky model that's interesting to Parapraxis and dichro [20:35] <wtay> omega_rr: you implemented it? [20:35] <omega_rr> yup [20:35] <wtay> well, me too now :) [20:36] <omega_rr> hmm [20:36] <omega_rr> I started implementing other stuff in the queue as well [20:36] <omega_rr> like a start at size-by-bytes [20:37] <wtay> in fact buffers should be dropped in an intelligent way, like when you have a queue depth of 5 and 10 buffers have been put into it you'll have 1 3 5 7 9 in the queue [20:37] <omega_rr> wtay: why? what purpose would that serve? [20:37] <omega_rr> a decimating queue has to be data-specific [20:37] <steveb> is this the same as QOS? [20:37] <omega_rr> it's a small part of the QoS solution [20:37] <wtay> omega_rr: I guess, although for audio it could make sense [20:38] <omega_rr> right. data-specific <g> [20:38] <omega_rr> for audio it only makes sense if you decimate every other sample [20:38] <wtay> omega_rr: a leaking queue is also data specific [20:38] <omega_rr> and that's the *wrong* way to do it anway (even if it's the fast way) [20:38] <wtay> yes [20:39] <omega_rr> wtay: not really. it works for any data that follows a few basic guidelines, like no interdependencies, latest-data matters most, and each buffer is one 'unit' [20:39] <taaz> user beware with this sort of thing... if the receiver expects continuity in the data stream... [20:39] <omega_rr> taaz: of course. you only use this when you know what you're doing [20:40] <wtay> ok [20:40] <omega_rr> now, here's a question: [20:40] <omega_rr> the queue will have the ability to set the max size based on buffer count, byte total, and time [20:41] <ChiefHighwater> yes. whew I got one 8-] [20:41] <omega_rr> do we want to allow more than one of these limits to be in place at the same time? [20:41] <omega_rr> ChiefHighwater: ? [20:41] <wtay> neat [20:43] <omega_rr> wtay: if we do, how do we go about configuring that? bunches of args? [20:43] <ChiefHighwater> omega_rr: I had a 50:50 8-] [20:43] <omega_rr> ChiefHighwater: of what? [20:43] <ChiefHighwater> omega_rr:getting the question right, what else [20:43] <wtay> omega_rr: there should be an event to control this buffer size at runtime (say mpeg2parse knows about the buffer size) [20:43] matth_ (ma...@bo...) joined #gstreamer. [20:43] <wtay> yo [20:44] <taaz> time? like difference between first and last buffer timestamps? that would be nice [20:44] <omega_rr> taaz: yup [20:44] <omega_rr> sizing queues in time was a major, major feature at OGI [20:44] <wtay> omega_rr: I would say args for now, yes [20:45] <wtay> or we make the queue *realy* core and make methods to control it.. [20:47] <wtay> omega_rr: an enum arg for buffer size type (bytes, time, count) and an arg for size [20:48] <taaz> any reason not to support all types at once? [20:48] <wtay> sizeing in buffers is kinda lame [20:49] Nick change: omega_rr -> omega_phone [20:49] <wtay> taaz: how would that go? [20:49] <wtay> MIN (size, time, count) or something [20:50] <taaz> wtay: multiple limits? if count > maxcount then drop; if size > maxsize then drop; if time > maxtime then drop; [20:50] <taaz> just seems it wouldn't be too hard to do all at once [20:51] <wtay> looks like overkill [20:51] <taaz> so? ;) [20:51] <wtay> dunno :-) [20:53] <taaz> let's say you're working in an embedded system like, oh, a phone ;) and have limited space for the queue. so you want to limit the max size. but also maybe want to do basic QoS via the max time limit too... having 2 queues seems silly since logic could be all put in one queue [20:53] Nick change: omega_phone -> omega_rr [20:53] <omega_rr> QWorst is mangling the local CO in pdx, most DSLs in the area have been on&off for the last 24++hrs [20:54] <omega_rr> I had to make an eBay bid for my Dad from here because of that... [20:54] <wtay> pff [20:54] <omega_rr> taaz: right, the idea is to MIN() them somehow [20:54] Action: wtay wonders why QWorst aren't out of business yet [20:54] <Parapraxis> are the *_save_thyself functions passed their own (externally created) XML node or are they give their parent node? [20:55] <omega_rr> basically say: if ((size_buffers && (level_buffers >= size_buffer)) || (size_bytes && (level_bytes >= size_bytes))) .... [20:55] <wtay> Parapraxis: parent node [20:55] <wtay> I thin [20:55] <wtay> k [20:56] <Parapraxis> k, thanks [20:56] <wtay> * @parent: the xml parent node [20:56] <Parapraxis> is there an api ref? [20:57] <Parapraxis> oh, nm [20:57] <Parapraxis> I found it [20:58] Action: taaz ponders such limited yet quick & dirty design... would be cooler to do gst_queue_add_policy() and pass in a GstQueuePolicy object that has a method like gboolean gst_queue_policy_validity_check(GstQueue*).. and have checkers for size, time, etc... OOP design overkill is fun ;) [20:59] <omega_rr> well, I do want to be able to specialize the chain/get functions eventually [20:59] <omega_rr> so the excess code isn't sitting there when it's not used (which will be the common case) [20:59] <omega_rr> we can add a custom API to the queue, but I'd much rather not [21:00] <omega_rr> the reason is that we may have different varieties of 'default' queues depending on what threading model, etc. [21:08] <omega_rr> committed [21:09] <taaz> btw, should the IS_COTHREAD_STOPPING be renamed? seems like plugins should be unaware that they happen to be running in cothreads [21:19] <wtay> hmm great ximian packages in stable are more recent than those in unstable... [21:20] <dobey> haha [21:20] <dobey> cool [21:20] <Parapraxis> anyone here understand the difference between g_type_register_static and g_type_register_dynamic? [21:21] Action: taaz probably would have used "lossy" instead of "leaky", not that it matters [21:21] <wtay> Parapraxis: not me.. [21:21] <Parapraxis> I don't understand what libs should use (especially if they are linked to plugins) [21:21] <taaz> Parapraxis: dynamic types can be changed, static cant. ;) I have no idea... [21:24] <taaz> omega_rr: maybe want to use typedef on that leaky enum. helps for binding automatic type scanning [21:28] <wtay> taaz: g(tk)object introspection will handle that fine [21:29] <taaz> ok, then it helps for compile type type checking ;) [21:30] <omega_rr> ok, all hands! <g> [21:31] <omega_rr> I need to do a survey of INCSCHED1 programs, to determine which ones work, which don't, then figure out why those don't work [21:31] <omega_rr> there are two options: they're coded wrong, or they're coded right and the core is failing [21:31] <omega_rr> who has a built (or nearly built) copy of INCSCHED1? [21:31] Action: taaz runs away ;) [21:32] <omega_rr> taaz: if you have it, you just have to pick some programs and run them, see if they work [21:32] <omega_rr> you don't have to try to figure out why they fail, if they do <g> [21:32] <omega_rr> taaz: I do have a gtkarg built for the leaky arg [21:32] <wtay> I have [21:33] <omega_rr> and yes, COTHREAD_STOPPING should be renamed, to what I have no idea [21:33] <taaz> well, i'm not at home now... what can i test remotely? [21:33] <omega_rr> taaz: you have a copy of incsched? [21:33] <taaz> as of a few days ago... [21:34] Action: taaz updates & rebuilds [21:35] <omega_rr> basically, need a) decide which programs should work right now at all, b) which of those fail, c) whether they fail because of core bugs, and d) what those bugs are [21:36] <taaz> oh jeez... autoconf 2.50 errors poping up now :( [21:36] <omega_rr> mu [21:36] <omega_rr> this is why /me doesn't upgrade to the latest&greatest [21:36] <taaz> wimp ;) [21:37] <omega_rr> I want to keep getting work done [21:37] <taaz> yeah but new tools are better... get rid of acconfig.h for one thing [21:38] <omega_rr> better & works is good, better & fails isn't ;-) [21:42] <omega_rr> so, who's taking which programs? <g> [21:43] <wtay> omega_rr: is the purpose testing wether the test[s]/ dir work? [21:43] <omega_rr> any program in incsched cvs that's "supposed" to work [21:43] <wtay> or should we try to create a pipeline that breaks things when it shouldn't? [21:43] <omega_rr> that too [21:44] <omega_rr> so someone should try to make -launch pipelines that break [21:44] <omega_rr> even better: make a list of all the pipeline in the wiki somewhere, for future regression testing [21:44] <omega_rr> (finally) <g> [21:44] <taaz> ok, i dunno what is wrong and my connection to home machine is crap right now. i'll help out in about 5-6 hours [21:44] <omega_rr> ok [21:44] <omega_rr> anyone else around with incsched built? [21:44] <wtay> ok [21:45] <wtay> omega_rr: are your latest changes in CVS (mad, cothread stuff?) [21:45] <omega_rr> checking [21:46] <wtay> hint: they aren't <g> [21:46] <omega_rr> nope [21:46] <omega_rr> ok, will commit shortly [21:46] Nick change: ajbusy -> aj_uni [21:47] <omega_rr> btw, anyone seen zaheer? [21:47] <wtay> no, it's been a long time since I've seen him around here [21:47] <omega_rr> ok, I'll have to email him [21:49] <taaz> new NGPT out [21:50] <omega_rr> beta 2? [21:50] <taaz> y [21:50] <omega_rr> a week late, like beta 1 [21:50] <omega_rr> not bad, for a big company [21:51] <omega_rr> ok, update INCSCHED1 and build [21:55] <omega_rr> I'll focus on examples/autoplug if you want to start with -launch scenarios [21:56] <wtay> mpeg2parseX where X == 2,3, X == 4 has a huge mem leak [21:56] <omega_rr> oh? [21:56] <omega_rr> you're putting this in an email or something, I hope? <g> [21:57] <wtay> no, I'm opening up a new blank file in vi :-) [21:57] <omega_rr> ok [21:57] <omega_rr> same thing ;-) [21:57] <Parapraxis> are any of the save_thyself's implimented yet? [21:57] <wtay> Parapraxis: all of them [21:57] <omega_rr> only if an element has custom state that it has to save (that isn't already represented by the gtkargs) does it need to implement it [21:58] <Parapraxis> ah, so where could I find one? [21:58] <omega_rr> I don't think there are any <g> [21:58] <Parapraxis> lol [21:58] <omega_rr> you can -inspect plugins to see what their save_thyself function is [21:59] <omega_rr> if it's gst_element_save_thyself, ignore it. otherwise, it's got its own [21:59] <Parapraxis> I was just wondering why the load_thyself passes the parent node, instead of the element's node [21:59] <omega_rr> ask wtay <g> [21:59] <Parapraxis> wtay: ? [21:59] <dobey> speaking of wtay [21:59] <dobey> ... [22:00] <wtay> Parapraxis: cause then it can append something to the parent node... [22:00] <omega_rr> ./gstreamer-inspect | cut -d: -f2 | xargs -n 1 ./gstreamer-inspect 2>&1 | grep save_thyself [22:00] <omega_rr> load_thyself or save_thyself? [22:01] <omega_rr> grrr suddenly autoplug isn't working for what it handled yesterday [22:03] <Parapraxis> wtay: even in the load_thyself? [22:04] <wtay> Parapraxis: well, maybe it could be changed.. I dunno [22:05] Action: omega_rr dubs his computer: "Parade of the Chipmunks", or "What happens when rate&channels mismatch a lot" [22:05] <wtay> lol [22:06] <omega_rr> esdsink needs help [22:06] <wtay> omega_rr: fixed in HEAD [22:06] <omega_rr> ok [22:07] <omega_rr> um, why does thx.mp2 (a mpeg2 video stream) get played via mad?? [22:08] <wtay> maybe because it's mp3 inside [22:08] <omega_rr> no, mpeg2 video [22:08] <omega_rr> elementary [22:08] <wtay> oops [22:08] <omega_rr> thx.mp2 + thx.ac3 = thx.vob [22:09] <omega_rr> we really need to find a solution to -launch creating windows for xvideosink [22:09] <omega_rr> I've had zero luck getting anything to show [22:09] <wtay> sdlvideosink works [22:09] <omega_rr> ok [22:09] <omega_rr> I'll hack autoplug.c to use that [22:09] <wtay> opens a window in READY state [22:09] <omega_rr> er, I mean use it with -launch [22:10] <omega_rr> that's not in INCSCHED though ;-( [22:10] Action: omega_rr hacks it in temporarily [22:11] <omega_rr> wow, that was easy [22:11] <omega_rr> cp -rav gst/plugins/sdlsink gst.incsched1/plugins;cd gst.incsched/plugins/sdlsink;make c... [truncated message content] |
From: <wim...@ch...> - 2001-05-25 04:30:19
|
[06:43] Nick change: ajmitch -> ajbusy [07:16] <Parapraxis> sleep time, bye y'all [07:16] Parapraxis (Sp...@us...) left irc: Client Exiting [07:57] Nick change: taaz -> taazzzz [08:21] taazzzz (dlehn@66.37.66.32) left irc: Ping timeout for taazzzz[66.37.66.32] [08:24] taazzzz (dlehn@66.37.66.32) joined #gstreamer. [09:22] ChiefHighwater (pa...@su...) joined #gstreamer. [09:28] arik (ar...@sd...) joined #gstreamer. [09:28] <arik> anyone in here use gphoto2? [09:33] arik (ar...@sd...) left #gstreamer. [09:40] steveb (st...@no...) joined #gstreamer. [10:13] steveb (st...@no...) left irc: [10:16] arik (ar...@sd...) joined #gstreamer. [10:16] <arik> anyone around? [10:34] <arik> aaaaaaarrrrrrrrggggggghhhhhhhh! [10:44] <ChiefHighwater> I'm here, but not much help [10:44] <arik> fucking ay [10:44] <arik> i need a usb expert ;-) [10:45] <arik> or a gphoto expert [10:45] <arik> or both [10:45] <ChiefHighwater> hehe...nope, sorry [10:46] <arik> so frustrated [10:46] <ChiefHighwater> i can tell [10:48] <ChiefHighwater> my current USB woes...I have DSL to my home using an Intel 3200 USB DSL device. My machine, runnnig win2k, shares the irq with USB and AGP. Other device don't care, but my dsl won't work cuz Intel's driver is busted so I have to bring my laptop hom from work to get online 8-[ [10:49] <arik> hehe [10:49] <arik> eek [10:51] <ChiefHighwater> I really want to put it in my linux box to masq all my other machines...good luck on that one [10:51] <arik> heh [10:51] <arik> thanks [10:51] <arik> all [10:51] <arik> i want is to figure out why gphoto2 can't write to my usb port [10:51] <arik> it knows it's there [10:51] <arik> it can query from it abit [10:51] <arik> but then i get weird errors [10:52] <ChiefHighwater> sorry, I'm not a programmer, I'm a Church Music Pastor [10:52] <arik> ;-) [10:52] <ChiefHighwater> I ask omega_ to fix my computers [10:53] <arik> hehe [10:53] <ChiefHighwater> it's not that bad, but I don't know squat next to him [10:54] <ChiefHighwater> well, I just changed something on this win98 box...so it needs to reboot. sorry I couldn't help [10:54] <arik> s'ok [10:54] ChiefHighwater (pa...@su...) left irc: Read error to ChiefHighwater[sub18-226.member.dsl-only.net]: Connection reset by peer [11:00] ajbusy (aj...@p4...) left irc: Ping timeout for ajbusy[p47-max4.dun.ihug.co.nz] [11:20] ajbusy (aj...@p2...) joined #gstreamer. [11:20] <arik> do you know anything about usb aj? [11:41] arik (ar...@sd...) left irc: leaving [11:58] Nick change: ajbusy -> ajmitch [12:27] <ajmitch> hmm, wassup ppl? [12:49] Nick change: wtay-sleep -> wtay [12:49] <wtay> yo [12:49] Action: wtay is learning RTP [12:52] <ajmitch> hey wtay [12:53] <ajmitch> i think i need to apply your patch to allow stuff to work with autoconf 2.50 [12:53] <wtay> yes [12:54] <ajmitch> i'm just learning what to put into a plugin before i can implement these filters ;) [12:54] <wtay> examples/plugins/example.c has lots of info [12:54] <ajmitch> how can i have it so that the filter type can be selected by an argument? [12:54] <ajmitch> ooh, ok ;) [12:55] <ajmitch> because there are four different filtering algorithms that will go into this plugin - make them separate elements? [12:55] <wtay> yes [12:56] <wtay> unless they are sufficiently similar and then you can use an enum to select the type [12:56] <ajmitch> one is a simple mean filter, the others are simple, conditional, and double median filters [12:56] <ajmitch> they are quite similar, i think [12:57] <wtay> make an enum then, like the median filter does with the filter size [12:58] <ajmitch> ok then [12:58] BBB-zZz (BB...@uc...) left irc: Ping timeout for BBB-zZz[ucu-105-116.ucu.uu.nl] [12:59] BBB-zZz (BB...@uc...) joined #gstreamer. [13:00] <ajmitch> i think i'll just implement the algorithms one at a time [13:00] <ajmitch> i'm not exactly awake so it may take awhile for stuff to sink in ;) [13:02] <ajmitch> partly because i'm not sure of how to handle stereo data [13:05] <ajmitch> the example plugin has no caps nego stuff? ;) [13:07] <wtay> nope [13:08] <wtay> anyway, chances are slim you'll need capsnego [13:09] <ajmitch> yeah [13:09] <ajmitch> so it's not mandatory for plugins that can accept basically any data? [13:10] <wtay> ajmitch: you sepcify what you can accept in the padtemplate, then you deal with it, if you need some specific value, like the number of channels, you can get that from the caps [13:10] <ajmitch> ah, ok [13:10] <wtay> the padtemplate will make sure your plugin is not fed with video data for example [13:11] <ajmitch> feel free to hit me with big cluestick at anytime ;) [13:11] Action: wtay takes the cluestick, just in case [13:11] <ajmitch> ;) [13:13] <ajmitch> i noticed the docs were a little sparse in places [13:13] <wtay> yes, I might update them today... [13:15] <ajmitch> oh, the power that nagging has... ;) [13:15] <wtay> right, then I wont update them today <g> [13:16] <ajmitch> ah well, prepare to be bombarded with questions then ;) [13:19] <ajmitch> a commit, i see a commit! ;) [13:19] <ajmitch> curses, it's not docs ;) [13:22] thomas_ (th...@ad...) joined #gstreamer. [13:22] <thomas_> hi all [13:22] <thomas_> omega's still watching star trek ? ;) [13:25] <ajmitch> wtay: hmm, how is this patch meant to be applied? [13:25] wtay (wi...@ca...) left irc: Ping timeout for wtay[cable-195-162-214-58.upc.chello.be] [13:25] <ajmitch> hmm, what an answer ;) [13:26] <ajmitch> how are you thomas_? ;) [13:26] <thomas_> ajmitch: heh, fine [13:26] <thomas_> I'm trying to get the tv-out to work on my tv on my ati rage pro [13:26] <thomas_> I would have thought it would be easier ! [13:26] <thomas_> it seems there aren't too many linux cards that handle tv out well [13:27] sienap (sy...@ip...) joined #gstreamer. [13:27] <thomas_> ajmitch: and you ? [13:27] <sienap> hmmm sun has gnome for solaris now great stuff.. too bad that they act like they made gnome them selfs.. [13:27] <thomas_> hi sienap [13:27] <ajmitch> thomas_: fine, playing with making plugins [13:27] <sienap> hi :) [13:27] <thomas_> ajmitch: which one ? [13:27] wtay (wi...@ca...) joined #gstreamer. [13:27] <wtay> whoa [13:27] <thomas_> ajmitch: if you learn stuff while doing, add them to the wiki ;) [13:27] <sienap> wtay! [13:27] <ajmitch> wtay: i think the patch wasn't applying because of evolution line wrapping ;) [13:27] <thomas_> sienap: do you know anything about tv out under linux ? [13:27] <wtay> yo [13:28] <ajmitch> hey sienap [13:28] <ajmitch> thomas_: add stuff to the wiki? ;) [13:28] <thomas_> ajmitch: that way I can learn from what you run into ;) [13:28] <sienap> thomas nope sorry.. [13:28] <thomas_> ajmitch: I added some stuff yesterday, hope others will too [13:28] <thomas_> wtay: what exactly is the gsm plugin going to do ? [13:28] <wtay> ajmitch: a correct configure.in is now in CVS, you'll need to rename it to though [13:29] <wtay> thomas: encode audio to/from gsm [13:29] <ajmitch> wtay: heh, ok [13:29] <ajmitch> thomas_:, ok, so if i run into anything nonobvious i should add it so that the masses can learn from me being a guinea pig? ;) [13:30] <thomas_> ajmitch: you can put a spin on it that makes you look good ;) [13:30] <thomas_> ajmitch: I didn't put up all of the stupid things I did while learning [13:30] <thomas_> ajmitch: that's why there isn't much stuff in the wiki yet, hehe [13:31] <ajmitch> hehe [13:31] <thomas_> ajmitch: but other than that, yeah, be a pig [13:31] <thomas_> wtay: ok, that was helpful <grin> [13:31] <thomas_> wtay: I'll put it this way: can I call gstreamer and ask it to play an mpeg through aalib ? [13:31] <ajmitch> http://gstreamer.net/wiki/CapsNegotiation looks helpful [13:32] <wtay> thomas_: yes <g> [13:32] <wtay> ./gstreamer-launch disksrc location=... ! mpeg1parse video_00! queue ! { mp1videoparse ! mpeg_play ! aasink } [13:33] Action: ajmitch sees if it crashes like usual [13:34] <thomas_> wtay: so where's the gsm in that ? [13:34] <wtay> thomas_: gsm? in mpeg? [13:35] Action: ajmitch listens to gstreamer-launch sinesrc freq=16000 ! osssink [13:36] sienap (sy...@ip...) left irc: sienap has no reason [13:38] <thomas_> wtay: I was asking what I can do with the gsm plugin ;) so the question was : can I call with my mobile to my gstreamer server component through the gsmsrc and ask it to play something ? [13:38] <wtay> thomas_: not yet :) [13:39] <wtay> ./gstreamer-launch disksrc location=/opt/data/mp3/files.cgi ! mad ! stereo2mono ! gsmenc ! gsmdec ! osssink [13:39] <ajmitch> config.status: error: cannot find input file: plugins/gsm/Makefile.in [13:39] <thomas_> wtay: I have better ways to make it sound like shit ;) [13:39] <wtay> ajmitch: cvs update -dP [13:39] <thomas_> wtay: so that works already ? [13:39] <wtay> thomas_: yup [13:39] <thomas_> wtay: I'll test it then [13:39] <ajmitch> wtay: -dP? [13:40] <ajmitch> wtay: already got those flags set in ~/.cvsrc [13:40] <wtay> ajmitch: I just commited the plugin [13:43] <ajmitch> ok, updating again... [13:57] arik (ar...@sd...) joined #gstreamer. [13:57] <arik> ajmitch: thanks [13:57] <arik> it worked ;- [13:57] <arik> ) [13:57] <ajmitch> arik: cool ;) [13:57] <ajmitch> wtay: did you bother to commit your plugins/Makefile.am changes? [13:57] <arik> hehe [13:57] <arik> so [13:57] <arik> my isp seems to be blocking port 25 [13:58] <arik> which means no smtp for me [13:58] <ajmitch> oh? [13:58] <ajmitch> that's bad... [13:58] <arik> at least that's what nmap seems to indicate [13:58] <arik> yes [13:58] <ajmitch> i like running an open mail server here ;) [13:58] <arik> makes no sense [13:58] <arik> i should try something [13:58] Action: ajmitch watches the number of connects to port 25 suddenly increase [13:58] <arik> hehe [13:58] <wtay> ajmitch: I haven't done any Makefile.am changes 'cause configure.in is toast [13:59] <arik> it's very annoying [13:59] <arik> i want to be able to send mail [13:59] <ajmitch> wtay: oh? how can i hope to compile the gsm plugin then? [14:00] <ajmitch> arik: oh, you can send mail just fine ;) [14:00] <wtay> ajmitch: yup, manually cd into it and build [14:00] <arik> i can? [14:00] <ajmitch> arik: a local mailserver is only required if you want to receive mail locally ( you need a domain & an MX record) [14:01] <arik> i think i'm confused [14:01] <ajmitch> arik: yep, there's a smarthost option in mailservers, which means that they just forward the mail to your isp [14:01] <arik> ok [14:01] <arik> how do i setup sendmail to do that? [14:01] <ajmitch> arik: *or* you could be tricky and use iptables to redirect connections to port 25 on your machine to port 25 at the ISP ;) [14:01] <wtay> my ISP blocks *ALL* ports [14:01] <wtay> incomming that is [14:01] <ajmitch> wtay: ouch, my ISP blocks none [14:02] <arik> heh [14:02] <arik> ok [14:02] <ajmitch> arik: hmm, i don't use sendmail, but i think the option is smarthost or something in sendmail.cf [14:02] <arik> what i really want to do is just have gnus send it [14:02] <arik> but i don't know how ;-) [14:03] <ajmitch> gnus? the mailreader? [14:03] <arik> es [14:03] <arik> er yes [14:03] <ajmitch> mail, or news? [14:03] <arik> mail [14:03] <ajmitch> hmm, ok, i should get gnus & look at it [14:03] <arik> ;-) [14:03] <arik> it's quite nice [14:04] <ajmitch> damn, 1.4MB to download ;) [14:04] <arik> hehe [14:05] <ajmitch> yay, compile breaks ;) [14:05] <arik> haha [14:05] <arik> 2.4 [14:05] <arik> so sweet [14:05] <arik> usb that actually works [14:06] <ajmitch> arik: and sound? [14:07] <arik> yes [14:07] <arik> sounds works too ;-) [14:07] <arik> oh [14:07] <arik> well [14:07] <arik> it did till now [14:08] <arik> do i need to recompile alsa or something? [14:08] <ajmitch> heh [14:09] <ajmitch> you rebuilt alsa drivers? [14:09] <ajmitch> lemme see, they are kernel modules... [14:09] <ajmitch> would you need to recompile 2.2 drivers for 2.4? ;) [14:09] <arik> oh [14:09] <arik> yes [14:09] <arik> ok [14:09] <arik> well [14:09] <arik> compile time [14:10] <ajmitch> hehe [14:10] Action: ajmitch is fixing up the major screwups he's making in his plugin atm [14:11] <arik> heh [14:11] <arik> i haven't coded on gstreamer in days [14:11] <arik> was to busy with mozilla 0.9.1 [14:13] <ajmitch> how's that going? [14:13] <arik> it's done [14:13] <arik> well [14:13] <arik> it froze anyway [14:14] <ajmitch> ok, how to use gnus? ;) [14:14] <arik> hehe [14:14] <arik> you make a .gnus file with lots of interesting things in it ;-P [14:14] <arik> with a few important bits [14:14] <ajmitch> ok, it's an emacs thing, right? ;) [14:15] <arik> yes [14:15] Action: ajmitch now understands why it was a 1.4MB download ;) [14:15] <arik> M-X gnus [14:15] <arik> hehe [14:15] <ajmitch> that means i need to use emacs? [14:16] <ajmitch> ah well, i was already using it... ;) [14:16] <arik> depmod: *** Unresolved symbols in /lib/modules/2.2.16-22/misc/snd-pcm.o [14:16] <arik> not good [14:16] <arik> not good at all [14:16] <arik> configure [14:17] <arik> thinks that it is still 2.2.16 for some reason [14:17] <arik> and i have no idea why [14:18] <ajmitch> rm config.cache [14:18] <arik> i did [14:18] <ajmitch> you installed kernel headers? [14:18] <arik> yes [14:18] <ajmitch> ran make clean? [14:18] <arik> no [14:18] <arik> ;-) [14:18] <arik> hehe [14:19] <ajmitch> reconfigured? ;) [14:19] <arik> yes [14:19] <arik> and configure still says 2.2.16 [14:19] <arik> checking for kernel version... 2.2.16-22 [14:19] <ajmitch> hmm, where are the kernel headers located? [14:20] <arik> /usr/include/linux [14:20] <ajmitch> hmm, usually they're under /usr/src/linux [14:20] <arik> um... [14:20] <arik> are you sure [14:20] <arik> ? [14:21] <ajmitch> /usr/include/linux is often glibc headers (although not the case with a lot of distros) [14:21] <arik> oh [14:21] <arik> i know why [14:21] <arik> cause i still have kernel-source and kernel-utils from 2.2.16 installed [14:21] <ajmitch> well, /usr/include/linux is often linked to /usr/src/linux/include ;) [14:21] <arik> heh [14:22] <arik> yeah [14:22] <arik> it is on my system as well ;-P [14:22] <ajmitch> take a look at /usr/include/linux/version.h [14:22] <arik> 2.4 [14:23] <arik> hmm [14:23] <arik> failed (probably missing /usr/src/linux/include/linux/version.h) [14:25] <ajmitch> hmm [14:26] <arik> i don't want to dl kernel-source [14:27] <arik> i tried making a link [14:28] <arik> from /usr/src/linux to /usr [14:28] <arik> but that didn't work [14:28] <arik> even though then the file was there [14:28] <arik> it still failed [14:28] <thomas_> arik: did you try the three options I gave you too for your mail problem ? [14:28] <arik> no not yet [14:28] <thomas_> arik: try them first, they might work ;) [14:28] <arik> been to busywith this crap [14:28] <arik> hehe [14:29] <thomas_> the option ajmitch is talking about is exactly what we tried yesterday [14:29] <arik> well this has nothing to do with that and i forgot ;-) [14:29] <arik> what? [14:29] <arik> oh [14:29] <arik> yeah but at the moment i am trying to get alsa built [14:29] <thomas_> the smart host, that's the DS option, it lets you relay mail through you iSP server [14:29] <arik> right [14:29] <thomas_> arik, well you can do that at the same time, it's multitasking baby ;) [14:29] <arik> hehe [14:29] <arik> ok [14:29] <arik> what were your ideas? [14:29] <ajmitch> thomas_ of course i prefer iptables & transparent proxying ;) [14:29] <arik> sorry ;-( was very tired lat night [14:30] <thomas_> arik: no problem [14:30] <thomas_> arik: in the DS line, you now have something like DSmail.earthlink.net, right ? [14:30] <arik> right [14:30] <ajmitch> especially if you're on a network doing ip masq, you can setup transparant proxying for http, irc, etc [14:31] <thomas_> ajmitch: yeah, but I always do iptables config by chance ;) [14:31] <thomas_> arik: you should try these three instead : [14:31] <thomas_> DSmailer:mail.earthlink.net [14:31] <thomas_> DSsmtp:mail.earthlink.net [14:31] <ajmitch> thomas_: i'm just helping someone with an iptables problem by ICQ now ;) [14:31] <thomas_> DSesmtp:mail.earthlink.net [14:31] <arik> ok [14:31] <thomas_> put in the first, save, restart sendmail, send mail, watch for the relay line, watch if it gets delivered [14:32] <arik> ajmitch: any idea about the kernel version thing? do i need to dl 21mb of kernel source? ;-( [14:32] <thomas_> I'm not sure they might work but a lot of people suggest doin that [14:32] <arik> right [14:32] <arik> what was your address [14:32] <arik> ? [14:32] <thomas_> anyone tried mplayer ? is it *supposed* to give a black screen and have the audio playing ? [14:32] <thomas_> arik: th...@ur... [14:32] <arik> heh [14:32] <thomas_> though any outbound address should do [14:33] <thomas_> wtay: do we have wireless network running in belgium somewhere ? [14:33] <thomas_> now that my laptop has gstreamer, it would be nice to work in the park ;) [14:33] <arik> buildaddr: unknown mailer mailer: Error -30991 [14:33] <arik> so i'm guessing that failed [14:33] <thomas_> arik: can you post the line before and after ? [14:34] <arik> that was the only line [14:34] <wtay> thomas_: I doubt it :) [14:34] <arik> May 24 05:37:04 localhost sendmail[9481]: f4OCb0e09481: from=root, size=48, class=0, nrcpts=1, msgid=<200105241237.f4OCb0e09481@localhost.localdomain>, relay=root@localhost [14:34] <arik> this is with smtp: [14:35] <thomas_> wtay: mplayer, you mean ? yeah, well I don't have Xv running, I use X11 output [14:35] <thomas_> black screen [14:35] <thomas_> same thing with DGA output [14:36] <arik> i _really_ don't want to dl the kernel-source rpm [14:36] <arik> May 24 05:39:14 localhost sendmail[9483]: f4OCb0e09 [14:36] <arik> be, ctladdr=root (0/0), delay=00:02:14, xdelay=00:0 [14:36] <arik> relay=xwing.aoltw.net. [204.29.187.150], dsn=4.0.0 [14:36] <arik> imed out with xwing.aoltw.net. [14:37] <ajmitch> arik: ok, so what do you have installed? [14:37] <arik> ajmitch: everything but that ;-) [14:37] <thomas_> ok, gl works [14:37] <ajmitch> the kernel headers should be enough, i think ;) [14:37] <thomas_> that's a relief [14:38] <ajmitch> it's been awhile since i used any premade kernel packages ;) [14:38] <arik> hmm [14:38] <arik> heh [14:38] <arik> it doesn't work atm [14:38] <arik> that's all i know [14:38] <ajmitch> on debian, people generally make their own packages using make-kpkg ;) [14:38] <arik> cause configure isn't finding version.h [14:38] <arik> rigt [14:38] <arik> er right [14:38] <ajmitch> all the symlinks are nicely setup? ;) [14:38] <arik> i have built my own in the past [14:38] <arik> hmm [14:39] <arik> no [14:39] <arik> i don't think they are [14:39] <arik> and i tried to make them myself but i guess i failed [14:40] <arik> i have /usr/src/linux-2.2.16/include [14:40] <arik> but that's all [14:42] <arik> ok [14:42] <arik> stmp: and esmtp: did nothing [14:42] <arik> and mailer: caused an error message [14:43] <arik> oh well [14:45] sienap (sy...@ip...) joined #gstreamer. [14:45] <sienap> Hi ppl [14:46] <arik> hey\ [14:46] <sienap> hej arik! [14:46] <sienap> how is gstmediaplay going ? [14:46] <sienap> 4 weeks to go huh >:) [14:46] <arik> hej [14:46] <arik> heh [14:46] <arik> it has had to wait for mozilla 0.9.1 [14:46] <arik> 4 weeks? [14:46] <sienap> 4 weeks and 0.2.0 will be released [14:46] <sienap> or something in that direction [14:47] <sienap> so a better gstmediaplay would be nice *G* :) [14:47] <arik> heh [14:47] <arik> ok ;-) [14:47] <sienap> so arik gonna make that ? [14:47] <arik> yes [14:47] <arik> theoretically [14:47] <ajmitch> hi sienap ;) [14:47] <sienap> hej ajmitch [14:47] <sienap> arik he bah :) [14:47] <ajmitch> sienap: you volunteering to help arik out? [14:47] <sienap> ajmitch hmm nope :) [14:48] <arik> ;-) [14:48] <arik> oooh [14:48] <arik> volunteeer! [14:48] <sienap> ajmitch i will be doing stuff for paranormal (effects that is) [14:48] <ajmitch> sienap: you know you want to, and Uraeus wants you to... ;) [14:48] <sienap> arik i am not that good at coding anyway :) [14:48] <arik> hehe [14:48] <ajmitch> arik: victim ;) [14:48] <arik> heh [14:48] <arik> cheep excuse [14:48] <arik> yyes [14:48] <sienap> ajmitch Uraeus wants be to do some other favors for him.. [14:48] <arik> victim [14:49] <ajmitch> arik: 4 weeks from now i'll be sitting my last exam :( [14:49] <arik> eeek [14:49] <sienap> ARIK in 4 weeks you are able to make a little decent player :) [14:49] <arik> 4 weeks from now i'll be done working at netscape [14:49] <arik> that will be nice ;-) [14:49] <arik> hehe [14:49] <ajmitch> hehe [14:49] <ajmitch> where will you work? [14:50] <arik> nowhere [14:50] <sienap> he [14:50] <sienap> bleh [14:50] <sienap> :(( [14:50] <arik> gstreamer! [14:50] <arik> i will be back in school [14:50] <sienap> GSTREAMER! [14:50] <arik> heh [14:50] <arik> yes [14:50] <ajmitch> arik: ah, i see ;) [14:50] <ajmitch> arik: how old are you? [14:50] <arik> 20 [14:50] <arik> younger then you might have thought [14:50] <ajmitch> yeah ;) [14:51] Action: ajmitch is just a little kid, only 18 [14:51] <arik> *grin* [14:51] <arik> hehe [14:51] <ajmitch> ah well ;) [14:51] Action: sienap is 17 [14:51] <arik> soon [14:51] <sienap> also a little kid :) [14:51] <arik> soon we will all be older [14:51] <ajmitch> yeah ;) [14:51] <arik> heh [14:51] <arik> how old are you wtay? [14:51] <sienap> he aah shit happens [14:51] <sienap> 28 [14:51] <sienap> almost 29 [14:51] <ajmitch> well, i'm old enough to be at uni ;) [14:51] <arik> heh [14:52] <arik> i am at uni ;-P [14:52] <sienap> dus [14:52] <sienap> but ehm [14:52] <ajmitch> yeah, i guessed that ;) [14:52] <sienap> there are many items on the roadmap [14:52] <ajmitch> which one? [14:52] <sienap> that won't get it made on time.. [14:52] <ajmitch> arik: which uni? [14:52] <ajmitch> sienap: volunteer ;) [14:52] <arik> i'm glad i don't have a roadmap for gst [14:52] <arik> mp [14:52] <arik> ajmitch: um, antioch college [14:53] <sienap> arik but gstmediaplay / moz thingy ARE on the roadmap [14:53] <arik> hmm [14:53] Action: ajmitch hasn't heard of it [14:53] <arik> yes [14:53] <ajmitch> what's it like? [14:53] <arik> unfortunatly [14:53] <arik> liberal [14:53] <arik> very [14:53] <arik> but [14:53] <arik> for the next term i will be in sf [14:53] <arik> at an art school [14:54] <ajmitch> art school? [14:54] <sienap> art school? [14:54] <sienap> =] [14:55] <arik> yes [14:55] <arik> art school [14:55] <sienap> he [14:55] <arik> working! [14:55] <arik> 2.4.2 baby [14:55] <sienap> as long you work on gstreamer 6 ours a day we won't send nukes to you.. [14:56] <arik> heh [14:56] <ajmitch> arik: why 2.4.2? [14:56] <arik> at the moment [14:56] <arik> i barely have a free minute to look at it [14:56] <ajmitch> 2.4.2 has some bad bugs that could ruin your day ;) [14:56] <arik> really? [14:57] <ajmitch> well, like the nice loopback bug [14:57] Action: ajmitch just downloads patch files now [14:57] <sienap> ajmitch what kind of loopback bug ? [14:57] <arik> hmm [14:57] <ajmitch> sienap: try and mount an iso image, mounts hangs [14:57] <arik> i don't think alsa 0.5 likes the new kernel [14:57] <ajmitch> i've also heard of it resetting [14:58] <arik> maybe i should try alsa 0.8 [14:58] <arik> er 0.8 [14:58] <ajmitch> arik: nah, i'm running alsa 0.5.10 with 2.4.4 [14:58] <arik> 0/9 [14:58] <arik> hmm [14:58] <ajmitch> it ran fine with 2.4.2 [14:58] <arik> /usr/src/linux/include/linux/modversions.h:1:2: #error Modules should never use kernel-headers system headers, [14:58] <arik> usr/src/linux/include/linux/modversions.h:2:2: #error but headers from an appropriate kernel-source [14:58] <thomas_> right, I'm starting to hate ATI !!!! [14:58] Action: thomas_ subscribes to livid-gatos [14:58] <arik> card-dt0197h.c:92:60: warning: nothing can be pasted after this token [14:58] <ajmitch> arik: that's why i prefer to compile my own kernels ;) [14:59] <arik> thomas_: so nothing worked, oh well, thanks again ;-) [14:59] <arik> ajmitch: ;-P [14:59] <thomas_> and I hate people on deja who, when asked how to get tv-out working on xfree, answer how to get *only* the card running under *without* tvout [14:59] <sienap> http://www.linuxrising.com/files/gstreroadmap2.html << ROADMAP [14:59] <arik> heh [15:00] <arik> damnit [15:00] <arik> i finally had sound working [15:00] <arik> and now i have nothing [15:01] <ajmitch> what happens now? [15:02] <arik> stuff above [15:02] <arik> it craps on compiling [15:02] <arik> /usr/src/linux/include/linux/modversions.h:1:2: #error Modules should never use kernel-headers system headers, [15:02] <arik> /usr/src/linux/include/linux/modversions.h:2:2: #error but headers from an appropriate kernel-source [15:02] <arik> hmm [15:03] <arik> i bed i know why actually [15:03] <arik> er bet [15:03] <ajmitch> yeah? [15:03] <arik> well [15:03] <arik> maybe [15:04] <arik> somehow [15:04] <arik> i'm guessing i'm gonna need kernel-source in the end [15:04] <arik> damnit [15:05] <ajmitch> just go with the stock kernel & compile your own ;) [15:06] <arik> hah [15:06] <arik> thanks dude [15:06] <ajmitch> saves a lot of hassle, when a new release comes out just grab the patch & recompile ;) [15:06] <arik> heh [15:09] hadess (ha...@pc...) joined #gstreamer. [15:09] <ajmitch> hey hadess [15:09] <sienap> hadess! [15:09] <hadess> heya ajmitch [15:09] <hadess> hey sienap [15:10] <arik> hey hadess ;-) [15:10] <thomas_> arik: one more question: did you ever try to put an IP address after the DS directive ? [15:10] <arik> i think we did [15:10] <arik> yeah [15:10] <sienap> so arik... gstmediaplay improvents won't make it for 0.2.0 :( ? [15:10] <arik> heh [15:10] <arik> no they will [15:11] <arik> just not all of them ;-P [15:11] <sienap> which on them will ? :) [15:11] <hadess> hey arik, thomas_ [15:11] <thomas_> hi hadess [15:11] <arik> heh [15:11] <ajmitch> arik: you want helpers? [15:12] Action: ajmitch isn't volunteering, btw... ;) [15:12] <ajmitch> fscking exams ;) [15:12] <arik> ajmitch: yes i do ;-) [15:13] <sienap> arik what kind of simple issues do you have left? [15:13] <sienap> *SIMPLE* [15:13] <sienap> :) [15:13] <ajmitch> arik: talk nicely to hadess, maybe? ;) [15:13] <arik> hehe [15:13] <arik> hadess has he's own project [15:13] <sienap> i think we really should try to get all the roadmap items done >:) [15:13] <thomas_> arik: if you're still up to it, there's one thing left you can try and that's forcing it through the rulesets [15:13] <sienap> his :) [15:13] <thomas_> arik: you wanna ? ;) [15:14] <arik> if you want ;-) [15:14] <sienap> arik i maybe can do some simple things and gui thingies.. [15:14] <thomas_> arik: ok, first get the ip of your mailserver [15:14] <arik> ok [15:14] <arik> the earthlink one? [15:14] <thomas_> if you're sure it's mail.earthlink.net (if that was the one that had port 25 open) then that's the one [15:14] <sienap> so you have simple things left ? [15:14] <arik> i think that was it [15:14] <thomas_> if you're not sure try it again first [15:14] <arik> sienap: maybe, i will think about it ;-) [15:14] <arik> ok i have an ip [15:14] <hadess> arik: actually, i think i'll be able to give you a playlist algo if you do the gui for it [15:15] <sienap> arik ok [15:15] <arik> hadess: oooh, ok [15:15] <arik> hadess: that sounds good [15:15] <thomas_> ok, now look in sendmail.cf for a line that says: use unqualified mailer [15:15] <thomas_> arik: (you have a backup right ? not that it matters since it doesn't work) [15:15] <arik> yes i do [15:15] <thomas_> paste the second of these two lines so I can see [15:16] <arik> R< [ IPv6 $+ ] > $* $#relay $@ $(dequote $1 $) $: $2 use unqualified mailer [15:16] <arik> R< $+ > $* $#relay $@ $1 $: $2 use unqualified mailer [15:16] <thomas_> ok, in the second line you see $@ $1 $: $2, right [15:17] <arik> yes [15:17] <sienap> ok off nowe [15:17] <sienap> bye [15:17] <arik> bye [15:17] sienap (sy...@ip...) left irc: sienap has no reason [15:17] <thomas_> you should only change the $1, and you should change that to $[full ip address] [15:17] <thomas_> the right brackets are important ! [15:18] <arik> ok [15:18] <thomas_> damn, my xchat is screwed up... it holds on line in a buffer before showing it [15:18] <thomas_> [15:18] <thomas_> fill in the ip of your mail server [15:18] <arik> done [15:18] <thomas_> when you've changed that, restart sendmail, send mail, paste log output [15:19] <thomas_> btw: I checked, the first line with relay=root@localhost is normal [15:19] <arik> ok [15:19] <thomas_> it's the second line that we have to worry about [15:19] <arik> buildaddr: no user [15:19] <arik> is that cause i took out the masking? [15:20] Nick change: BBB-zZz -> BBB-zZz-[away] [15:20] Action: BBB-zZz-[away] is away: zon [15:21] <arik> ajmitch: i give up on alsa for now, i'll download kernel-source tomorrow prob [15:21] <thomas_> arik: possible; I would leave in DSmail.earthlink.net anyway [15:21] <thomas_> but put DS(ip address) instead [15:21] <thomas_> without the braces [15:21] <arik> where? [15:21] Action: ajmitch should either get back to plugin hacking or do homework due tomorrow (well, today, it's after 1am ;) ) [15:21] <thomas_> arik: where it was, or are you talking about something else ? [15:21] <arik> ok [15:21] dobey (do...@dr...) joined #gstreamer. [15:22] <thomas_> hey dobey [15:22] <arik> so replace DSmail.earthlink.net [15:22] <dobey> hey [15:22] <arik> with DS(ip) [15:22] <thomas_> arik: yeah [15:22] <ajmitch> hi dobey [15:22] <thomas_> no brackets, just DS190.x.y.z [15:22] <arik> ok [15:22] <thomas_> (I don't know what it is) [15:22] <dobey> hey aj dude yo [15:22] <dobey> oh by the way... [15:22] <arik> do ximian hired you? [15:22] <dobey> did i mention how much i rule? [15:22] <arik> congrats [15:22] <dobey> yes [15:22] <dobey> thanks [15:23] <thomas_> arik: btw could you just paste those two sendmail lines anyway ? [15:23] <arik> it's an interesting place to work [15:23] <arik> sure [15:23] <thomas_> dobey: woohoo ! nice going [15:23] <dobey> thomas: i've been here since January [15:23] <dobey> ;-P [15:23] <thomas_> dobey: ok, then what are we celebrating ? [15:23] <arik> May 24 06:22:31 localhost sendmail[16108]: f4ODMU116108: SYSERR(root): buildaddr: no user [15:23] <arik> May 24 06:22:31 localhost sendmail[16108]: f4ODMU116108: from=root, size=48, class=0, nrcpts=0, msgid=<200105241322.f4ODMU116108@localhost.localdomain>, relay=root@localhost [15:24] <dobey> thomas: i'm celebrating my ruling i guess [15:24] <thomas_> dobey: you rule ! ;) [15:24] <arik> a very interesting place [15:24] <thomas_> arik: ok that first line, we didn't have that yesterday did we ? [15:24] <thomas_> was that because of the DM directive ? [15:24] <dobey> thomas: since i did some ruling with encompass last night [15:24] <arik> yes [15:24] <hadess> j00 don't rUl3z j00 sUx0r [15:25] <dobey> hadess: uh [15:25] <dobey> hadess: -hi- [15:25] <thomas_> hadess: wrong list ;) [15:25] <hadess> dobey: i rule because i got multiple playlists loading [15:25] <arik> thomas_: i think so [15:25] <arik> thomas_: but that first line is bad [15:25] <dobey> hadess: i got neon working [15:25] <dobey> :-/ [15:26] <arik> thomas_: same thing happened when i replaced DSearth with DSip btw [15:26] <hadess> dobey: heh [15:26] <ajmitch> i wish colorgcc worked right :P [15:26] <arik> thomas_: i think that means it doesn't know my smtp user name [15:26] <thomas_> arik: yeah, but you should get that masquerading line out anyway [15:26] <dobey> ajmitch: use mandrake? [15:26] <thomas_> arik: you never specified an smtp user name [15:26] <hadess> dobey: when is s/gtk_html_/gtk_html2_/ ? [15:26] <arik> thomas_: which one? [15:26] <arik> thomas_: no i didn't [15:26] <thomas_> arik: the one with DM [15:26] <arik> ok [15:26] <dobey> hadess: when s/gtk/gtk2/ happens [15:26] <arik> i removed that i thikn [15:26] <thomas_> you should replace it with what it was originally [15:26] <thomas_> not just delete it [15:26] <dobey> ajmitch: like, mandrake 7.x? [15:27] <arik> it was orriginall nothing afaik [15:27] <hadess> dobey: and probably s/gnome/gnome2/ [15:27] <thomas_> if you're confused, just restart from your fresh sendmail.cf, add ip to DS, and change that one ruleset [15:27] <ajmitch> dobey: yeah, it's also in debian [15:27] <arik> thomas_: that is the way it is [15:27] <arik> thomas_: there is no DM thing in there [15:27] <ajmitch> ii colorgcc 1.3.2-4.1 Colorizer for GCC warning/error messages [15:27] <dobey> hadess: probably [15:28] <thomas_> arik: well, the line should be there, but nothing after DM [15:28] <dobey> ajmitch: oh, it seems ok in manrake 8 [15:28] <arik> ]yes [15:28] <dobey> ajmitch: aside from the fact that it is -required- [15:28] <arik> that is what i have [15:29] <dobey> and that other browser is ugly and sickly [15:29] <ajmitch> dobey: heh, yeah, it's one of the things i like about mandrake [15:30] <dobey> ajmitch: you -like- mandrake? [15:30] Zeenix (s_l...@ho...) joined #gstreamer. [15:30] <arik> i give up [15:30] <dobey> ajmitch: do you -like- crack as well? [15:30] <arik> sleep time [15:30] <dobey> ;-P [15:30] <thomas_> arik: sorry again ;) [15:30] <arik> talk to you all later [15:30] Action: thomas_ is taking a shower [15:30] <arik> thomas_: s'ok ;-) we'll get it eventually [15:31] <arik> or i'll stop using earthlinlk [15:31] <arik> when i leave my parents house// [15:31] <thomas_> arik: yeah, switch isp [15:31] <arik> to go back to school [15:31] <arik> later all [15:31] <ajmitch> dobey: no, i like some things about mandrake [15:31] <thomas_> arik: get a dialin at school/uni [15:31] <arik> yees [15:31] <arik> ;-) [15:31] arik (ar...@sd...) left #gstreamer. [15:31] <ajmitch> dobey: heh, it helps if i read the colorgcc docs... ;) [15:31] <dobey> thre are docs? [15:31] <dobey> i hate colorgcc [15:31] <dobey> it's a piece of shit [15:31] <ajmitch> dobey: yeah, the source ;) [15:32] <ajmitch> dobey: useful when i use small fonts in Eterm [15:32] <ajmitch> i just prefer to be able to see the errors easily at times [15:32] <dobey> ajmitch: yeah, well [15:32] <dobey> ajmitch: it fucks up stuff [15:32] <ajmitch> (apart from the fact that i can use make in emacs) [15:33] <ajmitch> holy crap, it really screws up running make in emacs ;) [15:33] Action: ajmitch disables colorgcc [15:33] <dobey> haha [15:34] <dobey> FUCK [15:34] <ajmitch> ok, you were right, it does screw things ;) [15:34] <dobey> the dns for one of the free pop servers i use is down :-( [15:35] <dobey> ajmitch: on mandrake 7 it causes stuff to use cpp when it should just be c, so, i had to rebuild guile and everything after :-/ [15:35] Zeenix (s_l...@ho...) left irc: Ping timeout for Zeenix[host-64.netzone.net.pk] [15:36] <ajmitch> dobey: oh, that is not good [15:36] <dobey> ajmitch: on mandrake 8, i couldn't build cpp stuff without it :-/ [15:36] Zeenix (s_l...@ho...) joined #gstreamer. [15:36] <ajmitch> dobey: cpp doesn't compile some looser (yet valid) c code the same [15:36] <Zeenix> sorry got disconnected for a min. [15:37] <dobey> ajmitch: but anyway, [15:37] <Zeenix> plz answer my question: [15:37] <dobey> encompass is going to ownz all [15:37] <Zeenix> is mp3 encoding & decoding done at the same speed ? [15:37] <ajmitch> Zeenix: no [15:38] <ajmitch> Zeenix: i can playback an mp3 at about 2.5x playback speed, but using some mp3 encoders it takes 20min to encode a 3min song ;) [15:38] <hadess> decoding is much faster [15:38] Action: ajmitch has a slow computer, btw [15:38] <hadess> i can encode at 1.5x on my g3 400 [15:39] <ajmitch> hadess: this was with bladeenc, 192kbit ;) [15:39] <ajmitch> i can get about 2.5x using gogo [15:39] <hadess> ajmitch: this was with lame VBR, 128 kpbs mini [15:39] <ajmitch> heh, ok [15:40] <big_T> it is a bit harder for the encoders, they have to do some serious signal-analysis using fourier-transforms and so on... [15:40] <hadess> i wish there were some g3 optimisations for lame... [15:40] <dobey> heh [15:40] <hadess> but some ppl have lame optimisations for g3 :P [15:41] <hadess> i need to get some drinks [15:41] Nick change: hadess -> hds-town [15:45] Nick change: wtay -> wtay-gone [15:54] <Zeenix> what is x in 1.5x ? [15:54] <ajmitch> yay, finally got this thing to compile, now to add the mean & median algorithms ;) [15:56] <ajmitch> just means 1.5 times faster than normal playback speed [15:57] <Zeenix> oh no [15:58] <Zeenix> not the playback [15:58] <ajmitch> what? [15:59] <Zeenix> just the row sound data [15:59] <Zeenix> what is encoding & decoding speeds at you computer in KB/s [16:00] <ajmitch> i don't know, it would depend on the file (whether VBR or not, what bitrate, stereo, etc) [16:00] <ajmitch> plus this computer is just crap ;) [16:01] <ajmitch> why are you asking? [16:03] <Zeenix> i need to do it in my voice chat app [16:03] <Zeenix> real time encoding at one end & decoding at the other [16:03] <ajmitch> hmm, ok [16:04] <Zeenix> i am using the lowest speed i could get from OSS( like 5KB/s ) [16:04] <ajmitch> what sort of sound quality are you looking at? [16:04] <Zeenix> but we use internet line thats 2 KB [16:04] <Zeenix> i mean 2KB/s [16:04] <dobey> ogg [16:05] <Zeenix> is ogg fast [16:06] <Zeenix> lowest quality sound [16:07] <ajmitch> i think ogg can go to lower bitrates [16:08] <Zeenix> how does it compress ? [16:08] <ajmitch> with bitrates that low, encoding speed won't be much of an issue [16:08] <Zeenix> how much does it compress ? [16:08] <ajmitch> www.xiph.org/ogg/vorbis for more info, i dunno much about it [16:14] <ajmitch> hmm, amazing, my plugin is still compiling after i've made changes ;) [16:14] Action: ajmitch thinks this shouldn't be too hard [16:15] <dobey> yeah, well, what i'm doing "shouldn't" be too hard either [16:17] <ajmitch> hehe [16:17] <ajmitch> encompass stuff? [16:17] <dobey> no [16:17] <dobey> mandrake 8 [16:17] <ajmitch> ah, ok ;) [16:17] <ajmitch> packages? [16:18] <dobey> fuck, i'm tired too [16:18] <ajmitch> only 2:20am here [16:18] <dobey> because it is dark outside and it's 10:30 AM [16:18] <Zeenix> www.xiph.org/ogg/vorbis says it goes to low bit rates only in theory yet [16:18] <dobey> split libraries suck. [16:19] <ajmitch> Zeenix: i guess it just hasn't been tested ;) [16:19] <dobey> fuck [16:20] <ajmitch> Zeenix: it looks like wtay-gone is vorbisenc maintainer, he *might* be able to help you with that, but afaik, he's working on a gsm plugin designed for voice data [16:22] Nick change: hds-town -> hadess [16:22] <ajmitch> hadess: what do you know of the gsm plugin? ;) [16:22] <dobey> ajmitch: maybe you should talk to cactus? [16:22] <Zeenix> is 128kbps = 16KB/s [16:23] <hadess> yep [16:23] <ajmitch> dobey: i suppose, i just remember wtay talking about the gsm plugin earlier ;) [16:23] <hadess> gsm uses 64kpbs voice data and compresses it [16:23] <dobey> ajmitch: i meant about bonobo [16:24] <ajmitch> dobey: ah, ok ;) [16:24] Action: ajmitch didn't know that dobey was referring to bonobo... [16:24] <Zeenix> but i dont need lowest sound quality( no stereo either ) to make it fast [16:24] <Zeenix> i meant the opposite [16:24] Nick change: wtay-gone -> wtay [16:24] <dobey> ajmitch: ;-) [16:24] <hadess> Zeenix: the main problem won't be the speed, it'll be the bandwidth [16:25] <Zeenix> i am too week at english da... [16:25] <ajmitch> dobey: i'm working on a plugin tonight ;) [16:25] <hadess> it's 15:25 and it's goddam hot outside, i love it [16:25] <Zeenix> yes if i talk of low speed that i would save a lot of time for encoding/decoding [16:25] <dobey> ajmitch: ? [16:25] <ajmitch> dobey: you wanting bonobo stuff done soon (within next 4 weeks?) [16:26] <dobey> ajmitch: no, not really [16:26] <dobey> ajmitch: i am just confused by bonobo-media [16:26] <hadess> Zeenix: no, you need something like the gsm plugin [16:26] <wtay> Zeenix: yes, an RTP/RTCP [16:26] <ajmitch> dobey: tonight, i'm playing with making a gstreamer plugin for noise reduction (algorithms shamelessly ripped from gramofile) [16:26] <wtay> s/an/and [16:27] <Zeenix> someone told me the other day that i can get a 56KB/s average encoding/decoding speed out of MP3 [16:27] <dobey> ajmitch: oh, ok, i thought you meant a bonobo thing you were working on ;-P [16:27] <ajmitch> dobey: no, that might have to wait until holidays or some time when i can look at how to do it right ;) [16:27] <wtay> Zeenix: yup, you can but 128Kbps is too high a bandwidth for most net connections [16:27] <ajmitch> holidays start in 4 weeks ;) (i get a month off) [16:28] <wtay> Zeenix: you'll need to lower the quality and then mp3 really sucks [16:28] <dobey> ajmitch: heh [16:28] <Zeenix> what should i use for the lowest quality sound [16:28] <Zeenix> then ? [16:29] <wtay> Zeenix: gsm IMO [16:29] <wtay> Zeenix: you can interface with other phone apps too then [16:29] <ajmitch> dobey: hopefully i'll have a job coding stuff for GNUe, or stuff around Dunedin, so i should be reasonably confident with C, at least ;) [16:29] <Zeenix> tell me that thing which can give me 56KB/s encoding/decoding speed atleast [16:30] <wtay> Zeenix: gsm can do that easily [16:30] <Zeenix> thats the thing i need than [16:30] <ajmitch> wtay: amazingly, this plugin skeleton is compiling, i just need to fit the algorithms into it, and tweak it a bit ;) [16:30] <wtay> Zeenix: anyway, that's my opinion, feel free to use mp3 if you want [16:31] <Zeenix> would it be able to cope with 5 KHhz, mono, unsigned 8bit data [16:31] <wtay> Zeenix: gsm is tailored for speech, mp3 isn't [16:31] <Zeenix> i am slowly being attracted to gsm [16:31] <dobey> ajmitch: hehe, it's cool, i just need to figure out what the hell is going on [16:32] <Zeenix> wtay : my question above ? [16:32] <ajmitch> dobey: coding it is usually easier than understanding how it should be designed & put together ;) [16:32] <dobey> ajmitch: yeah, i just don't know exactly what bonobo-media does [16:33] <ajmitch> dobey: talk to wtay then ;) [16:33] <wtay> Zeenix: gsm operates on XKhz mono 16 bits signed audio [16:33] <dobey> ajmitch: yeah, i did, i discovered i should talk to cactus [16:34] <ajmitch> dobey: ah, ok, have you seen him on irc? [16:34] <dobey> ajmitch: i just wondered if you would be the one, since i can't seem to be online the same time he is [16:34] <Zeenix> wtay: why not 8 bit [16:34] <wtay> Zeenix: have you looked at PreViking or gphone? [16:34] <wtay> Zeenix: actually 13 bit :) [16:34] <hadess> dobey: that's cactuz you want to talk to [16:34] <ajmitch> dobey: hmm, he seems to visit #gnome on irc.gnome.org (according to Xachbot) [16:35] <dobey> ajmitch: yes [16:35] <ajmitch> hadess: oh, ok ;) [16:35] <dobey> with an s [16:35] <hadess> hmm, actually, no ;) [16:35] <hadess> -XachBot- I last saw cactus (~ca...@is...) 1d 16h 3m ago [16:35] <Zeenix> that mean i'll get 10 KB/s not 5KB/s when using 5Khz data with mono [16:36] Action: dobey thinks hadess go s/z crossed from cactuS.ruleZ.org ;-P [16:36] <hadess> dobey: yup [16:36] <wtay> Zeenix: gsm takes 160 signed shots in and outputs 33 unsiged butes as output [16:36] <ajmitch> dobey: we shall have to lie in wait for him ;) [16:36] <wtay> s/butes/bytes [16:36] <dobey> heh [16:37] <wtay> s/shots/shorts [16:37] <wtay> doh [16:37] <ajmitch> 160 shots is a lot ;) [16:37] Action: hadess fires [16:37] Action: ajmitch imagines wtay after 160 shots... [16:37] <ajmitch> hadess: how about a shot of whisky? ;) [16:37] Action: wtay is glad he didn't type 'buts' [16:38] <hadess> wtay: s/buts/butts/ :) [16:38] <wtay> doh [16:38] <hadess> lol [16:38] <wtay> 160 shots and 33 butts then :-) [16:38] <ajmitch> hehe [16:39] <wtay> that's what g-SM does [16:39] <Zeenix> what the hell are you all at ? [16:39] <hadess> rotfl [16:39] <Zeenix> is GSM in that ogg [16:39] <dobey> GNU Sado Masochism? [16:39] <dobey> Free as in spanking? [16:39] <wtay> Zeenix: vorbis is something completely different [16:39] <wtay> lol [16:40] <Zeenix> i said ogg not vorbis [16:40] <hadess> dobey: hehe [16:41] <wtay> Zeenix: http://kbs.cs.tu-berlin.de/~jutta/toast.html [16:41] <wtay> Zeenix: ogg is a shell around vorbis [16:41] <wtay> ogg is a transport stream [16:41] <Zeenix> then what for GSM then [16:41] <hadess> Zeenix: gsm is what's used for mobile phones [16:41] <ajmitch> wtay: actually vorbis is a subproject of ogg, from www.xiph.org/ogg [16:42] <Zeenix> i meant which rpm it is from [16:42] <wtay> ajmitch: yup, 'cause they needed an equivalent for a system stream too [16:42] <ajmitch> currently there's ogg vorbis (useful), and ogg tarkin (planning stage) [16:42] <hadess> ogg is a wrapper around different formats, like quicktime or avi [16:42] <wtay> hadess: exactly [16:42] <dobey> ogg ogg [16:42] <wtay> Zeenix: rpmfind.net, look for libgsm [16:43] <ajmitch> hadess: so the codecs are like vorbis & tarkin, but the file/stream format is ogg? [16:43] <wtay> ajmitch: yes [16:43] <ajmitch> ah, that clears it up ;) [16:43] <Zeenix> is it a pluggin for gstreamer [16:43] <hadess> ajmitch: yes, like avi has mjpeg, indeo and all that... [16:43] <ajmitch> wtay: go ahead, hit me with the cluestick ;) [16:44] <wtay> ajmitch: why, you didn't say anything stupid yet :) [16:44] <ajmitch> wtay: hehe [16:44] Action: wtay pats ajmitch on the back [16:45] <Zeenix> wtay: then which one do you recemmend for a 1KB/s internet connectivity line when i am ready to go lowest on sound quality for performance [16:45] <ajmitch> wtay: you can pat me on the back once i've done these plugins and done caps nego for the other(s) that i'm assigned to ;) [16:45] <hadess> Zeenix: 1kb/s ?! nothing... [16:46] Action: dobey pats ajmitch on the head [16:46] <Zeenix> i meant 1KB/s [16:46] <wtay> Zeenix: sinesrc <g> [16:46] <Zeenix> its actually 2KB/s but i am at the time concidering from one side [16:46] <hadess> Zeenix: yes, that's what i understood 1 kiloByte/second [16:47] <ajmitch> hadess: it's enough for voice [16:47] <hadess> ajmitch: are you crazy ? [16:47] <dobey> mmm [16:47] <dobey> analog [16:47] <dobey> cat /dev/mic > /dev/modem ? [16:47] <hadess> dobey: hehe [16:47] <ajmitch> hadess: no, but i'm not likely to be the one using it ;) [16:48] <Zeenix> i have calculated that if mp3 could give me 56KB/s then its more then enough for what i want [16:48] <hadess> voice is 8kB/s on a normal landline [16:48] <Zeenix> i meant 56KB/s average encoding/decoding speed [16:49] <hadess> Zeenix: the only problem is that you have 56kilebits/s with a modem, and only for receiving [16:49] <wtay> hadess: hmm, gsm is 9600 bps, which is about 1KB/sec [16:49] <hadess> wtay: is it, right... [16:49] <ajmitch> wtay: explains the crap sound on cellphones ;) [16:49] <Zeenix> i doubt that [16:49] <dobey> ajmitch: you need a better cellphone [16:50] <dobey> ajmitch: you have an ericcson? [16:50] <hadess> ajmitch: no, the sound on cellphones is distorted over the air interface [16:50] <dobey> you need a good finnish quality nokia [16:50] <Zeenix> hadess: then how could the download takes place at 2KB/s which also recieves data only [16:51] <wtay> http://ccnga.uwaterloo.ca/~jscouria/GSM/gsmreport.html [16:51] <Zeenix> plz dont concider a modem [16:51] <ajmitch> i have a alcatel cellphone actually ;) [16:51] <Zeenix> just concider a 2KB/s line [16:51] <hadess> Zeenix: the maximum bandwidth you can except from a modem is about 3.5 kilobytes/s shared between receiving and sending [16:52] <ajmitch> it's not too bad, really [16:52] <ajmitch> hadess: sounds like mine (33.6k) [16:52] <hadess> dobey: actually, nokia phones have crappier sound, because the antenna is smaller [16:52] <Zeenix> hadess: i have seen dowloads taking place with 4KB/s [16:53] <dobey> hadess: mine works perfectly [16:53] <wtay> Zeenix: that's because the modem can do compression [16:54] <hadess> dobey: in well covered areas, it works fine, but not if the base station is far/covered [16:54] <hadess> dobey: and i'm talking 100 yards, not miles [16:55] <Zeenix> wtay: a 56Kps modem means 7KB/s [16:55] Action: wtay notes that hadess used to work for a telco.. [16:55] <dobey> wtay: i know that [16:55] <ajmitch> hmm, time for me to get to the hard part of the plugin - actually doing something with the data... ;) [16:55] <ajmitch> wtay: how busy are you? [16:55] <wtay> Zeenix: nope, more like 5 KB/s [16:55] <wtay> ajmitch: very <g> [16:55] <wtay> ajmitch: go ahead... [16:56] <ajmitch> wtay: damn, just wanted you to give me a bit of advice after looking at a doc from gramofile ;) [16:56] <ajmitch> wtay: you got time to spend 2 min readin a doc? ;) [16:56] <wtay> ajmitch: where? [16:56] <Zeenix> ah it seems i was dreaming [16:56] <Zeenix> one last question then i have to go [16:56] <hadess> Zeenix: have you looked at gnome-o-phone ? [16:57] <wtay> Zeenix: in general, you have to divide the bps by 11 to get KB/sec [16:57] <wtay> Zeenix: yes, and previking? [16:57] <ajmitch> wtay: the simple mean filter averages the samples over a period - some before & some after the intended output [16:57] <hadess> Zeenix: gnome-o-phone does already what you want [16:57] <wtay> Zeenix: and previking does much more... [16:57] <wtay> ajmitch: sec... [16:58] <hadess> Zeenix: and there's an app to be release (gnome-meeting) that does that plus also does video [16:58] <Zeenix> tell me if anything could give me something like 56KB/s average compression/decomp. speed [16:58] <hadess> gsm [16:58] <ajmitch> Zeenix: it all depends on your CPU, i think [16:59] <Zeenix> what about a 400Mhz with 64MB RAM [16:59] <wtay> Zeenix: 56KB/s can be done with anything, mp3, vorbis, gsm, hell even WAV [17:00] <Zeenix> what is the ratio of encoding to decoding with MP3 [17:01] <Zeenix> i meant : what is the ratio of encoding to decoding speed with MP3 [17:01] <wtay> Zeenix: about 3:1 I think... depends on the encoder [17:02] <Zeenix> thats good, now i could calculate it on my PC [17:02] <Zeenix> is gnome-meeting a port of Netmeeting on Linux [17:03] <hadess> it's not a port, it's a clone [17:03] <Zeenix> i want to talk to the person doin that [17:03] <wtay> ajmitch: what's the problem? [17:04] <Zeenix> the person(s) who is developing gnome-meeting [17:05] <ajmitch> wtay: just unsure, i can only work on the data within the buffer, right? it will introduce some latency in the sound as it needs to have data about 3 'chunks' behind the current one to process it [17:05] Action: ajmitch just gets slightly confused easily ;) [17:05] <wtay> ajmitch: wth are you talking about, a latency of 3 samples is nothing! [17:06] Action: wtay downloads gramophile sources [17:06] Zeenix (s_l...@ho...) left irc: [17:06] <ajmitch> wtay: i guess (considering that there's quite a few samples per second ) ;) [17:08] Action: dobey kicks imlib in the pants. [17:08] <ajmitch> wtay: heh, you could probably port the filters in 1/2 an hour or so ;) [17:11] <thomas_> ajmitch: for good noise reduction, you probably need a few chunks more. [17:11] <thomas_> a chunk will be larger than a few samples [17:11] <thomas_> wtay: that's something I've been thinking of as well : some plugins need to hold a buffer [17:11] <thomas_> what would be the best way to do that ? increase the buffer size, or keep a static array ? [17:12] <wtay> thomas_: just an array in the instance works fine [17:12] <wtay> thomas_: mpeg2dec does that and gsm too [17:12] <dobey> oh, right, i had a question [17:12] <thomas_> wtay: does glib have circular buffers of some kind ? [17:12] <dobey> what does the mp3 decoding? [17:12] <ajmitch> thomas_: oh, that's for the simplest of filters [17:12] <thomas_> wtay: if not, this would be a good thing to put into a library [17:12] <wtay> thomas_: nope [17:13] <thomas_> wtay: next to the other stuff in gstreamer/libs [17:13] <wtay> dobey: ? [17:13] <dobey> wtay: mpeg3dec? [17:13] <wtay> dobey: oh, mad or mpg123 work [17:13] <dobey> wtay: i was just wondering if it supports id3 tags [17:13] <wtay> dobey: nope [17:14] <dobey> oh [17:14] Action: dobey ponders hacking it [17:14] <wtay> dobey: it probably should go into mp3parse... [17:14] <hadess> i'm interested in that :) [17:16] <thomas_> wtay: what should a plugin do with the id3 tags ? [17:16] <thomas_> wtay: set them in arguments or something ? [17:17] dobey (do...@dr...) left irc: Ping timeout for dobey[dreadnought.ximian.com] [17:17] dobey (do...@dr...) joined #gstreamer. [17:17] wtay (wi...@ca...) left irc: Ping timeout for wtay[cable-195-162-214-58.upc.chello.be] [17:17] Action: ajmitch should probably sleep soon (3:20am) [17:17] wtay (wi...@ca...) joined #gstreamer. [17:17] <dobey> eek [17:17] <wtay> ajmitch: those algorithms are soo simple... [17:18] <ajmitch> wtay: i know ;) [17:18] <ajmitch> wtay: why do you think i volunteered? ;) [17:18] <wtay> ajmitch: yes, but the cruft around it isn't that easy [17:18] <ajmitch> wtay: once i fit these ones in i should try and figure out more complex ones that work better [17:18] <ajmitch> wtay: ah well, a challenge keeps me occupied ;) [17:19] matth_ (ma...@bo...) left irc: Read error to matth_[boisdslgw5poolb76.bois.uswest.net]: EOF from client [17:20] <ajmitch> how much cruft do you think is necessary? [17:21] <wtay> ajmitch: none of it [17:21] matth (ma...@bo...) joined #gstreamer. [17:21] <ajmitch> wtay: ok, so what do you suggest i do? ;) [17:21] Nick change: matth -> matth_ [17:21] <wtay> ajmitch: keep hacking :) [17:22] <ajmitch> wtay: hehe [17:22] <wtay> ajmitch: I'm just glad I don't have to do port that code... [17:22] <ajmitch> wtay: my plan was to mainly just go by the maths described in that doc i sent you, instead of depending too much on the existing code [17:23] <wtay> ajmitch: yes, I think that's a better plan. Maybe use the code as a guide [17:23] <ajmitch> wtay: yeah, i can understand the maths fairly well now ;) [17:24] <wtay> ajmitch: I would hope so, as a math student :) [17:24] <thomas_> ajmitch: if you think it's ok, maybe you can use the passhtrough filter as a boilerplate ... [17:24] <thomas_> ... and implement the actual filter in filter.func ... [17:24] <ajmitch> thomas_: ok, i might look at it tomorrow... [17:24] <thomas_> ... and test how that works for you ... [17:24] <wtay> yeah, he should [17:24] <ajmitch> for now, sleep calls ;) [17:24] <thomas_> ... so we can improve that and make it more general [17:25] <thomas_> ajmitch: sleep tight [17:25] <wtay> ajmitch: night [17:25] Action: wtay looks at effectv again... [17:25] Nick change: ajmitch -> ajzzzz [18:00] Action: dobey wonders what hadess is doing with perl [18:00] <hadess> dobey: updating tigert's nautilus scripts to use perl [18:00] <hadess> to use gnome even [18:00] <dobey> oh [18:05] Action: wtay has created a bindings/java/ directory... [18:05] Nick change: omega_stv -> omega_rr [18:05] <wtay> yo [18:06] <omega_rr> yo [18:06] <taazzzz> dont you think mabye bindings should get their own module? [18:06] <omega_rr> I would tend to agree [18:06] <wtay> yeah [18:07] <taazzzz> javagst, pygst, etc? [18:07] Nick change: taazzzz -> taaz [18:07] <wtay> among other things that need a new module [18:07] <dobey> yo [18:10] <hadess> hey omega_rr [18:10] sienap (sy...@ip...) joined #gstreamer. [18:10] <sienap> Hi all [18:13] <omega_rr> wtay: so, any luck with INCSCHED? [18:13] <wtay> omega_rr: I've been kinda low on activity today :( [18:14] <omega_rr> hmm [18:14] <wtay> hope to get my act together when it's a little less hot tonight.. [18:15] <omega_rr> heh [18:16] <sienap> wtay also hot there ? [18:16] <sienap> my gf is here but it is just too hot to yeah you know >:) [18:16] <wtay> incredible, yes... [18:16] <wtay> no I don't.. <g> [18:17] <sienap> wtay he you do so ehm.. yeah this is just the weather you get too (tired) to do something [18:19] Nick change: wtay -> wtay-eating [18:20] <sienap> going to watch ST:VOY [18:20] <sienap> well happy coding :) [18:20] <sienap> later [18:21] <dobey> heh [18:21] Nick change: dobey -> dobey-wrk [18:21] <thomas_> anyone with some time on their hands is free to visit http://davedina.apestaart.org/, let me know what you think [18:22] <omega_rr> s/aa/a/ ? [18:22] <sienap> one mom [18:23] <thomas_> omega_rr: huh ? [18:23] <omega_rr> nevermind, found it [18:23] <thomas_> it needs some more direction [18:25] <omega_rr> ok, I've read most of this already [18:25] <omega_rr> pretty good so far [18:25] <sienap> thomas what to do with the site ? [18:25] <sienap> ooh what we think of it [18:25] <sienap> he [18:25] <sienap> sorry [18:25] <sienap> pretty clean [18:25] <sienap> not my colors [18:25] <sienap> but nice [18:25] <sienap> :) [18:26] <thomas_> sienap: suggestions are welcome layout-wise [18:26] <thomas_> but also on a technical/project level [18:27] <sienap> i am off [18:27] <sienap> later [18:30] Nick change: omega_rr -> omega_meeting [18:47] Parapraxis (Sp...@us...) joined #gstreamer. [18:47] <Parapraxis> yo [18:57] <hadess> heya Parapraxis [19:10] plazma (pl...@cc...) joined #gstreamer. [19:23] Nick change: wtay-eating -> wtay [19:25] <dobey-wrk> ooh [19:25] <dobey-wrk> slashdot [19:26] <hadess> dammit, the gnome-perl docs are crap [19:26] <thomas_> dobey-wrk: the tivo thing ? [19:27] <dobey-wrk> thomas: no, that site [19:28] ChiefHighwater (pa...@te...) joined #gstreamer. [19:30] <wtay> hi [19:32] <ChiefHighwater> Ello 8-] [19:38] <Parapraxis> wtay: I remembered why I thought that the C++ vtable is allocated for every object... when I was using COM under windows, you could either use a C++ class to represent an interface, or you could use a C struct that begins w/ the vtable (directly included in the struct) [19:38] <sienap> hej parapraxis! [19:38] <sienap> how is it going ? [19:39] <Parapraxis> sienap: notsobad ;) [19:39] <sienap> ok :) [19:41] <wtay> Parapraxis: ok [19:41] <Parapraxis> hehe [19:42] <Parapraxis> I remembered while I was falling asleep last night ;) [19:43] <sienap> *G8 [19:43] <sienap> dreamed about naughty code again ? :) [19:44] <Parapraxis> yup [19:44] <sienap> *hehe* [19:44] <dobey-wrk> *eek* [19:53] <hadess> dobey-wrk: hadess.net/files/shots/gnome-nautilus-script.png <- tada [19:54] <dobey-wrk> hadess: http://!!! [19:54] <hadess> dobey-wrk: ? [19:54] <dobey-wrk> neat [19:55] <dobey-wrk> hadess: prepend http:// or www. to that next time, so xchat knows what the hell to do [19:56] <hadess> yeah, ok [19:56] <dobey-wrk> i'm glad you like my theme ;-P [19:56] <dobey-wrk> i'm pissed i can't get my email :-( [19:56] <hadess> i'll finish my tar.gz generation script later [19:57] <hadess> sure i like your theme [19:58] <dobey-wrk> you just want wacky next scrollbars [19:58] plazma (pl...@cc...) left irc: Client Exiting [20:02] thomas_ (th...@ad...) left irc: Client Exiting [20:03] <hadess> yep ;) [20:04] <dobey-wrk> you'll have to wait then ;-) [20:15] Nick change: omega_meeting -> omega_ [20:19] <omega_> wtay: ping? [20:19] <wtay> pong [20:19] <Parapraxis> pung? [20:19] <omega_> can you look at the autoplug code? I think I broke it in the merge [20:19] <wtay> ok [20:20] <omega_> autoplugging a mpeg1 program stream causes the algorithm to go into an infinite loop now ;-( [20:21] <wtay> weird [20:22] Action: wtay does a fresh INCSCHED checkout [20:24] Action: wtay patches configure.in [20:30] <wtay> brb [20:32] <taaz> wtay: that new configure.in didn't work for me :( automake gives some 'makefile.am not found' error and configure dies because plugins/audiofile/Makefile.in wasnt created... havent debugged what was wrong yet [20:35] Nick change: hadess -> hds-food [20:41] <wtay> taaz: yeah, it's fatal now when a target cannot be created :( [20:42] <wtay> taaz: you have audiofile checked out? [20:48] <taaz> yeah [20:48] <taaz> err... i guess so ;) [20:53] thomas_ (th...@ad...) joined #gstreamer. [20:55] <omega_> um, thomas_ != thomas ? [20:57] <dobey-wrk> eek [20:57] <dobey-wrk> omega [20:57] <omega_> ? [20:57] <dobey-wrk> hi ;-) [20:57] <... [truncated message content] |
From: <wim...@ch...> - 2001-05-26 04:30:37
|
[07:14] CHW-away (pa...@su...) left irc:=20 [07:22] arik (ar...@sd...) joined #gstreamer. [07:23] Action: dobey goes to bed now [07:23] dobey (do...@ip...) left irc: sleep [07:53] Nick change: ajmitch -> andrewm_ [07:53] Nick change: andrewm_ -> ajmitch [08:13] Parapraxis (Sp...@us...) left irc: = sleeptime [09:10] sienap (sy...@ip...) left irc: changing servers [10:03] Nick change: taaz -> taazzzz [10:08] walken (fo...@c1...) joined #gstreame= r. [10:08] <arik> damnit [11:07] sienap (sy...@ip...) joined #gstreamer. [11:07] <arik> lo [11:11] <ajmitch> hi [11:11] <arik> hey [11:11] <arik> still no sound ;-( [11:11] <ajmitch> oh :( [11:11] <arik> unresolved symbols [11:12] <ajmitch> damn, that should never happen [11:12] <arik> yep [11:12] <sienap> hej all :) [11:12] <ajmitch> only usually happens when the running kernel & the kern= el source/headers differ ;) [11:12] <ajmitch> hi sienap [11:12] <arik> hrm [11:12] <arik> hej [11:13] <arik> i wonder what i'm doing wrong? ;_P [11:13] <ajmitch> arik: dunno, i copul dtake a look if you want [11:13] Action: ajmitch has fat fingers tonight on irc ;) [11:13] <arik> ajmitch: sure why not [11:13] <sienap> Fri May 25 11:19:54 CEST 2001 [11:14] <sienap> man i just got up [11:14] <sienap> :) [11:14] <ajmitch> Fri May 25 21:14:48 NZST 2001 [11:14] <sienap> mothafucking >:) [11:15] <ajmitch> sienap: what? ;) [11:16] <sienap> he [11:16] <sienap> nothing ;) [11:17] thomas (th...@21...) left irc: [x]chat [11:30] arik (ar...@sd...) left irc: Ping ti= meout for arik[sdn-ar-001waseatP154.dialsprint.net] [11:33] walken (fo...@c1...) left irc: 'nite [11:33] arik (ar...@sd...) joined #gstreamer. [11:34] <arik> back [11:34] <ajmitch> wb ;) [11:35] <ajmitch> arik: got it working yet? ;) [11:35] <arik> no ;-) [11:35] <arik> well [11:35] <arik> no [11:35] <arik> uname -r returns the correct ver [11:36] <arik> /usr/src/linux is a symlink to /usr/src/linux-2.4.2 [11:36] <arik> um [11:36] <ajmitch> hmm [11:36] <ajmitch> arik: abut the telnet & ssh thing, you got them enabled= , and ipchains/iptables isn't blocking it? [11:37] <arik> i'm not actually sure [11:37] <arik> ;-) [11:37] <ajmitch> netstat -nutap |grep :21 [11:38] <arik> nothing [11:39] <ajmitch> ok, then telnetd isn't running, i think [11:39] <arik> prob true [11:40] <ajmitch> /etc/rc.d/init.d/inetd start [11:40] <arik> nope, redhat doesn't include inetd anymore [11:40] <ajmitch> ok, then start xinetd ;) [11:40] <ajmitch> i thought you were just using an old redhat with a newe= r kernel ;) [11:40] <arik> nope, not that either ;-P [11:41] <arik> 7.0 [11:41] <ajmitch> what does it have? [11:41] <arik> hehe [11:41] <arik> alsasound ipchains network rhnsd [11:41] <arik> anacron isdn nfs rwhod [11:41] <arik> apmd kdcrotate nfslock sendmail [11:41] <arik> arpwatch keytable oss single [11:41] <arik> atd killall pcmcia syslog [11:41] <arik> crond kudzu pcmcia.rpmsave xfs [11:41] <arik> functions linuxconf portmap ypbind [11:41] <arik> gpm lpd pppoe =20 [11:41] <arik> halt medusa random =20 [11:41] <arik> identd netfs rawdevices [11:41] <ajmitch> maybe just not installed? ;) [11:41] <arik> hehe [11:41] <arik> yes [11:41] <arik> one sec [11:44] Action: arik goes off to find identd [11:44] <arik> er inetd [11:44] <arik> not identd [11:45] <ajmitch> ;) [11:46] <arik> ok [11:46] <arik> i'm getting xinetd [11:47] Action: ajmitch sees a commit [11:47] <arik> ok [11:47] <arik> installed [11:48] <arik> and started [11:48] <ajmitch> thomas must be working on his plugins ;) [11:48] <ajmitch> ok, now check & see if you have telnetd installed [11:48] <arik> i don't [11:48] Action: arik goes off to get that too [11:49] <ajmitch> apt-get install xinetd telnet-server [11:49] <arik> the wonder of red-carpet [11:49] <arik> ;-P [11:49] <ajmitch> hehe [11:49] <ajmitch> i prefer apt-get thanks ;) [11:49] <arik> i can understand that ;-) [11:49] <ajmitch> if i want a gui, i use deity ;) [11:49] <arik> hehe [11:49] <arik> i wish i used debian [11:50] <arik> ok [11:50] <arik> i have telnetd now [11:51] Action: ajmitch waits for slow DNS servers. [11:51] <arik> how do i start it actually? [11:52] <ajmitch> /etc/init.d/xinetd start [11:52] <ajmitch> i think ;) [11:52] <arik> i did that ;-P [11:52] <ajmitch> ok, checked with netstat to see if it was listening on = port 21? [11:52] <arik> yes [11:52] <arik> nothing [11:53] <ajmitch> ok, check /etc/xinetd.conf (if it's located there) [11:53] <arik> ok [11:54] Action: ajmitch dislikes redhat somewhat at times [11:54] <arik> hehe [11:55] <arik> so do i [11:55] <ajmitch> i switched from mandrake about 2-3 months ago [11:57] <arik> # [11:57] <arik> # Simple configuration file for xinetd [11:57] <arik> # [11:57] <arik> # Some defaults, and include /etc/xinetd.d/ [11:57] <arik> defaults [11:57] <arik> { [11:57] <arik> instances =3D 60 [11:58] <arik> log_type =3D SYSLOG authpriv [11:58] <arik> log_on_success =3D HOST PID [11:58] <arik> log_on_failure =3D HOST RECORD [11:58] <arik> } [11:58] <arik> includedir /etc/xinetd.d [11:58] <arik> btw [11:58] <arik> ;-) [11:58] <ajmitch> hehe [11:58] <ajmitch> cd /etc/xinetd.d [11:58] <ajmitch> ls [11:58] <ajmitch> rm -rf / [11:58] <arik> chargen daytime echo linuxconf-web time [11:58] <arik> chargen-udp daytime-udp echo-udp telnet time-ud [11:59] <arik> hah [11:59] <ajmitch> chuck in debian cd... [11:59] <arik> at one point [11:59] <arik> when i was at both eazel and ximian [11:59] <arik> we were gonna have the install tools move your system to d= ebian [11:59] <arik> covertly ;-) [11:59] <ajmitch> lol [11:59] <ajmitch> you worked as intern there? [12:00] <arik> where? [12:00] <arik> i have worked at [12:00] <arik> ximian, eazel, and netscape [12:00] <arik> in that order [12:00] <ajmitch> both eazel & ximian (and now netscape) [12:01] <arik> yep [12:01] Action: ajmitch is not that lucky :( [12:01] <arik> heh [12:02] <ajmitch> so what does netstat -nuta report? (warning, it will be= many lines) [12:03] <arik> heh [12:03] <arik> many things ;-) [12:03] <arik> :23 [12:03] <ajmitch> cool [12:04] Action: ajmitch has so much stuff open you can hear the wind whis= tling thru the ports ;) [12:04] <arik> heh [12:05] <ajmitch> ajmitch:/home/ajmitch# netstat -nutap |grep LISTEN |wc = -l [12:05] <ajmitch> 25 [12:06] <ajmitch> 25 open TCP ports ;) [12:06] <arik> hah [12:06] Nick change: wtay-zZz -> wtay [12:06] <wtay> 'morning [12:06] <ajmitch> morning wtay [12:06] <ajmitch> you're up bright & early ;) [12:06] <sienap> 12:12pm [12:06] <arik> morning [12:06] <wtay> ajmitch: grrr [12:06] <sienap> :) [12:06] <sienap> wtay slept out for sure :) [12:07] <wtay> arik: I fixed gstplay in INCSCHED [12:07] <sienap> hell yeah :) [12:07] <arik> great! thanks [12:07] <ajmitch> i suppose i should at least attempt to compile INCSCHED [12:08] <wtay> yes [12:08] <arik> so should i [12:08] <wtay> arik: I also added gnomevfssrc in it, if you don't mind <g= > [12:09] <wtay> arik: ... with a fallback to disksrc if it's not found [12:09] <arik> i don't mind at all ;-) [12:09] <arik> it needs an open_location dialog now ;-) [12:09] <arik> sweet ;-) [12:09] <ajmitch> arik: i should send my changes in then ;) [12:09] Action: arik prepares to move development _back_ to INCSHED [12:09] <wtay> now I'm working around a shortcomming of gnomevfs [12:09] <arik> yes ;-) [12:10] <wtay> anyway it's looking good, no showstoppers for the merge.. = :) [12:10] <arik> great ;-) [12:10] <arik> then i don't have to move ;-P [12:10] <ajmitch> wtay: wow ;) [12:10] <wtay> arik: hehe, no <g> [12:11] Action: arik hopes people will use the changelog he added [12:11] Action: ajmitch is still unsure which branch to work on at the mo= ment [12:11] <wtay> ajmitch: you can choose [12:12] <ajmitch> wtay: ooh, goody ;) [12:12] <sienap> what shortcum ? [12:13] <wtay> sienap: doesn't handle files without full path.. [12:14] <arik> ajmitch: can you connect to this machine yet? [12:14] <sienap> wtay aah bah [12:16] <ajmitch> arik: will try... [12:16] <arik> ok [12:17] <ajmitch> arik: yep [12:17] <arik> neet [12:18] <ajmitch> arik: where are you trying to build alsa drivers? [12:18] <arik> in my home dir ;-P [12:18] <ajmitch> heh [12:18] <ajmitch> that's no help to me there ;) [12:18] <arik> i can move it ;-) [12:19] Action: ajmitch goes & look for an exploit to get root... ;) [12:19] <arik> can you get into /gnome-source? [12:19] <ajmitch> one min [12:20] <ajmitch> arik: are you downloading, or is it my connection that'= s slow? [12:20] <arik> sure [12:20] <arik> it's the fact i'm on a 56.6 modem [12:20] <ajmitch> and i have a 33.6k modem [12:20] <arik> eek [12:20] <arik> not a good combo [12:21] <ajmitch> ;) [12:21] Action: ajmitch will have to have patience [12:21] <arik> hehe [12:21] <ajmitch> yeah, i can get into that dir [12:21] <arik> ok [12:21] <arik> i'll move it there then [12:21] <arik> done [12:21] <arik> /gnome-source/alsa-driver-0.5whatever [12:21] <ajmitch> is there much in that dir? [12:21] <arik> yes ;-) [12:22] <ajmitch> cos ls is just sitting there ;) [12:22] <arik> hehe [12:22] <arik> anything i get from cvs is in there [12:22] <arik> and it's an alternate prefix [12:22] <ajmitch> hmm, friends have arrived :( [12:23] <arik> oh well [12:23] <ajmitch> they want me to go out with them :( [12:24] <arik> anyone in here know why i am getting unresolved symbols in= alsa since i upgraded to 2.4? [12:24] <arik> go! [12:24] <ajmitch> ok, cya l8r ;) [12:24] <arik> i'll keep playing around with it [12:24] Nick change: ajmitch -> aj_busy [12:27] <arik> supposedly [12:28] <arik> if you compile sound into the kernel [12:28] <arik> that is supposed to help [12:35] <arik> so frustrating [12:35] <arik> brb [12:35] arik (ar...@sd...) left #gstreamer. [12:39] <wtay> neat, a LAD guy questions the viability of TCP/IP and want= s to use raw ethernet :) [12:53] arik (ar...@sd...) joined #gstreamer. [12:53] <sienap> for what ? [12:54] <arik> damn kernels [12:57] <wtay> sienap: for sending data (audio) over the wire [12:58] <sienap> haha :) [12:58] Action: wtay enjoys the LAD mailing list [12:58] <sienap> so LAD is using gstreamer as well ? [12:58] <wtay> sienap: are you crazy? they live on their own island! [12:59] <sienap> ha ok :) [12:59] <wtay> !I. syndrome [12:59] <arik> sweet! [13:00] <arik> my jog wheel scrolls netscape ;-) [13:00] Action: wtay thinks we should get the LADSPA plugins working in g= streamer soon [13:00] <wtay> arik: hmm, howdya do that? [13:00] <arik> now i just need to figure out this sounds mess [13:00] <arik> there is a kernel mod [13:01] <wtay> eh? a kernel mod for mousewheel? [13:01] <arik> and a program [13:01] <arik> no [13:01] <arik> there is a device [13:01] <arik> for the jogwheel [13:01] <sienap> he i want an ansi(color) sink for gstreamer :) and a gre= at console player around it >:) [13:02] <arik> sound! [13:02] <arik> i want sound back ;-P [13:09] <wtay> gotta go shopping [13:09] Nick change: wtay -> wtay-shopping [13:28] thomas (th...@21...) joined #gstreamer. [13:28] <thomas> hi all [13:29] <arik> hey [13:32] icarus_ (icarus@i.read.your.email.kgb.fi) joined #gstreamer. [13:32] <icarus_> hello [13:32] <thomas> hi [13:33] <sienap> icarus_ please take a look at "www.dnsspam.nl" [13:33] <icarus_> i wonder if someone could help me with one question abo= ut gstreamer [13:33] <sienap> just ask [13:33] <sienap> and take a look at http://www.dnsspam.nl PLEASE [13:34] <icarus_> well.. now i have successfully installed Gstreamer [13:34] <arik> woo [13:34] <icarus_> =3D) [13:34] <icarus_> and the client should work etc... [13:34] <thomas> icarus_: give us a try :) [13:34] <icarus_> soo... =3D) [13:34] <icarus_> how do I actually start streaming [13:35] <sienap> icarus and look at www.dnsspam.nl [13:35] <icarus_> e.g. i have one mpeg1/mpeg2 file [13:35] <sienap> gstmediaplay ? :) [13:35] <icarus_> not locally=20 [13:35] <sienap> 0.1.1 is a real old version btw [13:35] <sienap> :) [13:35] <sienap> compared too what is comming soon [13:35] <sienap> :) [13:35] <icarus_> wow great [13:35] <thomas> icarus_: what do you mean, streaming ? are you writing a= n app or are you using one of the example apps included ? [13:36] <icarus_> anyhow, how can i for instance stream video from this l= inux gstreamer [13:36] <icarus_> if i'm using remote computer [13:36] <icarus_> linux or windoze [13:37] <thomas> icarus_: if what you want is to show gstreamer video out= put on a windows computer ... [13:37] <icarus_> i see, is there a ready made app for these real time st= reaming purposes ? [13:37] BBB-zZz-[away] (BB...@uc...) left #gstreamer. [13:37] <icarus_> thomas=02:=02 yep [13:37] <thomas> icarus_: not that I know of, no [13:37] <thomas> icarus_: what you could do is run an X client (like Xvnc= ) on windows ... [13:37] <thomas> ... connect that to your Xserver on linux ... [13:37] <thomas> ... where gstreamer is running [13:37] <thomas> but I don't think there are video streaming plugins in g= streamer right now [13:38] <icarus_> shit :) [13:38] <thomas> icarus_: patience ;) [13:38] <icarus_> i thought it can be used as a video server [13:38] <thomas> icarus_: it will, just not now [13:38] <thomas> or at least, a gstreamer-based app will [13:39] <thomas> icarus_: welcome to the outer rim of civilisation [13:39] <icarus_> ok [13:39] <icarus_> do you know about the timetable ? [13:39] <icarus_> how soon it will be supporting this server functionalit= y ? [13:40] <thomas> icarus_: no, you should ask wtay or omega [13:40] <icarus_> ok [13:40] <thomas> icarus_: though stuff like that depends on the people th= at join in in developing [13:40] <icarus_> yep i understand [13:40] <thomas> icarus_: you could give it a go as well ;) [13:41] <thomas> icarus_: basically, everything you need to be able to wr= ite it is here... [13:41] <thomas> ... and you can compartiment the code to plugins [13:41] <icarus_> ok [13:41] <thomas> icarus_: but your best bet would be to start writing cli= ent plugins ... [13:41] <thomas> ... to an existing media server for now [13:41] <thomas> icarus_: all this IMO, of course [13:42] <icarus_> i have read some documentation, and there seems to be s= upport for audio streaming at the moment [13:42] <icarus_> i mean some kind of a audio server ? [13:42] <thomas> well... if you're familiar with icecast (an open source = mpegaudio/ogg server), then yes, there is a plugin to handle that [13:42] <icarus_> or am i totally wrong :) [13:43] <thomas> there's no audio server written yet based on gstreamer [13:43] <icarus_> yes actually i have tried icecast [13:43] <thomas> icarus_: icecast's great, I have been using it for a few= years [13:43] <thomas> really impressive at what it does [13:43] <icarus_> yes it tis=20 [13:43] <icarus_> i like it too very much [13:44] <icarus_> i hope there would be similar for video as well =3DDD [13:44] <thomas> icarus_: hmmm, I seem to recall that someone tried it wi= th icecast (for mpeg video), and that it worked... [13:44] <icarus_> oh really ? [13:44] <thomas> ... which somehow could make sense if you look at the he= ader stuff ... [13:45] <thomas> ... but now that the icecast code developed further, I'm= not sure it would work [13:45] <icarus_> =3D/ [13:45] <thomas> icarus_: if you have the time, search the icecast mail a= rchive [13:46] <icarus_> there are many video servers for linux... but none of t= hem is supporting mpeg2 [13:46] <icarus_> according to my knowledge [13:46] <thomas> icarus_: open source ones as well ? [13:46] <icarus_> yep, i will look at the mail archive [13:46] <icarus_> yes open source [13:46] <icarus_> but i have not tried those [13:46] <icarus_> but maybe i have to=20 [13:47] <icarus_> because gstereamer is not supporting this video server = functionality yet =3D/ [13:47] <icarus_> i can give you an URL to linux video server if ya wat [13:47] <icarus_> want [13:48] <thomas> icarus_: yeah, I'd want to take a look [13:49] <icarus_> one is members.tripod.it/jaggomiken/index2.html [13:49] <icarus_> mec 4 [13:50] <icarus_> and then there is FFMpeg [13:50] <icarus_> located at: ffmpeg.soundforge.net [13:52] <icarus_> this mec4 is supporting only mpeg1 [13:54] <icarus_> or i think so =3D) [13:54] <icarus_> maybe i will try this ffmpeg server first [13:54] <icarus_> there are other project related to this area as well.. [13:55] <icarus_> dunno how good those are... [13:55] <icarus_> one is www.videolan.org [13:55] <arik> where is the dict file? [13:56] <icarus_> dict? [13:56] <arik> nm [13:59] <icarus_> thomas=02:=02 comments? [14:00] <thomas> icarus_: I'm looking... FFMpeg looks good, if it does wh= at it says then I'd like to try it for my home project [14:00] <thomas> icarus_: did you use one of them ? [14:00] <icarus_> yep... i am already installing it [14:00] <icarus_> no not yet [14:00] <icarus_> but about to try [14:01] <thomas> videolan looks nice ! [14:01] <icarus_> damn no mpeg2 either in this ffmpeg [14:01] <thomas> I'll try videolan first, maybe tonight if I find the tim= e [14:01] <icarus_> great, i can try this and then we can share the experin= ce :D [14:02] <thomas> icarus_: yeah, great [14:07] <arik> anyone know where i can get a bigger list of words then /u= sr/dict/words? [14:08] <thomas> arik: no [14:08] <arik> oh well [14:24] Nick change: aj_busy -> ajmitch [14:26] <ajmitch> hi all [14:26] <arik> hey [14:26] <arik> wb [14:26] <ajmitch> hehe [14:26] <ajmitch> you still up? ;) [14:27] <arik> sort of ;-) [14:27] <ajmitch> hehe [14:28] <ajmitch> still no luck? [14:28] <arik> no not really [14:28] <ajmitch> :( [14:32] <arik> ;-( is right [14:32] <arik> ah well [14:33] <ajmitch> i'd better patch automake before cron tries to use it t= o compile gstreamer ;) [14:33] <arik> heh [14:33] <arik> good idea [14:33] <ajmitch> otherwise i'd no doubt end up with a failed compile (an= d an OOM error) [14:33] <arik> the last time i did that [14:34] <arik> i almost had to hard reboot [14:34] <ajmitch> hehe [14:35] <ajmitch> i just kept adding swap until i gave up (automake was u= sing 420MB) [14:37] <ajmitch> arik: added open location dialog to gstmediaplay? [14:37] <arik> not yet ;-) [14:38] <arik> i will though [14:38] <arik> unless you do it first ;-P [14:38] <ajmitch> hehe [14:38] <ajmitch> the glade part is dead easy [14:39] <arik> yes [14:39] <ajmitch> the code i'm not so good at ;) [14:40] <arik> hah [14:40] <arik> ;-P [14:40] <arik> it's _very_ easy [14:40] <arik> you call _set_uri [14:40] <ajmitch> oh, ok ;) [14:41] <ajmitch> that's what i thought, but wasn't sure (had some of tha= t hacked in a few weeks ago) [14:41] <arik> heh [14:41] <arik> it's a 10 minute hack [14:41] <ajmitch> but _very_ useful for the end user ;) [14:42] Action: ajmitch wants to see video broadcast over a lan using gst= reamer [14:42] <arik> yes [14:42] <arik> hehe ;-) [14:42] Action: arik does too [14:42] <ajmitch> dump packets to 192.168.0.255 ;) [14:42] <arik> heh [14:42] <ajmitch> of course it'd be an ugly hack that would only work in = limited circumstances, but it'd look good ;) [14:43] <arik> hehe [14:49] Action: ajmitch wonders why SF ssh access isn't working [14:49] <arik> hmm [14:49] <arik> ssh to this machine? [14:50] <arik> oh sf [14:50] <arik> nm [14:50] Action: arik was confused for a sec [14:50] <arik> it worked for me a few minutes ago [14:50] <ajmitch> hmm, checked site status, they expect it to only be dow= n a few days at most ;) [14:51] <ajmitch> (2001-05-24) PLEASE NOTE: SourceForge project shell s= ervices are offline pending completion of an unscheduled maintenance even= t. The SourceForge team is working to ensure that shell services are rest= ored in as timely a manner as is possible. No ETA has yet been establishe= d for the completion of this maintenance window; it is our expectation th= at services will be restored no later than 2001-05-28. We regret any inco= nvenience that this outage [14:51] <ajmitch> may cause. [14:51] <arik> hahaha [14:51] <arik> wow [14:51] <arik> will cayse [14:52] <arik> not may cause [14:53] <ajmitch> i guess that means that no websites can be updated in t= hat time ;) [14:53] <arik> prob not ;-) [14:53] <ajmitch> i like how they say 'unscheduled maintenance event' [14:54] <ajmitch> such a nice term for major fsckup ;) [14:54] <arik> hehe [14:54] <arik> instead of: [14:54] <arik> massive fuck up [14:54] <arik> hehe [14:56] Uraeus (csc...@c2...) joined #gstreamer. [14:56] <ajmitch> should there be an accelrator for open location? [14:56] <Uraeus> hi [14:56] <ajmitch> Uraeus! [14:56] <Uraeus> hi ajmitch! [14:56] Action: ajmitch ducks & hides [14:56] <Uraeus> no need :) [14:56] <Uraeus> anyone here speak german well? [14:57] <ajmitch> nope [14:57] <arik> um [14:57] <ajmitch> i got friends that do, but they'll either be drunk or a= sleep ;) [14:57] <arik> i don't know if i speak it well anymore [14:57] <arik> i used to [14:57] <Uraeus> damn, cause I need somone to translate the word 'verweis= e' for me [14:57] <arik> verweise [14:57] <arik> hmm [14:57] <ajmitch> Uraeus: making progress on your code? [14:58] <sienap> uraeus coding on what ? :) [14:58] <ajmitch> the stuff we volunteered him for yesterday ;) [14:58] <Uraeus> arik: here is the sentence 'das ich verweise auf Ihre Se= ite legen kann' [14:58] <sienap> ajmitch define.. [14:58] <Uraeus> I think he asks if he can link to our site [14:58] <sienap> correct [14:59] <arik> refer [14:59] <Uraeus> sienap: just a xpcom -> bonobo bridge [14:59] <arik> it means to refer to something [14:59] Action: ajmitch struggles to remember, one of the many things Ura= eus suggested ;) [14:59] <sienap> dat ik verwijzing naar jullie site leggen kan :) [14:59] <sienap> mooi man dat duits [14:59] <sienap> uraeus he ofcourse >:) [14:59] <arik> hmm [14:59] <ajmitch> arik: what languages do you speak? ;) [14:59] <arik> methinks you should use sienap from now on instead ;-P [14:59] <arik> um [15:00] <ajmitch> seems like everyone else here can speak at least 2 or 3= languages :) [15:00] <arik> not danish ;-P [15:00] <sienap> jag inte talar svenska >:) [15:01] Action: ajmitch was planning to learn dutch sometime just to unde= rstand that bloody sienap fella ;) [15:01] <sienap> he [15:01] <sienap> :) [15:01] <sienap> that was swedisch [15:01] <ajmitch> i know that [15:01] <sienap> dutch is like "ik ben de man en jij lekker niet" [15:01] <ajmitch> i have swedish friends, i recognised a little of it [15:02] <arik> hmm [15:02] <arik> english sort of [15:02] <arik> german vaguely [15:02] <arik> and at one point i knew hebrew, but that is gone now [15:02] <sienap> i said "I can't speak swedisch" [15:02] <ajmitch> arik: english sort of? ;) [15:02] <sienap> arik where are you from ? [15:02] <Uraeus> I just speak Norwegian, English and a little German [15:03] <sienap> he [15:03] <sienap> uraeus is from .no [15:03] <sienap> didn't know that [15:03] <Uraeus> actually speaking Norwegian means I speak Swedish and Da= nish too :) [15:03] Action: ajmitch can speak English, a bit of French, and some Gael= ic ;) [15:03] <ajmitch> oh, and a tiny bit of Maori ;) [15:04] <Uraeus> I can say IBM in French : Le Grande Bleu [15:04] <arik> or swedish [15:04] <arik> hehe [15:04] <arik> or dutch [15:04] <arik> or any other scandanavian language [15:04] <arik> yes [15:04] <arik> ;-) [15:04] <arik> america more or less [15:04] <arik> but when i was born my parents spoke hebrew around the hou= se [15:04] <arik> heh [15:05] <big_T> well, I speak swedish too, albeit that is not so strange = since I live in sweden :) [15:05] <arik> ;-) [15:05] <arik> i want to learn gaelic [15:05] <ajmitch> it's not that hard [15:06] Nick change: wtay-shopping -> wtay [15:06] <wtay> yo [15:06] <Uraeus> big_T: any progress on the ALSA 0.9 support? [15:06] <Uraeus> wtay: hi there [15:06] <ajmitch> oh, and now we have wtay as well ;) [15:06] <ajmitch> hi wtay [15:06] <arik> yo [15:07] <wtay> verweisung =3D link [15:07] Action: thomas considers taking his laptop out on the roof terrac= e at work to work on gstreamer [15:07] <arik> damn alsa [15:07] <big_T> a friend of my learned esperanto in school as a "pet proj= ect" [15:07] <arik> u gate alsa right now [15:07] <ajmitch> hmm, i must make the choice, sleep or hack? [15:07] <arik> er i hate [15:07] <thomas> wtay: can a plugin use pipe and fork ? [15:07] <arik> i hate alsa right now [15:07] <sienap> hack [15:07] <arik> hack! [15:07] <Uraeus> hack! [15:07] <thomas> ajmitch:sleep, hack with a fresh mind ! [15:07] Action: thomas ducks [15:07] <ajmitch> but /me has no coffee :( [15:07] <big_T> Uraeus, nope - haven't had the time... school is pretty n= asty right now, a lot of reports and a few exams coming up... but at the = end of next week freedom awaits - so then I will have the time (I hope ;) [15:08] <arik> damn alsa!!! [15:08] <Uraeus> big_T: you plan on redoing the OSS sink like the alsa pl= ugin too, right? [15:08] <ajmitch> such a pity that i'll only really have time for gstream= er after 0.2.0 is meant to be released ;) [15:08] <arik> heh [15:09] <arik> me too really [15:09] <Uraeus> ajmitch: well we do plan further releases so no need to = panic :) [15:09] <arik> heh [15:09] <arik> no [15:09] <arik> 0.2 is it [15:09] <arik> done [15:10] <Uraeus> arik: so the plan is for 0.2.0 to be the ultimate perfec= t solution then? [15:10] <ajmitch> Uraeus: yes, but it'll move to glib 2.0 then, i dunno i= f i can handle running such an experimental lib on my computer ;) [15:10] <arik> yep [15:10] <arik> perfect [15:10] Action: wtay attempts to create java bindings for gstreamer [15:10] <ajmitch> Uraeus: that is why we need your help [15:10] <arik> ajmitch: use the eazel hacking scripts ;-P [15:10] <Uraeus> ajmitch: I can make some nice RPMS for you of the new ex= perimental system :) [15:10] <wtay> thomas: what do you mean 'use a pipe'? [15:10] big_T (th...@fy...) left irc: Ping timeout for big_= T[fysgr763.sn.umu.se] [15:11] <ajmitch> Uraeus: no thanks, i'll just use the copy from cvs i co= mpiled yesterday ;) [15:11] <thomas> wtay: I mean using the "pipe" command so that I can wrap= that around afsink to catch the fd output... [15:11] <thomas> ... since unwrapping the audiofile lib to make an af "en= coder" looks a bit too difficult [15:11] <wtay> thomas: yes, you can, xmms does this and so does gstoss [15:11] <thomas> wtay: ok, and is it a good idea or is it overkill ? [15:11] <wtay> thomas: sounds like a good idea [15:12] Action: ajmitch thinks sleep is good option when one has to get u= p before 9am [15:12] <ajmitch> night all ;) [15:12] <thomas> wtay: ok, I'll try that then. btw, afsrc seems to work = as well. [15:12] <wtay> cool [15:12] Nick change: ajmitch -> ajzzzz [15:13] <thomas> LADTheres nothing to stop you using a novel protocol ove= r ethernet as well as [15:13] <thomas> TCP/IP. There not using low latency though, so there wou= ld be no [15:13] <thomas> particular reason for using a novel protocol." : " [15:13] Action: thomas is going to shoot every LAD poster [15:13] <wtay> thomas: yeah, shoot them twice :) [15:14] <thomas> gstreamer-launch attacksrc=3Dmachinegun ! peoplesink loc= ation=3DLAD-posters [15:15] <wtay> heh=20 [15:16] big_T (th...@fy...) joined #gstreamer. [15:16] <arik> alsa is pissing me off [15:17] <wtay> yo [15:17] <thomas> wtay: what is the "right" gstreamer way of outputting de= bug info and error info in plugins ? I want to clean up some of my stuff = ... [15:18] <wtay> GST_DEBUG (GST_CAT_PLUGIN_INFO, "something %d\n", an_int); [15:20] <thomas> anyone know what's up with the gstreamer Wiki ? styleshe= et seems missing... [15:24] <thomas> hrmph, now that I have something to contribute ;) [15:24] <arik> hehe [15:32] <big_T> arik, what problems are you having with ALSA? [15:32] <arik> oh [15:32] <arik> i upgraded to 2.4.2 [15:32] <arik> and rebuild alsa [15:32] <arik> and now i get unresolved symbol errors [15:35] <big_T> strange... what version of alsa do you have? [15:35] <big_T> did you do a 'make install'? [15:35] <arik> 0.5.11 [15:35] <arik> yes ;-) [15:36] <big_T> what modules are causing the error? [15:36] <arik> snd.o [15:38] <big_T> try doing a 'make clean', then removing config.cache and = rerunning configure [15:38] <arik> i did that ;-) [15:38] <big_T> okey [15:39] <big_T> what kernel-version did alsa's configure report? [15:39] <arik> 2.2.4 [15:39] <arik> er [15:39] <arik> 2.4.2 [15:39] <big_T> hmm... [15:39] <thomas> wtay: is there also a way of generating errors for plugi= ns ? [15:39] <wtay> gst_element_set_error I think [15:39] <big_T> arik, might be some alsa-issues with 2.4.2... I'm running= 0.5.10 with 2.4.4 without problems [15:40] <big_T> arik, what card do you have? [15:40] <arik> hmm [15:40] <arik> ymfpci [15:43] <big_T> arik, are there many unresolved symbols or just a few? [15:43] <arik> hmm [15:44] <arik> 10 or so [15:45] <big_T> okey... [15:46] <big_T> try a 'modprobe snd' [15:46] <arik> ok [15:46] <big_T> or 'modprobe snd-card-ymfpci' [15:46] <arik> the second one [15:46] <arik> is what caused the problem [15:46] <arik> snd does the same thing [15:46] <thomas> wtay: is it normal for a plugin to have GST_STATE_PENDIN= G (element) =3D=3D GST_STATE_READY be true at the *end* of it's run ? [15:47] <wtay> thomas: I don't think so, no... [15:47] <wtay> thomas: but then again, state changes are not yet fully im= plemented I think [15:48] <big_T> arik, this is strange... I'm sorry, but I can't think of = anything to solve this without even more info... and I'm no kernel/alsa-h= acker so... :( [15:48] icarus_ (icarus@i.read.your.email.kgb.fi) left irc: Ping timeout = for icarus_[i.read.your.email.kgb.fi] [15:48] <arik> hehe ;-) [15:49] <arik> that's ok [15:49] <arik> maybe i should try the beta alsa [15:50] ajzzzz (aj...@p2...) left irc: Ping timeout f= or ajzzzz[p28-max4.dun.ihug.co.nz] [16:11] <thomas> wtay: I noticed gstosssrc doesn't set caps on the src pa= d... does that mean bitrotten or does it mean something else ? [16:11] <wtay> bitrotten for sure [16:11] <thomas> wtay: ok, since I need it I'll try to fix it ;) [16:12] <wtay> cool [16:13] <big_T> hmm... anonymous cvs-access doesn't work... anyone know a= nything about this? [16:15] <wtay> hmm nope [16:16] <big_T> (sorry for this spam) [16:16] <big_T> cvs update: authorization failed: server cvs.gstreamer.so= urceforge.net rejected access to /cvsroot/gstreamer for user anonymous [16:16] <big_T> could it be sourceforge that is doing something or? [16:16] <arik> sourceforge is down i believe [16:16] ajzzzz (aj...@p2...) joined #gstreamer. [16:16] dobey (do...@dr...) joined #gstreamer. [16:17] <big_T> okey, got worried you was sick of "outsiders" or somethin= g ;-) [16:18] <arik> ;-) [16:18] <wtay> INFO(23357:-1): Initializing GStreamer Core Library [16:18] <wtay> INFO(23357:-1): CPU features: (c1c7f9ff) MMX 3DNOW MMXEXT=20 [16:18] <arik> we are, but we just kill on site ;-P [16:18] <wtay> /opt/src/sourceforge/gstreamer-HEAD/bindings/java/ jre Te= st [16:18] <arik> ooh [16:18] <arik> nice wtay=20 [16:18] <wtay> heh [16:19] <big_T> arik, then it's best to avoid meeting the developers face= -to-face then ;) [16:19] <arik> hehe [16:19] <arik> yes [16:20] <dobey> hrmm [16:20] Action: thomas tries to force ATI to hand over card docs & info t= o linux developers through mail [16:21] <thomas> a crontab "* * * * *" entry should do for the mailing ;) [16:21] <big_T> and they said mass-mail wouldn't accomplish anything! [16:21] <thomas> big_T: one to one works better ;) [16:23] <big_T> :) [16:24] <arik> bbl [16:24] arik (ar...@sd...) left #gstreamer. [16:29] <thomas> damn, ati doesn't have an e-mail support, they use forms= ... smart move ;) [16:29] <dobey> View->source [16:31] Action: big_T is reading up on oss [16:31] <big_T> hmm... oss-developers document is about 120pages... [16:32] <big_T> might take me a while ;) [16:44] <sienap> dobey=02:=02 i am sure you know were to get this checkin= g for gal >=3D 0.8.99.0... 0.7 found [16:44] <sienap> the gal version that is [16:44] <dobey> sienap: configure.in [16:45] <sienap> aah i need the package ;) [16:46] <sienap> not more then that [16:46] <dobey> ? [16:46] <dobey> cvs gal [16:47] <sienap> hmm [16:47] <sienap> let's see [16:47] <sienap> btw it seems that a lot of stuff in the evolution snapsh= ot channel has an brokken download url [16:48] Nick change: thomas -> thomas-ontheroof [16:48] <sienap> he [16:48] <sienap> :) [16:49] <dobey> i don't care, that's not my problem [16:58] <sienap> what a great motivation [16:59] <dobey> eh, if the snapshots were my responsibility, i might care [17:07] <Uraeus> hehe, I think I make midi support in GStreamer this week= s GNOME Summary POW [17:08] <dobey> Uraeus: not encompass? [17:10] <Uraeus> dobey: cause there still is a gtkhtml browser being deve= loped ;) [17:11] <dobey> Uraeus: huh? [17:12] <dobey> hrmm, brb [17:17] <taazzzz> Uraeus: perhaps you should let Omega work on the event = system first. that may end up turning into a general solution for comman= d data like midi [17:17] <dobey> Uraeus: ok, i'm back [17:23] <Uraeus> taazzzz: ok, if you think that is a good idea I drop it [17:23] <Uraeus> dobey: heh, I was just refering to the other browser I s= howed you [17:23] <dobey> Uraeus: oh, the one that doesn't work? [17:26] <Uraeus> dobey: heh, ok do you want me to request a new maintaine= r for Encompass as the POW? [17:26] <dobey> Uraeus: uh, i'm still the maintainer [17:26] <dobey> Uraeus: i just need help [17:26] <Uraeus> dobey: ok, I just got the impression you where stopping = development [17:27] <Uraeus> dobey: ok, mail me a specific task I can put into the su= mmary [17:27] <dobey> Uraeus: if i were stopping development, there wouldn't be= http://sourceforge.net/projects/encompass/ [17:27] <dobey> ;-) [17:27] <dobey> Uraeus: hrmm, i have one on there, look at that [17:28] <Uraeus> http://sourceforge.net/projects/encompass [17:28] <Uraeus> opps, wrong window :) [17:28] <dobey> heh [17:28] <dobey> dude, context menus [17:28] <dobey> that's what they are for [17:29] <Uraeus> hi, there is already 2 more developers on Encompass [17:29] <dobey> Uraeus: i need more people [17:29] <Uraeus> dobey: huh: what task? [17:29] <dobey> Uraeus: they are friends, and one of them is only writing= docs, part time [17:30] <dobey> http://sourceforge.net/people/?group_id=3D27705 [17:31] <Uraeus> dobey: context menues in X-chat don't work with Galeon [17:31] <dobey> Uraeus: haha [17:32] <dobey> galeon <uri> DOESNT WORK?! [17:32] Nick change: thomas-ontheroof -> thomas [17:33] <Uraeus> dobey: ok, I just hadn't bothered to configure x-chat :) [17:33] <big_T> Uraeus, check out http://gtkhtml2.codefactory.se :) [17:33] <dobey> Uraeus: i just suggested because putting POW as encompass= might help get me some people [17:33] <big_T> you could use that instead of the gtkhtml-widget [17:34] <dobey> Uraeus: and i'm going to implement it a help browser in i= t too [17:34] <dobey> big_T: uhm [17:34] <dobey> big_T: how do you mean? [17:34] <Uraeus> dobey: you want em to make a bonobo wrapper around gtkht= ml? [17:34] <dobey> Uraeus: yes, i've already started it, i need to finish cl= eaning it up and put it in cvs [17:35] <big_T> dobey, well - gtkhtml is an old widget, and gtkhtml2 is a= newer, better widget with smaller memory footprint and so on... [17:35] <big_T> jo...@co... is the maintainter [17:35] <big_T> unfortunately it is based on gtk1.3(cvs-version) [17:35] <dobey> big_T: no, gtkhtml2 has a larger memory footprint, and re= quires cvs glib, gtk, and libxml [17:35] <dobey> and that is not unfortunate [17:36] <dobey> that is perfect [17:36] <big_T> dobey, sorry... just learned it was the other way around = - perhaps one should ask before assuming :) [17:37] <dobey> big_T: it's ok [17:38] <dobey> big_T: and jonas is one of the two maintainers [17:39] <big_T> dobey, yeah - i know... Anders Carlsson is the other one [17:39] <big_T> just thought that jonas would appreciate "spam" more ;) [17:39] <dobey> big_T: besides i knew about gtkhtml2, and what it could d= o before the general public ;-) [17:40] <big_T> dobey, okey... well, I'm not the general public either so= ... [17:40] <dobey> big_T: eh [17:40] <big_T> dobey, do you work over at ximian? [17:40] <big_T> do...@dr... [17:40] <dobey> big_T: yes [17:41] <big_T> dobey, okey... not many have heard of gtkhtml2... but if = you work that it explains it ;) [17:41] <dobey> big_T: eh, well, gtkhtml2 is pretty public now [17:42] <dobey> big_T: i only started a few months ago (here at ximian) [17:42] <big_T> dobey, perhaps... no official release though...=20 [17:42] <big_T> accordin to jonas it'll be when gtk2 is closer to a relea= se [17:43] <dobey> big_T: yeah, it doesn't matter to me really, it will be u= sed when gtk2 is released [17:44] <big_T> okey [17:44] <dobey> big_T: i just need something now [17:46] <big_T> ??? [17:46] <dobey> big_T: a component [17:46] <dobey> something i can ship, and will work with what we ship [17:50] <sienap> looks at his ximian monkey at his monitor [17:50] <dobey> huh? [17:50] <Uraeus> dobey: have Ximian stopped selling t-shirts and monkeys? [17:51] <dobey> Uraeus: no [17:51] <Uraeus> dobey: I can't find any links for it anymore on the Ximi= an pages [17:52] <dobey> Uraeus: mail webmaster@ or hello@ or something then [17:54] <Uraeus> dobey: why don't you just peak out of your cubicle over = to the web guy and scream: hi, braindead fix the ordering pages or get a = dos attack on your desktop! [17:54] <dobey> Uraeus: uhm [17:54] <Uraeus> that probably works better than a mail from me [17:54] <dobey> blah [17:54] <dobey> brb [17:57] <big_T> dobey, okey... you lost me there for a while, but now I'm= with you... [17:58] <dobey> heh [17:58] <sienap> working in cubes must be boring :) [17:58] <sienap> so lonely and stuff :) [17:58] <big_T> dobey, and yes - I agree with you there, better to use so= mething people can actually use themself... gtkhtml2 depends on the cvs-v= ersions of a lot of libs... which makes it's use somewhat limited for non= -bleeding-edge-people [18:01] <Uraeus> sinap: don't think most people work in cubicles anymore = it was a early 90'ies thing [18:01] <dobey> Uraeus: depends on how you define cubical [18:02] <Uraeus> dobey: you actually have cubicles at Ximian? [18:02] <dobey> Uraeus: yes [18:02] <Uraeus> oh, hmm [18:02] <dobey> Uraeus: i'm sitting on a beanbag right now though [18:03] <sienap> dobey relaxing :) [18:03] <Uraeus> hehe [18:03] <dobey> sienap: trying to, have you seen mandrake 8? [18:03] <sienap> dobey why not making some photo's at ximian and show the= m on internet :) [18:03] <sienap> dobey nope sorry.. [18:03] <dobey> sienap: i don't have a camera [18:04] <sienap> dobey is the work ambience nice their ? [18:04] <dobey> sienap: good for you, you don't want to see it [18:04] <dobey> sienap: yes [18:04] Action: big_T goes home for the day [18:05] Action: taazzzz thinks you all need a #gstreamer-talk channel for= this non-gstreamer chatter... [18:05] Action: taazzzz needs more zzz's [18:06] <sienap> dobey he cool :) [18:06] <Uraeus> taazzzz: well, we usually shut up when someone want to t= alk shop ;) [18:06] <sienap> uraeus can you explain me what this rubberband selection= is in nautilus ? [18:07] <dobey> sienap: the pretty box you see when you click and drag to= select a bunch of things [18:08] <sienap> ooh [18:08] <sienap> that is just normal selection [18:08] <sienap> however a little bit too pretty [18:09] <sienap> it slows the stuff way too much [18:11] <dobey> Uraeus: let me know when the status report goes up [18:11] <dobey> sienap: it's faster than non-pretty on my box [18:12] <sienap> ooh wait now i got it.. [18:12] <sienap> i though you meant the selection box it self [18:12] <sienap> that is way slow [18:12] <dobey> eh, hrmm [18:12] <dobey> i dunno [18:13] <dobey> i was talking about the selection rubber band [18:13] <sienap> he [18:13] <sienap> :) [18:15] <dobey> Uraeus: you can get monkies and shirts at thinkgeek [18:17] <sienap> i still need to get a shirt [18:17] <sienap> when we do the next giant thinkgeek order :) [18:25] Action: Uraeus prepares to be flamed in case his CVS mail is not = welcome [18:25] <sienap> where what how ? [18:25] omega_away (om...@bo...) left irc: Pi= ng timeout for omega_away[boisdslgw5poolb76.bois.uswest.net] [18:26] <Uraeus> sienap: gnome-devel, galeon-devel and gnome-i18n [18:26] <sienap> what did you post ? [18:26] <sienap> btw i am off [18:27] <sienap> will read in some hours [18:27] <sienap> :) [18:36] mattsm (ma...@ad...) left irc= : Read error to mattsm[adsl-64-123-206-39.dsl.hstntx.swbell.net]: EOF fro= m client [18:46] ChiefHighwater (pa...@te...) joined #gstreamer. [18:46] #gstreamer: mode change '+o ChiefHighwater' by ChanServ!s@ChanSer= v [18:46] #gstreamer: mode change '-o ChiefHighwater' by ChiefHighwater!pau= l...@te... [18:47] <dobey> hmm [18:56] thomas (th...@21...) left irc: Ping ti= meout for thomas[212-100-172-175.adsl.easynet.be] [19:03] <Uraeus> who here has GStreamer CVS acccess? [19:03] <dobey> i would [19:03] <Uraeus> dobey: can I DCC you the new MAINTAINER file so you can = add it to GStreamer CVS? [19:03] omega_rr (om...@bo...) joined #gstrea= mer. [19:04] <Uraeus> hi omega_rr [19:04] <omega_rr> yo [19:04] <wtay> yo [19:04] <dobey> Uraeus: i can't add it [19:04] <dobey> Uraeus: i don't think wtay will add me [19:04] <omega_rr> I reordered PLAYING and PAUSED in INCSCHED last night [19:04] <Uraeus> omega_rr: can I send you the Maintainer file for inclusi= on in GStreamer CVS? [19:04] <omega_rr> and reworked the thread interlock froms cratch [19:04] <omega_rr> Uraeus: yup [19:04] <wtay> oh [19:05] <wtay> I made JAVA bindings :) [19:05] <Uraeus> wtay: cool :) [19:05] <Uraeus> wtay, omega_rr: any recent updates that I should reflect= in the roadmap? [19:05] <wtay> I can create elements and add them to a bin :-) [19:06] <omega_rr> Uraeus: um, sec.. [19:06] <omega_rr> Uraeus: INCSCHED->HEAD merge is today [19:07] Action: omega_rr gets laptop out [19:09] <omega_rr> Uraeus: I'd leave GObject list empty until we get the = Pioneer guys onboard, since they're planning on doing that [19:10] <Uraeus> omega_rr: we are getting some Pioneer hackers onboard? [19:11] <wtay> Uraeus: yup, the real ones :) [19:13] Parapraxis (Spluzz@165.247.57.116) joined #gstreamer. [19:13] <Parapraxis> yo [19:13] <wtay> ho [19:19] Action: Uraeus just mailed RMS for some license advice [19:19] <omega_rr> Uraeus: yup [19:20] <omega_rr> anyone awake have a dual machine? [19:20] <dobey> omega_rr: depends on how you mean that [19:20] <omega_rr> dual-processor, SMP kernel [19:21] <dobey> omega_rr: yes, "what for?" [19:21] <omega_rr> thread-lock testing [19:21] <omega_rr> stressing the locking primitives used by the thread [19:22] Action: taazzzz is awake [19:22] <omega_rr> taazzzz: you can test thread interlock <g> [19:23] <dobey> oh [19:23] <dobey> i don't have an smp machine to do that on [19:24] <dobey> damn [19:24] <dobey> sourceforge is slow [19:25] <Uraeus> hehe, has anyone ever tried mailing RMS? he hasn't a rat= her funny auto-reply :) [19:27] omega_rr (om...@bo...) left irc: Ping= timeout for omega_rr[boisdslgw5poolb76.bois.uswest.net] [19:27] <dobey> Uraeus: no [19:27] <dobey> Uraeus: when do the summaries go up? [19:28] <Uraeus> dobey: monday, but the template is already in CVS [19:28] <dobey> ahh, ok [19:29] <Uraeus> RMS email autoreply: I am not on vacation, but I am at t= he end of a long time delay.=A0 I am [19:29] <Uraeus> located somewhere on Earth, but as far as responding to = email is concerned, [19:29] <Uraeus> I appear to be well outside the solar system. [19:29] omega_rr (om...@bo...) joined #gstrea= mer. [19:29] <omega_rr> stupid network [19:30] <dobey> heh [19:34] <omega_rr> ok, INCSCHED has state reorder [19:34] Nick change: omega_rr -> omega_lunch [19:34] <wtay> omega_rr: you think some plugins will break? [19:35] <omega_lunch> I fixed 3 or 4 that referred to the wrong states [19:35] <omega_lunch> check them out [19:35] <omega_lunch> test test test while I eat [19:35] <omega_lunch> brb [19:35] <Uraeus> need food, might be back later [19:35] Uraeus (csc...@c2...) left irc: syntax error -= user imploded [19:35] Zeenix (pro...@ho...) joined #gstreamer. [19:35] <wtay> yo [19:37] <Zeenix> i want to ask on sockets: which channel i should go [19:37] <Zeenix> ? [19:37] <wtay> no idea... #unix maybe? [19:38] <big_T> Zeenix, rtfm ;) [19:38] <big_T> Zeenix, just joking [19:38] <Zeenix> can i discuss it here [19:39] <sienap> try :) [19:39] <Zeenix> can i control the connection speed of the socket [19:40] <Zeenix> i mean data transfer speed [19:40] <wtay> you mean: "does the socket do that for me"? [19:40] <wtay> no [19:41] <Zeenix> is there a way round [19:41] <wtay> yes, make sure to only send data to it at the rate you wan= t [19:41] <Zeenix> you mean i should keep another layer for that [19:42] <wtay> hmm, yes I suppose [19:46] Zeenix (pro...@ho...) left #gstreamer. [19:51] Action: wtay listens to an mp3 pipeline constructed in JAVA <g> [19:53] <sienap> hell yeah :) [19:53] <dobey> hrmm [20:12] maYam (ma...@ca...) joined #gstreame= r. [20:12] <maYam> hellooh [20:12] <wtay> hey! [20:13] <wtay> maYam: what's up? [20:13] <maYam> let the evening begin [20:14] <wtay> yes :) [20:14] <dobey> WHOOHOO! [20:14] <maYam> dobey! ..oceans of love.. [20:15] <dobey> heh [20:15] <dobey> hey baby [20:15] <maYam> how's life these days? [20:15] <maYam> sad i may hope? [20:15] <dobey> can i get a hug? [20:16] <maYam> hmm.. depends.. [20:16] <maYam> i'll need something in return [20:16] <dobey> what? [20:16] <maYam> let me think about that a minute.. [20:17] <dobey> ok [20:17] <maYam> ah i know! [20:17] <dobey> yeah? [20:17] <maYam> the sawfish theme selector needs to be fixed [20:17] <maYam> i need it [20:17] <wtay> lol [20:18] <dobey> oh, ok, that's planned [20:18] <maYam> now! [20:18] <dobey> fly here and give me a hug, and i'll get right on it [20:18] <maYam> you planned it?? plans plans.. we want to see results [20:19] <maYam> i need a date first :) [20:19] <dobey> date for what? i'll be your date [20:20] hadess (ha...@pc...) joined #gst= reamer. [20:20] <maYam> so is that what you usually do, dobey, when you get a hug= from a girl you start hacking a bit? ;) [20:20] <dobey> maYam: no, i need a hug [20:20] <hadess> hey gizzahs [20:20] <dobey> yo hadess [20:20] <wtay> hadess: yo dude [20:20] <maYam> hey hadess [20:20] <hadess> hey wtay, dobey, maYam [20:22] <maYam> hey hadess i hear you're working on a new XMMS? [20:22] <dobey> maYam: so... [20:22] <maYam> that's a Good thing [20:22] <maYam> *claps* [20:22] <hadess> maYam: a replacement actually [20:23] <maYam> even better! [20:23] Action: dobey feels no love [20:23] <hadess> you don't like xmms ;) [20:23] <maYam> aaargh no! [20:24] <maYam> dobey, isn't it more fun for you to get sponatenous hugs? [20:24] <hadess> i must say that the interface is dodgy [20:24] <maYam> in stead of having to ask for it? [20:24] <dobey> maYam: it would, yes, if i could get one [20:24] <maYam> hadess: dodgy is a nice word for what it is ;) [20:25] <dobey> maYam: you are more than welcome to fly here spontaneousl= y and hug me [20:25] <maYam> ok dobey.. [20:25] <maYam> (dobey) [20:25] <dobey> hrmm [20:25] Action: hadess watches lotr's new trailer [20:26] <ChiefHighwater> hehe...is on my desktop 8-] [20:26] Nick change: omega_lunch -> omega_rr [20:26] <dobey> maYam: i'm not feeling it :-( [20:26] <omega_rr> anyone find problems with incsched state swap? [20:27] <maYam> dobey new in cyberworld? [20:27] <maYam> yes.. you'll have to use your imagination a bit, dobey [20:27] <dobey> maYam: no [20:27] <dobey> maYam: i'm tired of "imagination" [20:27] <wtay> omega_rr: yes, I have problems [20:28] <wtay> omega_rr: gstplay locks up 50% of the time [20:28] <maYam> dobey: don't they have women in the us? [20:28] <dobey> maYam: i would think so, where are they? [20:28] <omega_rr> wtay: ? [20:28] <omega_rr> which media types? [20:29] <wtay> omega_rr: all [20:29] <omega_rr> um [20:29] <ChiefHighwater> so many women, and none as good as mine 8-] [20:29] <wtay> looks like a thread state change lock [20:29] <maYam> dobey: maybe you live in the middle of cornfields? [20:29] <dobey> maYam: there are no cornfields in boston [20:29] <maYam> ooh ChiefHighwater.. that's nice :) [20:29] <ChiefHighwater> maYam: txs 8-] [20:29] Action: dobey hurts chw [20:29] <omega_rr> wtay: hrm, ya think? <g> [20:30] <wtay> omega_rr: :-) [20:30] <wtay> omega_rr: hmmm [20:30] <dobey> you aren't helping here [20:30] <maYam> lol [20:30] <omega_rr> wtay: what's the last line you see before the lock? [20:30] <wtay> omega_rr: too bad it doesn't happen with debug on [20:30] <maYam> dobey, you recognize women, right? [20:30] <wtay> PLAYING->PAUSED: element "mad" has parent "thread_mp3parse= " and sched 0x8114490 [20:30] <wtay> PLAYING->PAUSED: element "thread_mp3parse" has parent "new= _element" and sched 0x8114490 [20:31] Action: ChiefHighwater gives dobey a big chocolate chip cookie [20:31] <omega_rr> no, with debug [20:31] <dobey> maYam: yes [20:31] <maYam> well then, get them!! [20:31] <dobey> maYam: HOW! [20:31] <wtay> omega_rr: it doesnt happen with debug on [20:31] <maYam> go out! [20:31] <omega_rr> neat [20:31] <wtay> plays fine then.. [20:31] <hadess> dobey: fishnet i think [20:32] <dobey> hadess: *smack* [20:32] <maYam> dobey: i tried to tell you about bars the other time.. [20:32] <maYam> *he just won't listen to me* sigh.. [20:32] <dobey> maYam: we have this line here in america called age, it's= something about 21 [20:32] <dobey> maYam: and i don't want an alcoholic [20:33] <maYam> so you're looking for a single sober female.. in Boston? = good luck! [20:33] <omega_rr> taazzzz: you fully awake yet?? [20:33] <wtay> omega_rr: it doesn't seem to happen with mp3/ogg [20:33] <dobey> maYam: whatever [20:34] <omega_rr> what does it occur most with? mpeg2 system? that's wha= t I'm using (CLIP04.VOB) [20:34] <wtay> omega_rr: mpeg1 [20:34] <dobey> maYam: not everyone is > 21 [20:34] <omega_rr> ok, trying AlienSong [20:34] <ChiefHighwater> dobey: you may think it sounds dumb, but it work= s very well...the best way to get someone to invest in you is to invest i= n them. i.e. Don't look to take, look to give. [20:34] <maYam> dobey: i thought you wanted a woman, you didn't ask for s= pecifics.. [20:34] <wtay> ChiefHighwater: wise words... [20:34] <dobey> ChiefHighwater: i'm not looking to take [20:34] <dobey> ChiefHighwater: i just haven't found anyone who will let = me give [20:35] <omega_rr> wtay: does it happen during playout, or when you hit p= ause? [20:35] <wtay> omega_rr: playout [20:35] <omega_rr> hmm [20:35] <maYam> dobey: how old are you? [20:35] <wtay> omega_rr: at startup [20:35] <dobey> maYam: 20 [20:35] Action: omega_rr has no troubles at all [20:35] <wtay> omega_rr: uhm [20:35] <maYam> damn i had a whole speech ready in case you were only 16 [20:35] <omega_rr> wtay: dblcheck how you match incsched [20:36] <omega_rr> maYam: hehehe [20:36] <dobey> maYam: hah, what was that? [20:36] <wtay> omega_rr: doing that [20:36] <maYam> 20 should be a beautiful age.. [20:36] Action: omega_rr is running INCSCHED 100%, no local changes, just= built [20:36] <dobey> maYam: "should" [20:36] Action: hadess readies to go to the pub [20:36] <wtay> omega_rr: same here, only mad has a fix and mixer.c [20:36] <dobey> hadess: :-( [20:36] <maYam> yeah! oh.. but for men i think it comes a bit later.. [20:36] <omega_rr> what are those? [20:36] <maYam> 23 or 24 maybe.. [20:37] <hadess> maYam: like 22 ;) [20:37] <maYam> so you just have to wait, dobey ;) [20:37] <dobey> maYam: what does that have to do with 16? [20:37] <wtay> omega_rr: checking.. [20:37] <dobey> maYam: i'm tired of fucking waiting [20:37] <maYam> yeah hadess.. because you're so mature.. lol [20:37] <hadess> maYam: no laughing allowed :P [20:37] <wtay> omega_rr: I removed the layer porperty on the sink pad [20:38] <omega_rr> ok, shouldn't affect anything [20:38] <maYam> nah 22 is great... [20:38] <maYam> and i hear hadess has a girlfriend? [20:38] <dobey> maYam: so what's your rant about 16? [20:38] <hadess> maYam: had actually [20:38] <wtay> omega_rr: lemme checkout mad again [20:39] <maYam> dobey: at 16 it's demanded that you're depressed.. now yo= u're too old for that.. [20:39] <omega_rr> hehehe [20:39] <dobey> maYam: at 16 i wasn't depressed [20:39] <maYam> hadess: wow! heheh.. that didn't last for long! [20:39] <wtay> omega_rr: same [20:39] <dobey> maYam: or was less so than i am now [20:39] <hadess> maYam: not my fault, that's hers [20:40] <maYam> ah hadess you'll find someone at the pub tonight :) [20:40] <maYam> all women are bitches i say [20:40] <dobey> no [20:40] <wtay> omega_rr: lemme attach gdb to it when it is locked... [20:40] <omega_rr> taazzzz: ping??? [20:40] <dobey> maYam: is that why you aren't a lesbian? [20:40] <hadess> maYam: she spitted me all the bullshit about her still n= ot being over her old boyfriend, and yadda, yadda [20:41] <ChiefHighwater> isn't it amazing how wtay and omega_rr keep work= ing amidst all this 8-] [20:41] <maYam> hadess: oh dear.. that's terrible.. best to get rid of h= er then.. [20:41] <maYam> (i hate it when they do that) [20:41] <maYam> dobey: exactly [20:41] Action: wtay looks for the gdb option to attach to a running proc= ess [20:42] <dobey> ChiefHighwater: no, i'm working too [20:42] <omega_rr> gdb -p pid [20:42] <dobey> maYam: so.. then...=20 [20:42] <wtay> gdb: unrecognized option `-p' [20:42] <hadess> wtay: -P ? [20:42] <omega_rr> hmm [20:42] <maYam> yes dobey? [20:42] <wtay> nope [20:42] <hadess> wtay: gdb [options] [executable-file [core-file or proce= ss-id]] [20:43] <hadess> wtay: so it's "gdb PID" [20:43] <dobey> maYam: i still don't see how all this is supposed to help= me [20:43] <wtay> This GDB was configured as "i686-pc-linux-gnu"...21291: No= such file or directory. [20:43] <hadess> wtay: or gdb /path/to/bin PID [20:43] <maYam> dobey: it isn't.. i'm no good at helping people [20:43] <omega_rr> gdb\nattach pid\ [20:44] <dobey> maYam: you seem to help wtay a lot [20:44] <wtay> omega_rr: hum, that works... but... [20:44] <wtay> omega_rr: no symbols [20:44] <maYam> dobey: i do? lol.. [20:44] <dobey> maYam: he's not depressed [20:45] <maYam> dobey: wtay is not the kind of guy to get depressed.. not= even slightly sad [20:45] <omega_rr> attach pid progname ? [20:45] <omega_rr> libtool gdb [20:45] <omega_rr> progname =3D .libs/lt-gstmediaplay [20:45] <dobey> maYam: of course not, he has you [20:46] <maYam> dobey: he's always been like that.. there's no way i coul= d change him [20:46] <dobey> maYam: then come over here [20:46] <dobey> ;-P [20:46] Action: hadess is gone [20:46] <maYam> bye hadess [20:47] <hadess> cya later, i'm going to watch a nifty football match :P [20:47] <maYam> dobey: lol [20:47] hadess (ha...@pc...) left irc: m= ooooh! [20:47] Action: maYam is gone too [20:47] <maYam> buh bye! [20:47] <dobey> :~( [20:47] maYam (ma...@ca...) left #gstreamer. [20:47] <wtay> omega_rr: ok got something [20:49] <wtay> #0 0x405efbde in sigsuspend () from /lib/libc.so.6 [20:49] <wtay> #1 0x403cb6eb in __pthread_wait_for_restart_signal () fro= m /lib/libpthread.so.0 [20:49] <wtay> #2 0x403c80c2 in pthread_cond_wait () from /lib/libpthrea= d.so.0 [20:49] <wtay> #3 0x40037cd2 in gst_queue_get (pad=3D0x8115310) at gstqu= eue.c:410 [20:49] <wtay> #4 0x4003124b in gst_pad_pull (pad=3D0x80a8188) at gstpad= .c:1475 [20:50] <omega_rr> hmmm [20:50] <omega_rr> matth and I are gonna walk through the locking rework = on paper [20:50] <wtay> yup [20:51] <wtay> looks like a state change when the queue is holding the lo= ck again [20:51] <omega_rr> right, interruptability didn't happen somehow [21:02] Nick change: taazzzz -> taaz [21:04] <omega_rr> taaz..... [21:04] <omega_rr> need you to try INCSCHED1 on a dual machine ASAP [21:04] <taaz> yeezzzz? [21:05] <taaz> hmm... challenge will be to do testing remotely ;) [21:05] <omega_rr> hmm [21:05] <taaz> whats up with this broken sf anon cvs access? [21:05] <omega_rr> no idea [21:05] <omega_rr> ask sf [21:05] <dobey> whee [21:05] <omega_rr> I have no problems <g> [21:06] <wtay> omega_rr: I was wondering wheter the _create_plan and _sch= edule are still needed in GstBin [21:06] <omega_rr> nope [21:06] <wtay> good [21:06] Action: wtay removes that code [21:08] <wtay> uhm, it's still used [21:08] <taaz> hmmm... sf shell dead and expected back no later than the = 28th [21:08] Action: wtay smacks himself for looking in HEAD [21:08] <omega_rr> I thought SF has mega hardware?? [21:09] <taaz> even mega hardware breaks [21:09] <omega_rr> no, I mean mega hardware as in failover machines [21:09] <ChiefHighwater> but not usually all of it at the same time [21:10] <taaz> speaking of mega hardware... someone should look into tha= t /. article tidbit about IBM giving access to 10 processor mainframes...= you know you want a S/390 gstreamer port ;) [21:11] <wtay> taaz: yeah :) [21:14] <taaz> is it okay to have elements that process different parts o= f a pipeline? for instance one chain func pad that does some stuff and f= orwards to output and another chain sink that takes data from another par= t of the pipeline? [21:15] <omega_rr> wtay: update gstthread.c and try again [21:15] <wtay> sec.. [21:16] <wtay> looks fixed.. [21:16] <wtay> yup, can't cause it to hang now.. [21:16] <omega_rr> taaz: any chance of a vfb X server to test on dual ? [21:16] <taaz> hmm? [21:16] <taaz> i can run X apps... i just wont be able to see them ;) [21:17] <omega_rr> that's fine [21:17] <omega_rr> just run with --gst-mask=3D0x00200000 [21:17] <omega_rr> if it wedges inside a thread (thread_main_loop), it's = broke [21:24] <taaz> oops, gotta deal with new configure stuff before i can bui= ld. (sorry if i'm slow about this... trying to do other things at same t= ime ;) [21:39] <omega_rr> taaz: getting there? [21:39] <taaz> building... [21:39] <taaz> my automake 1.4-p2 bug was closed... guess the debs should= show up soon [21:40] <taaz> wtay: that will fix the AC_CONFIG_FILES vs AC_OUTPUT bug [21:43] <wtay> finally.. [21:45] <taaz> anyone want to debug makefiles? test/Makefile.am is causi= ng those apps to link with same libs many times. this works of course bu= t its makes the linking really slow [21:46] <taaz> and i dont think that -lXv and -Xxf86dga stuff is needed a= nymore? [21:47] <wtay> taaz: yup, seen that, fixing it now [21:53] <omega_rr> taaz: getting it compiled? <g> [21:54] <taaz> yeah [21:55] <wtay> taaz: what's so good about a dual when it's that slow? <g> [21:56] <omega_rr> hehe [21:56] <taaz> its not that slow, its just failing on certain things [21:56] <omega_rr> we might get a dual for just this purpose at rr [21:57] <taaz> like alsa wont compile with 0.9 headers i have installed [21:57] <taaz> and that stuff you just fixed up wtay [21:57] <omega_rr> right [21:57] <taaz> i started to convert alsa... then realized i couldnt find= any of the pcm api calls in headers i had and gave up ;) [21:58] <omega_rr> heh [21:58] <omega_rr> whoops [21:58] <wtay> omega_rr: is a dual all they can afford? <g> [21:58] <omega_rr> heh [21:58] <wtay> somthing with 32 CPUs seems more appropriate :-) [21:59] <omega_rr> it'd be nice to have a quad, just because the 2-way co= ntention case isn't the only one... [21:59] Action: taaz thinks this is amusing... 2 people from different co= mpanies with budgets asking me to test out SMP on my <$1k machine ;) [21:59] <omeg... [truncated message content] |
From: <wim...@ch...> - 2001-05-27 04:31:03
|
[06:31] omega_ (omega@216.99.212.251) joined #gstreamer. [06:31] <omega_> wow. I managed to smack the gnome panel into submission finally!! [06:32] <omega_> erm, sorta [06:33] <omega_> one more try [06:33] omega_ (omega@216.99.212.251) left irc: [x]chat [06:33] omega_ (om...@om...) joined #gstreamer. [06:33] omega_ (om...@om...) left irc: [x]chat [06:34] omega_ (omega@216.99.212.251) joined #gstreamer. [06:34] <omega_> better, but now it's stupid in another way... oh well, I can live with that [06:37] omega_to_pdx (om...@bo...) left irc: Read error to omega_to_pdx[boisdslgw5poolb76.bois.uswest.net]: EOF from client [06:38] <taaz> omega_: what's next on your todo list? [06:38] <omega_> um, sleep is near the top [06:38] <omega_> I ended up going through Oakland [06:40] <omega_> taaz: yes, there's gonna be a cothread limit for the main thread, but we can increase the stacksize on new threads [06:42] <omega_> taaz: the 3-pad nav should work, even if it was a single loopfunc [06:42] <omega_> since that's what the sched is gonna do to it right now anyway <g> [06:43] <taaz> i imagine there'd be problems if inputs come from different threads... would have to do local element data lockign or somethnig then [06:43] <omega_> true [06:43] <Parapraxis> how are the gst docs generated? [06:43] <omega_> keep in mind that you can create separate instances of entirely different elements, and as long as they have the same plugin, they can share data [06:43] <omega_> Parapraxis: via gtkdoc [06:44] <taaz> huh? [06:45] <omega_> you have a plugin struct [06:45] <omega_> from each instance of every type of element in that plugin, you can walk pointers to get to that one struct [06:45] <omega_> so you can even have a navigator element and a privstreamsink element [06:45] <omega_> and they can share data and locks [06:46] <omega_> walken: 'sup? [06:48] <walken> yop [06:48] <walken> hmmmm not much [06:48] <walken> low energy today [06:48] <walken> need to go out and try to have fun [06:50] omega_ (omega@216.99.212.251) left irc: Read error to omega_[216.99.212.251]: Connection reset by peer [06:50] <walken> need to find a new GF [06:50] <taaz> long distance attempt failed? [06:52] <walken> heh [06:52] <walken> not that its really a surprise but heh... :-/ [06:56] omega_ (om...@om...) joined #gstreamer. [06:56] <omega_> hrm, that was smart. [06:56] <omega_> note to self: always check which machine is showing on the problem before typing 'halt' [06:57] <taaz> been there, done that [06:58] <walken> hehe [06:58] <walken> once I halted a server at work [06:58] <walken> ooops [07:04] <omega_> hrm, that's a new one: gimme a good one-word name for 'high-energy, short-duration event' [07:05] <Parapraxis> burst? [07:05] <omega_> the word that came to mind was 'explosion', but yeah [07:06] <omega_> burst is the word these people use... less of a loaded term [07:07] Action: walken is used to the burst word [07:07] <Parapraxis> is there a nice way to handle GObjects in asm, or should I just pass asm code a pointer to the first field of the class (skipping its parent classes) [07:07] <Parapraxis> er... GObject derived classes that is [07:07] <omega_> um, dunno [07:07] <omega_> ask walken <g> [07:08] <Parapraxis> walken: ? :) [07:09] <walken> you should do it in cobol [07:09] <omega_> er [07:09] Action: omega_ will veto cobol bindings for gst [07:09] Action: walken just realized none of his ac3 test streams use delta bit allocation [07:09] <omega_> oops [07:09] <walken> its bad since I'm hacking the delta bit allocation code right now [07:09] <omega_> um, yeah [07:10] <walken> (not that I have any chance to break it of course. *cough*) [07:10] <taaz> you think you should start using cvs for all this stuff? [07:10] <walken> I do [07:10] <walken> locally :) [07:10] <omega_> mu [07:10] <walken> guess I need to upload it sometime though [07:10] <omega_> bad walken, not releasing early and often [07:11] Action: omega_ hides [07:11] <omega_> taaz: now that you're home, have you verified the threadlocking? [07:11] <walken> anyone has low-bandwidth ac3 streams ? [07:11] <omega_> at least, I assume you're home, it's 1:12am [07:11] <walken> say <256 at least [07:12] <omega_> walken: um, does ac3dec 0.6.1 print out the bw ? [07:12] <walken> a stereo stream would be nice too [07:12] <omega_> hrm, not tha tI have it [07:12] <omega_> try gstreamer.net/media/thx.ac3 [07:12] <walken> export AC3_DEBUG=1 [07:12] <omega_> no ac3dec [07:12] <walken> should do it [07:13] Action: omega_ waits for 0.6.2 [07:14] <taaz> omega_: yeah i tested a bit, things seemed to work [07:14] <omega_> did you test several times? thread lock problems tend to take a while to show [07:15] <taaz> is there an automagic stress test app i could run? (hint hint) [07:15] <omega_> working on it [07:16] <taaz> need to fix gstmediaplay so you can run it from outside gstplay dir too [07:16] <omega_> yeah [07:17] <omega_> tests/threadlock.c: update,build,run [07:18] <taaz> the lat queue test still doesnt work... not sure if it even should though ;) [07:18] <omega_> hmm, I'll have to check [07:19] <taaz> tl is dumping lots of text [07:19] <omega_> yeah [07:19] <omega_> let me know if it ever stops [07:21] Nick change: omega_ -> omega_sleepy [07:22] <omega_sleepy> try adding a sleep(1) after each of those two state changes in the while(1) [07:23] <omega_sleepy> that will probably make it more likely that the main process and thread really are running at the same time, and not just lockstepping [07:23] <taaz> ok... runs much more efficiently on a console... else X takes too much cpu scrolling ;) [07:23] <omega_sleepy> heh [07:23] <omega_sleepy> you can turn off the debug if it's too much of a problem (could be, contention and all) [07:24] <taaz> ya know... it is getting slowly bigger [07:24] <omega_sleepy> hmm [07:25] <omega_sleepy> update threadlock.c [07:25] <omega_sleepy> mine isn't changing size [07:26] <omega_sleepy> oh, wait [07:26] <omega_sleepy> right.... there's going to be a leak in some cases caused by debug [07:26] <taaz> uh oh... it stopped [07:26] <omega_sleepy> um [07:26] <omega_sleepy> last line? [07:26] <taaz> err.. just a sec... need to copy from console ;) [07:27] <taaz> DEBUG(30707:-1)gst_thread_change_state:398: peer element isn't DECOUPLED [07:27] <taaz> DEBUG(30707:-1)gst_thread_change_state:419: [thread] sync(30721): waiting for thread to stop spinning [07:28] <omega_sleepy> hm [07:28] <omega_sleepy> what state is it changing to? (look up) [07:28] <taaz> DEBUG(30721: 0)gst_thread_main_loop:571: [thread] sync-main(30707): thread in PLAYING state, waiting for PAUSED [07:28] <taaz> ? [07:28] <omega_sleepy> hmm [07:28] <omega_sleepy> ok, this is the most likely point for problems.... [07:28] <taaz> keep in mind these sorts of tests may expose bugs in the hardware ;) [07:28] <omega_sleepy> heh [07:29] <omega_sleepy> I doubt it [07:29] <omega_sleepy> but underclock if you're not positive [07:30] <taaz> this new one isnt as fun to watch [07:30] <omega_sleepy> new one? [07:30] <taaz> with the sleep [07:30] <omega_sleepy> oh [07:31] <omega_sleepy> to limit debug to what's relevant, add --gst-mask=0x00200040 [07:31] <omega_sleepy> then if you can repeat the lock reliably, dump the entire debug to a file [07:33] <taaz> el halto againo [07:33] <omega_sleepy> hmmm, sec [07:33] <taaz> DEBUG(31228:-1)gst_thread_change_state:398: peer element isn't DECOUPLED [07:33] <taaz> DEBUG(31228:-1)gst_thread_change_state:419: [thread] sync(31242): waiting for thread to stop spinning [07:33] <omega_sleepy> yup [07:33] <taaz> INFO (31228:-1)gst_thread_change_state:266: [thread] sync(31242): changing state from PLAYING to PAUSED [07:33] <taaz> DEBUG(31242: 0)gst_thread_main_loop:556: [thread] sync-main(31228): removed spinning state due to failed iteration! [07:33] <taaz> INFO (31228:-1)gst_thread_change_state:341: [thread] sync(31242): pausing thread [07:33] <omega_sleepy> um [07:33] <omega_sleepy> failed? [07:34] <taaz> up a little bit from where it stopped [07:34] <omega_sleepy> update again and start recording [07:34] <omega_sleepy> 2>&1 | tee lock.out [07:35] <omega_sleepy> with 00200040 mask [07:35] <taaz> i made the fake* elements silent... [07:35] <omega_sleepy> so did I ;-) [07:39] <taaz> sent failure log to you [07:40] <omega_sleepy> ok [07:40] <taaz> run from before your last commit [07:41] <omega_sleepy> hrm, can you attach it instead? [07:41] <omega_sleepy> all the colors get fragged, makes it very hard to read [07:42] <taaz> heh... thought of that 1s after sendnig ;) [07:42] <omega_sleepy> :419: ^[[04m[thread]^[[00m^[[00m sync(^[[00;36m32675^[[00m): waiting for thread [07:42] <omega_sleepy> to stop spinning [07:42] <walken> whats the .wav files in audio-test ? [07:42] <walken> is it compressed at all ? [07:42] <omega_sleepy> nope [07:42] <omega_sleepy> just wavs [07:42] <walken> no fun then [07:43] Action: walken wants low-bitrate ac3s [07:43] <walken> bouhouhoubouuuu [07:43] <omega_sleepy> thx.ac3 no use? [07:43] <omega_sleepy> or 404? [07:43] <walken> 404 ? [07:43] <omega_sleepy> was it there? [07:44] <walken> yeah.... [07:45] <omega_sleepy> ok [07:45] Action: walken looking [07:45] <walken> had to cut 48 bytes from it [07:45] <omega_sleepy> oh? [07:45] <walken> 448 kbps [07:45] <walken> = too much [07:45] Action: omega_sleepy has fun bashing Eliza's brains in [07:46] <omega_sleepy> see bottom of http://aimovie.warnerbros.com/html/flash.html [07:46] <omega_sleepy> walken: ok, darn [07:46] <omega_sleepy> neat ac3 though. /me wants real 5.1 output from libac3 [07:46] <walken> coming... wait [07:46] <walken> do you have >16bit output too ? [07:46] Action: walken thinks it would be worth supporting [07:47] <omega_sleepy> walken: you could output 32-bit floats [07:47] <walken> yes [07:49] <omega_sleepy> um, taaz: the problem seems to be that the iterator is failing [07:49] <omega_sleepy> any chance you can track down why that might be, while I fix the threadlock case it triggers? [07:50] <taaz> um... [07:52] <omega_sleepy> in the meantime, try this patch: [07:52] <omega_sleepy> while (GST_FLAG_IS_SET (thread, GST_THREAD_STATE_SPINNING)) { [07:52] <omega_sleepy> if (!gst_bin_iterate (GST_BIN (thread))) { [07:52] <omega_sleepy> - GST_FLAG_UNSET (thread, GST_THREAD_STATE_SPINNING); [07:52] <omega_sleepy> - THR_DEBUG_MAIN("removed spinning state due to failed iteration!\n"); [07:52] <omega_sleepy> +// GST_FLAG_UNSET (thread, GST_THREAD_STATE_SPINNING); [07:52] <omega_sleepy> +// THR_DEBUG_MAIN("removed spinning state due to failed iteration!\n"); [07:52] <omega_sleepy> + THR_DEBUG_MAIN("iteration failed, something very wrong, spinning to let parent sync\n"); [07:52] <omega_sleepy> + while (GST_FLAG_IS_SET(thread, GST_THREAD_STATE_SPINNING)) ; [07:52] <omega_sleepy> } [07:53] <omega_sleepy> to gstthread.c, line 553 and on [07:57] arik (arik@168.191.238.37) joined #gstreamer. [07:57] <arik> lo [07:57] <ajmitch> hi [07:57] <omega_sleepy> yo [07:58] <arik> sup? [07:58] <omega_sleepy> see topic [07:58] <arik> yes i already knew about that ;-) [07:58] <taaz> omega_sleepy: it's hitting that print :( [07:58] <ajmitch> heh [07:58] <taaz> brb [07:58] <omega_sleepy> taaz: right, it's supposed to, but it shouldn't lock any more [07:59] <ajmitch> omega_sleepy: how long do we have before 0.2.0? [07:59] <omega_sleepy> I'd like to get it out in the next week [07:59] Action: arik is gonna add open location support tonight [07:59] <omega_sleepy> then we can restructure cvs, and start on other stuff like events [07:59] <ajmitch> arik: that 10-min hack? ;) [07:59] <arik> yes [07:59] <arik> unless you did it? [07:59] <omega_sleepy> in the meantime, I need to sleep [07:59] Nick change: omega_sleepy -> omega_sleep [07:59] <ajmitch> nah, didn't get around to finishing it off ;) [08:00] <arik> sleep [08:00] <arik> wuss ;-( [08:00] <arik> hehe [08:00] <ajmitch> night omega_sleep [08:00] <arik> night [08:00] Action: ajmitch must go soon... [08:00] <omega_sleep> arik: I came home via Oakland.... and it's midnight for me still [08:00] <ajmitch> gotta ring friend ;) [08:00] <arik> ah [08:00] <arik> well, sleep well ;-) [08:00] <omega_sleep> hope to... [08:00] <arik> (i did get up at 7pm today) [08:00] <omega_sleep> yeah, I got up at, um, nevermind <g> [08:00] Action: arik begins his update [08:01] <arik> hehe [08:01] <omega_sleep> I'll be up earlier tomorrow to work on stuff though [08:01] <omega_sleep> so.... l8r all [08:01] <arik> later [08:01] omega_sleep (om...@om...) left irc: killall -SLEEP omega [08:01] <arik> damnit [08:02] <arik> wtay didn't add a ChangeLog entry [08:02] <arik> when he updated gstplay [08:09] <ajmitch> heh [08:10] <taaz> arik: its easier to just use cvs2cl before a release [08:10] <arik> heh [08:10] <arik> i write longer changelogs then i do commit messages [08:11] <ajmitch> hehe [08:11] <ajmitch> arik: what's worse is when the changelog is bigger than the patch ;) [08:11] <arik> ;-) [08:12] <ajmitch> i prefer highly descriptive messages duch as 'committed fix' [08:12] <arik> that is the sign of too complicated code [08:13] <ajmitch> why is that? ;) [08:13] <ajmitch> usually it is sign of lazy coder when the changelog has very little in it ;) [08:14] <arik> heh [08:15] <arik> if the changelog is bigger then the patch [08:15] Nick change: taaz -> taazzzz [08:15] <arik> it means there is to much to explain [08:16] <ajmitch> yes [08:16] <ajmitch> and the code should be rewritten in a way that is easier to understand if possible [08:16] <arik> yes [08:18] <ajmitch> wow, HEAD is so colourful now ;) [08:18] <arik> heh [08:18] Action: arik is till trying to compile it [08:19] <arik> and has failed [08:19] <arik> to do so [08:19] <arik> libtool: link: error: cannot link shared libraries into libtool libraries [08:19] <ajmitch> heh [08:19] <ajmitch> so far gstmediaplay crashes on everything i play ;) [08:20] <arik> great ;-) [08:20] <Parapraxis> nightinight ;) [08:20] Parapraxis (Spluzz@165.247.57.116) left irc: sleep [08:21] <ajmitch> hmm, i can play my mp3s with gstreamer-launch [08:21] <arik> goo [08:21] <arik> d [08:21] <ajmitch> so it's not all dead ;) [08:23] <ajmitch> good to see the Indian govt getting behind FSF-I [08:24] <arik> heh [08:24] <arik> brb [08:25] arik (arik@168.191.238.37) left irc: leaving [10:01] dobey (do...@dr...) joined #gstreamer. [10:01] <dobey> whee [10:03] <ajmitch> heh [10:36] walken (foobar@24.176.233.236) left irc: zzzZzZZzZzzzzzz [11:07] <dobey> bleh [11:27] greg_ (gr...@ho...) joined #gstreamer. [11:27] <greg_> hello all ! [11:28] <ajmitch> hi [11:28] <dobey> hola [11:28] <dobey> greg_: oh, you don't know anyone at certum.pl would you? [11:29] <ajmitch> night all [11:29] Nick change: ajmitch -> ajzzzz [11:32] <greg_> dobey: no - I haven't the need to go for certificate yet. Why do You ask ? [11:33] <dobey> greg_: because about a month ago, i got an email from them, about making their ssl certs compatible with encompass, and i wasn't able to reply back due to some mail issue [11:33] <dobey> (on their end) [11:36] icarus_ (icarus@i.read.your.email.kgb.fi) joined #gstreamer. [11:36] icarus_ (icarus@i.read.your.email.kgb.fi) left irc: changing servers [11:36] icarus_ (icarus@i.read.your.email.kgb.fi) joined #gstreamer. [11:48] Nick change: dobey -> dobey-slp [12:02] arik (ar...@sd...) joined #gstreamer. [12:02] <arik> ek [12:02] <arik> anyone here? [12:26] arik (ar...@sd...) left irc: Ping timeout for arik[sdn-ar-001waseatP213.dialsprint.net] [12:27] icarus_ (icarus@i.read.your.email.kgb.fi) left irc: [BX] OH SHIT!\15 [12:47] thomas (th...@ad...) joined #gstreamer. [13:42] sienap (sy...@ip...) joined #gstreamer. [13:47] <sienap> hej all :) [13:51] arik (ar...@sd...) joined #gstreamer. [13:52] <sienap> arik ! [13:52] <arik> hey [13:52] <arik> how goes it? ;-) [13:53] <sienap> he all ok [13:53] <sienap> done some eleet gstmediaplay coding ? [13:53] <arik> glad to hear it ;-) [13:54] <arik> heh [13:54] <arik> i was planning on adding open location tonight [13:54] <arik> got sidetracked somehow [13:54] <arik> need to do that asap [13:54] <sienap> he [13:54] <sienap> ok :) [13:54] <sienap> maybe make an roadmap [13:54] <sienap> and some TODO items [13:55] <arik> heh [13:55] <sienap> so it is more clear for other people what they can do [13:55] <arik> doubful, atm i can't even get the merged tree to compile [13:55] <arik> at some point i will make a todo [13:55] <sienap> he [13:55] <sienap> :) [14:04] Nick change: wtay-zZz -> wtay [14:04] <wtay> yo [14:05] <wtay> mozilla nightlies are getting very fast.. [14:05] <arik> ;-) [14:05] <arik> we think so [14:05] <arik> i can't compile the merged tree [14:05] <wtay> can't wait 'till I can use galeon with the new core [14:05] <arik> heh [14:05] <wtay> arik: ? [14:05] <arik> just wait [14:05] <arik> blizzard is rewriting gtkmozembed [14:06] <arik> to make it much much better [14:06] <wtay> how, If I dare ask? [14:06] <sienap> :) [14:06] <arik> how is he gonna make it better? [14:06] <arik> libtool: link: error: cannot link shared libraries into libtool libraries [14:06] <sienap> wtay how come so that they are getting faste r? [14:07] <arik> i know why they are getting faster [14:07] <wtay> better image lib, better cache [14:07] Action: wtay reads the weekly reports [14:07] <wtay> yay JMF is out [14:08] <arik> some rewrites [14:08] <arik> it's about to get even better [14:08] <arik> the cs stylesystem's been rewritten [14:09] <arik> libtool: link: error: cannot link shared libraries into libtool libraries <-- this is the error i get when compling [14:09] <wtay> arik: where? [14:09] <arik> elements [14:10] <wtay> ok, sec, fixing... [14:12] <wtay> libtool --version? [14:12] <arik> 1.3.5 [14:12] <wtay> hm [14:15] <wtay> cvs update in gst/elements [14:15] <arik> doing [14:16] <wtay> I had to link against libgst for my java bindings to work... [14:16] <arik> oh [14:16] <wtay> see if this fix solves your problem, if so I'll try to find another solution [14:17] <arik> yes i think it did [14:17] <wtay> weird [14:17] <arik> shame [14:18] <arik> anyway bed time i think [14:18] <arik> night [14:19] <wtay> night [14:19] arik (ar...@sd...) left #gstreamer. [14:21] <sienap> cvs checkint gstreamer [14:24] <wtay> ? [14:26] <sienap> ing [14:26] <sienap> ben gstreamer aanut cvs downloaden [14:26] <sienap> ffeks proberen te compilen enzo [14:41] hadess (ha...@pc...) joined #gstreamer. [14:42] <hadess> hey guys [14:42] <wtay> hey hadess [14:42] <hadess> heya wtay [14:42] <sienap> hadess! [14:42] <hadess> hey sienap [14:42] <hadess> the lack of EOS breaks even gstmediaplay [14:43] <hadess> that sucks [14:43] <wtay> yes [14:43] <wtay> we could make a quick workaround for that [14:43] Action: wtay wonders why gmodule always uses RTLD_GLOBAL... [14:43] <hadess> by having an eos signal from the src ? [14:44] <wtay> up [14:44] <wtay> yup [14:45] <wtay> damn damn damn [14:50] <hadess> cvs upd takes ages, bbiab [14:54] <sienap> hmm [14:54] <sienap> why is the autogen.sh going into an infinity loop " [14:55] <wtay> sienap: did you patch autoconf? [14:55] <wtay> automake even [14:55] <sienap> nope [14:55] <sienap> he nope [14:55] <wtay> do you have 512MB ram? [14:55] <sienap> where to get the patch ? [14:55] <wtay> in the REAME file [14:56] <sienap> mwha not that much [14:56] <sienap> but i can say [14:56] <sienap> VM errors were getting really close [14:56] <sienap> :) [14:56] <sienap> so i killed the thingy [14:57] <sienap> he [14:57] <sienap> bah [14:57] <sienap> then i have to reinstall automake as well [14:57] <sienap> blergh [14:57] <wtay> right :-) [14:57] <sienap> :) [14:57] <sienap> he [14:57] <wtay> nope, just do vi /usr/bin/autoconf and edit it around line 2380 [14:57] <sienap> will do later today [14:57] <sienap> i think :) [14:57] <wtay> automake even [14:58] <sienap> is the patch submitted to the automake author as well ? [14:58] <sienap> he [14:58] <sienap> ooh [14:58] <sienap> :) [14:59] <wtay> yes [15:00] <sienap> eens zien of hij nu werkt [15:00] <sienap> :) [15:01] <thomas> libtool: link: error: cannot link shared libraries into libtool libraries [15:01] <thomas> morning all [15:02] <thomas> what's that libtool error when I compile gstreamer ? [15:05] <wtay> no idea, just checout my fix for it [15:06] <wtay> I had to link libgstelements.so to libgst.so because alse my java bindings didn't work... [15:06] <sienap> het werkt nu wim :) bedankt [15:06] <thomas> wtay: where's the fix ? [15:07] <wtay> in cvs gst/elements [15:12] <hadess> there are a lot of warning: `return' with no value, in function returning non-void [15:13] <hadess> in nearly every plugin in fact [15:13] <hadess> btw wtay, what does mp12mp1 do ? [15:14] <wtay> hadess: plugin? [15:14] <wtay> hadess: testapp, yes [15:14] <wtay> mp2tomp1 converts mpeg1 to mpeg1 (on a good day) [15:14] <wtay> mp12mp1 trancodes mpeg1 [15:14] <hadess> video ? [15:15] <wtay> yup [15:15] <thomas> wtay: random question about xine: why does fullscreen xine only work with Xv extension ? [15:16] <wtay> thomas: fullscreen without is is dog slow... [15:16] <wtay> I guess they just didn't want to implement their own scaling routines [15:17] <thomas> wtay: hm, right now I can only use either framebuffer device or vesa device for Xfree on TV... [15:17] <thomas> ... neither has Xv it seems [15:17] Action: thomas wants to watch DVD's under linux, not windows ! [15:17] <wtay> that sucks [15:17] <thomas> wtay: I'd think stuff like Xv would be bloody useful to get stuff working on TV [15:17] <thomas> wtay: but ATI is difficult about it it seems [15:18] <wtay> yes, that's why I always buy matrox cards, they have good linux support [15:18] <sienap> yeah [15:18] <sienap> matrox rules [15:18] <sienap> :) [15:18] <thomas> wtay: but does tv out work with them ? [15:19] <wtay> yes [15:19] <thomas> wtay: which card do you use for tv out ? [15:19] <hadess> my ATI works fine, with Xv et all [15:19] <wtay> G450 [15:19] <thomas> hadess: with TV/out or not ? I can get it to work fine as well with Xv [15:19] <wtay> I don't use TV out myself but I know it's supported [15:20] <thomas> wtay: how do you know ? I've searched all over, didn't find anything definite [15:20] <hadess> thomas: dunno, i know that mirror works fine [15:20] <wtay> lemme check the docs... [15:23] <thomas> wtay: can xine use DGA ? that does work on my card at the moment [15:23] <wtay> thomas: I think so.. [15:24] <wtay> I can't find it... [15:25] <thomas> wtay: I did find some mention of some G450 versions having a passhtrough cable converting VGA out to SVHS... [15:25] <wtay> thomas: yup, I have such a cable [15:26] <thomas> wtay: well, if you ever have the time, you should check what it does ;) [15:26] <wtay> thomas: maybe I can try to get the output of the G450 into the input of my BTTV card <g> [15:28] <thomas> wtay: well, that would work too ;) [15:28] <thomas> if anyone has a g400 left over, let me know ;) [15:28] <thomas> ok, xine doesn't do dga. is there any other dvd player that works well ? [15:28] <wtay> gvlc [15:31] <hadess> wtay: if gstmediaplay uses gnomevfssrc for local files, how come i can seek ? [15:32] <wtay> hadess: it doesn't if there is no :/ in the uri [15:32] <hadess> right [15:33] <thomas> wtay: gvlc, is that omi/oms ? [15:33] <wtay> thomas: nope, just another mpeg player [15:33] <hadess> wtay: i'll implement seek sometime [15:34] <wtay> hadess: cool [15:35] <hadess> gstmediaplay should use fdsrc if the file is local... [15:35] <wtay> hadess: not disksrc? [15:36] <hadess> wtay: fdsrc so it can mmap [15:37] <wtay> hadess: right... so should disksrc IMO [15:38] <hadess> you think disksrc should mmap, or ? [15:38] <wtay> hadess: I think it should, yes [15:39] <hadess> fair enough [15:45] Nick change: hadess -> hds-town [15:46] <wtay> uh? evolution went berserk... [15:51] Procule (Ac...@cn...) joined #gstreamer. [15:51] <Procule> howdy [15:52] <wtay> yo [15:52] <Procule> hi wtay [15:52] <Procule> 2.4.5's out [15:54] <sienap> hej procule [15:54] <sienap> :) [15:54] <Procule> hello sienap [15:55] <sienap> you are working on an gstreamer project ? =] [15:56] <Procule> here's my situation: At home, i'm on the phone for the net. Since 6 years. At the moment, i'm at my job and internet-cable have just been installed [15:56] <Procule> I'm my laptop and a network card [15:57] <Procule> What should i download ? [15:57] <Procule> :P [15:57] <sienap> gstreamer [15:57] <sienap> :) [15:57] <Procule> I have it [15:57] <sienap> in a short time their will be a new release [15:57] <sienap> i think it is better if you wait for that [15:58] <sienap> or download cvs if you want to test with us [15:58] <sienap> :) [15:58] <Procule> I have the CVS version of 2 days ago [15:58] <sienap> he [15:58] <sienap> check out again [15:58] <sienap> :) [15:58] <sienap> incshed1 is in now [15:59] <Procule> what is that ? [15:59] <Procule> I've found gstreamer 2 days ago [15:59] <Procule> ;) [15:59] <sienap> he [15:59] <sienap> it is the sheduling system [15:59] <sienap> :) [16:00] <Procule> oh [16:00] <Procule> well [16:01] <Procule> gstreamer apart, give me something I won't download on the phone cause of the size ? :P [16:01] <Procule> I'm all excited with that toy [16:01] <sienap> naughty avi's ? [16:01] <sienap> >:) [16:02] <Procule> haha no [16:07] <greg_> hey ! may I ask if gstmediaplay is fixed yet ??? [16:08] <sienap> he [16:08] <sienap> :) [16:08] <sienap> not really i think.. [16:11] <wtay> it's fixed [16:12] <sienap> he [16:12] <sienap> ooh [16:12] <sienap> ;) [16:12] <Procule> lol [16:13] <wtay> avi is broken though [16:13] <greg_> great. gonna do update from CVS [16:13] <greg_> any possibilities that f***d M$ formats like asf axf and other will be supported ? [16:14] <wtay> greg_: sure, when someone writes a plugin [16:14] <greg_> wtay: Oh - I know. But for someone to write a plugin there should be some little docs... are there any ? [16:15] <wtay> asf is docuented and implemented in avifile IIRC [16:15] <sienap> greg_ avifile supports them [16:15] <wtay> axf I dunno [16:15] <sienap> by using the win dll's [16:15] <greg_> I don't think so. I seen ATM stopped project related to avifile reading asf, [16:15] <sienap> (lagging) [16:15] <greg_> but others - NO. [16:15] <wtay> asf is not implemented with dll's [16:15] <wtay> native C++ [16:16] <greg_> I thought only codecs can be taken from dll's ? [16:16] <sienap> he [16:16] <wtay> yup [16:16] <sienap> didn't know that [16:16] <greg_> not the "input file stream" type ? [16:19] <wtay> bleagh, java doesn't like to cooperate with gstreamers thread mechanisms :( [16:20] <sienap> he [16:32] Nick change: hds-town -> hadess [16:39] <greg_> anyone using oms here on debian woody ? [16:39] <greg_> I apt-get [16:39] <greg_> 'ed oms, but it segfaults on any mpeg ;-( [16:39] <wtay> great [16:40] Nick change: wtay -> wtay-bath [16:42] <hadess> greg_: oms is not for playing mpegs [16:43] <dobey-slp> or mpeg2s [16:44] <hadess> is for playing dvds [16:44] Action: hadess watches dobey talk in his sleep [16:44] Action: dobey-slp watches his life suck [16:45] <hadess> go get strawberries, some whipped cream and a video ;) [16:45] <dobey-slp> uh [16:49] <hadess> i think that's what i'm gonna do, don't feel like coding right now [16:49] Nick change: hadess -> hds-afk [16:49] <dobey-slp> i don't feel like * right now [17:15] greg_ (gr...@ho...) left irc: Ping timeout for greg_[home.sente.pl] [17:17] Procule (Ac...@cn...) left irc: Read error to Procule[cnq23-175.cablevision.qc.ca]: Connection reset by peer [17:27] arun (arun@202.88.232.42) joined #gstreamer. [17:29] <wtay-bath> yo [17:29] Nick change: wtay-bath -> wtay [17:37] Nick change: dobey-slp -> dobey [17:37] <dobey> eh [17:44] arun (arun@202.88.232.42) left #gstreamer. [17:52] <sienap> back [18:04] Nick change: hds-afk -> hadess [18:04] <dobey> hey sie/had [18:04] <hadess> yo [18:04] <sienap> hej dob/had [18:05] <sienap> :) [18:07] <sienap> he the new debug is way cool [18:07] <dobey> eh [18:08] <sienap> he still only a sink via Xv ? [18:09] <dobey> eh [18:09] <dobey> <- sad [18:09] <wtay> dobey: ? [18:10] <dobey> <- lonely [18:10] <hadess> wtay: why doesn't fdsrc have a get_region function ? [18:10] <wtay> hadess: getregion is deprecated [18:11] <dobey> hah [18:11] <hadess> wtay: so i should use the (offset) bit when i'm asked for get() ? [18:11] <dobey> i read that as <wtay> getreligion is deprecated [18:11] <wtay> hadess: yep [18:11] <wtay> dobey: that too, but that's only my opinion [18:12] <hadess> wtay: and mine :P [18:12] <dobey> s/opinion/fact/ [18:13] <dobey> now [18:13] <dobey> blah [18:13] <dobey> fuck it [18:13] <dobey> i'm too tired to do anything [18:23] Action: hadess ran indent on gnomevfssrc.c and can actually read the code now [18:23] <sienap> Are you worried about GNOME? [18:23] <sienap> No. We're way ahead of them. When GTK 2.0 is released sometime this year, they'll [18:23] <sienap> reach where we were, at version 1.4. QT 3 is going to come out later this year and [18:23] <sienap> that will take us even further. [18:23] <sienap> he [18:23] <sienap> man [18:23] <sienap> stupid kde people [18:23] <sienap> :) [18:23] <sienap> The GNOME developers are working on a native port of Open Office. Why [18:23] <sienap> aren't the KDE guys doing the same? [18:23] <sienap> This is new for me ? [18:24] <hadess> sienap: you must the vocal user... [18:24] <sienap> he :) [18:25] <thomas> hmm... where can I find info about SDL ? or *good* info on all of the different X extensions for that matter ? xfree doesn't have much... [18:26] <sienap> www.libsdl.org ? [18:26] <sienap> they have docs [18:27] <dobey> bah [18:28] <dobey> fuck it all [18:29] <sienap> what ? [18:29] <dobey> all the bullshit [18:29] <dobey> life [18:29] <dobey> fuck it [18:31] <sienap> he [18:31] <sienap> man [18:31] <sienap> get a grip on your self [18:32] <hadess> wtay: i just found out that i was never free'ing the uri :P [18:32] <wtay> hadess: was that causing a huge leak? [18:32] <dobey> i'm fine [18:32] <dobey> it's the rest of the world that's the problem [18:33] <hadess> wtay: no, nearly not detectable [18:33] <wtay> heh [18:33] <sienap> dobey yeah it is the rest.. [18:33] <thomas> Is SDL fast ? [18:33] <hadess> maybe 50 bytes for each open/close [18:33] <wtay> thomas: somewhat [18:34] <thomas> I'm getting sick off all these different versions and cards and drivers and accel stuff and video and tv [18:34] <dobey> WHY?!? [18:34] Action: thomas heads off to a deserted islands in the bermudas [18:34] <hadess> thomas: that's the same under any OS dude [18:34] <dobey> why me?! [18:35] Action: wtay is coding in JAVA [18:36] <thomas> hadess: depends... when my friends in the house get full-screen dvd and video playing running in half an hour under Windows ME ... [18:36] <thomas> ... then I'd say there's a difference between OS'es ... [18:36] <dobey> why can't i fucking find someone within 10 miles of me?!? [18:36] <thomas> ... for practical reasons if nothing else [18:36] <thomas> much as I hate to admit it [18:39] Procule (Ac...@cn...) joined #gstreamer. [18:47] Nick change: taazzzz -> taaz [18:48] <sienap> hi procule [18:48] <Procule> hi hi [18:48] <Procule> i'm eating a poutine [18:48] <hadess> yo taaz [18:48] <hadess> a poutine ? [18:49] <dobey> uh [18:49] <Procule> yeah [18:49] <hadess> wtay: dude, is there any source that uses get instead of get_region ? [18:49] <sienap> what is an poutine ? [18:49] Action: dobey dies [18:49] <wtay> hadess: all of them use get, no? [18:50] <Procule> a poutine is a french canadians meal [18:50] <sienap> dobey rather don't [18:50] <hadess> wtay: fdsrc doesn't seek it seems and disksrc uses get_region [18:50] <Procule> fries+cheeze+gravy [18:50] <hadess> Procule: you're disgusting [18:50] <Procule> it's the most popular meal here [18:50] <wtay> hadess: no it doesn't, and if it does, it'll never be called anyway [18:51] <Procule> fast food [18:51] <sienap> hehe [18:51] <sienap> gravy == ? [18:51] <hadess> wtay: huh... [18:51] Nick change: dobey -> dobey-ded [18:51] <sienap> sleep well [18:51] <dobey-ded> sleep? [18:52] <Procule> gravy = brown sauce [18:53] <hadess> wtay: ok, fdsrc is useless, disksrc uses mmap(), is that all right ? [18:54] <Procule> can you use fdsrc to get through a socket ? [18:54] <wtay> hadess: yup [18:54] <sienap> ehm o i though it was dobey-bed [18:54] <sienap> sorry dopey :) [18:54] <dobey-ded> no [18:54] <dobey-ded> ded [18:54] <dobey-ded> as in death [18:55] <sienap> procule ehm.. wutever ;) [18:55] <hadess> wtay: ok, i think i have added local file fallback for gnomevfssrc, with mmap'ing [18:55] <wtay> hadess: nice! [18:55] <wtay> Procule: yes [18:55] <Procule> gravy = liquid material wich is commonly used commbined with food products, Has a neat aura! [18:56] <dobey-ded> hmm [18:57] <sienap> dobey man come an [18:58] <dobey-ded> sienap: bring me some happiness and i will [18:58] <sienap> dobey man happiness doesn't come to you you have to go to it.. [18:59] <dobey-ded> yeah well [18:59] <dobey-ded> that doesn't fucking work [18:59] <sienap> so go out in city.. [18:59] <sienap> go party and stuff [18:59] <sienap> go watch some chicks [18:59] <sienap> or better make them watch you [18:59] <dobey-ded> go jump off the pier... [18:59] <sienap> and ehm yeah i bet you'll enjoy [19:00] <dobey-ded> no i won't enjoy [19:00] <sienap> man dobey [19:00] <sienap> don't [19:00] <sienap> it won't solve things [19:01] <sienap> or actually it is an real cheap way [19:01] <sienap> to solve things [19:01] <dobey-ded> at least i'll save money [19:01] <sienap> he [19:01] <sienap> watch query [19:02] <hadess> wtay: hmmm, how do i get debug from gnomevfssrc ? [19:02] <hadess> wtay: i added a GST_DEBUG(0, ...) [19:02] <wtay> hadess: --gst-mask=-1? [19:03] <hadess> wtay: that's a bit too much ;) [19:03] <wtay> hadess: oh, right, then you need GST_CAT_PLUGIN_INFO and the appropriate mask [19:04] <wtay> hadess: you can't filter out gnomevfssrc on its own for now [19:04] <hadess> wtay: i just want it to have some output when it's running and debug is not disabled... [19:04] <wtay> hadess: ok, use g_print then [19:05] <hadess> wtay: gah, that's nasty ;) [19:05] <wtay> hadess: life is nasty :) [19:05] <hadess> wtay: hmm, i think it works =) [19:07] <Procule> humph [19:07] <Procule> viva los bananas [19:08] <hadess> "file:///opt/mp3s/rap/Run DMC feat. Aerosmith - Walk This Way.mp3 is a local file" [19:09] <wtay> cool [19:10] <Procule> how can you specify a file with space ? ic location= ? [19:10] <Procule> in [19:10] <wtay> "\ " [19:11] <hadess> wtay: but gstmediaplay segfaults :/ [19:11] <Procule> like: "/home/mp3s/A\ cool\ mp3.mp3" [19:11] <Procule> ? [19:12] <wtay> yup [19:12] <hadess> Procule: the quotes _or_ the \ [19:12] <wtay> if you use quotes it has to be like \"location=some\ cool.mp3\" [19:12] <Procule> ok [19:13] <hadess> i use location="blah blah.mp3" [19:13] <wtay> hadess: that works? [19:13] <hadess> wtay: yep [19:13] <wtay> hmm, I didn't know that [19:15] <hadess> wtay: can you test the just committed gnomevfssrc ? [19:15] <hadess> i think i forgot something... [19:16] <wtay> ok [19:16] <hadess> gstmediaplay crashes with some weird locking problem... [19:17] <wtay> oops [19:17] <hadess> what ? [19:18] <wtay> nothing.. just updating now [19:19] <wtay> wow, you added the mmap code to gnomevfs? [19:19] <hadess> no, to the local fallback [19:19] <wtay> in gnomevfs? [19:19] <hadess> no [19:20] <hadess> in the plugin [19:20] <wtay> that's what I mean... [19:20] <hadess> oh, yes, i did [19:20] <wtay> isn't that a bit too much for this plugin? [19:20] <wtay> I feel this is an app issue [19:21] <hadess> this way you can use gnomevfssrc as the only src in gstmediaplay, without having to care if it's local or not [19:21] <wtay> I mean, the mmap code in fdsrc is broken and I don't intend to fix it in N places [19:21] <hadess> right... [19:22] <wtay> it works fine though [19:22] <wtay> file:/opt/data/armageddon1.mpg is a local file [19:22] <hadess> can you try with gstmediaplay ? [19:23] <wtay> I did, it works fine [19:23] <hadess> weird, it breaks here [19:23] <wtay> hadess: you have the latest CVS version? [19:23] <hadess> i think so yes [19:24] <hadess> yep, it is [19:24] <wtay> a thread lockup? [19:24] <hadess> did you open the file on the command line or in the gui ? [19:25] <Procule> "macro: AM_PATH_ESD not found in library" [19:25] <wtay> cmd line [19:25] <Procule> what is that ? [19:25] <hadess> #0 0xfed41d0 in pthread_mutex_unlock () from /lib/libpthread.so.0 [19:25] <hadess> #1 0xffaa558 in cothread_switch (thread=0x100b0638) at cothreads.c:319 [19:25] <hadess> #2 0xffaa38c in cothread_stub () at cothreads.c:232 [19:25] <hadess> #3 0xffaa7c8 in cothread_switch (thread=0x0) at cothreads.c:358 [19:25] <hadess> cothread code, oopsie [19:25] <hadess> Procule: install the esd-devel packages [19:25] <wtay> Procule: you need esd dev packages [19:26] <Procule> oh [19:26] <wtay> hadess: need to tell omega about that [19:26] <hadess> ok [19:26] <wtay> hadess: is that a segv? [19:26] <hadess> wtay: yep [19:26] <wtay> hadess: try lowering the cothread stacksize [19:26] <hadess> how ? [19:27] <wtay> in gst/cothreads.c [19:27] <wtay> line 39 [19:27] <wtay> make it like this: [19:27] <wtay> #define COTHREAD_STACKSIZE 32768 [19:27] <wtay> #define COTHREAD_MAXTHREADS 64 [19:27] <wtay> I had this problem too [19:28] <wtay> I'll commit a patch [19:28] <wtay> err, rather not, I changed something else... [19:28] <hadess> works fine now [19:28] <wtay> ok [19:29] <wtay> when omega arrives, tell him that the cothread stacksize is too small [19:29] <hadess> okie [19:29] <hadess> seek doesn't work yet though... [19:30] <wtay> nope [19:31] <hadess> will do that next [19:31] <hadess> is the mmap code in disksrc better ? because that's the one i used [19:31] <wtay> hadess: you don't use mmap to read the bytes yet [19:32] <hadess> is a bit broken ;) [19:32] <wtay> disksrc is better [19:33] <hadess> <- stupid [19:33] <hadess> at least, it skips the gnome-vfs overhead for local files [19:34] <wtay> I still think it's an app issue to detect local files and insert a disksrc when needed [19:34] <wtay> at most gnomevfssrc can have an arg "is_local" [19:35] <hadess> and you'd need to add a signal to gnomevfssrc, and the code to translate the uri into a local filename, etc. [19:36] <wtay> nope, just when you set the uri, the arg is set [19:36] <hadess> you'd need to have the code, either in disksrc or the app to translate the uri into a local filename [19:37] <wtay> yes [19:38] <hadess> i don't think the amount of code added in gnomevfssrc is so big [19:38] <wtay> true [19:38] <wtay> but it doesn't feel right... [19:38] <hadess> the only problem i could see is maintenance [19:39] <hadess> i see what you mean [19:39] <wtay> for the same reason you could decide to use your own implementation of http is you detect http:// [19:39] <wtay> s/is/if [19:39] <hadess> i wouldn't, i would patch gnome-vfs in such a case [19:40] <sienap> where is zaheer lately btw ? [19:40] <hadess> nope, not lately [19:41] <hadess> wtay: the only perfomance problem in gnome-vfs, as far as gst is concerned, is local file access [19:41] <hadess> wtay: in this case, i just make the app developer's job easier [19:42] <wtay> hadess: yeah, I suppose.. better fix gnomevfs for the local case IMO [19:42] <hadess> wtay: mmap support in gnome-vfs will never be possible, at least in its current form [19:43] <wtay> oh? [19:43] <hadess> wtay: but i understand your point [19:43] <wtay> well, I'm not overly concerned about it... [19:45] <hadess> cool, then ;) [19:46] <hadess> adding mmap reading and seek to it now [19:55] <hadess> seek works for local files now ;) [19:56] greg_ (greg@213.77.64.92) joined #gstreamer. [19:58] <greg_> uh-oh autogen.sh just killed my machine ... while I was reading README with info about "bug" ;-)) [19:59] <hadess> heh [20:01] <wtay> hehehehe [20:01] <wtay> hum, shell.sourceforge.net is down... [20:05] <hadess> hmm, fdsrc doesn't seek, and disksrc seeks using mmap, so i don't have any example code... [20:09] <wtay> hadess: lseek is your friend [20:09] <hadess> gnome_vfs_seek is [20:09] <wtay> doh, yes :) [20:09] <hadess> i just need to figure out what the args means [20:09] <hadess> s/means/mean [20:10] <hadess> GnomeVFSResult gnome_vfs_seek(GnomeVFSHandle *handle, GnomeVFSSeekPosition whence, GnomeVFSFileOffset offset); [20:10] <hadess> handle, no problem i understand [20:10] <hadess> /* This is used to specify the start position for seek operations. */ <- hehe easy [20:11] <wtay> whence is the same as lseek probably [20:11] <hadess> wtay: the offset is from the start right ? [20:11] <wtay> gotta go now [20:11] <wtay> hadess: yup [20:11] <hadess> ok, cya [20:11] Nick change: wtay -> wtay-snooker [20:11] <wtay-snooker> cya [20:11] Action: hadess prepares surprise for wtay [20:12] Action: Procule is download Visual C++ [20:12] <Procule> ign [20:12] <Procule> ing [20:15] omega_ (om...@om...) joined #gstreamer. [20:15] <sienap> erik! [20:15] <omega_> yo [20:15] <hadess> yo omega_ [20:15] <hadess> <wtay> when omega arrives, tell him that the cothread stacksize is too small [20:16] <omega_> I saw that in the logs, yeah [20:16] <hadess> so i tell you [20:18] greg_ (greg@213.77.64.92) left irc: Ping timeout for greg_[213.77.64.92] [20:19] <thomas> omega_: I ran gstreamer-compreg as root, gstreamer-complete doesn't do much, how does it work ? [20:19] <omega_> hadess: I wonder if you can simply replace self with a disksrc in the fallback case, rather than duplicating the mmaping code [20:20] <omega_> thomas: how'd you set up complete? [20:20] <thomas> omega_: that's the point : how do you have to do it ? [20:20] <omega_> complete -C gstreamer-complete gstreamer-launch [20:20] <omega_> a bash command [20:21] <hadess> omega_: if you think that can work... [20:21] <thomas> ok, can you put that in .bashrc ? [20:21] <omega_> hadess: no reason it shoudln't [20:21] <thomas> or do you have to alias it [20:21] <omega_> thomas: yeah, though I have mine in .bash_profile and it doesn't seem to hold ;-( [20:21] <hadess> omega_: can you give it a try ? i don't have a clue how you can do that... [20:21] <omega_> hadess: I'll give it a try on my machine, see what happens [20:21] <thomas> omega_: cool ! [20:21] <thomas> it works [20:21] <omega_> see gstautoplugger.c for an idea of how it might work [20:22] <thomas> omega_: it does take over file completion however [20:22] <omega_> thomas: it has a lot of limitations still [20:22] <omega_> thomas: that's one of them ;-( [20:22] <hadess> omega_: the thing is for example if the app assumes that it has a gnomevfssrc, and puts the state to pause, replaces the filename by an http:// uri, how would that work ? [20:23] <omega_> hadess: good point [20:23] <hadess> omega_: you'd have to have a fallback in disksrc as well... [20:23] <omega_> ick [20:23] <hadess> and that's messy... [20:24] <omega_> hadess: right. hmm [20:24] <omega_> we should find some way to avoid duplicating code, though [20:24] <omega_> I need to rewrite disksrc to use a better mmaping scheme anyway, it'll get a lot more bulky and you really don't want to copy it... [20:25] <hadess> omega_: have a #include'd file in both of them ? [20:25] <omega_> maybe [20:25] <omega_> another option would be to make gnomevfssrc into a pseudo-Bin and instantiate disksrc internally [20:26] <omega_> I'll have to play with that idea [20:26] <hadess> sounds interesting [20:26] <omega_> trick is simulating enough of the Bin to make sure that the scheduler can maintain track of the disksrc [20:26] <omega_> but I think it'd work [20:27] <hadess> as long as i can still understand the code of this plugin ;) [20:27] <omega_> heh [20:28] <omega_> hmmm, did you copy the set_state(NULL) on filename == NULL from somewhere? [20:29] <omega_> also, your log message mentioned mmap, which you're doing, but your not using [20:30] <hadess> i'm using it now, it works [20:30] <omega_> not in cvs [20:30] <hadess> i know [20:30] <omega_> ok [20:30] <hadess> i'm finishing adding seek to the gnome-vfs case, and committing right after [20:30] <omega_> ok [20:30] <omega_> I'll wait then [20:31] greg_ (gr...@ho...) joined #gstreamer. [20:31] <hadess> omega_: but i can commit right now if you want [20:31] <omega_> don't rush [20:31] <thomas> omega_: is it now possible to implement the mixer right ? meaning, have the app pause the pipeline, request a new sink pad from adder, connect new thread and go back to playing ? [20:31] <omega_> I've got a gstfilesrc.c in the works that I need to work on before it makes much sense to try this with gnomevfssrc [20:31] <omega_> thomas: should be, yes [20:32] <omega_> that's the idea ;-) [20:32] <thomas> omega_: ok, so for the timestamps, what was a good example ? that way I can try it out soon without having to stay up all night asking you about it :) [20:32] <omega_> mpeg1 should do it, dunno how well [20:34] <thomas> ok, thanks, I'll check that [20:34] <thomas> have to go now, bye all [20:34] thomas (th...@ad...) left irc: [x]chat [20:37] dobey-ded (do...@dr...) left irc: home [20:42] <hadess> omega_: dude ? [20:42] <omega_> ? [20:42] <hadess> i'm going to commit it, but seek doesn't work on gnome-vfs [20:43] <omega_> ok [20:43] <omega_> gnome-vfs bug? [20:43] <hadess> don't think so [20:43] <hadess> bug's in my code [20:43] <omega_> ok [20:43] <hadess> so it's broken, but can you tell me what i'm doing wrong ? [20:43] <omega_> ok, I'll check it out [20:44] <omega_> erm, are you on g-h ? [20:44] <hadess> no, why ? [20:44] <omega_> just got a post supposedly from Alan Cox that is very odd [20:45] <omega_> re gnotices being broken: [20:45] <omega_> That isnt the problem. In fact if gnotices fell down irrepairably it would be [20:45] <omega_> a major plus point for the gnome project. Gnotices consists of nothing but [20:45] <omega_> libel, defamation and actionable hate speech. [20:45] <omega_> news to me, and to Uraeus as well I would think..... [20:45] <hadess> sopwith and uraeus had a chat about this some time ago [20:46] <hadess> i committed the broken gnomevfssrc [20:46] <hadess> try it with gstmediaplay [20:46] <omega_> will do [20:47] <omega_> maybe I don't read Gnotices enough, but I've never seen anything that would fall into either of those three categories [20:47] <hadess> it's the comments that have these problems [20:47] <omega_> hmm [20:47] <omega_> well, can gnotices still claim common-carrier status? [20:48] <hadess> what is common-carrier ? [20:49] <omega_> telcos and others (like /., presumably) can claim that because they do not touch the data they pass through, thus they are immune from requirements that they *fully* censor the data to match local decency, etc. laws [20:49] <omega_> basically, if you pass data without modification, you cannot be held liable for the data's contents [20:50] <omega_> the moment you as a carrier start to censor any bit of the data, you become responsible/liable for censoring all of it [20:50] <hadess> right, so gnotices is in this case, for the comments at least [20:50] <omega_> ok [20:50] <omega_> so from a legal standpoint it's fine as is [20:50] <hadess> well, there's censor in the comments [20:50] <omega_> oh? [20:50] <omega_> then it's *not* a common carrier [20:51] <hadess> yep [20:51] <hadess> ppl posting huge photos, or really abusive stuff [20:51] <omega_> neat [20:51] <omega_> then I'd say: ditch comments, just make it a headline system [20:51] <hadess> yep [20:51] <hadess> that's what i told them [20:52] <omega_> if they want to comment, have someone else run a slash with RDF download to mirror the headlines, and comments can go on that site instead [20:52] <hadess> or have moderators [20:52] <omega_> way too much work [20:52] <hadess> like linux.com does [20:52] <hadess> no, not really [20:52] <omega_> is there really enough useful data going into comments that it's worth that much energy? [20:52] <hadess> sometimes yes [20:53] <omega_> data that can't be obtained another way? [20:53] <hadess> even developers take part in the comments usually [20:53] <omega_> hmm [20:54] <omega_> well, I'd say the best bet is to find some other mechanism for the comments [20:55] <hadess> that's what the moderator are for, on #linux.com there's a bot that says how many comments are up for moderation, and you just click on a button to have it appear on the site [20:56] <omega_> hmmm [20:56] <omega_> and M2 ? [20:56] <hadess> M2 ? [20:56] <omega_> meta-mod [20:56] <omega_> who moderates the moderators? [20:57] <hadess> hmmm... [20:57] <omega_> one potential advantage in just using Slash for gnotices..... [20:57] <omega_> already has that stuff [20:57] <hadess> slash is a pita to install [20:58] <omega_> even the latest rev? claimed to be a lot easier [20:58] <hadess> might be... [20:59] <hadess> can you try out the gnomevfssrc please :) [20:59] <omega_> reading [20:59] <Procule> who wrote the plugins for gstreamer ? [20:59] <omega_> wtay and I wrote most of them [20:59] <omega_> check the roadmap for the rest [21:00] <hadess> omega_: line 345, if i disable this block, it works, but doesn't seek [21:00] <Procule> from scratch ? [21:00] <hadess> Procule: http://www.linuxrising.com/files/gstreroadmap2.html [21:00] <omega_> Procule: some of them are reworks of existing programs, some of them are based around existing libraries, and some are written from scratch [21:00] <omega_> probably more than half written from scratch [21:01] <Procule> woh I'm not able to understand how you could build a driver for a video dec/enc [21:01] <omega_> hadess: if that's the only bit that deals with seek, I don't see anything wrong with it [21:01] <Procule> ;) [21:02] <Procule> too low level for me hehe [21:02] <omega_> Procule: what do you mean by 'driver for video enc/dec' ? [21:02] <hadess> omega_: me neither... [21:02] <Procule> like mpeg decoder and encoder [21:02] <Procule> to display it to screen [21:02] <omega_> hadess: try a standalone program that does a read,read,read,seek,read,read and see if it works properly [21:02] <hadess> there's already one in there [21:03] <omega_> Procule: all the mpeg stuff we have is based on existing programs right now [21:03] <hadess> omega_: ok, will ask yakk as well if he's around [21:03] <omega_> but I have my own mpeg1 video decoder that I wrote from scratch that I want to put in here eventually [21:03] <omega_> hadess: ? just write one... [21:03] <omega_> should be about 20 lines [21:04] <Procule> is that all algorythm stuff and the rest handled by the X server ? [21:04] <hadess> omega_: i want to ask yakk if it's supposed to work [21:04] <omega_> hadess: ok ;-) [21:04] <omega_> Procule: no, all the mpeg stuff is done by the decoder plugins [21:04] <omega_> the X server just handles output, and can be used for anything else that feeds just raw video [21:08] <Procule> so the plugin converts the mpeg stream in raw video and outputs it to X ? [21:09] <omega_> yup [21:09] <omega_> video/mpeg in, video/raw out [21:09] <Procule> oooh I see [21:10] <Procule> same thing for the audio i guess ? [21:10] <omega_> yup [21:10] <omega_> unless of course the sound card can accept mp3 directly... [21:10] <omega_> but I know of no such card under linux at least... [21:10] <omega_> the sblive should be capable of it with the right firmware, though... [21:11] <Procule> so If my card handles MPEG compression, can I pass direct mpeg2 video ? [21:11] <Procule> like DVD [21:11] <omega_> yup [21:11] <omega_> I have a dxr2 that I want to write a plugin for at some point [21:12] <Procule> the X server handles all that stuff of supported thing with the video card ? [21:13] <omega_> there are no X extensions or drivers for such cards (like the ATI Rage) yet [21:13] <omega_> the dxr2 has its own driver interfaces [21:13] <Procule> nvidia ? [21:13] <omega_> same [21:14] <omega_> no extension, no drivers [21:14] <omega_> all of that hardware was bought as silicon IP by ati,nvidia,etc. from companies that refuse to let the video card manuf's even release binary-only stuff [21:15] <Procule> I guess that when such drivers will be available, the DVD will be more easy the integrate in linux [21:15] <omega_> we might have a roundabout chance of getting something done, but I wouldn't hold my breath. loki games was supposedly working on that two years ago and nothing ever happened [21:15] <omega_> yup [21:16] <Procule> but nvidia features drivers for linux [21:17] <omega_> only 2d and 3d [21:18] <omega_> and binary only anyway, which is why I will not by an nvidia board [21:18] <omega_> I will not put myself at the mercy of a company who doesn't even consider Linux a first-class platform anyway [21:19] <omega_> not to mention one that ignored all the developing standards like DRI [21:19] <Procule> true [21:23] <Procule> Does gstreamer handle correctly things from v4l ? [21:23] <Procule> like a camera [21:23] <Procule> or tn tuner [21:23] <Procule> tv [21:23] <hadess> omega_: dude ! [21:24] <hadess> omega_: i think the problem is when issuing seeks too fast, i had the same problem when opening and closing files fast [21:28] mattsm (ma...@ad...) joined #gstreamer. [21:29] Action: omega_ goes to help move the trailer to the backyard [21:29] <omega_> brb [21:30] <hadess> ok [21:57] ChiefHighwater (paul@63.105.18.226) joined #gstreamer. [22:01] ajzzzz (aj...@p2...) left irc: Ping timeout for ajzzzz[p20-max1.dun.ihug.co.nz] [22:10] Nick change: hadess -> hds-busy [22:19] dobey (do...@ip...) joined #gstreamer. [22:20] Procule (Ac...@cn...) left #gstreamer. [22:23] ajzzzz (ajmitch@203.173.236.129) joined #gstreamer. [22:23] Nick change: ajzzzz -> ajmitch [22:24] <dobey> yo [22:25] <ajmitch> hi [22:26] <dobey> hrmm [22:26] <dobey> i wonder why xterm isn't having zsh login work [22:28] <ajmitch> dunno [22:28] <dobey> you use xterm? [22:29] <ajmitch> nah, Eterm usually [22:30] <dobey> oh [22:30] <dobey> i would use eterm [22:30] <dobey> but it was sucking worse than this [22:30] <dobey> i guess i gotta write a terminal myself [22:30] <ajmitch> yeah, Eterm is broken on here at the moment, it comes up as unknown terminal when i use telnet or ssh ;) [22:36] <dobey> hrmm [22:36] <dobey> you used powershell? [22:37] sienap (sy...@ip...) left irc: sienap has no reason [22:38] <ajmitch> nope [22:38] <dobey> hrmm [22:38] <dobey> i'll try it [22:40] <dobey> fuck that [22:40] <dobey> the spec file for it is broken [22:40] <omega_> aren't they all?? [22:41] <omega_> that's one of the first things gleblanc is gonna have to tackle in gnome cvs, and I need to explicitly offer to help out with that... [22:41] <dobey> eh [22:41] <dobey> i'll probably fix it sometime [22:41] <dobey> but it's fucked for now [22:42] <dobey> and no, my spec files work just fine ;-) [22:42] <omega_> heh [22:43] <omega_> I discovered recently that several of the spec files I wrote a long time ago became the official spec files for several projects [22:43] <omega_> I didn't know that <g> [22:43] <dobey> heh [22:43] <dobey> it's not hard to get that done [22:43] <dobey> just ask hadess ;-P [22:55] <ajmitch> l8r [22:55] Nick change: ajmitch -> ajbusy [23:06] Nick change: dobey -> dobey-out [23:15] dobey-out (do...@ip...) left irc: Read error to dobey-out[ip131.bedford3.ma.pub-ip.psi.net]: EOF from client [23:19] dobey-out (dobey@38.32.11.131) joined #gstreamer. [23:19] Nick change: taaz -> taaz-away [23:22] Kuroyi (ri...@ub...) left irc: Read error to Kuroyi[ubr-35.82.223.wmelbourne.cfl.rr.com]: EOF from client [23:25] Kuroyi (ri...@ub...) joined #gstreamer. [23:27] ChiefHighwater (paul@63.105.18.226) left irc: [23:35] Uraeus (csc...@c2...) joined #gstreamer. [23:35] <Uraeus> hello [23:35] <dobey-out> yo [23:35] Nick change: dobey-out -> dobey [23:36] <Uraeus> dobey: anything fun happened today? [23:36] <dobey> haha [23:36] <dobey> that's a good one [23:37] <Uraeus> dobey: ok, anything not boring? [23:37] <dobey> oh man [23:37] <dobey> you are funny [23:38] <Uraeus> well after being out on the sea today with no wind, the sailboat engine stopping and having to pull it with a small rowboat I feel entitled :) [23:38] <Uraeus> believe me it is hard work rowing that little boat with a 3 tonn sailboat hanging after it [23:39] <omega_> whoops [23:39] <dobey> heh [23:40] <Uraeus> my arms feel like somebody drove over them [23:40] <omega_> then stop napping in the street... ;-) [23:41] <Uraeus> hmm, what is this a new kind of spam, ...want to be TV commercial star... [23:43] <Uraeus> if I reply yes I guess I will be on 500 spam mailing lists in a second [23:44] <omega_> no, more like 500msec [23:46] <Uraeus> hehe [23:47] Nick change: hds-busy -> hadess [23:47] <hadess> buya [23:48] <dobey> hmm [23:48] <dobey> i need something [23:48] <dobey> eek [23:48] <hadess> omega_: you seen my mail ? [23:48] <omega_> yup [23:49] <hadess> i think the problem is pretty much the same i had with switching between PAUSE and PLAY states fast [23:49] <omega_> why does speed have anything to do with it? [23:49] <omega_> that indicates a design flaw in gnome-vfs.. [23:50] <hadess> we're not using any mutexex right now [23:50] <omega_> uhm [23:50] <hadess> mutexes even [23:50] <omega_> ok, I guess that would matter then [23:50] <omega_> you should use them internally then, in gnomevfssrc [23:51] <hadess> wtay is going to help me with that soon [23:51] <omega_> ok [23:51] <dobey> *sigh* [23:51] <hadess> or if you want to give it a try... [23:52] <omega_> no, I should get this filesrc working ;-) [23:52] Nick change: ajbusy -> ajmitch [23:52] <hadess> heh, ok [23:52] <omega_> besides, actually you shouldn't need any changes in the long run [23:52] <omega_> once we have cothread locking for set_arg, it should cease to be a problem [23:55] <hadess> oh, ok [23:55] <omega_> that'll serialize the chainfunc and set_arg code [23:58] <hadess> and when do you expect to have that ? [23:58] <omega_> dunno, you wanna give it a try? it's pretty easy, more bulk code changes than anything [23:58] <omega_> I think it's mostly implemented already, you just have to make use of it [23:58] <omega_> checking [23:58] <omega_> yup, it's implemented already <g> [23:59] <hadess> hehe [23:59] <omega_> instead of setting gtkobject->set_arg, set gstelement->set_arg [23:59] <hadess> where ? [00:00] --- Sun May 27 2001 [00:00] <hadess> in my app ? [00:00] <omega_> class_init [00:00] <omega_> your element [00:00] <hadess> oh, right [00:00] <omega_> also, if you have any good ideas for s/bytes_per_read/xxxx/, let me know <g> [00:05] <hadess> gstelement_class->set_arg = gst_gnomevfssrc_set_arg; <- that ? [00:05] <omega_> yup [00:05] <hadess> didn't work [00:05] <omega_> compile or run? [00:05] <hadess> run [00:06] <omega_> how did it fail? [00:06] <hadess> i replace that: gtkobject_class->set_arg = gst_gnomevfssrc_set_arg; [00:06] <hadess> Gtk-ERROR **: file gtkobject.c: line 808 (gtk_object_arg_set): assertion failed: (oclass->set_arg != NULL) [00:06] <omega_> um [00:06] <omega_> odd [00:07] <omega_> uh oh [00:07] <omega_> set_arg conveniently can't be overridden [00:07] <omega_> I guess that makes sense [00:08] <omega_> sorta [00:08] <hadess> heh [00:08] <omega_> um, lemme try some stuff [00:21] Nick change: Uraeus -> Ura_brb [00:21] <dobey> blee [00:21] <dobey> hadess: how is xbox coming along? [00:22] <hadess> dobey: what xbox ? i don't do no xbox [00:22] <dobey> heh [00:22] <dobey> rhythmbox or whatever [00:22] <dobey> (x)box [00:22] <hadess> *smack* [00:22] <omega_> *box [00:22] <dobey> heh [00:22] <dobey> LoveBox [00:22] <omega_> \W+box [00:23] <dobey> well [00:25] <dobey> how is it? [00:25] <omega_> um, /me is confused [00:25] Action: omega_ dives into the gtk code some more [00:25] <dobey> heh [00:27] <hadess> not worked on it yesterday, nor today. [00:27] <hadess> worked on gst instead [00:27] <hadess> today,... [truncated message content] |
From: <wim...@ch...> - 2001-05-29 04:31:37
|
[06:52] ChiefHighwater (pa...@su...) left irc:=20 [07:17] Nick change: ajmitch -> ajbusy [07:37] sienap (sy...@ip...) left irc: sienap has no rea= son [07:48] steveb (st...@no...) joined #gstreamer. [07:49] Nick change: ajbusy -> ajmitch [07:51] steveb (st...@no...) left irc: [x]chat [08:43] Nick change: ajmitch -> ajzzzz [08:47] ajzzzz (aj...@p1...) left irc: Ping timeout fo= r ajzzzz[p1-max3.dun.ihug.co.nz] [08:50] big_T (th...@fy...) left irc: Ping timeout for big_= T[fysgr763.sn.umu.se] [08:56] _gst_newt_ joined #gstreamer. [08:57] big_T (th...@fy...) joined #gstreamer. [09:04] omega_tv (om...@om...) left irc: Read error to omega_tv[ome= gacs.net]: Connection reset by peer [09:07] omega_ (om...@om...) joined #gstreamer. [09:07] matth_ (ma...@bo...) left irc: Ping t= imeout for matth_[boisdslgw5poolb76.bois.uswest.net] [09:07] matth_ (ma...@bo...) joined #gstreame= r. [09:13] ajzzzz (aj...@p5...) joined #gstreamer. [09:34] Nick change: taaz -> taazzzz [09:42] mourad (foobar@D5E003D7.kabel.telenet.be) left irc: Ping timeout = for mourad[D5E003D7.kabel.telenet.be] [09:50] steveb (ste...@21...) joined #gstreamer. [09:50] <omega_> yo [09:51] <steveb> yo [09:51] Action: omega_ is building an rpm for pkgconfig, including writin= g the spec file, fixing the build, etc. [09:52] <steveb> gstplay bails because I don't have Xv - could we have fa= llbacks to sdlsink -> aasink -> fakesink? [09:52] <omega_> check with wtay/arik [09:52] <omega_> I dunno the player code [09:52] <steveb> ok [09:53] <omega_> but colorspace should be much better than it is [09:53] Action: omega_ notices that =02make distcheck=02 is still working= , even [09:56] matth_ (ma...@bo...) left irc: Ping t= imeout for matth_[boisdslgw5poolb76.bois.uswest.net] [09:57] mourad (foobar@D5E003D7.kabel.telenet.be) joined #gstreamer. [09:57] <omega_> yo [10:00] big_T (th...@fy...) left irc: Ping timeout for big_= T[fysgr763.sn.umu.se] [10:05] big_T (th...@fy...) joined #gstreamer. [10:18] capuchin (jb...@or...) joined #gstreamer. [10:18] <omega_> yo [10:18] <capuchin> Yo. [10:18] <capuchin> Where's the css code for gstreamer 0.1.1 ? [10:19] <capuchin> Anyone? [10:19] <omega_> the css plugin is very old, it doesn't interface with an= y currently available css packages afaict [10:19] <capuchin> Bah. [10:19] <omega_> gstreamer is not a dvd player [10:19] <capuchin> I've got libcss working fine... xine plays dandily. [10:19] <capuchin> I know, I know... but gstmediaplay should work, no? [10:20] <capuchin> I'm just reviewing the quality of the library... I'm w= orking on a couple of projects where it looks like it might be handy. [10:20] <omega_> depends on what you try. CVS in the last couple days ha= s started working again for most common media types [10:20] <omega_> a major branch was merged Friday [10:20] <capuchin> Hmmm. [10:20] <capuchin> I only just heard of gstreamer five minutes ago. [10:20] <omega_> a release is being worked on right now [10:20] <omega_> heh [10:21] <omega_> 0.2.0 should be out within the week [10:21] <omega_> and work a lot better than 0.1.1.... [10:21] <capuchin> Do you know if there's a css plug-in for mplayer? [10:21] <omega_> no idea, I wouldn't doubt that one exists somewhere [10:21] <capuchin> Yeah, I've seen vidcaps that lead me to believe one ex= ists. [10:21] <capuchin> Screenshots, that is. [10:21] <omega_> else they ripped offline [10:21] <omega_> which is what I did for my mpeg2 sample streams ;-) [10:22] <omega_> what kind of projects are you working on? [10:23] <capuchin> Well, for tonight, I just wanted to rip some DVDs for = an editor I'm toying with... I need some large MPEG2 files in a hurry... [10:23] <capuchin> I was going to rip them with mplayer using the divx ou= tput... but it won't read the css natively. [10:23] <omega_> best bet is to get libcss and use the apps that come wit= h it [10:23] <capuchin> I don't know where to turn to rip some big mpegs. [10:23] <capuchin> I've got libcss... Bah... there are sample apps? [10:23] <capuchin> I'll look again. [10:23] <omega_> the README gives the steps to just do the decrypt, yes [10:24] <capuchin> Recall what they're called? I'll unpack the source ag= ain. [10:24] <omega_> maybe not in the package, but the tarball should have th= em, last I checked [10:24] <capuchin> Oh, everything on this system is built from source... = even init and getty. [10:24] <omega_> um, I forget what they were called, I haven't used them = in over a year [10:26] <capuchin> Do you know where the libcss homepage is? I want to c= heck my revisions. [10:26] <omega_> linuxvideo.org somewhere [10:26] <omega_> link on the left side [10:26] <capuchin> Heh... I was just checking there. [10:37] <omega_> blegh. make distcheck is a pain, it takes way to long [10:37] capuchin (jb...@or...) left irc: Ping timeout for= capuchin[orifice.bitmine.net] [10:44] capuchin (jb...@or...) joined #gstreamer. [10:44] <capuchin> Whoops. [10:44] <omega_> heh [10:45] <capuchin> Well, I got css_cat working. Phew. [10:45] <omega_> getting valid mpeg2 streams out of it? [10:45] arik (ar...@sd...) joined #gstreamer. [10:45] <omega_> yo [10:45] <arik> my sounds finally works!!! [10:46] <arik> ;-) [10:46] <omega_> oooh [10:46] <arik> and hey [10:46] <capuchin> What was wrong with it, arik? [10:46] <arik> needed to upgrade from 2.4.2 to 2.4.3 [10:46] <arik> which took so damn long [10:46] <omega_> not 2.4.5 ? [10:46] <arik> not yey [10:46] <arik> went the rpm route this time ;-P [10:46] <omega_> ah [10:47] <arik> man though [10:47] <capuchin> Is there are 2.4.5? [10:47] <arik> it sucks downloading 21mb source files on a 56.6 [10:47] <omega_> capuchin: as of yesterday, yeah [10:47] <arik> it is just nice to have sound again [10:48] <omega_> since you're mr. player, it seems like it'd help.... ;-) [10:48] <arik> hehe [10:48] <arik> yes [10:48] <arik> ;-) [10:49] <capuchin> Well, hell.. I built 2.4.4 just about two days ago. [10:49] <steveb> arik: gstplay bails because I don't have Xv - could we h= ave fallbacks to sdlsink -> aasink -> fakesink? [10:49] <omega_> I've been running 2.4.4 for most of a month... [10:49] <arik> hmm [10:49] <arik> that would be a good idea [10:50] <steveb> maybe not aasink, but fakesink would be good - at least = mp3s would work (and audio tracks of movies) [10:50] <capuchin> I've never gotten Xvideo to work properly. No idea wh= y. [10:50] <omega_> what video card? [10:50] <capuchin> mga g400 [10:51] <omega_> odd, I've never had any problems [10:51] <capuchin> XFree 4.0.1? [10:51] <omega_> yeah [10:51] <omega_> stock rh7.0 [10:52] <capuchin> Yeah, when I do mplayer -vo xv file.mpg, I get a blue = overlay, but no actual video. [10:52] <omega_> hmm [10:52] <omega_> what about smpeg ? [10:52] <capuchin> But I can use the mga_vid device overlay fine... and f= b [10:52] <capuchin> Never used it. [10:52] <omega_> give it a try, it might work [10:53] <capuchin> That's in the SDL library? [10:53] <omega_> it's a separate app, but yeah, based on sdl [10:53] <arik> i should at least let the user choose which one to use [10:53] <omega_> ships with recent redhat [10:53] <capuchin> I don't think I have sdl installed. [10:53] <capuchin> My system is from scratch, omega_. [10:53] <omega_> libsdl.org [10:56] <capuchin> Well, we'll see how the build goes. [10:56] <capuchin> Gotta love make -j [10:56] <omega_> you have a dual? [10:57] <capuchin> Yup. PIII 850s. [10:57] <omega_> nice [10:57] <capuchin> I dig it. 512MB as well... looking to put another in = soon to double it. [10:58] <capuchin> I had a hell of a time with a badly manufactured A-bit= BP6 (dual celeron)... but now I have the VP6 with no problems. [10:59] <omega_> hmm, I have a dead BP6 [10:59] <omega_> I'm waiting for the second-gen 760MP boards though [10:59] <capuchin> I burned a hole in mine trying to solder on a differen= t capacitor to fix their manufacturing flaw. [10:59] <omega_> whoops [10:59] <capuchin> Whoops. [11:00] <capuchin> Yup. :) [11:07] Action: omega_ must sleep now [11:07] <arik> night dude [11:07] <omega_> l8r [11:07] <capuchin> Night. [11:07] omega_ (om...@om...) left irc: killall -SLEEP omega_ [11:13] big_T (th...@fy...) left irc: Ping timeout for big_= T[fysgr763.sn.umu.se] [11:20] capuchin (jb...@or...) left irc: Leaving [11:42] steveb (ste...@21...) left irc: quit has st= eveb, yes [12:42] thomas (th...@21...) joined #gstreamer. [13:41] greg_ (gr...@ho...) joined #gstreamer. [13:42] big_T (th...@fy...) joined #gstreamer. [14:32] arik (ar...@sd...) left irc: if you = had such a dream... [14:41] sienap (sy...@ip...) joined #gstreamer. [14:41] <sienap> hi all [16:01] greg_ (gr...@ho...) left irc: Ping timeout for greg_[home.= sente.pl] [16:19] conny (co...@f-...) joine= d #gstreamer. [16:20] sienap (sy...@ip...) left irc: sienap_ has no re= ason [16:25] thomas (th...@21...) left irc: [x]chat [16:27] conny (co...@f-...) left = irc: [x]chat [17:03] dobey (do...@ip...) joined #gstreamer. [17:16] sadnap (sy...@ip...) joined #gstreamer. [17:16] <sadnap> Bah.. [17:16] <sadnap> a #linux.nl dude died :((((( [17:16] <sadnap> BAH [17:17] <dobey> hrmm [17:17] <dobey> this site fuckin RULES [17:18] <sadnap> fucking non-sense [17:18] <sadnap> bAH [17:18] sadnap (sy...@ip...) left irc: Disconnecting [17:18] <dobey> sadnap: raverporn.net should cheer you up ;-) [17:18] <dobey> aww [17:18] <dobey> shit [17:22] <dobey> oh well [17:22] Action: dobey assimilated a developer [17:43] Uraeus (csc...@c2...) joined #gstreamer. [17:43] <Uraeus> hi [17:44] <dobey> * dobey assimilated a developer [17:46] <Uraeus> dobey? [17:46] <dobey> yeah [17:46] <dobey> that -other- web browser... [17:46] <dobey> btw [17:47] <Uraeus> aha [17:47] <dobey> i think this is a web site i could pay for [17:47] <dobey> raverporn.net [17:47] <dobey> heh [17:47] <dobey> IT RULES! [17:47] <Uraeus> dobey: you know what what nick Dan Mueth use online? [17:47] <dobey> muet [17:49] steveb (st...@no...) joined #gstreamer. [18:00] <Uraeus> bloody hell, DanM has wrecked the GNOME summary archive [18:00] <dobey> NOOOOOOOOOO [18:01] <dobey> i'm gonna get some food [18:01] <dobey> brb [18:10] <dobey> mmmm [18:15] <Uraeus> ok, things are somewhat fixed now and the summaries are = out [18:17] <dobey> new one?! [18:17] <dobey> where? [18:19] hadess (ha...@pc...) joined #gst= reamer. [18:19] <Uraeus> http://developer.gnome.org/news/summary/2001_May20-May25= .html [18:20] <hadess> yo [18:21] <dobey> yo [18:21] <Uraeus> yo [18:21] <dobey> wheeeeeee [18:33] <Uraeus> dobey: with your newly aquired hacker and the summary an= nouncement you will soon have a team big enough to take on anything :) [18:33] <dobey> Uraeus: heh [18:34] <dobey> i'm building me kernel now [18:34] <dobey> i hope mozilla doesn't fuck with it... [18:34] <Uraeus> your own kernel; is dobux on the way? <g> [18:35] <dobey> as in 2.2.19 [18:35] <dobey> so, no, not a package =3DP [18:36] <Uraeus> dobey: hmm, you don't even have your first release out o= f Dobux and already inflating version numbers [18:36] <dobey> huh? [18:36] <Uraeus> dobey: just kidding [18:36] <dobey> i'm building kernel 2.2.19 fool [18:36] <dobey> =3DP [18:38] <dobey> damn [18:38] <dobey> i need X4 [18:41] <Uraeus> interesting: http://freej.dyne.org/ [18:42] <dobey> maybe my visor will work properly now [18:42] <dobey> ... [18:42] <dobey> weird [18:43] <dobey> hmm [18:43] <hadess> Uraeus: these guys need to come over... [18:43] <dobey> encompass will have junkbuster support [18:43] <dobey> hehe [18:44] <hadess> dobey: proxy you mean ? [18:45] <dobey> hadess: i mean Tools->JunkBuster Config [18:45] <dobey> i love bonobo [18:45] <hadess> dobey: n33t [18:45] <dobey> yeah [18:45] <Uraeus> dobey: what is the command line string to get evolution = to open a composer window? [18:46] <hadess> the config is not complicated anyway [18:46] <dobey> Uraeus: there isn't one yet [18:46] <dobey> hadess: well, gijbman or whatever it is will be the confi= g tool [18:46] <Uraeus> dobey: buuu :), another missing feature is BCC [18:47] <dobey> Uraeus: no, i think bcc is there [18:47] <hadess> Uraeus: you need to make it show the field that's all [18:47] <Uraeus> found it, thnx [18:47] <hadess> Uraeus: in View, select the BCC: field [18:47] <dobey> haha [18:48] <dobey> i gots an email already [18:48] <dobey> damn [18:48] <dobey> Subject: RE: HTML Rendering Component Developer [18:49] <Uraeus> ok, and the content? [18:49] <dobey> > Hi!!! [18:49] <dobey> >=20 [18:49] <dobey> > I want try to help your project! [18:49] <dobey> >=20 [18:49] <dobey> > If you need me, my SouceForge login is dpinto [18:49] <dobey> heh [18:49] <Uraeus> what can I say, the power of the press if great :) [18:50] <dobey> yeah [18:50] <hadess> hehe [18:51] <hadess> as soon as i'm fully awake, i will start making an iceca= st module for gnome-vfs [18:51] <hadess> icecast://foo.net:800 [18:52] <Uraeus> dobey: to return the favour could you go and kick Iain i= n the arse until he begs for mercy? [18:52] <dobey> Uraeus: haha, why? [18:52] <dobey> hadess: eek? [18:52] <dobey> hadess: doesn't it just use http? [18:52] <Uraeus> dobey: he is flaming me for posting stories about XFCE [18:53] <dobey> oh [18:53] <dobey> heh [18:53] <hadess> dobey: because icecast and shoutcast are very very broke= n http servers [18:53] <Uraeus> so please kick hard and long :) [18:53] <dobey> hadess: heh [18:54] Action: dobey wonders how many http monikers will be made for gno= me-vfs [18:55] <hadess> this is not a moniker, this is a method [18:56] <dobey> it's a method on top of http? [18:56] <dobey> haha [18:56] <dobey> the skills profiles on sourceforge are totally useless [18:57] <Uraeus> hmm, what I find worrying is that there has been no Naut= ilus updates for a couple of days, we currently have more development on = GMC than nautilus [18:57] <hadess> no, this is a gnome-vfs method, like cdda access has a m= ethod [18:58] Nick change: hadess -> hds-food [18:58] <dobey> hmm [18:58] Action: dobey is confused now [18:58] <dobey> Uraeus: really? [18:58] <Uraeus> dobey: yup [18:59] <dobey> Uraeus: heh, that's quite amusing [18:59] <Uraeus> dobey: yes, but it is also kinda sad, hope things pick u= p as the Eazel hackers get resettled [19:00] <dobey> yeah, they should [19:00] <dobey> hrmm [19:00] <dobey> i'll be right back [19:00] dobey (do...@ip...) left irc: new kern= el [19:06] sadnap (sy...@ip...) joined #gstreamer. [19:06] <sadnap> hi.. [19:07] <Uraeus> hi sadnap [19:07] <sadnap> a mate of my/us past away :( [19:07] <Uraeus> someone I know? [19:07] <sadnap> nope [19:07] <sadnap> a #linux.nl dude ircnet [19:07] <sadnap> 31 years. [19:07] <sadnap> hart atck [19:07] <sadnap> attack [19:08] <Uraeus> bad [19:08] <sadnap> he had a little one some weeks ago [19:08] <sadnap> and this one was fatal.. [19:08] <sadnap> man.. [19:08] <sadnap> i feel like shit about it [19:09] dobey (do...@ip...) joined #gstreamer. [19:09] Action: dobey kicks linux in the pants. [19:09] <Uraeus> wb [19:09] <sadnap> i am off again [19:09] sadnap (sy...@ip...) left #gstreamer. [19:09] greg_ (gr...@ho...) joined #gstreamer. [19:09] <dobey> ack [19:09] <Uraeus> hi greg_ [19:09] <greg_> hello all ! [19:09] <dobey> kernel panic's suck [19:10] <dobey> god damn [19:15] dobey (do...@ip...) left irc: take 2 [19:20] dobey (do...@ip...) joined #gstreamer. [19:20] <dobey> whee [19:21] Nick change: taazzzz -> taaz [19:29] Nick change: taaz -> taaz-food [19:32] dobey (do...@ip...) left irc: take 3 [19:41] Nick change: steveb -> steveb-away [19:43] dobey (do...@ip...) joined #gstreamer. [19:43] <dobey> whee [19:43] <dobey> now [19:43] <dobey> i gotta get my trackpad working too [19:51] omega_ (om...@om...) joined #gstreamer. [19:51] <dobey> y0 [19:52] <omega_> yo [19:52] <dobey> i got's kernel 2219 now [20:00] <dobey> eek [20:00] <dobey> i wonder if it supports /dev/shm [20:03] Nick change: wtay-zZz -> wtay [20:03] <wtay> hello [20:03] <dobey> y0 [20:03] <omega_> yo [20:04] Action: dobey wonders what else he needs to get working with this= kernel... [20:04] <wtay> H1, d0b3y, omega_ [20:04] <dobey> h3h [20:04] <wtay> d0h [20:11] <omega_> wtay: turn off your 31337 filter.... [20:11] <dobey> man [20:11] <dobey> i'm freqin tired [20:11] <Uraeus> dobey: any more mails yet? [20:11] <dobey> Uraeus: nope [20:11] <dobey> Uraeus: just that one [20:12] Nick change: hds-food -> hadess [20:12] <hadess> yo guys [20:12] <omega_> ok, distcheck fails in xvideosink: [20:12] <Uraeus> hi hadess [20:12] <omega_> ../../../plugins/xvideosink/imagetest.c: In function `id= le_func': [20:12] <omega_> ../../../plugins/xvideosink/imagetest.c:36: structure ha= s no member named `sbuf' [20:12] <omega_> . . . .=20 [20:12] <omega_> during make check [20:12] <Uraeus> hey cool, new developer voluntering to help out with hav= ocs suggestions on gnotices :) [20:13] <omega_> which is that? [20:14] Nick change: wtay -> wtay-brb [20:14] <Uraeus> omega_: read my latest summary to find out :) [20:14] <Uraeus> omega_: I can do the license audit btw. [20:14] <omega_> Uraeus: you sent me a pre-release of it for checking? <g= > [20:14] <Uraeus> weeell [20:14] <omega_> mu [20:15] <Uraeus> actually I planed on giving you the link before release,= but Dan Mueth found it on CVS and thought it was unrelased due to a mist= ake so he released it [20:15] <omega_> doh [20:16] <Uraeus> omega_: the summary reaches suprisinly many people, I th= ink get got around 100 new developers siging onto the GNOME-love list whe= n we announced it last week [20:16] <omega_> wow [20:17] <Uraeus> omega_: so hopefully many people will see out list of pl= ugins needing maintainers and sign on :) [20:17] <omega_> heh [20:17] <Uraeus> s/out/our/ [20:18] <omega_> s/togheter/together/ [20:18] <hadess> Uraeus: how many of them are already gnome developers or= just lurkers , hmm ? [20:19] <Uraeus> hadess: well, I think around 5-6 announced themselves at= least when I looked last time, but even if we just get 10 new developers= from this it is a huge gain [20:19] <hadess> yeah [20:20] <Uraeus> omega_: should we try to assimilate this: http://freej.d= yne.org/ [20:20] <hadess> omega_: what do you think the latency would be like if i= destroy my pipeline between each song ? [20:22] <omega_> hadess: depends... [20:23] <hadess> omega_: the thing is that i'm wondering how to support m= ultiple types of media (ogg, mp3, mod...) with a single pipeline [20:23] <omega_> if you can pre-assemble the new pipeline in parallel, it= shouldn't be significant [20:25] <hadess> that would be tricky [20:25] <hadess> except if i have a pipeline for each type of supported m= edia, but that would be a mess [20:25] <omega_> but I wouldn't expect it to be significant [20:25] <omega_> the latency of autoplugging, that is [20:25] <hadess> ok [20:26] <omega_> try it, of course... [20:26] <hadess> will do [20:26] <omega_> Uraeus: some mail got lost, it had changes to the roadma= p [20:27] <omega_> linuxrising dropped it, what addr should I use instead? [20:27] <Uraeus> ur...@li... [20:27] <omega_> sent [20:28] <omega_> great, shell1 is still down [20:35] <Uraeus> omega_: do we have the quicktime plugin in now? I saw a = qt directory with some code? [20:35] <omega_> apparently, yes [20:36] <omega_> dunno it status, it requires openquicktime, don't think = there are any codecs built in [20:36] <omega_> then again openquicktime is still very much like qt4l, a= nd we some energy must be put into it to make it not suck [20:37] <omega_> but rr and maybe pioneer might have some corporate inter= est in working on that, so.... [20:37] <omega_> since qt is the official mpeg4 file format now.. [20:38] <Uraeus> aha [20:39] Action: Uraeus is busy poluting his system with a unholy gui lib = made by some trolls [20:40] <omega_> ? [20:40] <omega_> um, not quite [20:40] <Uraeus> testing some cruddy app which depend upon it [20:40] <omega_> hmm, ok. I thought you were joking qt =3D=3D qt [20:40] <Uraeus> omega_: would be very cool if you could get Pioneer to i= ssue a statement saying that they will be sponsoring GStreamer developmen= t [20:40] Nick change: taaz-food -> taaz [20:41] <Uraeus> or at least allow us to announce it [20:41] <omega_> Uraeus: as soon as we get the developers in here, I'll w= ork on that with them [20:41] <omega_> I'm hoping tomorrow, since they were targetting Friday f= or mergability with HEAD [20:41] <Uraeus> omega_: cool, since there is nothing like namedroping to= get more attention=20 [20:41] <omega_> heh [20:42] <omega_> esp two Senior Pioneer Research engineers [20:42] <Uraeus> yup [20:42] <Uraeus> omega_: talked more with Sopwith btw? [20:42] <omega_> not in the last few days, no [20:42] <omega_> but he's been here a bit [20:43] Nick change: wtay-brb -> wtay [20:43] <wtay> checking for libasound headers version >=3D 0.9.0... not p= resent. [20:43] <wtay> configure: error: Sufficiently new version of libasound no= t found. [20:43] <wtay> --*(root@smash =DB*- Mon May 28 20:43:42=20 [20:44] <Uraeus> wtay: yeah, Gman complained about that too [20:44] <omega_> wtay: I've commented out xvideosink/imagetest for now, i= t's not compiling [20:44] <omega_> wtay: hrm, checking [20:44] <omega_> ok, that's good [20:44] <wtay> omega_: ok [20:44] <omega_> the warning for alsa is not nice though [20:45] <omega_> we're actually checking for alsa =3D=3D 0.5.x [20:45] <omega_> so I check for >=3D 0.9.0 second, and turn alsa *off* if= it's there [20:45] <omega_> in the meantime alsa.m4 complains that it's not new enou= gh [20:45] <omega_> even though it's correct [20:45] <omega_> we should print out a message in the not found case expl= aining that [20:45] <omega_> I'll do that [20:45] <wtay> but it stops the configure with an error now [20:45] <omega_> oh? [20:45] <omega_> how nice of them [20:46] <omega_> then we need an alternate way of checking [20:46] <wtay> yes, notice how I pasted my prompt :) [20:46] <omega_> ah [20:46] <wtay> else I wouldn't mind [20:46] <omega_> you'd think that with alsa being so horribly incompatibl= e between releases that their m4 would handle this case [20:46] <omega_> but I'm not entirely impressed with the alsa development= process [20:47] <wtay> nope [20:47] <wtay> so, it's the .m4 that errors and aborts.. [20:47] <omega_> yes [20:47] <omega_> checking if it can be disabled [20:48] <omega_> um. [20:48] <wtay> btw, freej should be assimilated ASAP [20:48] <omega_> I don't have an 0.5.x alsa.m4, you'll have to look at it= to see if it can be avoided [20:48] <omega_> wtay: yup [20:48] <omega_> anyone send mail to him yet? [20:49] <wtay> Uraeus: ? [20:49] <taaz> let Uraeus do it ;) [20:49] <wtay> AM_PATH_ALSA(0.9.0...) aborts here... [20:49] <omega_> check /usr/share/aclocal/alsa.m4 [20:50] <wtay> AC_MSG_ERROR(Sufficiently new version of libasound not = found.)] [20:50] <omega_> yup, but no indication of how it fails from that point [20:51] <wtay> that must be the line... [20:51] <omega_> oh, right, _ERROR [20:51] <omega_> wonderful [20:52] <wtay> --disable-alsatest would work... [20:52] <wtay> trying [20:52] <wtay> nope [20:52] <Uraeus> wtay: ok I mail em [20:52] <omega_> Uraeus: did you cc to gst-devel ? [20:53] <wtay> It's C++ project... [20:53] <Uraeus> omega_: since I haven't sent the mail yet I haven't cc'e= d either :) [20:53] <omega_> ok, but don't forget then [20:54] Nick change: ajzzzz -> ajmitch [20:54] <Uraeus> morning ajmitch [20:54] <ajmitch> morning [20:54] <wtay> hmmm, I vote for removing the alsa check and leave it as i= s [20:54] <wtay> hi aj [20:54] <wtay> s/is/it was/ [20:54] <ajmitch> ah yes, i had a problem with that too [20:54] <omega_> wtay: except it fails if you have 0.9 installed, alsa sp= ews 100+ lines of compile errors [20:55] <ajmitch> i just did apt-get install libasound2-dev [20:55] <wtay> omega_: bleagh [20:55] <omega_> we must have some way of only compiling for 0.5 [20:55] <ajmitch> didn't see if it failed ;) [20:55] <wtay> omega_: can we ifdef in the source on the version number? [20:55] <omega_> yeah, use SOUNDLIB_VERSION_MINOR somehow [20:55] <omega_> checking feasibility [20:55] <wtay> it's not nice.. [20:56] <omega_> really really ugly, and the alsa people should have anti= cipated this [20:56] <wtay> alsasrc/sink has quite a lot of code [20:56] <omega_> since 0.2, 0.3, 0.4, 0.5, and 0.9 are all completely inc= ompatible [20:56] <wtay> pff [20:58] <wtay> it looks like they could use someone with a bit of managem= ent skills [20:58] <ajmitch> like Uraeus? ;) [20:58] <wtay> ajmitch: we need him ourselves :) [20:58] <ajmitch> hehe [21:00] <Uraeus> ;) [21:01] <wtay> anyone who understands that a configure script is supposed= to *check* for a particular version, as opposed to error on everything i= t doesn't understand, will do :) [21:01] <ajmitch> Uraeus: is it good to feel wanted? ;) [21:01] <omega_> heh [21:01] <omega_> of course, with this method we'll get an empty plugin [21:01] <wtay> hum, yeah [21:01] <omega_> I'd much much rather find a way to kill the buid of it [21:01] <wtay> idef out plugin_desc [21:02] <omega_> btu we still have an empty .so [21:02] <wtay> we could write a little testapp that prints the version nu= mbers... [21:02] <Uraeus> ajmitch: I feel all warm inside :) [21:02] <omega_> maybe grep alsa.h for the version and check [21:02] <wtay> omega_: or grep for it, yes [21:02] <wtay> basically, write our own check [21:03] <omega_> ok, I'll give it a shot [21:03] <ajmitch> bbl [21:03] Nick change: ajmitch -> aj_uni [21:03] <taaz> doesnt the alsa.m4 have that test app? [21:03] <omega_> yes, but it's broken [21:04] <Uraeus> poor qt, once again it failed to gain a foothold on my c= omputer and was killed by a angry rpm -e passing by [21:04] <omega_> hehehe [21:06] <Uraeus> what is the difference between: make clean, make realcle= an and make distclean? [21:06] Action: wtay runs make distcheck [21:07] <omega_> wtay: I've got a few more things to check in, sec. [21:07] <wtay> ok [21:07] <omega_> ok, some makefile changes, still working on alsa fix [21:08] <omega_> gimme a min [21:12] <omega_> libtool: link: `../../../libs/winloader/libwinloader.la'= is not a valid libtool archive [21:12] <wtay> ugh [21:12] <omega_> probably a $(srcdir) vs. $(builddir) thing again [21:13] <wtay> that's for the codectest app... [21:13] <wtay> it probably doesn't even work... [21:13] <omega_> then it should be commented out [21:13] <wtay> yes [21:17] Nick change: steveb-away -> steveb [21:19] <wtay> omega_: are you doing that? [21:19] <Uraeus> brb [21:19] <omega_> wtay: I'm attempting to build a test [21:20] <wtay> ok [21:22] Uraeus (csc...@c2...) left irc: Ping timeout f= or Uraeus[c224s9h5.upc.chello.no] [21:31] <omega_> ok, update configure.in, try that [21:32] <omega_> and plugins/avi/Makefile.am to solve winloader.la proble= m [21:35] <wtay> ok [21:35] <omega_> works? [21:35] <wtay> hold on.. still running autogen.sh [21:36] <wtay> plugins/avi/Makefile.am:25: invalid unused variable name: = `codectest_SOURCES' [21:36] <wtay> plugins/avi/Makefile.am:24: invalid unused variable name: = `codectest_LDFLAGS' [21:36] <wtay> doesn't fail though [21:36] <omega_> I didn't change that [21:37] <omega_> hmm, checking [21:37] <omega_> odd [21:37] <wtay> ok, my fault, I uncommented check_PROGRAMS [21:37] <omega_> ah, that would kinda do it ;-) [21:39] <wtay> running make distcheck again [21:41] <omega_> I'm gonna let you do that, since it takes for =02ever=02= on my machine [21:41] <wtay> it takes some time here too :( [21:44] thomas (th...@ad...) joined #gstreamer. [21:44] <thomas> hi everyone [21:45] <wtay> hi [21:45] <thomas> uhm, if people don't respond to propositions I've made, = should I take that as a) they don't care b) they don't agree c) they don'= t have time to think about it ? [21:46] <aj_uni> which propositions? ;) [21:46] <omega_> probably a mix of the above ;-| [21:46] <thomas> too bad. I'm really thinking about that control pads is= sue; I could really use them for what I want to do right now. [21:47] <aj_uni> probably too close to 0.2.0 to think about it right now = ;) [21:48] <dobey> lo [21:48] <wtay> too close to implement it, not to think about it [21:48] <aj_uni> anyway, i should go to uni now ;) [21:48] <omega_> btw, I noticed something... [21:48] <omega_> we haev =023=02 Thomas's here [21:48] <omega_> thomas, thomas_, and bit_T [21:48] <thomas> hm, I've been thomas_ a while as well, I left xchat open= at work [21:48] <dobey> heh [21:49] <dobey> thomas =3D=3D thomas_ [21:49] <omega_> hmmm, ok [21:49] <thomas> why are you omega_ ? [21:49] <wtay> make distcheck still runnning fine [21:49] <thomas> and not omega ? [21:49] <wtay> thomas: 'cause else he'll be kicked :) [21:50] <omega_> thomas: some itiot romanian kid (who likes death threats= ) has it [21:50] <dobey> hahah [21:50] <steveb> thomas: what do you want to do with control pads? [21:50] <dobey> --- omega :No such nick/channel [21:50] <dobey> hurry! take it back! [21:51] <thomas> steveb: when you consider src and sink pads controlling = the actual flow of real data, going from left to right in diagrams ... [21:51] <thomas> ... control pads would go from top to bottom; they would= act on data to generate temporary control information to other plugins [21:52] <thomas> case in point : a level detector plugin would take in da= ta on it's sink pad ... [21:52] <thomas> ... send out a volume envelope on it's control out pad .= .. [21:52] <thomas> ... which could be taken in on the control pad by a nois= e gate, or a compressor ... [21:52] <wtay> TMPDIR is DBTOHTML_OUTPUT_DIR1789 [21:52] <wtay> Cannot read "gstreamer-manual.sgml". Exiting. [21:52] <thomas> ... which uses this envelope to modify it's sink data in= to it's src data out [21:52] <omega_> wtay: would not surprise me at all if the doc build syst= em can't handle builddir.... [21:52] <thomas> all of this could be done with the current src's and sin= k's structure and a new data type, but it would violate the actual differ= ence between multimedia data and control data [21:52] <thomas> besides, it would keep most of the plugins in the simple= one src/one sink case. [21:53] <wtay> omega_: nope, it doesn't. gst.sgml is there... [21:53] <thomas> steveb: does that give a good picture ? [21:53] <steveb> thomas: yep... [21:53] <wtay> omega_: can I disable it someway? [21:54] <steveb> thomas: last things first, the noise gate would get its = control information from the dynparms api (which i am currently working o= n) [21:54] <omega_> wtay: it's disabled by default, but make dist probably f= orces it on [21:54] <wtay> uhm [21:54] <wtay> omega_: then we can assume everything works [21:54] <thomas> steveb: what's dynparms ? do you have some src or doc on= that ? [21:55] <omega_> wtay: no, we need to get doc building to survive distche= ck, I believe [21:55] <steveb> thomas: but the dynparms api will be able to get its dat= a from elements (or straight from the GUI) [21:55] <omega_> though we can put out pre0... [21:55] <wtay> yup [21:56] <omega_> um, /me adds ABOUT-NLS to cvs [21:56] <wtay> can we put extra strings in the version name? [21:56] <omega_> I think so, yes [21:56] <omega_> micro=3D0-pre0 [21:56] <steveb> er, there is a wiki explaining it all but wiki seems to = be b0rked [21:56] <thomas> steveb: yeah, I noticed that a few days ago as well [21:56] <thomas> omega_: any idea why the wiki's gone pot ? [21:56] <omega_> shell1 is toast (DNE), that probably did it [21:56] <omega_> though that's very very screwed, SF should never have le= t this happe [21:57] <thomas> omega_: do you have a copy of the style sheet somewhere,= maybe you could put it back through cvs, no ? [21:57] <omega_> it's not in cvs, and I can't touch it without shell1 bei= ng there [21:57] <omega_> sf just plain screwed up badly [21:58] <omega_> and there's no news item about it, either [21:58] <steveb> thomas: anyway if what you want is just a stream of numb= ers at a constant rate (much lower than sample rate) then i can't see why= the existing audio types can't be used [21:58] <omega_> wtay: you making the tarball and tarbzll? [21:59] <wtay> omega_: yup, checking out fresh copy now [21:59] <thomas> steveb: I'm not saying I can't, but it feels wrong. Fir= st of all, it *isn't* audio, and second, if you compare it to real-world = examples, that's just how it's done. [21:59] <steveb> thomas: and in the future there will be a need for a con= trol type which could contain discrete events (and this may map directly = to MIDI) [22:00] <thomas> steveb: plus, an element should be able to do something = useful *without* control data, but it probably wouldn't do anything usefu= l without audio in. [22:00] <wtay> omega_: I'm going to commit the configure.ac patch first [22:00] Nick change: hadess -> hds-afk [22:00] <steveb> or the other approach is to just have a control type whi= ch has events and low sample rate numbers [22:00] <omega_> wtay: ok, you're gonna bring configure.ac up to date wit= h configure.in ? [22:01] <wtay> config.status: error: cannot find input file: gst/meta/Mak= efile.in [22:01] <thomas> steveb: well, I'll have to try it that way first anyway,= so I'll concentrate on that for now [22:01] <omega_> wtay: yeah, it's not supposed to be there [22:01] <omega_> wtay: you're out of sync with configure.in [22:01] <wtay> ok [22:02] <steveb> thomas: the idea is that an element will never take cont= rol data on an input pad and produce audio - that is what the dynparms ap= i is for [22:03] <steveb> elements that take control data on a pad will just be tr= ansforming it into more control data [22:03] <omega_> sourceforge.net/projects/jahshakafx/ [22:03] <omega_> I love how these projects are created (and "moving into = high speed") with zero code and a "full everything" description [22:04] <omega_> there's a release, at least [22:04] <omega_> but that's a target for assimilation as well [22:05] <wtay> omega_: sec, I need some time with this ac stuff.. [22:05] <omega_> ok [22:06] <omega_> neat, the build environment for jahshaka is fscked [22:06] <omega_> INCLUDES =3D $(all_includes)\ =20 [22:06] <omega_> -I/root/jahshaka/glf-1.4/ [22:06] <omega_> um [22:09] <wtay> omega_: great, my GSM check is gone... [22:09] <omega_> you put a gsm plugin in cvs? [22:09] <wtay> yes [22:10] <omega_> oh, you did [22:10] <omega_> heh, MENOCLUE [22:10] <wtay> hmm, I thought I had all the checks for it... [22:12] <wtay> omega_: uh, there's quite a large diff between ac and in := ( [22:12] <omega_> yup [22:13] <wtay> mostly the warnings.. [22:15] <taaz> hmm. i have alsa 0.9 stuff installed, so alsa wont build,= but i have a debian alsa plugin... the alsa build dependencies fail caus= e i list old libs... but listing new libs is silly since that just means = the code wont build... grrr. [22:15] Topic changed on #gstreamer by ChanServ!s@ChanServ: INCSCHED1->HE= AD merge COMPLETE!!! [22:17] <wtay> taaz: configure.ac works, I put it in cvs [22:19] <dobey> hahahahha [22:19] <wtay> config.status: error: cannot find input file: plugins/vide= osink/Makefile.in [22:19] Action: dobey reads mail from sourceforge [22:20] <dobey> READ IT NOW [22:20] <dobey> ITS FUNNY! [22:20] <dobey> ** Until you change your password, you will not be able t= o login to the [22:20] <dobey> SourceForge website, use the CVS server or the shell serv= er. *** [22:20] <dobey> how lame [22:20] <omega_> huh? [22:20] <omega_> where is this? [22:22] <dobey> an email they just sent i guess [22:22] <dobey> if you used sf in the past week [22:23] <wtay> omega_: need to create a GSM rule... [22:23] <omega_> I don't have anything yet [22:23] <dobey> oh wow [22:24] <Kuroyi> yeah, I just got that too [22:25] <steveb> omega_: i'm having another crack at ladspa [22:25] <omega_> ok [22:26] <steveb> while there is still a chance it can get into 0.2.0 [22:34] <wtay> omega_: ok, I think I'm ready... doing a fresh checkout [22:34] <wtay> omega_: can you check my gsm test? [22:35] <omega_> ok [22:35] <taaz> i have a suggestion [22:35] <taaz> maybe we should start to use sourceforge features [22:35] <omega_> hah. if they worked reliably [22:35] <omega_> we can't keep our wiki running.... [22:35] <taaz> like tasks and bugs and feature request things [22:35] <omega_> yes, we should [22:36] <wtay> yes [22:36] <omega_> it's quite cumbersome though [22:36] <wtay> omega_: I patched configure.ac to have a 0-pre0 version nu= mber but forgot that in configure.in [22:36] <taaz> and what did the sf people say when you told them this? (h= int hint) [22:37] <omega_> I've suggested several improvements, heard nothing back [22:37] <wtay> me too, I got a T-shirt, but thats all <g> [22:37] <omega_> you got a tshirt?? [22:37] <wtay> yes, I'm wearing it now [22:38] <omega_> hmm [22:38] <wtay> that's not fair if you didn't get one [22:38] <wtay> make dist? [22:38] <omega_> never got one [22:38] <omega_> sec. [22:38] <omega_> getting libgsm [22:38] <wtay> right [22:41] <omega_> ok, it worked [22:41] <omega_> but there's other cleanup we need to do to configure.in [22:41] <omega_> but snap HEAD as pre0 now, tag it even [22:42] <omega_> RELEASE-0_2_0-PRE0 [22:42] <wtay> let's try to make dist first and then tag it... [22:43] <omega_> right [22:43] <wtay> 'cause I still doubt it'll work... [22:43] <wtay> right, I'm doing make dist now [22:43] <wtay> ok, failure [22:45] <wtay> configure.ac error... checking in [22:45] <omega_> huh? [22:46] <wtay> again... [22:46] <wtay> configure.ac was incomplter [22:46] <wtay> incomplete even [22:46] <steveb> omega_: would it be a problem if ladspa only had a loop = func? [22:46] <omega_> it would be OK, but I'd like to have both [22:46] <omega_> start with a loopfunc, it's easier [22:47] <omega_> lemme send you what I have, which is in progress [22:47] <steveb> would chain funcs even be practical with stereo plugins? [22:47] <steveb> 2 input pads + 2 output pads [22:47] <omega_> depends on whether everyone else cooperates and keeps in= sync [22:48] <wtay> make dist.. [22:48] <steveb> ok, need to sleep now [22:49] steveb (st...@no...) left irc: [x]chat [22:49] <wtay> uh oh... building docs... [22:50] <wtay> great, it failed... :( [22:50] <omega_> neat [22:50] <wtay> gstreamer-scan needed libgst.la... doing that now [22:51] <wtay> building libs/ too, jst in case [22:52] <wtay> make dist again [22:54] <wtay> ...going better now... [22:54] <wtay> YES! [22:54] <wtay> -rw-r--r-- 1 root root 2359780 May 28 22:54 gs= treamer-0.2.0-pre0.tar.gz [22:55] <wtay> trying yo build it [22:55] <dobey> rpm -tb?! [22:55] <omega_> ? [22:56] Action: taaz ponders why you don't just use 0.1.{2,3,4,5...} inst= ead of this 'pre' crap that is a serious PITA for those of us trying to f= igure out how to name packages such that automatic updating via version n= umber checking works when an actual 0.2.0 comes out [22:56] <wtay> config.status: error: cannot find input file: plugins/gsm/= Makefile.in [22:56] <wtay> :((( [22:57] <omega_> taaz: the pre tarballs are not going to be released publ= icly, just to get -devel people to check it out and find any stupid stuff= before the real 0.2.0 [22:57] <wtay> oh f*ck I forgot to check in the plugins/Makefile.am :((( [22:57] <omega_> in fact, maybe we shouldn't even change configure.in for= the pre [22:57] <omega_> wtay: yeah, leave the version 0.2.0 in configure.in, to = limit confusion [22:58] <omega_> taaz: all you need to do is grab the latest -preN and st= rip the pre from the name, try building that as a package [22:58] <omega_> the pre's go away forever once 0.2.0 is OK'd [22:59] <taaz> omega_: ok, well, i'm going to just use 0.2.0-{1,2,3,4,5} = etc and at some point it'll just be the actuall 0.2.0. i have no clue h= ow else to deal with this. sore spot in the debian versioning that it ca= nt handle negative versions like 0.2.0.-1 as a pre version... [23:00] <omega_> taaz: then don't handle it, just ignore the pre part of = the version and act as if it's 0.2.0 [23:00] <omega_> the only thing we need to use the '-preN' string for is = IDing the tarballs [23:00] <taaz> that's what i just said ;) [23:00] <omega_> no, I mean do not use any aux version at all in your pac= kaging [23:00] <omega_> because it's not relevant [23:01] <omega_> wget gstreamer-0.2.0-pre0.tar;gz;mv gstreamer-0.2.0-pre0= .tar.gz gstreamer-0.2.0.tar.gz;<build package> [23:01] <taaz> oh... X in 0.2.0-X is the debian package version, not the = upstream version [23:01] <omega_> repeat with pre1, then the real 0.2.0 comes out and all = the pre's go away [23:01] <omega_> no, not even that [23:01] <omega_> 0.2.0-X is for packging revs *after* the release [23:02] Uraeus (csc...@c2...) joined #gstreamer. [23:02] <taaz> so what do i name it now? [23:02] <Uraeus> hi [23:02] <wtay> ho [23:02] <omega_> taaz: *UTTERLY* IGNORE the preN [23:02] <omega_> the ONLY thing you have to do is strip the preN string f= rom the tarball before you try to build it [23:02] <wtay> it's just the filename of the tar.gz [23:03] <taaz> omega_: you are contradicting yourself. what version numb= er do i use now such that 0.2.0 is greater than todays version? [23:03] <omega_> you DON'T! [23:03] <omega_> every single preN tarball is for internal testing only, = and DISAPPEARS when 0.2.0 comes out [23:03] Action: taaz is confused... we are miscommunicating [23:04] <omega_> the ONLY thing that is different is that the tarball has= preN [23:04] <omega_> every package you build from the pre series gets rm -rf'= d when 0.2.0 is ready [23:04] <omega_> so you have zero versioning worries. from your point of= view, pre0 *is* 0.2.0 except that it won't build as well [23:04] Action: wtay runs make=20 [23:04] <omega_> when you find build problems, you tell us, we make preN+= 1, you try again, and we eventually get to 0.2.0 final release [23:05] <omega_> taaz: the sequence is get the pre tarball, rename to 0.2= .0, try to build a package [23:05] <omega_> when you find problems, we fix it, make preN+1, you repe= at that sequence [23:05] <wtay> omega_: we disabled something last time (static compile or= something)? [23:05] <omega_> yeah, we turned off static for devel, we need to turn it= back on before release [23:06] <omega_> dnl AM_DISABLE_STATIC [23:06] <wtay> what does it do? [23:06] <omega_> it saves compile time by not building static .o's [23:06] <wtay> ok [23:07] <wtay> can I upload 0.2.0-pre0.tar.gz somewhere? [23:07] <omega_> heh, um... [23:07] <omega_> question is where [23:07] Nick change: aj_uni -> ajmitch [23:07] <ajmitch> hi all [23:07] <wtay> that's why I ask :-) [23:07] <omega_> where's a server (!sf) that has bandwidth [23:07] <omega_> ? [23:07] <ajmitch> hey, you could host it on my ftp server (sure it's a 33= .6k modem) [23:08] <ajmitch> 33.6k ought to be enough, right? ;) [23:08] <omega_> 2.3MB x several dozen people [23:08] <ajmitch> hmm, maybe not ;) [23:09] <ajmitch> i could imagine more than several dozen wanting to down= load 0.2.0 release [23:09] <wtay> -pre0 is building fine... [23:09] <ajmitch> you've tagged it in CVS? [23:09] <wtay> not yet, first see if it compiles [23:09] <omega_> ajmitch: this is only pre0, doesn't even hit the mailing= list [23:09] <ajmitch> what's left for us mere mortals to do for you before 0.= 2.0? [23:10] <omega_> later pre's (I hope to keep it a very very short series)= will hit -devel [23:10] <wtay> test! [23:10] <ajmitch> omega_: ok, not even announcing with great fanfare, tel= ling people to test & fix it? [23:10] <omega_> and all those things that were mentioned in my mail [23:10] <omega_> pre1 + will be announced on -devel, with a request for p= eople to test it, etc. [23:11] <ajmitch> ok, so all crashes, video not working, etc, i complain = about? [23:11] <taaz> here's my plan: next day or two i'll have debs ready to te= st. i want it such that people can apt-get upgrade the debs from today f= orward. so all new debs must have increasing version numbers. just rena= ming to the same thing doesnt fit with that. so i can a) make up a name = 0.1.99-X as the pre, b) just call it 0.2.0-X and increase the debian X ve= rsion, c) continue my current thing of 0.1.1.20010528-X CVS date thing. =20 [23:11] <omega_> as packages start to be built, we'll put those up as wel= l, with explicit notices that these are *NOT* release packages and should= be removed asap [23:11] <ajmitch> taaz: i think a) [23:11] <omega_> taaz: why do they need to be apt-get'able? [23:12] <omega_> these are 3rd-class releases [23:12] <ajmitch> taaz: people think that packages named like b) are rele= ase quality [23:12] <taaz> omega_: because i say so ;) [23:12] <omega_> can the debian release be 0.X ? [23:12] <ajmitch> omega_: for people like chillywilly [23:12] <taaz> omega_: yes, like 0.x.y-0.z [23:13] <ajmitch> what's the code name? [23:13] <ajmitch> 'Free The Mallocs'? [23:13] <omega_> ok, then use 0.2.0-0.N, where N is the pre number [23:13] <taaz> omega_: normal usage for that is non-maintainer updates [23:13] <omega_> meaning? [23:14] <taaz> NMU, ie, person not maintaining the package made a fix and= uploaded it to debian servers. like if a maintainer is off on vacation = and security problem found or something [23:14] <omega_> does release 0 ever exist in that case? [23:15] <ajmitch> taaz: you are a debian maintainer now? [23:15] <taaz> its all kind of fuzzy and not really rules [23:15] <taaz> ajmitch no [23:15] <omega_> it seems to me that release 0.x can safely be used as a = pre testing release [23:15] <omega_> taaz: if you're not a maintainer, how are you going to p= ut these up for apt-get ? [23:16] <ajmitch> omega_: they can be hosted on any http or ftp server [23:16] <wtay> compilation is 100% successfull... [23:16] <omega_> ok [23:16] <omega_> docs build? [23:16] <wtay> yup [23:16] <omega_> hmm [23:16] <ajmitch> omega_: eg, i can put up packages that people can apt-g= et, if they put in my ftp server in their source.list [23:16] <omega_> ok [23:17] <wtay> now checking if it runs... [23:17] <omega_> details [23:18] <wtay> yeah, let's not test it :) [23:18] <omega_> taaz: will that release naming scheme work? if so, I'll= use it for the rpms as well [23:19] <wtay> whhohoo gstmediaplay works! [23:19] <omega_> to build rpms I need to get my hands on some more libs [23:20] <taaz> the 0.2.0-0.x? [23:20] <omega_> yeah [23:20] <taaz> sure [23:20] <ajmitch> yay! [23:20] <omega_> ok, sounds good [23:20] <taaz> with 0.2.0-1 being real release i suppose? [23:20] <omega_> yup [23:20] <omega_> and -2 and -3 being packaging revs as usual [23:20] <wtay> I have weird debug colors :( [23:20] <omega_> oh? [23:20] <wtay> INFO:st_plugin_load_absolute:386:m00;36 plugin "gsm" loade= d: 2 elements, 0 types [23:21] <wtay> ^^^^^^^^^ [23:21] <omega_> um [23:21] <wtay> ^^^^^^^^^ [23:21] <omega_> that's screwed up [23:21] <wtay> how can that be? [23:21] <omega_> checking [23:22] <omega_> building [23:23] Action: wtay is building again [23:25] dobey (do...@ip...) left irc: brb kern= el [23:30] <omega_> ok, I get the same thing [23:30] <omega_> somehow some of the debug defines are mangling the code [23:32] <omega_> whoops [23:32] <wtay> uh? how come I can log into CVS despite the mail from sf s= aying I can't? [23:33] <omega_> no idea. I never got any mail [23:33] <ajmitch> time for me to try & build... [23:34] <wtay> hmm, it works fine here... [23:35] <omega_> which does? [23:35] <wtay> colors [23:35] <omega_> hmm, where? [23:35] <wtay> cvs dir [23:35] <thomas> still need some place to put the pre.tar.gz ? I can host= it ... [23:35] <wtay> using autogen.sh [23:35] <omega_> note the difference between gstinfo.c:285 and gstinfo.c:= 295 [23:35] <wtay> thomas: ok, where? [23:35] <omega_> and note line autogen.sh:88 [23:36] <thomas> wtay: does one of you need access to it or can I put the= m there ? [23:36] <thomas> I can put them at RUG (Gent University) for example [23:36] <thomas> or somewhere on a server hosted at KPN [23:36] <omega_> wtay: I wonder if the difference is relevant, turning of= f the pid and cid output when DEBUG isn't turned out [23:37] <omega_> I'd think that pid,cid is still useful [23:37] <omega_> esp since INFO is to be used by app writers [23:37] <wtay> they should be the same [23:37] <omega_> ok [23:37] <omega_> fixing [23:37] <wtay> thomas: can you build a tar.gz? [23:37] <ajmitch> autoheader: warning: both `configure.ac' and `configure= .in' are present. [23:37] <ajmitch> autoheader: warning: proceeding with `configure.ac'. [23:37] <ajmitch> useful... ;) [23:37] <wtay> ajmitch: yup :) [23:38] <omega_> ok, committing [23:38] <ajmitch> argh, am just compiling ;) [23:38] Nick change: taaz -> taaz-away [23:38] <thomas> wtay: you mean, create one from a source dir tree ? [23:38] <wtay> thomas: yup, using make dist [23:38] <ajmitch> by the time i finish compiling i have to update & recom= pile ;) [23:38] <omega_> ./autogen.sh;make dist [23:38] <omega_> hold off a sec [23:38] <thomas> as in "get it from cvs right now, make it clean, and mak= e dist ?" [23:39] <omega_> update gst/gstinfo.c [23:39] <omega_> thomas: yup [23:39] <omega_> wtay: or you could just email him the tarball <g> [23:39] <wtay> thomas: beware, the docs will be built too [23:39] <wtay> heh, yeah [23:39] <wtay> sec... [23:39] <omega_> update and make dist again [23:39] <wtay> lemme create a good one first [23:39] <thomas> how big is the tarball ? [23:39] <omega_> to get the gstinfo.c patch [23:39] <thomas> wtay: what's to beware about the docs ? [23:40] <ajmitch> what is the gstinfo.c patch for? ;) [23:40] <omega_> thomas: the build enviro sucks [23:40] <wtay> omega_: there is no diff in gstinfo.c? [23:40] <omega_> ajmitch: it fixes the no-DEBUG case for info [23:40] <thomas> wtay, omega: build it myself now or wait for wtay to mai= l ? [23:40] <omega_> wtay: there better be, I just committed one [23:40] <omega_> thomas: wait for wtay [23:40] <omega_> easier that way <g> [23:40] <wtay> running make dist again [23:41] Parapraxis (Sp...@us...) joined #gs= treamer. [23:42] <wtay> yo [23:42] <ajmitch> hi [23:42] <wtay> thomas: mail comming up... [23:42] <thomas> wtay: how big is it ? [23:42] <omega_> 2.3MB [23:43] <wtay> it'll take a while [23:44] <wtay> I only have about 15K/sec upload speed [23:44] <wtay> ajmitch: yes, *only* :) [23:45] <ajmitch> wtay: hehe [23:45] <wtay> omega_: ok, lemme try to build that again [23:46] <Parapraxis> howdee [23:48] <wtay> thomas: sent [23:49] <omega_> In file included from gstgsm.c:21: [23:49] <omega_> gstgsmdec.h:28:21: gsm/gsm.h: No such file or directory [23:49] <omega_> In file included from gstgsm.c:22: [23:49] <omega_> gstgsmenc.h:28:21: gsm/gsm.h: No such file or directory [23:49] <wtay> oh fsck! [23:49] <wtay> >>> MAIL From:<wim...@ch...> SIZE=3D3224854 [23:49] <wtay> <<< 552 Message size exceeds maximum permitted [23:49] <wtay> 554 <tho...@us...>... Service unavailabl= e [23:49] <omega_> mu [23:49] <wtay> damn! [23:49] <thomas> wtay: send it to th...@ur... [23:49] <omega_> wtay: gsm/gsm.h dne [23:50] <thomas> not to sourceforge [23:50] <wtay> omega_: ? [23:50] <omega_> /usr/include/gsm.h in my rpm [23:50] <wtay> hmm [23:50] <wtay> it's in gsm/gsm.h [23:50] <wtay> doh! [23:50] <omega_> not in my rpm [23:51] <wtay> hmm [23:51] <wtay> -rw-r--r-- 1 root root 1681 Aug 28 1999 /u= sr/include/gsm/gsm.h [23:51] <omega_> is there any other file in /usr/include/gsm ? [23:51] <wtay> yes [23:52] <omega_> oh? [23:52] <wtay> config.h gsm.h private.h proto.h toast.h unproto.h [23:52] <omega_> eesh [23:52] Action: omega_ looks for a differnt rpmp [23:53] <omega_> what version? [23:53] <wtay> ii libgsm1 1.0.10-10 Shared libra= ries for GSM speech compressor. [23:54] <wtay> maybe I should check for gsm/gsm.h too [23:54] <omega_> guess so ;-( [23:55] <wtay> thomas: sending again [23:57] <thomas> wtay: ok [23:57] Nick change: hds-afk -> hadess [23:57] <thomas> btw I just downloaded a fresh copy, ran autogen.sh, and = make dist fails with : [23:57] <thomas> mcopidl -t gst_artsio.idl -I /usr/local/include/arts -I = /usr/include/kde/arts [23:57] <thomas> make[2]: mcopidl: Command not found [23:57] <thomas> make[2]: *** [gst_artsio.cc] Error 127 [23:57] <thomas> make[2]: Leaving directory `/tmp/gstreamer/plugins/arts' [23:57] <thomas> make[1]: *** [distdir] Error 1 [23:57] <thomas> make[1]: Leaving directory `/tmp/gstreamer/plugins' [23:57] <thomas> make: *** [distdir] Error 1 [23:57] <thomas> [thomas@thomas gstreamer]$=20 [23:57] <thomas> what's mcopidl ? [23:57] <omega_> mu [23:57] <ajmitch> from arts [23:58] <omega_> arts fun&games [23:58] <wtay> locate mcopidl? [23:58] <ajmitch> mcop is their corba replacement ;) [23:58] <thomas> ok I don't have that. Can I expect more of these or is = this just one I should download for now ? [23:58] <omega_> no, mcop is stw's dcop replacement, which is kde's corba= replacement [23:59] <omega_> we have to get a decent arts check [00:00] --- Tue May 29 2001 [00:00] <wtay> we could do a AC_CHECK_PROG (mcopidl) [00:00] <omega_> of debatable usefulness [00:01] <wtay> chances are it that if that exists, idl will be present to= o [00:01] <omega_> maybe [00:01] <hadess> i'm starting to understand xmms' http code, and i'm afra= id of my mental health [00:02] <omega_> hadess: we were afraid of your mental health a long time= ago [00:02] <wtay> hadess: heh [00:02] <omega_> what took you so long? [00:02] <hadess> *smack* [00:02] <omega_> ;-) [00:02] <wtay> ok, debug colors are sane now.. [00:02] <ajmitch> hadess: xmms code is not pretty? ;) [00:02] <hadess> i already stripped the udp code in there, starting to st= rip the prebuffer code [00:03] <wtay> *pffew* I'd hate to have such a crucial new feature all me= ssed up :-) [00:03] <hadess> then i'll have something (quite) sane [00:03] <hadess> wtay: heh [00:03] <hadess> ajmitch: you wouldn't believe [00:03] <omega_> hadess: how do you plan on implementing prebuffer at thi= s point? [00:03] <wtay> thomas: go mail? [00:04] <wtay> s/go/got/ [00:04] <thomas> wtay: not yet [00:04] <thomas> wtay: hope it doesn't backfire as well [00:04] Action: wtay plays AlienSong with 0.2.0 [00:05] Action: ajmitch waits for gstreamer to compile *again* [00:06] <thomas> wtay: maybe it'd be smarter if you'd ftp it ? [00:08] <wtay> thomas: yes [00:09] Action: Uraeus is busy with a XFCE & GNOME flamewar [00:09] <ajmitch> hehe [00:09] <ajmitch> yes, i saw that on the mailing list [00:10] <Uraeus> ajmitch: it makes me a little angry, he couldn't even ha= ve read the interview to make those statements [00:11] <ajmitch> Uraeus: ah well, flamewars keep you warm in winter ;) [00:11] <Uraeus> ajmitch: true, but it is summer now..... [00:12] <ajmitch> Uraeus: not here ;) [00:12] <hadess> omega_: prebuffer will be in the gst pipeline, not in th= e gnome-vfs icecast module [00:12] <Uraeus> ajmitch: do you have winter in NZ, or is it 'winter'? [00:12] <ajmitch> Uraeus: what do you mean? [00:12] <ajmitch> NZ is a wonder tropical pacific island ;) [00:13] <Uraeus> ajmitch: so you don't have winter in other words -> trop= ic !=3D winter [00:14] <omega_> ajmitch: what latitude are you at? [00:14] <ajmitch> Uraeus: well, we had snow here a few days ago, not as m= uch os .no but enough... [00:14] <ajmitch> omega_: 45 south [00:14] <omega_> just about exactly where I am, north [00:14] <thomas> gstreamer pre is available from http://urgent.rug.ac.be/= thomas/download [00:15] <ajmitch> but totally different climate... [00:15] <thomas> can someone please get it to check ? [00:15] <omega_> got it [00:15] <ajmitch> there is no land between NZ & antarctica, and we get a = lot of southerly weather [00:15] <omega_> err, except for NS crashing [00:16] <thomas> lynx it [00:16] <omega_> no, ns is just being stupid, it put the save dialog in a= different virtual desktop [00:17] Gman (gm...@be...) joined #gstreamer. [00:17] <Uraeus> hi gman [00:17] <omega_> yo [00:17] <ajmitch> hi Gman [00:17] <wtay> hi [00:17] <Gman> heya [00:17] <Gman> jesus, it's a possey [00:17] <ajmitch> hehe [00:17] <Gman> In file included from gstarch.h:39, [00:17] <Gman> from cothreads.c:36: [00:17] <Gman> gstsparc.h:28: parse error before string constant [00:17] <Gman> cothreads.c: In function `cothread_create': [00:17] <Gman> cothreads.c:126: warning: long int format, int arg (arg 2) [00:17] <Gman> cothreads.c: In function `cothread_switch': [00:17] <Gman> cothreads.c:357: parse error before `;' [00:17] <Gman> cothreads.c:359: parse error before `;' [00:17] <Gman> cothreads.c: At top level: [00:18] <Gman> cothreads.c:218: warning: `cothread_stub' defined but not = used [00:18] <Gman> enjoy :) [00:18] <ajmitch> blame omega_ ;) [00:18] <Uraeus> Gman: with help like that who can complain :) [00:18] <Gman> heh [00:18] <Gman> that must have been a recent alteration... [00:19] <thomas> ok, the pre got configure'd nicely ... [00:19] <thomas> can I go to bed now ? ;) [00:19] <wtay> thomas: no <g> [00:19] <omega_> it implies that GST_ARCH_SET_SP didn't get properly defi= ned [00:19] <omega_> thomas: build it ;-) [00:19] <Gman> well I hacked the configure.in [00:19] <Gman> so maybe that's the problem... [00:19] <omega_> to do which? [00:19] <thomas> damn. make'ing ... [00:19] Action: Gman does cvs update [00:20] <Gman> omega_: for the alsa stuff.. [00:20] <omega_> oh, ok [00:20] <thomas> oh well since you're against it I might as well rack you= r brain a little more... [00:20] <omega_> we've fixed that up a bit in cvs [00:20] <thomas> ... I was thinking about the plugin factory today. [00:20] Action: ajmitch sees the latest GNOME summary, by Uraeus... [00:20] <omega_> Gman: hmm, I thought i fixed this.... [00:20] <thomas> ... I started something simple using perl to replace spe= cific placeholder tags in source code templates... [00:20] <omega_> fixing gstsparc.h in cvs [00:20] <thomas> ... but that would probably be a bit limited an approach= , right ? [00:20] <thomas> what do you want from a good plugin factory ? [00:21] <ajmitch> ooh, Uraeus mentioned gstreamer in the summry, what a s= urprise ;) [00:21] <omega_> Gman: update gstsparc.h, try again [00:21] <wtay> thomas: a wizard, I want a wizard :) [00:21] <thomas> wim: mail arrived ;) [00:22] <thomas> wtay: frontend comes last, what should it actually *do* = ? [00:22] <thomas> you must have discussed it before, no ? [00:22] <wtay> actually, no [00:22] <thomas> heh [00:23] <thomas> so I'm guessing, as a first stab, my perl replacing impl= ementation isn't even that bad [00:23] <wtay> but omega_ has been keeping a secret for all of us [00:23] <omega_> oh? [00:23] <thomas> I would like it to start from xml, being largely compati= ble with the registry format [00:23] <wtay> omega_: your idea about a filterfactory... [00:23] <omega_> oh, hmmm [00:24] <wtay> omega_: you never explained what it should do [00:24] <omega_> make filters [00:24] <omega_> duh ;-) [00:24] <thomas> omega_: you'd get us to do it a lot more quickly if you'= d elaborate a little bit ;) [00:24] <ajmitch> hehe [00:25] <omega_> brb, need food [00:25] <Gman> oh this is good...checking for blah....no....checking for = blah.....no...... :) [00:25] <wtay> heh [00:25] <wtay> Gman: use a *real* unix :-) [00:25] <thomas> btw wtay: where do I subscribe for the cvs logs again ? [00:26] <Gman> wtay: yeah yeah [00:26] <wtay> thomas: there are links on the sf page [00:26] <Gman> wtay: I'm only here because Uraeus promised to give me lot= s of free beer and free accomodation in Norway :P [00:26] <hadess> chopp... [truncated message content] |
From: <wim...@ch...> - 2001-05-30 04:31:38
|
[07:15] <ChiefHighwater> omega_:ping [07:16] Nick change: ajmitch -> ajbusy [07:39] capuchin (jb...@or...) joined #gstreamer. [07:39] <capuchin> Hey, all... anyone know anything about Xvideo in XFree= 4? [07:40] <ChiefHighwater> ello [07:40] <capuchin> Yo. [07:40] <ChiefHighwater> there's not many folks awake right now [07:40] <capuchin> I see that. [07:40] <ChiefHighwater> hehe [07:40] <capuchin> Any suggestions on where to turn for Xv issues? [07:41] <capuchin> omega_ was quite helpful last night. [07:41] <ChiefHighwater> omega_ is quite helpful in a lot of arenas [07:41] <capuchin> Well, he's apparently out now. [07:41] <ChiefHighwater> not sure why he's not here [07:41] <ChiefHighwater> i pinged him about 20 mins ago [07:41] <capuchin> I just can't think of where a nice place to turn might= be. [07:42] <ChiefHighwater> sorry, best i can do is ask omega...big help I a= m, sorry. [07:42] <capuchin> S'cool. [07:48] <capuchin> This is incredibly stupid, but what's the best way to = check which version of X I'm running? I built it quite a while ago and c= an't recall if it's 4.0.3 or no. [07:48] capuchin (jb...@or...) left #gstreamer. [07:55] Nick change: ajbusy -> ajmitch [07:58] Nick change: taaz -> taazzzz [08:04] Nick change: ChiefHighwater -> CHW_away [08:14] steveb (st...@no...) joined #gstreamer. [08:14] <ajmitch> hi stevebq [08:14] <ajmitch> bah [08:14] <ajmitch> hi steveb [08:14] Action: ajmitch hits stupid keyboard [08:16] <steveb> yo [08:34] arik (ar...@sd...) joined #gstreamer. [08:34] <arik> lo [08:34] <ajmitch> hey arik [08:34] Nick change: arik -> arik|afk [08:39] Nick change: steveb -> steveb-away [09:00] ajmitch (aj...@p5...) got netsplit. [09:00] big_T (th...@fy...) got netsplit. [09:00] steveb-away (st...@no...) got netsplit. [09:04] CHW_away (pa...@su...) left irc:=20 [09:04] ajmitch (aj...@p5...) returned to #gstreamer. [09:05] <omega_> yo [09:05] <ajmitch> hi [09:05] Action: omega_ is in asm hell [09:06] <omega_> but creeping out of it [09:06] <ajmitch> hehe [09:07] <omega_> but I think I've finally conquered inline asm [09:08] <ajmitch> that is good... [09:08] <omega_> it even does exactly what I want [09:08] <omega_> it's not pretty, though [09:11] big_T (th...@fy...) got lost in the net-split. [09:11] steveb-away (st...@no...) got lost in the net-split. [09:11] <omega_> I wonder if gcc has any concept of mmx registers as a sp= ecial case of inline asm [09:11] <omega_> that would let the inline case be even faster maybe [09:11] <walken> no [09:11] <walken> gcc doesnt know about mmx [09:11] <omega_> darn [09:11] <walken> well 2.9x at least [09:11] <omega_> I was hoping I could put a "mx"(currentbits) [09:11] <walken> I think gcc 3.0 should know about mmx intrisics [09:11] <walken> (if they went with their plan) [09:12] <omega_> then it could decide whether or not to save it to memory= in the first getbits and load it from memory in the second [09:12] <walken> hehe [09:12] <walken> you wish :) [09:12] <omega_> yes, I do. [09:12] <walken> well at the same time it might make it easier to keep bi= t parsing variables in mmx register accross function calls [09:12] <omega_> well, at least this is cooperating, that's better than n= othing [09:12] <walken> something gcc can not do with regular variables [09:13] <omega_> walken: right, that's the idea [09:13] <omega_> it *can* do it across calls to the same inline function [09:13] <walken> (I suspect this hurts our parsing performance a bit) [09:13] <omega_> if there's no register pressure in between [09:13] <walken> hmmm [09:13] <omega_> which for mm7 there wouldn't be, in dense parsing [09:13] <walken> I dont trust inline functions too much [09:13] steveb-away (st...@no...) joined #gstreamer. [09:13] <omega_> this is exactly why I want an mmx register class that in= lining can be smart enough to handle loading/saving of [09:13] <walken> they are basically fine if you onle need to return one t= hing [09:14] <omega_> walken: I'll email you what I just wrote up [09:14] <walken> but if not you will pass a pointer to return more variab= les and gcc will not be able to get rid of the pointer [09:14] <walken> yes I could look into this :) [09:14] <omega_> sent [09:14] Action: walken looking at the ac3 downmix games [09:14] <omega_> you'll see what I mean by 'conquering inline asm' [09:15] <omega_> sent you another email with the compiled .S results of t= hat [09:16] <walken> got them [09:17] <walken> in my experiments I found out it was a gain to not need = to pass a number of bits to the needbits function [09:18] <omega_> could be [09:18] <omega_> I want to find a way to mix this with what you do, keepi= ng the same data structure [09:19] <omega_> I'm pretty certain it can be done, though mmx makes it m= essier [09:19] <omega_> but.. if it can be machine-generated..... [09:19] <walken> so basically you want to fit some of my bit parsing vari= ables in mmx registers and generate the code for that ? [09:19] <omega_> something like that [09:20] <walken> omega. bewere of reasoning about getbits too much [09:20] <omega_> question is: is your style of checking the bit_buf for w= alking the huffman tree compatible with mmx [09:20] <omega_> walken: hehehe, too late ;-) [09:20] <omega_> muhahahaha [09:20] <walken> for huffman parsing you have to be able to look at the b= its before you know how many of them you want [09:20] <omega_> yup [09:20] <omega_> have you read the code vlcgen spits out? [09:21] <walken> no :) [09:21] <omega_> codecs.org cvs [09:21] <walken> omega. it should ? [09:21] <omega_> libmpeg/libmpeg/mpeg1_ something or other [09:21] <omega_> it should what? [09:21] arik|afk (ar...@sd...) left #gstream= er. [09:21] <walken> I guess we want to transfer the high-orde bits to an int= eger register at some point [09:22] <walken> and then do basically like I do currently [09:22] <omega_> depends on the method, but yeah [09:22] <omega_> problem is that you can'd movd the upper half of an mm r= eg [09:22] <walken> I'm still not quite clear on the advantages of the mmx s= tuff though [09:22] <omega_> the shifter [09:22] <omega_> it's 1 cycle, no matter what [09:22] <walken> bit_buf could be moved to an mmx register [09:22] <omega_> the x86 shifter is really slow [09:22] <walken> oh ? [09:22] <walken> I thought it was 1 cycle too [09:22] <ajmitch> how often is this piece of inline asm called? [09:23] <omega_> 3 cycles minimum, from ap527 [09:23] <walken> damn [09:23] <omega_> ajmitch: this one is called many many many times [09:23] <ajmitch> you seem to be going to a lot of effort to optimise thi= s code [09:23] <omega_> ajmitch: intel's numbers show that the average bits per = read for an mpeg1 stream is 4.86 [09:23] <ajmitch> omega_: ok ;) [09:23] <walken> ajmitch. parsing takes something between 25% and 45% of = the time in libmpeg2 [09:23] <omega_> now, consider how many bits are in a 19.8MBps digital ca= ble stream [09:23] <ajmitch> hmm, not good [09:23] <omega_> do the math <g> [09:24] <ajmitch> what target are you going for? [09:24] <omega_> target? [09:24] <omega_> =02FAST=02 [09:24] <ajmitch> lots ;) [09:24] <ajmitch> hehe [09:24] <ajmitch> will we see it being faster than WMP ? ;) [09:24] <omega_> at the very least I want to be able to guarantee DVD pla= yback in software with YUV overlay on my laptop [09:25] <omega_> ajmitch: doubtful, since Intel wrote most of that [09:25] <walken> omega. OK so the idea is to keep "bits" and "buf_ptr" in= integer registers and just put bit_buf in an mmx register [09:25] <ajmitch> hmm [09:25] <ajmitch> i wish i could even get something to show up ;) [09:25] <omega_> walken: but how do you compare bit_buf against something= is the question [09:25] <omega_> if (bit_buf & 0x80000000) { [09:25] <walken> and take advantage of the faster shifter and the fact th= at reloads will be less common [09:25] <walken> right ? [09:25] <omega_> right [09:25] <omega_> did you check out the vlcgen code? [09:26] <walken> no [09:26] <omega_> check it out ;-) [09:26] <walken> omega. I think there is no instruction to move the 32 hi= gher order bits to an integer register ? [09:26] <omega_> it uses hierarchical tables, no asm generation yet, but = hand-tuned stuff in libdv turns into very few instructions [09:26] <omega_> exactly [09:26] <omega_> sse2 has it, I think [09:27] <omega_> but not sure [09:27] <ajmitch> sse2 is in pentium 4 only? [09:27] <omega_> and not-yet-existing Athlon-series chips, yeah [09:27] <omega_> Athlon 4 [09:28] <walken> omega. how do you adjust the size of the respective hier= archies ? [09:28] <walken> (for the hierarchic tables) [09:28] <omega_> (no, I don't know why both Intel and AMD have to extend = their own dumb names with numbers.... lazy marketing depts or something) [09:28] <ajmitch> i don't think the 3dnow in structions in my k6-2 are us= eful for much of this [09:28] <omega_> walken: it's a bit hokey right now, but it's manual [09:28] <omega_> so you can tweak it [09:28] <omega_> I plan on writing a timing routine to try various splits= , later [09:28] <walken> OK [09:28] <walken> this is something that matters a lot :) [09:29] <omega_> but there's plenty of room for that [09:29] <omega_> yes, the split affects cache levels and all sorts of oth= er things [09:29] <walken> yes. mainly cache and branch prediction [09:29] <omega_> down to things like trying to align tables and such to c= ache lines, even spanning depending on some things [09:29] <omega_> walken: there are no branches in that code [09:30] <omega_> at least, there aren't until gcc stupidly inserts them (= yes, I saw it happen) [09:30] <walken> no branches ? how do you switch from one hierarchy to th= e other then [09:30] <omega_> read it [09:31] <ajmitch> heh, and i thought that the black art of asm optimisati= on had died out in today's day & age of modern cpus & compilers [09:31] <walken> where :) [09:31] <omega_> http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/codecs/li= bmpeg/libmpeg/mpeg1_cbp.c?rev=3D1.1.1.1&content-type=3Dtext/vnd.viewcvs-m= arkup [09:31] <omega_> ajmitch: pffff [09:31] <ajmitch> hehe [09:32] <omega_> Buck Krasic came up with this, while reworking the parse= r for libdv [09:32] <ajmitch> i guess that with small pieces of code like this that a= re used a lot, hand optmisation still wins? [09:32] <omega_> vlcgen needs a lot of work, but it does the job (sorta) [09:32] <omega_> ajmitch: always [09:33] <walken> huh [09:33] <ajmitch> even with multipipelines & ordering of instructions to = minimise stalls, etc? [09:33] <walken> you use c syntax I didnt know about :) [09:33] <walken> is this a gcc extension ? [09:33] <omega_> uh? which? [09:33] <omega_> oh, the tables? [09:33] <omega_> yeah [09:33] <walken> the [n] =3D value [09:33] <omega_> makes it nice and non-portable to sucky archs like win32= <g> [09:34] <omega_> since it's generated code it can be done more portably j= ust by changing the generator [09:34] Action: omega_ gives walken some time to figure it out, since it = took him a long time too [09:35] <omega_> have the mpeg2 huffman table handy.... [09:35] Action: ajmitch shoudl try & figure out this screwy colorspace is= sue [09:36] <walken> why the class_mask thing ? [09:36] <walken> I think you can get rid of it [09:36] <omega_> oh, now you're gonna make me thing about it again.... [09:36] Action: omega_ tries to figure it out again [09:37] capuchin (jb...@or...) joined #gstreamer. [09:37] <walken> yes you can :) [09:37] <omega_> yo [09:37] <omega_> reading [09:37] <capuchin> Hey, all... [09:37] <capuchin> Anybody know where to get the good info on XVideo in X= Free 4? [09:37] <omega_> not really [09:37] <omega_> I haven't seen any good anything on it ;-( [09:38] <capuchin> omega_: I've done all the troubleshooting I can, and I= still just get solid blue where the video overlay should be. [09:38] <walken> well look [09:38] <omega_> hmmmm [09:38] <capuchin> gtv using SDL as well as mplayer. [09:38] <walken> for the class 0 - you had 9 bits, you shift by 4, you ha= ve 5 bits left, so the mask is a nop [09:38] <omega_> odd [09:38] <omega_> walken: ok [09:39] <omega_> capuchin: what X version? [09:39] <walken> class 1 - you had 9 bits, you shift by 2, but the 3 uppe= r bits are known (else you wouldnt be in class 1), so there are 4 unknown= bits, and you dont need the mask [09:39] <omega_> (no, I don't know how to find that except by restarting = X.... ;-( [09:39] <omega_> walken: good point [09:39] <walken> class 2 - 9 bits, 3 known because of the class, no shift= , that gives you 6 unknown bits, and mask does nothing [09:39] <omega_> walken: ok, let's take DV.... [09:39] <omega_> lemme find the huffman table, sec. [09:40] <ajmitch> omega_: xdpyinfo [09:40] Action: capuchin looks daffy for a moment... Um... I menat to ask= : what's a quick way to check? It's at least 4.0.1 [09:40] <omega_> capuchin: I read logs <g> [09:40] <ajmitch> capuchin: xdpyinfo, it's near the top [09:40] <walken> omega. overall its neat though. [09:40] <omega_> ajmitch: xdpyinfo tells you? hmmm [09:40] <ajmitch> vendor release number [09:40] <omega_> hmm, maybe I don't have the dv table typed out [09:40] <walken> I'm a bit worried about the fact that we need one array = access to find the class [09:41] <omega_> there it is [09:41] <omega_> walken: should be minimal compared to branching [09:41] <walken> I mean if you had a really tall and unbalanced huffman t= ree that would be a problem [09:41] <omega_> and it would be possible to group all the smaller tables= into a single cache line [09:41] <walken> but in practice I guess it must work well [09:42] <omega_> works pretty well for libd [09:42] <omega_> er, libdv [09:42] <ajmitch> omega_: hmm, the number reported by colorspace really i= sn't meaning much to me, who worked on it originally? [09:42] <omega_> what I'd like to try to do is take your block decoder co= mpletely out and put a perf harness around it [09:42] <walken> worth trying in libmpeg2 too [09:42] <walken> yay [09:42] <omega_> then try to optimize it to death [09:42] <walken> I like your idea :) [09:42] <omega_> ajmitch: I think it's a fourcc [09:43] <capuchin> Weird... it says 4.0.99.1 [09:43] <omega_> walken: how hard would that be to rip it out and constru= ct some test data? [09:43] <ajmitch> omega_: which means nothing to me ;) [09:43] <walken> hmmm - you want to test in a libmpeg2 context though [09:43] <omega_> ajmitch: convert it to hex, then take each byte and conv= ert to ASCII [09:43] <walken> else you'll be tempted to let it use more cache than it = should [09:43] <ajmitch> capuchin: where did you get this from (distro?) [09:43] <omega_> neat patterns should emerge [09:43] <omega_> walken: maybe, but a vacuum is a terrible thing to waste= ... [09:44] <omega_> walken: how many bytes are your dct coeff tables, total? [09:44] <capuchin> Now this is odd: xvinfo gives settings for screen 0 an= d adapter 0, but there's no xvideo line at all in my XF86config... is tha= t normal? [09:44] <ajmitch> 1CEB2F9 , then convert that to ascii? [09:44] <ajmitch> capuchin: yep [09:44] <omega_> ajmitch: hmm, i think so, yeah [09:44] <walken> have to look [09:44] <capuchin> OK, yeah... I see it now in xdpyinfo [09:44] <omega_> capuchin: sounds like you really do have 4.0.99.1 or som= esuch [09:45] <capuchin> Hmm... what does yours say? [09:46] <omega_> I have XVideo listed, but xdpyinfo doesn't tell you whet= her it's actually functioning or not [09:46] <omega_> you have gstreamer built? [09:47] <omega_> if you run ./gstreamer-register --gst-mask=3D-1 and look= for the output from xvideosink, you'll see whether it claims to have Xv = adaptors or not [09:47] <ajmitch> 1CEB2F9 doesn't mean much in ACSII does it? [09:47] <omega_> doesn't look like it [09:47] <omega_> I dunno, ask wtay <g> [09:47] Action: capuchin is pleased to see so much Chomsky on these openn= ap servers. [09:48] <ajmitch> omega_: thanks ;) [09:48] <omega_> ajmitch: I didn't write anything related to video.... [09:48] <omega_> oddly enough [09:49] <ajmitch> this video driver is crap, btw ;) [09:49] <walken> omega. rough estimate : 1.5K [09:49] <omega_> hmm, I have 2.7K, but I didn't try to optimize it much [09:49] <ajmitch> ajmitch:~# dga [09:49] <ajmitch> Xserver driver doesn't support DirectVideo on screen 0 [09:49] <capuchin> How do I set sdl to use aalib? [09:50] <walken> omega. half of it is unused when we parse the B14 table.= B15 is a bit weirder so the first level is 8 bits already which takes a = lot of space [09:50] <omega_> sdl in general or sdl as in sdlvideosink? [09:50] <omega_> walken: on, I'm couting my mpeg1 tables, btw [09:50] <walken> for B14 the first level is 5 bits only which is nicer [09:50] <capuchin> sdlvideosink [09:50] <ajmitch> so how much of a speedup do you have? ;) [09:50] <walken> mpeg1 would be b14 :) [09:50] <omega_> checking [09:50] <capuchin> Xlib: extension "XFree86-DGA" missing on display ":0.= 0" [09:50] <omega_> ok [09:51] <omega_> capuchin: looks like you just set an environment var as = per normal SDL, and go [09:51] <capuchin> Hmm... what's the var? [09:51] <omega_> checking [09:52] <walken> do you remember what processor got bswap first ? [09:52] <walken> is it 486 ? [09:52] <omega_> SDL_VIDEODRIVER =3D "aalib" [09:53] <omega_> um, I thought the 386 had it [09:53] <capuchin> Thanks. [09:53] <omega_> but I dunno [09:53] <omega_> walken: sec, I have a borland turbo asm ref somewhere th= at should say [09:53] <omega_> 486 [09:53] <walken> ok [09:54] <walken> hmmmm so theorically we have to detect it before we use = it [09:54] <omega_> that suck [09:54] <omega_> er, sucks [09:54] <walken> I guess we can puke when its not there though [09:54] <walken> :) [09:54] <omega_> I would be tempted to puke at configure time <g> [09:54] <walken> unless we want to decode mpeg2 on a 386 :) [09:54] <walken> ah no [09:54] <walken> you dont know at configure time where you'll run :) [09:54] <omega_> walken: actually... ac3 on an embedded 386, maybe.... [09:55] <walken> yeah ac3 is a different story [09:55] <omega_> remember the 386 is probably the world's most common mic= roproc, though you'd never know it [09:55] <walken> well maybe use bswap only if we do mmx parsing :) [09:55] <omega_> hmmmm [09:55] <walken> bswap probably is no help when you have no mmx anyway [09:56] steveb (ste...@21...) joined #gstreamer. [09:56] <capuchin> Anyone know how to check what drivers sdl is compiled = to use? aalib segfaulted on me, by the way. [09:56] <walken> ok so your bitstream_flush has no jump (basically just u= pdates the bit counter) and bitstream_show is similar to my needbits ? [09:57] <omega_> capuchin: not sure ;-( [09:57] <capuchin> Thanks. [09:58] <ajmitch> bbl [09:58] Nick change: ajmitch -> ajbusy [09:58] <omega_> capuchin: doesn't appear to be any way to ask SDL [09:58] <omega_> capuchin: maybe `nm libsdl.so | grep '_bootstrap'` ? [09:58] <omega_> walken: sounds about right [09:59] <omega_> walken: I think flush has to deal with the case where cu= rrent_bits < flush_bits, though [10:00] <omega_> hrm, actually, it does the psllq to physically drop the = bits, too [10:00] <capuchin> omega_: I think that did it... showed DGA, DMA, DSP, E= SD, FBCON, and X11... no AALIB. [10:00] <omega_> ok, there's the problem ;-( [10:01] <omega_> need to get aalib, then rebuild sdl [10:01] <capuchin> What's dga? [10:01] <capuchin> I've got aalib, I'm just rebuilding sdl now. [10:01] <steveb> omega_: thad ladspa code you send appears to be what is = in CVS [10:01] <omega_> XFree Direct something Access [10:01] <omega_> steveb: hmmm [10:01] <capuchin> is it " a good thing "? [10:01] <omega_> capuchin: for some uses, yes [10:01] <omega_> not for YUV video though, there's stupidly no overlap be= tween dga and xv [10:02] <omega_> Xv is a joke, btw, it's the absolute minimum necessary t= o use the overlay. much much more work needs to be done to that extensio= n [10:02] <omega_> steveb: you're right, lemme find the right one [10:02] <capuchin> I use the mga_vid module, but it doesn't seem to be re= cognized as supporting greater than 16bit color. [10:02] <walken> omega. oh ? [10:02] <omega_> walken: which? [10:03] <walken> omega. my flush will just not work if you dont have that= many bits to flush. [10:03] Action: capuchin wonders aloud which video output source xine is = using... [10:03] <omega_> ah. but remember your refill is half the size of bit_bu= f [10:03] <omega_> and you thus can't guarantee a flush of more than 16 bit= s [10:03] <walken> yes [10:03] <walken> dont do that :) [10:04] <walken> never flush more than you have [10:04] <walken> basically [10:04] <capuchin> Well, using aalib for SDL still segfaults, but now it'= s in the library, at least. :). [10:04] <omega_> I can flush up to 64 bits guaranteed, and maybe even 128= if you're really lucky [10:04] <omega_> heh [10:04] Action: capuchin smiles... omega_ funny. [10:05] <walken> it'd be nice if you could guarantee I'll be lucky [10:05] <omega_> steveb: sent the right ones this time <g> [10:05] <omega_> walken: no comment [10:05] <steveb> omega_: cheers [10:08] <walken> omega. do you know where I can find accurate instruction= timing info ? [10:08] <omega_> walken: hehehhehehe [10:08] <walken> like I'm soooooo worried about this 3-cycle shift thing = now :) [10:08] <omega_> walken: oh, you're serious? DNE [10:09] <omega_> walken: Intel holds that data tightly, and besides it's = very very dependent on pairing and uops these days [10:09] <walken> is it also true for constant shifts ? [10:09] <omega_> lemme see what the borland book says [10:10] <omega_> shld/shrd ? [10:10] <omega_> only goes to 486, but claims 2 cycles for r,r case, 3 fo= r r,m [10:11] <omega_> lemme see if I have info on uops and such [10:12] <omega_> ouch [10:13] <omega_> a lot of them are =024=02 uops [10:14] <omega_> for SHL/SAL, there are 13 variants with 4 uops and 10 wi= th 1 [10:14] <omega_> the 1 uops ones are those with register operands [10:14] <omega_> afaict [10:16] <omega_> according to ap527, the shift in the P5-MMX takes 4 cycl= es [10:17] <omega_> but the P6 can significantly change that [10:17] <walken> ok [10:17] <omega_> but all I can say is that I hope your code compiles down= the way you think it does [10:18] <walken> huh ? [10:18] <omega_> if it ever tries to shift a memory location, you're gonn= a get hit bigtime [10:18] <omega_> those are the 4uop cases [10:19] <walken> oh [10:19] <walken> well, not very surprising [10:19] <omega_> do you have all the relevant pdf's from intel? [10:19] <walken> I hope gcc does riscy things here :) [10:19] <omega_> one hopes [10:19] <walken> all, I dont know... most of them, probably [10:19] <omega_> like 245127-001, Intel Architecture Optimization Referen= ce Manual [10:19] <omega_> which is what I'm looking in, appendix C [10:21] <walken> hmmm [10:22] <walken> apparently the problem is only with variable shifts [10:22] <walken> constant shifts are fine [10:22] <walken> I prefer that [10:22] <omega_> that's not what I understood [10:22] <omega_> it seemed that the data being shifted was more important [10:22] <omega_> m vs. are [10:22] <omega_> er, stupid complete [10:22] <omega_> 'm' vs. 'r' [10:22] Action: omega_ turns that off [10:25] <walken> from the data I have integer register shifts are 1 or 4 = cycles on pentium depending of wether they are constant or not [10:25] <omega_> uops, not cycles [10:25] <omega_> biiiig difference [10:25] <omega_> ooooh, that can't be [10:25] <omega_> 7.4 cycles for 1000 5-bit getbits [10:26] <omega_> er, 7.4 cycles each [10:26] <walken> omega. pentium. so not uops. [10:26] <omega_> no, not correct, but close [10:26] <walken> hmmmm [10:26] <omega_> walken: ok [10:26] <omega_> but p6 core is what we're really targetting [10:27] <walken> but pii/ piii data I have says 1 uop for both constant a= nd cl register shifts [10:27] <omega_> ok [10:27] <omega_> well, I think it would be very good to construct a timin= g harness and test data [10:27] <omega_> and build several different varieties of dct block parse= rs [10:28] <omega_> to get a handle on what concepts work and what don't [10:28] <omega_> I don't have any concrete basis for deciding what might = or might not be faster, except what I've read in Intel's docs [10:28] <omega_> which aren't anywhere near complete enough for a full pi= cture [10:31] <walken> hmmm [10:31] <walken> I'm under the impression that mmx parsing was worth it o= n pentium only because the regular integer register shift (by a variable = amount) was so slow [10:31] <omega_> no, there's also the fact that the regs are twice the si= ze [10:32] <walken> well on ppro we may still have the advantage that we don= t reload as often [10:32] <walken> but we have to find an easy way to test the high order b= its though [10:32] <omega_> depends [10:32] <walken> probably has to add a shift there [10:32] <omega_> if what vlcgen does is fast, that combo may be faster [10:32] <walken> yes [10:32] <omega_> but we should construct and time them to find out [10:33] <walken> guess it has to be tested [10:33] <omega_> well probably have a dozen different variants before we = know what's going on <g> [10:33] <omega_> and just maybe we'll hit on a new one ;-) [10:34] <walken> heh [10:34] <omega_> that's how I managed to come up with that new startcode = search.... [10:34] <walken> that I still have to integrate too [10:34] <omega_> doh [10:35] <walken> I'm pretty slow these days [10:35] <walken> I need an upgrade :) [10:35] <omega_> of course, that startcode search isn't a good thing, as = I mentioned [10:35] <omega_> hrm, another idea: [10:36] <omega_> with sse, I think you can do something like movaps or va= riant that doesn't polute the cache [10:36] <omega_> a streaming load, which would be ideal for the bitstream= stuff, since it hits the reg, gets torn apart, and never used again [10:37] <walken> yes [10:37] <walken> I dont expect it to make a bit diff, but we can get it f= or free :) [10:38] <omega_> maybe [10:38] <omega_> it's part of the prefetch stuff, so it could suck badly [10:38] <walken> in piii there is the prefetch bits that can achieve the = same (prefetchnta) [10:38] <omega_> right, I've messed with that and it can =02obliterate=02= performance if not done perfectly [10:38] <walken> the amount of bitstream data is quite minimal compared t= o the mc data though [10:39] <walken> now MC *is* trashing the cache [10:39] <omega_> right [10:39] <omega_> I've had a lot of ideas on that.... [10:39] <omega_> I need to see about going to OLS [10:39] <omega_> cause we need a whiteboard and some pens.... [10:40] <omega_> or maybe some whiteboard=02s=02 and some pens [10:42] <walken> yay [10:42] <walken> it'd be very nice to meet you there [10:42] <omega_> I should just plan on being there whether rr sends me or= not [10:44] <walken> thats what I did [10:44] <walken> vmware doesnt send me [10:44] <walken> (bastards) [10:44] <omega_> mu [10:45] <omega_> wow. it's not entirely correct, but it's below 7.5 cycl= es per getbits [10:46] <omega_> about 7.36 avg [10:46] <walken> you should test with variable-size getbits [10:47] <walken> else the branch predictor task is too easy :) [10:47] <omega_> yeah [10:47] <omega_> problem is then accurately timing it [10:47] <omega_> I suppose I should just pre-build an array of random bit= lengths [10:47] <omega_> or maybe build a dct-coeff parsing test program <g> [10:49] <walken> I guess [10:56] big_T (th...@fy...) joined #gstreamer. [11:03] <omega_> hrm, /me has a problem that he doesn't want to try to so= lve right now [11:03] <walken> what ? [11:04] <omega_> inline code calls a function that can change memory, but= gcc isn't recognizing that it then just force a reload from the memory i= nto the register before calling that inline code again [11:04] <omega_> so the inline code gets old data after a while ;-( [11:04] <omega_> need to have the inline syntax somehow make sure that th= e compiler knows that that call can mangle %5 [11:05] <omega_> where %5 is the register version of it [11:05] <omega_> or.... [11:05] <omega_> there [11:05] <omega_> just don't let gcc use a register for it [11:05] <omega_> checking the resulting code though.... [11:05] <omega_> ah, no problem [11:05] <omega_> ooooh, problem solved [11:06] <omega_> getbits works perfectly now [11:07] <walken> omega. yet another bit parsing strategy: [11:07] <walken> (that one is from someone else) [11:07] <omega_> um, except for a segfault ;-( [11:07] <walken> when you need to have at least n bits, do something like [11:08] <walken> whatever =3D mem[bits >> 3] << (bits & 7); [11:08] capuchin (jb...@or...) left irc: Leaving [11:08] <omega_> uh? [11:08] <walken> and then when you need to dump bits you just do bits +=3D= dumpbits [11:08] <walken> basically you always reload [11:08] <omega_> ick [11:08] <omega_> doing byte ops [11:09] <walken> adds instructions but avoids the conditional jump [11:09] <omega_> mabye write it down for future investigation, but I'd pu= t it towards the bottom of the list ;-) [11:09] <omega_> that's what ac3dec *originally* did [11:09] <walken> well no, you could reload 32 bits, do a bswap, and (afte= r the shift) be sure you have at least 24 bits [11:09] <omega_> hmmm [11:10] <omega_> actually, no, what the original ac3dec bitstream code di= d was even dumber [11:10] <omega_> every get would grab and assemble 24 bits and pick out o= f that [11:10] <omega_> so every get required a massive amount of work [11:11] <omega_> hrm, now I have an odd bug [11:14] <omega_> hrm, there are even more potential gains if I could make= use of the conditionality of the inline stuff [11:15] <omega_> "ri"(num_bits), where the "r" case would movd %n,%mm4 an= d psllq %mm4,%mm7 [11:15] <omega_> and the "i" case would just psllq n,%mm4 (immediate) [11:15] <omega_> one of the disadvantages of a riscy insn-set [11:16] <omega_> but even alpha suffers from that particular problem, iir= c [11:18] <omega_> specifically, alpha has the ability to use immediate in = that particular way [11:21] <walken> I wouldnt call this a problem [11:21] <omega_> right [11:24] <walken> where are the intel appnotes now ? [11:24] <walken> I want to have a quick look at their mmx parsing appnote [11:24] <omega_> they've hidden them behind a login system [11:24] <omega_> it really pisses me off [11:24] <walken> rhaaaaaaaaaaaaaaaaaaaaaaaaaaaaa [11:24] <walken> whats the number ? [11:25] <omega_> I 'mirrored' ap527, but netscape still has a cow loading= it [11:25] <omega_> they need to just put the stupid things out as pdf's and= be done with it [11:28] <omega_> um, this is screwy [11:28] <omega_> if I printf num_bits before doing anything, everything i= s fine [11:28] <omega_> if I don't, I get num_bits totally mangled as it gets to= the bottom half [11:34] <omega_> #$^*%$@ [11:34] <omega_> I can see no reason for this to have this behavior [11:36] arik (ar...@sd...) joined #gstreamer. [11:36] <arik> lo [11:36] <omega_> yo [11:36] <arik> i'm actually working on gstplay ;-) [11:36] <omega_> cool [11:36] <arik> well [11:37] <arik> it would be if it was working [11:37] <omega_> heh [11:37] <omega_> I'd be making progress right now if =02I=02 was working,= but I'm not.... [11:37] <omega_> I need sleep [11:37] <arik> hehe [11:37] <arik> do you have any idea why the stop button doesn't work? [11:37] <omega_> unfortunately, this has me totally baffled [11:37] <omega_> um, not of the top of my head [11:37] <arik> heh [11:37] <omega_> esp not right now <g> [11:37] <arik> ;-) [11:37] <arik> enjoy the sleep ;-P [11:38] <omega_> yeah, I'll try [11:38] <omega_> walken: will you be around? [11:38] <walken> when ? [11:38] <walken> I'm about to go sleep [11:38] <omega_> whenever, like tomorrow evening [11:38] <walken> err, yes I guess [11:38] <omega_> or better question is: how soon would you be able to sta= rt building a perf harness? [11:38] <walken> huh [11:39] <arik> i wonder if there is a gnome program to catalog your cd's [11:39] <omega_> arik: of course there is ;-) [11:39] <walken> I'd rather work on the ac3 downmix :) [11:39] <omega_> I just dunno what it'd be called [11:39] <omega_> walken: hehe [11:39] <arik> heh [11:39] <arik> i'm not sure there is [11:39] <arik> ooh [11:39] <arik> cdkeeper [11:40] <arik> oh [11:40] <arik> it's for data cds [11:40] <walken> omega. in ac3 I also have a paper that describes another= implementation of the imdct that might be faster.... [11:40] <walken> have to try it though [11:40] <omega_> cool [11:40] <walken> hate it when they give equations but not the code [11:40] <arik> gmmusic [11:40] <omega_> well, I must sleep, these insn's are turning into mush [11:41] <omega_> l8r all [11:41] omega_ (om...@om...) left irc: killall -SLEEP omega_ [11:42] walken (fo...@c1...) left irc: l8r [11:51] steveb (ste...@21...) left irc: quit has st= eveb, yes [11:58] thomas (thomas@212.100.172.175) joined #gstreamer. [11:58] <thomas> hi [12:01] <arik> hye [12:11] Nick change: thomas -> thomas-eating [12:47] richardb (ri...@ix...) joined #gstreamer. [12:47] <richardb> Morning all. [13:11] Nick change: ajbusy -> ajmitch [13:12] <ajmitch> hi all [13:12] <arik> hey [13:12] <ajmitch> arik: workin on gstplay, eh? ;) [13:12] <arik> yeah more or less [13:12] <arik> it's not going all that well [13:12] <arik> but yeah [13:12] <ajmitch> you looked at my bug report(s)? ;) [13:12] <arik> yeah [13:12] <arik> thanks ;-P [13:12] <ajmitch> no probs ;) [13:13] <arik> hehe [13:13] <ajmitch> i was told to report crashes & stuff not working [13:13] <arik> oh [13:13] <arik> yeah [13:13] <arik> i need to figure out the stop button thing really [13:13] <ajmitch> yeah [13:13] <ajmitch> i dunno if it's a gstplay issue or something else [13:13] <arik> yeah me neither [13:14] <ajmitch> i haven't looked that hard at the gstplay source (and a= m not likely too right now) [13:14] <arik> hehe [13:14] <ajmitch> have just been doing about 3 hours or physics study, wi= th stuff like theory of relavtivity, etc... ;) [13:14] <arik> heh [13:14] <arik> i like that stuff [13:14] <ajmitch> yeah, it's not bad [13:17] <ajmitch> arik: what gstplay stuff you doing? [13:17] <arik> open location [13:17] <ajmitch> 10min hack? ;) [13:17] <arik> heh [13:17] <arik> mostly [13:17] <arik> the dialog is showing up [13:17] <arik> other then that it should work [13:17] <ajmitch> that takes 2 min ;) [13:17] <arik> heh [13:17] <ajmitch> well, maybe 5... [13:18] <arik> it's not working though [13:18] <arik> for some odd reason [13:18] <arik> i'm sure i would understand if i was less tired [13:18] <ajmitch> hmm, gonna commit the source? ;) [13:18] <arik> hmm [13:18] <arik> maybe [13:18] <arik> ;-) [13:18] <ajmitch> send me a diff, i might look at it (it's before midnigh= t here) [13:25] <arik> heh [13:26] <ajmitch> ah well [13:26] <ajmitch> only problem is i lack caffiene ;) [13:26] <arik> ;-) [13:30] Action: ajmitch should probably try & sleep soon - have lecture a= t 8am [13:30] <arik> yeah you should [13:30] <arik> go to bed! [13:30] <ajmitch> lol [13:30] <ajmitch> you should too then ;) [13:30] <arik> heh [13:30] <arik> yes [13:30] <arik> i should [13:31] <ajmitch> just a few minutes more on gstplay, he says... ;) [13:31] <arik> haha [13:31] <ajmitch> until the sun starts shining thru the curtains... [13:31] <ajmitch> or is it already? ;) [13:35] Nick change: ajmitch -> ajzzzz [13:35] <ajzzzz> night [13:43] ajmitch_ (aj...@p5...) joined #gstreamer. [13:44] ajzzzz (aj...@p5...) left irc: Ping timeout f= or ajzzzz[p53-max3.dun.ihug.co.nz] [13:44] Nick change: ajmitch_ -> ajzzzz [13:58] ajzzzz (aj...@p5...) left irc: Ping timeout fo= r ajzzzz[p5-max2.dun.ihug.co.nz] [14:01] <richardb> Hmm, anyone know which versions of autoconf are meant = to use configure.ac? [14:02] <arik> hmm [14:02] <arik> nope [14:02] <richardb> Because I have Autoconf version 2.13 [14:02] <richardb> and that's using it. [14:02] <arik> hmm [14:02] <richardb> I had assumed that wtay had put it there for autoconf = 2.5 only.. [14:09] ajzzzz (aj...@p5...) joined #gstreamer. [14:27] <thomas-eating> richardb: is it you who is working on the floatin= g point audio stuff ? [14:28] Nick change: thomas-eating -> thomas [14:29] <richardb> Not really [14:29] <richardb> I was just trying to document what the properties used= were. [14:30] <richardb> steve baker was actually writing code using them... [14:30] <richardb> Docs are in the PWG, if you're looking for that. [14:30] <thomas> ok, I'll ask him then. I could use some floating point = sample files to experiment with audiofile floating point stuff [14:31] <richardb> Um:http://www.gstreamer.net/docs/cvs/gst-plugin-writer= s-guide/cha-basic-types.html [14:32] <richardb> You should be able to create sample files easily enoug= h with the int2float plugin... [14:34] <thomas> ok, thanks [14:43] <richardb> Ahhh.... It was the cvs automake that I've got that wa= s using configure.ac, not autoconf-2.13. [14:43] <richardb> That clears that mystery up. [15:00] <arik> dum de dum [15:03] Nick change: hds-slp -> hadess [15:03] <arik> hey hadess=20 [15:03] <arik> how goes the project? [15:03] <hadess> damn it, i can't get my groceries online, and will have = to go out [15:04] <arik> heh [15:04] <hadess> because mozilla crashes Mr.... [15:04] Action: arik hides [15:05] <hadess> bah [15:05] <arik> does rythmbox play anything yet? [15:06] <hadess> yep, but there are other issues that won't be fixed in g= st until later and that's fscking annoying [15:06] <arik> heh [15:06] <arik> can i try it out? [15:06] <hadess> no [15:07] <arik> heh [15:07] <arik> why not? [15:07] <arik> ;-P [15:07] <hadess> because it's not usable [15:07] <arik> heh, doesn't bother me [15:07] <arik> i just wanted to play with it [15:08] <hadess> there's nothing to play with ! [15:08] <arik> ok ok ;-) [15:08] <arik> i won't bug you anymore [15:08] Nick change: hadess -> hds-town [15:09] <hds-town> getting the food that mozilla crashes on ;) [15:09] <arik> enjoy the grocery hunting ;-) [15:09] <arik> heh [15:09] <arik> sorry bout that [15:09] <hds-town> i think the nss plugin is flaky [15:10] <arik> prob [15:10] <arik> in fact [15:10] <arik> i know it is [15:26] matth (ma...@bo...) joined #gstreamer. [15:27] Nick change: matth -> matth-distracted [15:31] <thomas> anyone: know if it's ok to still add stuff to CVS now th= at the 0.2.0 pre version is out ? [15:31] <arik> yes it's fine [15:31] <arik> afaik [15:31] <thomas> ok, oss needs some work [15:32] <thomas> ;) [15:32] <arik> good deal [15:32] <arik> ;-) [15:32] Action: arik is struggling with gstplay [15:32] <thomas> what's it doing wrong ? [15:32] <arik> heh [15:32] <arik> well [15:32] <arik> stop doesn't work [15:32] <arik> but i'm actually adding an open location dialog atm [15:33] Action: thomas curses at sourceforge [15:33] <arik> heh [15:33] <thomas> ok, where do I reset my password ? [15:33] <arik> i have no idea [16:28] gilles (no...@ca...) joined #gstr= eamer. [16:39] <thomas> hi [16:40] Nick change: hds-town -> hadess [16:43] dobey (dobey@141.154.95.104) joined #gstreamer. [16:44] arik (ar...@sd...) left #gstreamer. [16:44] <dobey> whee [16:45] <hadess> helllooo [16:45] <dobey> sup? [16:45] <hadess> done my shopping and got a haircut [16:45] <dobey> eh, i guess writing a bonobo control for gstreamer is som= ething i dunt have to do [16:46] <hadess> cactus and wtay are doing that [16:46] <dobey> yeah, assholes [17:00] <thomas> anyone care to place a bet as to why oss plugin compiles= but fails at register with the warning undefined symbol: _gst_debug_hand= ler ... [17:01] <thomas> ... when that symbol is only in the binaries, not in the= oss src ? [17:01] <thomas> or is it something stupid on my part ? [17:01] <hadess> thomas: there might be a #define GST_DEBUG in the oss sr= c sources [17:03] <hadess> thomas: tell wtay about it [17:23] dobey (dobey@141.154.95.104) left irc: eh [17:32] dobey (dobey@141.154.95.104) joined #gstreamer. [17:34] gilles (no...@ca...) left irc: KV= irc 2.1.0 'Dark Star' [17:45] matth-distracted (ma...@bo...) left i= rc: Ping timeout for matth-distracted[boisdslgw5poolb76.bois.uswest.net] [18:16] ChiefHighwater (pa...@su...) joined #gstre= amer. [18:16] #gstreamer: mode change '+o ChiefHighwater' by ChanServ!s@ChanSer= v [18:16] #gstreamer: mode change '-o ChiefHighwater' by ChiefHighwater!pau= l...@su... [18:16] Action: dobey ^5 hadess [18:18] <hadess> ~~~ got icecast to work ~~~ [18:18] <hadess> la la la [18:19] <hadess> i need to test more though [18:20] <thomas> hadess: what are you testing ? [18:20] <hadess> thomas: an icecast gnome-vfs module [18:21] <hadess> wheee, works fine with 5 different radios ;) [18:21] <thomas> hadess: as in, get audio from an icecast server through = http in gnome-vfs ? [18:22] <hadess> thomas: but using a custom icecast module, icecast and s= houcast modules are badly broken http servers, and this module supports t= hem [18:23] <hadess> icecast and shoutcast servers are badly broken, even [18:23] <thomas> hadess: what do you mean, "broken" ? Are you talking abo= ut the upcoming 2.0 version ? [18:23] Action: thomas is still waiting for his sourceforge password [18:24] <hadess> thomas: if you use the stock http gnome-vfs module, you = won't be able to play half of the streaming radios [18:24] <hadess> thomas: they spit out stuff that is not http-compliant [18:25] Uraeus (csc...@c2...) joined #gstreamer. [18:25] <hadess> hey Uraeus [18:25] <Uraeus> howdy [18:26] <thomas> hadess: any idea what they do wrong ? you're not connect= ed to the metadata port, right ? [18:26] <dobey> hadess: custom methods are http compliant, the rfc allows= them i think [18:26] <hadess> Uraeus: dude, i got the icecast gnome-vfs module to work= ! [18:26] <dobey> Uraeus: !!!! [18:26] <hadess> thomas: headers broken for the most part [18:26] <Uraeus> hadess: le cool! [18:26] <Uraeus> dobey: got more replies to your hacker request? [18:26] <thomas> hadess: I'd like to try it anyway ;) [18:27] <dobey> Uraeus: heh, i got the 1 guy wanting to help, 1 from jfle= ck asking about using it for a help browser, and 1 guy telling me i shoul= d use EBrowser instead [18:27] <dobey> :-( [18:27] <hadess> thomas: i need to hack a little bit on the gnomevfssrc f= irst, then add authentication and proxy support [18:28] <dobey> hrmm [18:28] <Uraeus> EBrowser? [18:28] <dobey> i don't feel so good [18:28] <dobey> Uraeus: the control that comes in gtkhtml [18:28] <dobey> it's what it uses right now [18:28] <Uraeus> dobey: well, actually if you can use it as a help browse= r in nautilus instead of mozilla that would probably speed help reading u= p a bit [18:28] <dobey> Uraeus: well, hopefully it'll be used as a help browser l= ike IE is in windows [18:28] <dobey> ;-) [18:28] <Uraeus> dobey: yup, well I guess that with our [18:28] <Uraeus> err [18:29] <dobey> Uraeus: heh, and the barque dude sent me a patch that see= ms somewhat wrong :-/ [18:29] <Uraeus> dobey: wrong? [18:30] <dobey> Uraeus: he added "proxy" support, and set a default, and = did some weird thing to the about box, and made File-Exit "work" by just = doing gtk_main_quit() which doesn't kill the component processes [18:32] <hadess> dobey: tell him, he'll correct the patch [18:32] <dobey> hadess: i will, i'm just not feeling well right now [18:33] <Uraeus> dobey: your sick? [18:33] <dobey> Uraeus: headache [18:33] <Uraeus> dobey: heard of Paracet? [18:34] <dobey> no [18:35] <Uraeus> dobey: ok, the layman term is 'painkiller' drug [18:35] <dobey> Uraeus: yeah, they don't usually work for me [18:35] <dobey> Uraeus: unless i sleep as well [18:35] <dobey> Uraeus: and the sleep is what does it i think [18:35] <dobey> Uraeus: lunch will probably help too [18:36] thomas (thomas@212.100.172.175) left irc: [x]chat [18:36] <Uraeus> dobey: yes, to little food and to little sleep has a ten= dency to make people feel unwell [18:36] <hadess> whee, this is wicked ;) [18:37] <Uraeus> hadess: so what is the URL of your icecast broadcast? [18:37] Nick change: steveb-away -> steveb [18:37] <hadess> Uraeus: i'm not broadcasting, i'm listening [18:38] <Uraeus> so what do you think of my suggestion for a codename for= 0.2.0? [18:39] <dobey> i'm gonna eat [18:39] Nick change: dobey -> dobey-eat [18:39] <hadess> heh, twisted as usual ;) [18:40] <richardb> What's a svirfneblin? [18:40] <hadess> richardb: i told you he was twisted [18:40] Action: richardb hmmms [18:41] <hadess> richardb: probably some sort of goblin thingy from norwe= gian mythology [18:41] <Uraeus> no quite [18:41] <Uraeus> but if you find out you discover the humour of it too :) [18:41] <Uraeus> hint -> google [18:42] <hadess> heh, a gnome ! [18:42] <Uraeus> yup [18:43] <hadess> http://www.oec-sys.com/users/steve/ropforwhities/Races/s= virfneblin.htm [18:50] Zeenix (pro...@ho...) joined #gstreamer. [18:51] <Zeenix> hi everyone [18:52] <Zeenix> does gtv uses gstreamer [18:52] <Zeenix> ? [18:53] <Uraeus> Zeenix: no it uses smpeg if I remember correctly [18:54] <Zeenix> when i istalled gstreamer rpms for the first time yester= day, i found gstmediaplayer, which does not perform well [18:54] <Zeenix> atleast on my computer. But performance of gtv is bearab= le [18:54] <Uraeus> Zeenix: well, the 0.1.1 release is very alpha quality, b= ut we have a 0.2.0 release in a week or two which I think will be much mo= re userfriendly [18:55] <Zeenix> are you one of the developers of gstreamer ? [18:55] <Zeenix> Uraeus: asking from you [18:56] <Uraeus> Zeenix: depends on what you call a developer, I don't co= de :) [18:56] <hadess> hmm, bad, gstmediaplay plays shoutcast streams with mpg1= 23 [18:56] <Zeenix> then what do you do ? [18:57] <Uraeus> Zeenix: documentation and other such tasks [18:58] Zeenix (pro...@ho...) left irc: Read error t= o Zeenix[host-97.netzone.net.pk]: Connection reset by peer [19:01] <Uraeus> hadess: are we closer to a rhytmbox alpha release? [19:02] <hadess> Uraeus: no, not until the signals are fixed in gst [19:02] <Uraeus> hadess: solved the bonobo problem? [19:03] <hadess> Uraeus: yeah, ditched bonobo ;) [19:03] <Uraeus> whas mt support in orbit the problem? [19:04] Zeenix (pro...@ho...) joined #gstreamer. [19:04] <hadess> probably yeah [19:04] <Zeenix> sorry, got disconnected for a while [19:04] <Zeenix> hadess: talking to me [19:04] <Uraeus> no me [19:05] <Zeenix> what about my question ? [19:06] <hadess> what question ? [19:06] <Uraeus> to me? [19:07] <Zeenix> can i use mp3 compression/decompression for data other t= han voice/video [19:07] <Zeenix> ? [19:07] <hadess> mp3 is only sound, period [19:08] <Zeenix> but i want to use 5Khz/unsigned 8bit/ mono [19:08] <Zeenix> it doesnt seem to do that for me [19:10] <Uraeus> what does it do? [19:11] <Zeenix> i havent tried it practically but i could not find this = format on man page for lame [19:12] <Uraeus> Zeenix: maybe bladeenc supports it? [19:12] <Zeenix> whats that ? [19:13] <Uraeus> Zeenix: another mpe encoder, and alternative to lame [19:14] <Zeenix> how much does it compress & at what speed( in KB/s ) [19:15] <Zeenix> ? [19:15] <hadess> Zeenix: for heaven's sake, why do you want to use mp3 fo= r voice ? [19:15] <Uraeus> Zeenix: dont really know, you have to check it out yours= elf [19:16] <Zeenix> hadess: sorry, dont be so angry, i know i am very stupid= at all these [19:17] Parapraxis (Sp...@us...) joined #gs= treamer. [19:17] <Uraeus> Zeenix: hadess isn't angry, he is just a bit loud :) [19:17] <Zeenix> its just that i am very new to Linux world, specially gs= treamer [19:17] <Uraeus> Zeenix: np [19:18] <Uraeus> Zeenix: ask and we will answer as best as we can :) [19:19] <Zeenix> hadess: then give me a neet & clean solution for the com= pression/decompression of the lowest quality sound [19:19] <hadess> gsm [19:21] <hadess> for very low bitrates and voice, gsm is the best you can= find [19:22] Zeenix (pro...@ho...) left irc: Ping timeou= t for Zeenix[host-104.netzone.net.pk] [19:24] Nick change: wtay-zZz -> wtay [19:24] <Uraeus> hi wtay [19:24] <wtay> bonjour les mecs :) [19:24] <Uraeus> trebie [19:24] <wtay> tr=E8s bien? [19:24] <Uraeus> is that how its spelt? [19:24] <wtay> I think so... [19:24] Zeenix (pro...@ho...) joined #gstreamer. [19:24] <Uraeus> think it means great or wonderful or something [19:25] <wtay> yes, it means, "very good" [19:26] <hadess> wtay: salut ! [19:26] <Zeenix> sorry! my internet conection is too dump, i got disconne= cted again for while=20 [19:26] <hadess> <Zeenix> hadess: then give me a neet & clean solution fo= r the compression/decompression of the lowest quality sound [19:26] <hadess> <hadess> gsm [19:26] <hadess> <hadess> for very low bitrates and voice, gsm is the bes= t you can find [19:26] <wtay> hadess: du-u-ude <g> [19:27] <wtay> yes please [19:27] <hadess> wtay: i have an icecast gnome-vfs module [19:27] <wtay> hadess: yeah, I heard, does it work in gstmediaplay? [19:27] <hadess> wtay: the only problem is that the autoplug plays it wit= h mpg123, otherwise, yes, it works [19:28] <wtay> hadess: yes, I dunno how that happens yet.. [19:28] <wtay> s/how/why/ [19:28] <Zeenix> hadess: are u sure can support 5Khz/ 8bit/ mono [19:28] <Zeenix> ? [19:28] <wtay> Zeenix: yes [19:29] <wtay> Zeenix: actually 16bit/X-bitrate [19:29] <Zeenix> wtay: how much does it compress & how fast [19:30] <wtay> 160 shorts to 33 bytes (10:1) and it's very fast [19:30] <wtay> lemme check the speed [19:30] <hadess> you can even clone some features of the mobile phones wi= th gst [19:31] <hadess> like the 'comfort silence' [19:31] <Zeenix> wtay: what do ya mean by shorts ? [19:31] <wtay> Zeenix: 16 bit signed int [19:33] <hadess> i'm missing proxy support... [19:36] <Zeenix> waiting for wtay..... [19:38] <wtay> Zeenix: creating a test app [19:39] Zeenix (pro...@ho...) left irc: Ping timeout= for Zeenix[host-67.netzone.net.pk] [19:42] Zeenix (pro...@ho...) joined #gstreamer. [19:43] <Zeenix> got dissconnecte again for a while [19:43] <wtay> Zeenix: just a moment, I need to compile gstreamer... [19:43] <Zeenix> wtay: did you get that ? [19:43] <wtay> yes [19:43] <Uraeus> Zeenix: you need to get yourself a better line :) [19:44] <Zeenix> OK [19:44] <wtay> hmm, i need to add RGB24 to YUV conversion to stop ajzzzz = from complaining.. <g> [19:44] <Zeenix> wtay: whats the speed ? [19:45] <wtay> Zeenix: dunno yet, I'm still compiling [19:45] <wtay> Zeenix: I expect several MB/sec [19:45] <Zeenix> whats you estimation about its speed ? [19:45] <wtay> Zeenix: I expect several MB/sec [19:45] <hadess> wtay: do you know how i can have a #define that would pr= int more than one arg ? [19:46] <Zeenix> its the compression speed ? [19:46] <hadess> wtay: like #define FOO(x) g_print(x) but that would work= with more than one arg ? [19:46] <wtay> hadess: #define myprint(format, arg...) g_print ("somethin= g" format, ##arg) or something [19:47] <wtay> hadess: gstinfo has lots of examples [19:47] <hadess> wtay: ok, thanks [19:47] <wtay> Zeenix: gsm has equal speed in compression/decompression I= think [19:48] Action: richardb has spent enough time at computers for today. [19:48] <wtay> hi richardb [19:48] <richardb> Hi wtay [19:48] <wtay> hi Uraeus [19:48] <Zeenix> i installed gstreamer yesterday for the first time, how = can i check it myself [19:48] <Zeenix> ? [19:48] <richardb> wtay: do you find that autoconf 2.50 is much much slow= er that 2.13? [19:48] <wtay> Zeenix: write a little app [19:48] <richardb> or whatever you used before? [19:48] <wtay> richardb: oh yes :( [19:48] Action: richardb sighs [19:49] <wtay> richardb: did you notice the sier of the configure script = it creates? [19:49] <wtay> s/sier/size [19:49] <wtay> almost twice as big [19:49] <richardb> -rwxr-xr-x 1 richard richard 352734 May 29 18:1= 8 configure [19:49] <richardb> Eek. [19:49] <wtay> yup [19:49] <Zeenix> ha, i come to this channel to find out first that can gs= treamer work as i want it to do, then i'll learn & use it [19:49] <wtay> mine is a little larger I think [19:50] <wtay> Zeenix: that's why I'm writing the test app for you <g> [19:51] <Zeenix> wtay: im in a netcafe, i really need to go home or my mo= m will be very worried for me [19:51] <wtay> Zeenix: ok, give me 1 more minute [19:51] <hadess> #define PRINT(x...) g_print(## x) [19:52] <wtay> hadess: something like that, yes [19:52] <hadess> wtay: no, that works [19:52] <richardb> Time for me to go. [19:52] <richardb> Happy hacking everyone [19:53] <wtay> richardb: cya [19:53] richardb (ri...@ix...) left irc: =02[=02BX=02]=02 = I theenk I need a beeger box! [19:53] <wtay> real 0m1.605s [19:53] <wtay> user 0m1.340s [19:53] <wtay> sys 0m0.220s [19:53] <wtay> for a 3MB file [19:53] <wtay> decompression [19:53] matth-distracted (ma...@bo...) joined= #gstreamer. [19:54] <wtay> real 0m2.050s [19:54] <wtay> user 0m1.930s [19:54] <wtay> sys 0m0.090s [19:54] <wtay> compression [19:54] <wtay> so about 2MB/sec om my machine [19:54] <Zeenix> wtay: 2 sec for compression, 1 sec for decompression [19:55] <Zeenix> wtay check the size of compressed file [19:55] <wtay> Zeenix: 3MB [19:55] <wtay> actually, 3MB/10 =3D 300K [19:56] <wtay> input is 3MB [19:56] <wtay> Zeenix: that's on an Athlon 1.2GHz [19:57] <Zeenix> wont be too slow on 266Mhz, 64MB RAM [19:57] <wtay> nope, it even works on a cellular phone [19:57] <hadess> heh [19:58] <wtay> Zeenix: and mp3 hardly does 10:1 so gsm is the thing you w= ant for voice [19:58] <Zeenix> wtay: i could not understand, was the compressed file 30= 0K or 3MB [19:58] <wtay> Zeenix: input was 3MB, output compressed 300K, output deco= mpressed 30MB [19:58] <wtay> roughly [20:00] <Zeenix> 3MB to 30MB, why that ? [20:00] <wtay> 10:1 [20:01] <Zeenix> 3MB file when compressed & again decompressed, it should= be again 3MB [20:01] <hadess> wtay: what is weird is that input -> compression -> deco= mpression -> output give you different input and output sizes [20:01] <hadess> Zeenix: or roughly [20:02] <wtay> guys: I gave the decoder/encoder a 3MB input file [20:02] <wtay> the same file [20:02] <wtay> gsm is *always* 10:1 [20:03] <wtay> size (input -> comp -> decomp) =3D=3D size (input) [20:04] <Zeenix> wtay: 3MB input-> comp 300k -> decomp 30MB !=3D size inp= ut [20:05] Nick change: dobey-eat -> dobey [20:05] <wtay> Zeenix: read again: [20:05] <wtay> Zeenix: input was 3MB, output compressed 300K, output deco= mpressed 30MB [20:05] <Zeenix> wtay: 3MB !=3D 30MB [20:06] <hadess> Zeenix: no, he hasn't done that [20:06] <wtay> ok, lemme explain it in another way.. I used the following= pipeline: [20:06] <wtay> disksrc location=3Dsome3MBfile ! gsmdec ! fakesink [20:06] <wtay> and [20:06] <wtay> disksrc location=3Dsome3MBfile ! gsmenc ! fakesink [20:07] <wtay> case 1 (gsmdec) produces a 30MB file [20:07] <wtay> case 2 (gsmenc) produces a 300K file [20:07] omega_ (omega@216.99.212.251) joined #gstreamer. [20:07] <wtay> yo [20:07] <steveb> yo [20:08] <wtay> dammit, I can't change my PW on SF [20:10] <wtay> ok, changed it [20:11] <Zeenix> wtay: dont mind it, but a long way to get understand you [20:11] <Zeenix> your language [20:11] <wtay> uhm [20:12] <wtay> for those interested: gstreamer.net/java-gst.tar.gz [20:13] <Zeenix> what i could understand was 3MB , you compressed it to 3= 00k & after decompression you got a 30MB output [20:14] <wtay> Zeenix: nope, when I compressed the 3MB file and decompres= sed it again it would still be 3MB [20:15] <steveb> wtay: you sick monkey [20:15] <wtay> steveb: where? [20:15] <Zeenix> wtay: thats plain english [20:16] <steveb> so is that pure java or compiled cygnus stuff? [20:16] <Zeenix> wtay: plz speak plain eng. like this in future, as i am = too week at english [20:16] <wtay> steveb: pure java/jni [20:17] <wtay> steveb: I'm working on writing plugins in plain java now [20:17] <Zeenix> wtay: i have worked in Java but only at Win plateform, c= an Java byte code be translated to linux native [20:17] <Zeenix> ? [20:18] <steveb> one day I might have a crack at objective C bindings [20:18] <wtay> Zeenix: yup, that's done with the JVM [20:19] <Zeenix> wtay: can i have Gtk+ like i use swing [20:19] <wtay> Mp3.java is pretty cool IMO [20:19] <Zeenix> in java [20:19] <steveb> wtay: so what evil plans do you have in mind? [20:19] <wtay> Zeenix: there are some bindings for gtk, but they are not = very good IMO [20:20] <Uraeus> Zeenix: there is a compiler for linux to turn java into = linux bytecode [20:20] <wtay> steveb: nothing in particular <g> [20:20] <Uraeus> wtay: java-gnome? [20:20] <wtay> yes [20:20] <steveb> wtay: sure ;) [20:20] <wtay> that's what I used for inspiration [20:20] <Zeenix> Uraneus: where'll i get that ? [20:20] <wtay> Zeenix: I think gcc has a backend for that [20:21] <wtay> gcj or something [20:21] <wtay> although that voids the point of java bytecode [20:21] <Zeenix> doesnt matter [20:21] <Uraeus> Zeenix: http://gcc.gnu.org/java/ [20:22] <Uraeus> Zeenix: http://gnome-gcj.sourceforge.net/ [20:22] <Uraeus> Zeenix: http://java-gnome.sourceforge.net/ [20:22] <wtay> better run it in a JVM, they are getting pretty fast [20:23] <Zeenix> wtay: is Linux JVM faster than that for winxx [20:23] <dobey> jvm? [20:23] <dobey> where? [20:23] <Uraeus> wtay: did you ever get that GNOME snooker game working? [20:23] <wtay> Zeenix: probably not [20:23] <wtay> Uraeus: nope :( [20:24] <wtay> Uraeus: I tried to fix it but it seemed something ugly wit= h the guile bindings so I gave up [20:24] Action: Uraeus is playing with LeoCAD at the moment, the GTK+ Leg= o CAD program :) [20:24] <Zeenix> wtay: i have worked in that, the only thing i have in ja= va is its very low speed [20:25] <Zeenix> i meant hate [20:25] <Zeenix> not have [20:25] <Uraeus> wtay: yes, the GUILE bindings seems to get a lot of deve= lopment yet no releases, so it is hell to find what apps are based upon [20:25] <wtay> Zeenix: yes, gui stuff is very slow..=20 [20:25] <wtay> Uraeus: I suppose [20:26] <Uraeus> wtay: spent a full day almost trying to get the Galway (= html-editor) to run, but the GUILE bindings killed my effort :) [20:27] <wtay> Uraeus: :( [20:27] <wtay> so: I want to see other bindings too <hint> [20:27] <Zeenix> bi everybody, thanks for a lot of help, specially wtay [20:27] <wtay> Zeenix: cya [20:28] <Zeenix> wtay: special thanks, cant do anything else [20:28] <Zeenix> for you [20:28] <wtay> it's ok <g> [20:30] <Uraeus> Zeenix: you can make GStreamer java bindings :) [20:37] Nick change: omega_ -> omega_breakfast [20:37] <omega_breakfast> Uraeus: explain what 'Svirfneblins' are? [20:39] <wtay> http://www.oec-sys.com/users/steve/ropforwhities/Races/sv= irfneblin.htm [20:40] <omega_breakfast> hmm [20:41] <hadess> huh, dexter's lab special: a movie ;) [20:41] Nick change: hadess -> hds-tv [20:43] matth-distracted (ma...@bo...) left i= rc: Ping timeout for matth-distracted[boisdslgw5poolb76.bois.uswest.net] [20:46] Apoc (ep...@pp...) joined #gstreamer. [20:48] Nick change: taazzzz -> taaz [20:50] Zeenix (pro...@ho...) left irc:=20 [20:52] <wtay> yo [20:52] <Apoc> Hello all ... [20:53] heroines (root@205.229.113.88) joined #gstreamer. [20:53] <wtay> yo [20:53] <Apoc> There is gstreamer 0.2.0 pre ?? [20:54] <heroines> test [20:54] <wtay> hmm, yeah [20:59] <taaz> yo, dvdsrc isnt on the roadmap [20:59] <taaz> and speaking of that... is it supposed to work? ;) [21:00] <Apoc> wtay : I don't know if you remember ... I'm the guy who sp= oke with you about the ability to play just one track with cdparanoia plu= gin :-) [21:00] <wtay> taaz: nope [21:00] <wtay> Apoc: ah [21:00] <taaz> wtay: that would explain my oms thing not doing squat ;) [21:00] <wtay> Apoc: and? does it work? [21:01] <wtay> taaz: most probably, yes :-) [21:01] <wtay> taaz: If you could get l... [truncated message content] |
From: <wim...@ch...> - 2001-05-31 04:31:38
|
[06:32] ArthurDent (cf...@20...ble.rc= n.com) left irc: Read error to ArthurDent[209-6-86-78.c6-0.frm-ubr1.sbo-f= rm.ma.cable.rcn.com]: EOF from client [06:36] <taaz> ack... dent was here? [06:36] <ajmitch> must have been [06:37] Action: taaz just got done watching AntiTrust [06:37] <omega_> I don't think that was oms dent [06:37] <omega_> taaz: any good? [06:37] <taaz> i bet it was... he's in boston, the .ma. in host name... h= mm [06:38] <omega_> is his first name Arthur? [06:38] <taaz> no [06:38] <omega_> ok, then it was just a hitchhiker ;-) [06:38] <taaz> yup [06:39] <taaz> Anti Trust... entertained me. i doubt there will be too m= any movies that talk about open source ;) [06:40] <omega_> is it on video? [06:40] <taaz> yeah, i rented the dvd [06:40] <omega_> ok, I should do that when I'm in Boise next [06:40] <omega_> might be Thu/Fri [06:40] <taaz> watch it with other hackers [06:41] <omega_> heh [06:41] <omega_> why, to laugh at it? [06:41] <taaz> yeah ;) i think non-computer people may not get it [06:41] <omega_> heh [06:42] <taaz> well... there's a normal story there too... just bunch of = techno stuff too [06:42] <omega_> yup [06:46] <taaz> plus most people would seriously ridicule you for single f= raming the video just to read code ;) [06:46] <omega_> point [06:46] <omega_> recognize any? [06:46] <ajmitch> ;) [06:47] <taaz> its mostly html, just generic tags not really content [06:47] <omega_> neat [06:47] arik (ar...@sd...) joined #gstreamer. [06:47] <taaz> one scene with gnome desktop and some code but i couldnt r= eally read it [06:47] <omega_> taaz: you should at least be able to tell by the shape o= f it whether it's gnome code or some random junk [06:47] <arik> i can't remember how to pass two things as data to a signa= l callback [06:48] <omega_> you use marshal_NONE__POINTER_INT or somesuch [06:48] <omega_> and just keep listing arg types on the cmdline [06:48] <arik> are you talking to me? [06:48] <omega_> yup [06:48] <arik> hmm [06:48] <omega_> lemme see if there's an example [06:48] <arik> i wonder what the right way to do this is [06:48] <omega_> what are you doing? [06:48] <arik> open location dialog [06:48] <arik> the ok button callback [06:49] <arik> needs the entry to get the filename [06:49] <omega_> gstobjectc:620 [06:49] <omega_> er, gstobject.c:620 [06:49] <arik> and the gstmediaplay to play it [06:49] <taaz> the antitrust web site has some java in the background [06:49] Action: taaz is being a dork tonight [06:49] Action: omega_ is becoming an asm fiend tonight [06:49] <arik> hmm [06:50] <taaz> which asm? [06:50] <omega_> x86, of course [06:50] <omega_> w/mmx [06:50] <arik> not sure i follow that example 100% [06:50] Action: arik looks again [06:50] <omega_> marshal_NONE__POINTER_POINTER, NONE, 2, POINTER, POINTER [06:50] <arik> right [06:50] <arik> ok [06:50] <arik> i get it [06:53] <arik> hmm [06:53] <arik> too many arguments to gtk_signal_connect [06:53] <omega_> uh? [06:53] <arik> uh? [06:53] <omega_> you can only pass one pointer to _connect [06:53] <arik> yeah i know [06:53] <arik> oh i see [06:53] <arik> damnit [07:08] <taaz> omega_: http://www.newsforge.com/article.pl?sid=3D01/05/29= /215246 [07:08] <arik> night all [07:08] arik (ar...@sd...) left #gstreamer. [07:10] <awakened> linux has klingon laguage support in unicode! :) [07:12] big_T (th...@fy...) got netsplit. [07:12] iGN__ (ig...@lo...) got netsplit. [07:12] wtay-zZz (wi...@ca...) got netsplit. [07:12] omega_ (om...@om...) got netsplit. [07:12] omega_ (om...@om...) returned to #gstreamer. [07:12] wtay-zZz (wi...@ca...) returned to #gs= treamer. [07:12] iGN__ (ig...@lo...) returned to #gstreamer. [07:12] big_T (th...@fy...) returned to #gstreamer. [07:15] ajmitch (aj...@p5...) left irc: http://www.f= reedevelopers.net [07:16] Chw-away (pa...@su...) joined #gstreamer. [07:17] chillywilly (bau...@d1...) joined #gstr= eamer. [07:17] <omega_> awakened: but does pango draw it yet? [07:18] <chillywilly> re [07:20] awakened (awakened@A030-0722.PHL2.splitrock.net) left irc: Client= Exiting [07:32] Topic changed on #gstreamer by ChanServ!s@ChanServ: Try out 0.2.0= -pre0: http://www.urgent.rug.ac.be/thomas/download/gstreamer-0.2.0-pre0.t= ar.gz [07:33] big_T (th...@fy...) got netsplit. [07:35] big_T (th...@fy...) returned to #gstreamer. [07:42] Chw-away (pa...@su...) left irc: Read erro= r to Chw-away[sub18-226.member.dsl-only.net]: Connection reset by peer [07:44] ChiefHighwater (pa...@su...) joined #gstre= amer. [07:44] #gstreamer: mode change '+o ChiefHighwater' by ChanServ!s@ChanSer= v [07:44] #gstreamer: mode change '-o ChiefHighwater' by ChiefHighwater!pau= l...@su... [07:48] ajmitch (aj...@p5...) joined #gstreamer. [07:49] <ajmitch> hi again [07:49] <omega_> yo [07:49] <ajmitch> i should try gstmediaplay again after the alsa drivers = have been recompiled [07:50] <ajmitch> got much more memory free now (i think plex86 crashing = caused a big memleak) [07:50] <omega_> whoops [07:50] <ajmitch> no, more like oops ;) [07:51] <omega_> oh [07:51] <chillywilly> hehe [08:08] steveb (st...@no...) joined #gstreamer. [08:08] <omega_> yo [08:08] <steveb> yo [08:08] <ajmitch> yo [08:09] <steveb> has any of the lost sf stuff been recovered? [08:09] <ajmitch> lost sf stuff? [08:09] <omega_> haven't checked yet [08:10] <omega_> I had to explain to them *3* times which projects were n= eeding restore [08:10] <mwc> hi all. Should gstmediadisplay just core dump on me. (usin= g code from cvs) [08:10] <omega_> it appears it has been restored, es [08:10] <omega_> mwc: how recent is your copy? [08:10] <steveb> ajmitch: it seems sf was cracked [08:10] <omega_> hrm, maybe it hasn't been restored yet [08:11] Action: chillywilly saw the news about SF, gotta love nautilus's = new feature ;) [08:11] <chillywilly> s/new/news [08:12] <mwc> I just did a cvs co now [08:12] <ajmitch> chillywilly: or evolutions news feature? [08:12] <omega_> mwc: what kind of video are you trying to play? [08:13] <mwc> I have tried mpeg1. I have alo tried mp3 audio [08:13] <omega_> does it ever print out "RUNNING pipeline" ? [08:14] <mwc> No. 1 X "** WARNING **" and 4 X "** CRITICAL ***" [08:14] <omega_> um, hmm [08:14] <omega_> does it print a warning to run gstreamer-register? [08:15] <mwc> It did. I have now run gstreamer-register - now no warning= - but still 4 criticals [08:15] <omega_> what are those criticals? [08:16] <mwc> ** CRITICAL **: file gstbin.c: line 266 (gst_bin_add): asse= rtion `element !=3D NULL' failed. [08:16] <omega_> hmmm [08:16] <mwc> ** CRITICAL **: file gstelement.c: line 676 (gst_element_co= nnect): assertion `src !=3D NULL' failed. [08:16] <mwc> ** CRITICAL **: file gstelement.c: line 429 (gst_element_ge= t_pad): assertion `element !=3D NULL' failed. [08:16] <mwc> ** CRITICAL **: file gstelement.c: line 371 (gst_element_ad= d_ghost_pad): assertion `pad !=3D NULL' failed. [08:16] <omega_> did you do a `make install`? [08:16] <omega_> that causes problems when you're building from cvs [08:17] <mwc> Yes I did. What is the prefered method? [08:17] <omega_> don't, for a cvs build [08:17] <omega_> there are things that are designed for you to run from t= he builddir, and don't like it as much if it's installed too [08:18] <omega_> do a `make uninstall` and rerun -register [08:18] <omega_> cd tools;./gstreamr-register;cd ../gstplay;./gstmediapla= y file.mpg [08:18] <mwc> OK. Will do. [08:21] Action: ajmitch shoudl try mpeg video again after alsa drivers ha= ve compiled [08:21] <ajmitch> fscking alsa... [08:21] <mwc> Hmm - I think I need to recompile. When I installed I did = --prefix=3D/usr/local. [08:21] <omega_> steveb: see http://sourceforge.net/tracker/?func=3Ddetai= l&atid=3D200001&aid=3D428441&group_id=3D1 for the strange story of the re= store [08:22] <omega_> mwc: so? [08:22] <omega_> that's the default anyway, actually [08:24] <steveb> omega_: we can do that but only if you tell us what proj= ects to restore :) [08:25] <omega_> yeah. hello? [08:25] <steveb> weird [08:25] <omega_> and they do it again.... [08:25] <omega_> "uh, what projects were those again?" [08:25] <omega_> =02IT'S IN THE REQUEST, BUDDY=02 [08:25] <omega_> twice. [08:25] <omega_> now three times. [08:26] <steveb> maybe its the cracker and he is screwing with your head = <g> [08:26] <omega_> he's succeeded at that, at least... [08:39] <mwc> when I run -register I get 0 plugins with 0 elements and 0 = types [08:40] <omega_> hrm [08:40] <omega_> you used autogen.sh to configure gstreamer? [08:40] <omega_> or did you re-run ./configure later? [08:41] <ajmitch> omega_: only 50% cpu usage now instead of about 80-90% = ;) [08:41] <omega_> what did you change? [08:41] <ajmitch> rebooted, restarted X, got no swap in use now [08:41] <omega_> hmm [08:42] <ajmitch> X is crappy on here [08:42] <omega_> methinks you need an upgrade or three [08:42] <mwc> I think my problem is that I did ./autogen.sh and then ./co= nfigure -prefix=3D/usr/local [08:42] <omega_> ok, yeah, autogen.sh adds --enable-plugin-srcdir [08:42] <ajmitch> video card, CPU, and motherboard are the three i'd upgr= ade if i had the money ;) [08:44] <ajmitch> omega_: skips like hell now that i am compiling [08:44] <omega_> ajmitch: need to tune the queue sizes, or better do them= time-based and have a control panel in gstmediaplay [08:45] Action: ChiefHighwater loans ajmitch 'workhorse' [08:45] <ajmitch> ChiefHighwater: an old 486? ;) [08:45] <ChiefHighwater> p3 500 w/ 196M RAM and a viper v770 [08:45] <ajmitch> ChiefHighwater: this is a 400mhz cpu, and i got 384MB r= am in here (only 8mb i740) [08:46] <omega_> 8mb onboard or 8mb shared? [08:46] <ajmitch> onboard [08:46] Action: omega_ holds ChiefHighwater down to keep him from breakin= g things while twitching [08:46] <ajmitch> it's a slot-in AGP card, not like the i810 which is fou= nd on the motherboard [08:46] Action: omega_ tries not to mention shared video memory [08:46] <ChiefHighwater> down omega_, down [08:47] <ajmitch> hey, i used to have an S3 trio64v2 ;) [08:47] <omega_> ajmitch: at least your machine doesn't suck that badly [08:47] <omega_> but afaik the i740 has Xv support [08:47] <ajmitch> omega_: doesn't work tho [08:47] <omega_> oh? [08:47] <omega_> what XD? [08:47] <ajmitch> doesn't have DGA support [08:47] <omega_> er, X? [08:47] <ChiefHighwater> ajmitch has enough sense not to waste his machin= e on shared video ram [08:47] <ajmitch> 4.0.3 [08:48] <omega_> odd [08:48] <ajmitch> crappy neglected driver [08:50] walken (fo...@c1...) joined #gstreame= r. [08:50] <walken> yop [08:50] <ajmitch> i need a decent video card, what one is decent? [08:50] <omega_> yo walken [08:51] <walken> yo :) [08:51] <omega_> I'm getting more and more verse in deep inline extended = asm [08:51] Nick change: steveb -> steveb-away [08:51] <omega_> ajmitch: ati cards are good [08:51] <omega_> matrox too [08:51] <ajmitch> yeah, so i heard [08:51] <omega_> if you can get a G400 or a Rage128 [08:51] <walken> I got a g450 [08:51] <ajmitch> i would then chuck the i740 in the other computer ;) [08:52] <omega_> what I want to smack Matrox around for is the supposed p= lan to put yet another inferior secondary CRTC in the 550 [08:52] <omega_> if they're gonna do dual crtc's, they should make them =02= both=02 360MHz, not 360+230 [08:53] <taaz> if pcs didnt suck you could just use two agp cards [08:53] <omega_> yup [08:53] <omega_> but it's not pc's, it's the agp design [08:53] <omega_> it's inherently a point-to-point bus [08:54] <omega_> you'd have to add another couple hundred pins to the nor= thbridge to get a second agp port [08:54] <omega_> and there's just no way [08:55] Action: taaz uses pci for second head cause irc update bandwidth = is ~0 ;) [08:55] <omega_> heh [08:55] <omega_> I use a G400DH + Voodoo3 2000 PCI in my desktop [08:56] <omega_> the G400DH secondary crtc is only 160MHz or so, it's a j= oke [08:56] <taaz> i want the tv out to work on g400... does that work yet? [08:56] <omega_> yeah, has for a long time, but only hard-wired to certai= n modes [08:56] <taaz> could multicast dvd to my tv too that way [08:56] <omega_> XFree still has a long way to go [08:57] <ajmitch> heh, founf g450 dualhead, 32mb ddr for about $180US (st= ill way too high) [08:57] <taaz> hmmphf. in winders its really easy to use and configure [08:57] <omega_> ajmitch: check for a used g400dh or something [08:59] <ajmitch> could be a challenge [09:00] <ajmitch> will probably have to go on usenet or something ;) [09:00] <omega_> try ebay [09:01] chillywilly (bau...@d1...) left irc: Ph= ilosophers and plow men, each must know his part, to sow a new mentality = closer to the heart... [09:09] <walken> OK.... too tired to work [09:09] <walken> gotta sleep I guess [09:09] <omega_> heh. I'm using my reorder script [09:09] <walken> will do this downmix patch later :) [09:09] <walken> cya all [09:09] <omega_> timing things manually, finding the fastest subset of ge= tbits [09:09] <omega_> l8r [09:10] <taaz> that random inst ordering thing? [09:10] walken (fo...@c1...) left irc: zzzZZZ= Zzzzz [09:10] <omega_> not random, always in dependency order [09:10] <taaz> surely there is a better way [09:10] <omega_> gcc 3.0 looks like it might have some better tools and l= ibs for me to do this better [09:10] <omega_> taaz: yeah, only a deep understanding of the p6 core [09:11] <omega_> which I'm getting from a doc I just found [09:11] <omega_> one of these days I want to make a visual p6 core simula= tor, where you can *watch* uops going through the pipeline [09:11] <omega_> ouch: http://ars.userfriendly.org/cartoons/?id=3D2001053= 0 [09:11] <taaz> um... can you do gst event stuff first ;) [09:12] <omega_> yeah, that's a much later project [09:13] <omega_> this is interesting. various reorderings have 1/10th cy= cle differences over 1000 iterations [09:14] <omega_> gotta be the ifetch unit cycling around its 16-byte alig= nment stuff [09:14] <ajmitch> l8r [09:14] <omega_> cya [09:14] Nick change: ajmitch -> ajbusy [09:16] <taaz> you think a GA could improve your search? [09:16] <omega_> yes [09:16] Action: ChiefHighwater sleeps [09:16] <omega_> apparently I'm following some pretty well known people w= ith this stuff [09:16] <omega_> ChiefHighwater: you see the uf? [09:16] <ChiefHighwater> not yet [09:16] <omega_> you have to read yesterdays to get context [09:17] <ChiefHighwater> k [09:17] <omega_> taaz: you do GA's? [09:18] <taaz> not really, i just find the field interesting [09:18] <omega_> ok, wtay did his thesis on them.... [09:18] <taaz> haven't done any programming with it though [09:19] <taaz> i think genetic programming in particular is really neat [09:19] <omega_> yup [09:19] <ChiefHighwater> omega_:hehe [09:19] <omega_> it's =02so=02 true.... [09:19] <ChiefHighwater> yup [09:19] <ChiefHighwater> l8rz yall 8-] [09:19] <taaz> but my dreams of doing something interesting with it kinda= faded when i see that real projects require 1000 pc's running full time = for weeks to come up with an interesting result [09:20] <ChiefHighwater> time for sleep [09:20] <omega_> did you see the ST:V finale? [09:20] <ChiefHighwater> yes [09:20] <omega_> taaz: whoops [09:20] <omega_> so you know that the entire finale couldn't have happene= d, based on temporal mechanics.... [09:20] <taaz> not always the case i'm sure... its just the cost function= is usually computationally very expensive [09:20] <ChiefHighwater> details details [09:20] <omega_> taaz: depends on what you're going [09:21] <omega_> er, doing [09:21] ChiefHighwater (pa...@su...) left irc:=20 [09:21] <omega_> cost function for asm timings is relatively small if you= can get the assembler trimmed down enough [09:21] <omega_> you're looking at 10,000-100,000 cycle costs, on a 1,000= ,000,000-cycle/sec machine [09:21] <taaz> how would you have ended voyager? [09:22] <omega_> taaz: not sure, but I wouldn't have violated every tempo= ral mechanics rule on their books [09:22] <omega_> but then again, they do that all the time, so.... [09:22] <omega_> the whole idea of a temporal shock wave was ignored, eve= n though the idea is screwy anyway [09:23] <taaz> personally i would have had some off comment "I can't beli= eve that the main computer has run Windows all this time with any problem= s" then WHAM! they warp themselves straight through a star [09:23] <omega_> doh [09:23] <omega_> a blue star, of course [09:23] <taaz> heh [09:23] <omega_> Blue Star Of Death [09:24] <omega_> that's worth a spoof or two [09:24] <taaz> or better yet... they notice a moon in the distance. as t= hey approach... 'that's not a moon, its a space station' then the death= star blasts them out of the sky [09:24] <omega_> hehehe [09:26] <omega_> ooooh, timings validate my own analysis [09:26] <omega_> the original in-order sequence is fastest [09:27] <omega_> by .03 cycles/loop [09:27] <taaz> my brother just ordered a 1.7Ghz machine... i'm way out of= it i guess [09:27] <omega_> mu [09:27] <omega_> I'm still waiting for the 760MP [09:28] <taaz> 1,700,000,000 cyc/s... hmm [09:29] Action: omega_ wants 3,400,000,000 cyc/s [09:29] <omega_> and not a P4 either [09:29] <omega_> I want a P4, but not for that reason [09:29] <taaz> and why do you want all those cycles? [09:29] <omega_> it would be handy to have a single P4 and dual athlons o= f the same speed, just for comparison [09:30] Action: taaz has a point to make [09:30] <omega_> taaz: I eventually want to push MPEG-4 software decode t= o 1080p60 [09:30] <omega_> just to do it, cause that's a lot of stuff it'll teach m= e [09:30] <omega_> and for a sound processing machine, more cycles are *alw= ays* going to be needed [09:31] <taaz> well... not that i've done any real analysis on this but i= bet a fpga coprocessor close to the cpu running at even 50Mhz could offl= oad a very significant part of those cycles [09:31] <omega_> yup [09:31] <omega_> but try to get Intel to OK that one [09:32] <omega_> taaz: how reconfigurable would it be? [09:32] <omega_> could you allocate varying percentages to different proc= esses? [09:32] <omega_> and have it going on many threads at once, with pthreads= -type control on it? [09:32] <taaz> lots of options... hard to say what would be good to hook = on a general purpose uP [09:33] <omega_> we've got some stuff in the works at rr that's conceptua= lly very relevant to that [09:34] <omega_> taaz: how much work/research do you think it'd take to g= et it to that point? [09:35] <taaz> um.. to the point of, say, intel putting configurable logi= c of some sort on their cpus? [09:35] <omega_> that'd be ideal, yeah [09:35] <omega_> but no, say cpu + fpga + glue [09:36] <omega_> and being able to program it simply and easily, using a = standard threading model [09:36] <omega_> gate-count scheduling and such [09:36] <taaz> well, that's one of the main issues. reprogramming of mos= t devices is horribly slow [09:36] <omega_> ick [09:37] <omega_> how about area assignment, like giving 60% to one proces= s and 20% each to two others? [09:37] <taaz> right now (ie, during the day) i'm playing with an ageing = prototype of a chip that has multiple contexts and can switch between the= m in 1 cycle [09:37] <omega_> wow [09:37] <omega_> how many contexts? [09:38] <taaz> a whole 4 ;) [09:38] <omega_> not bad [09:38] <omega_> how many gates? [09:38] <taaz> i could leave it at "not enough" but for laughs i think it= s around 1000 ;) [09:38] <taaz> ie, nothing [09:38] <omega_> ick [09:39] <taaz> its really just a prototype... a few dozen made for resear= ch [09:39] <omega_> how close is it to the CPU? [09:40] <taaz> um.. well, on the board i'm using its not actually connect= ed. have PPC -> XC4085 -> [spaghetti mess of busses connecting fifos, 2 = of these prototype chips, memory, and other stuff] [09:40] <omega_> hmm, ok [09:40] <taaz> err.. that's not really accurate [09:41] <omega_> close enough ;-) [09:41] <taaz> its a big mess sort of designed for some specific applicat= ions [09:41] <omega_> so, how many gates do you think it might take to impleme= nt, say, mp3 decoding on an fpga? [09:41] Action: omega_ guesses 100K or more [09:41] <taaz> i have no clue ;) [09:42] <taaz> decoding is mucho smaller than encoding though [09:42] <omega_> what if you tied this close to memory and could switch h= ardware contexts really fast, treat each context like a kernel process? [09:42] <omega_> taaz: yeah, but still nontrivial [09:42] <omega_> you could pipeline data through fewer gates [09:43] <omega_> because you effectively have (gatecount * contexts - ove= rhead) [09:43] <taaz> this is where it gets hard [09:43] <taaz> you have lots of tradoffs [09:43] <omega_> right [09:44] <omega_> btw, can you reprogram one context while another is 'run= ning' ? [09:44] <taaz> in general you always end up with memory bottlenecks [09:44] <taaz> yeah [09:44] <omega_> taaz: what if you create registers in part of the fpga, = and don't change those between contexts ? [09:44] <omega_> do they retain their state ? [09:46] <mwc> I have recompiled and still get segfault from gstmediadispl= ay :( [09:46] <taaz> but something to remember about that... reprogramming whil= e another context is running does give you 'virtual' hardware of sorts on= ly if one context takes enough cycles to process data such that the progr= amming of other contexts finishes before the data processing is done [09:46] <omega_> hmm, does -register lists a bunch of elements now? [09:46] <omega_> taaz: of course [09:46] Action: taaz is getting sleepy... wonders if he's making sense ;) [09:46] <mwc> Yes: loaded 64 plugins with 78 elements and 10 types [09:46] <taaz> hmm... the docs for this chip seem to be offline now [09:47] <omega_> taaz: what about my prev question about not modifying a = set of gates between contexts for data storage? [09:47] <omega_> mwc: finding a good set of debug flags for you [09:48] <taaz> omega_: yeah, this design allows you to have bits that are= constant accross contexts [09:48] <omega_> taaz: so can't that alleviate your memory bw pressure so= mewhat? [09:48] <taaz> how so? [09:49] <omega_> cause you don't have to write to memory, switch, and rea= d back in order to transfer data from one context to another [09:49] <omega_> you just leave it in those registers [09:49] <taaz> yeah, in place processing is useful for some algorithms [09:49] <omega_> interesting, gstplay is failing for me too [09:50] <omega_> taaz: do newer fpga's have better memory interfaces? [09:50] <taaz> better how? [09:50] <omega_> faster, lower latency, internal caches, etc. [09:51] <omega_> seems that a small (8-16KB) of local sram in the memory = interface would blow your bw pressure out of the water [09:51] <taaz> i havent been dealing with latest fpgas much... you could = go read all about virtex stuff at xilinx page though=20 [09:51] <taaz> they do have blocks of ram and stuff in them now i think [09:52] <omega_> I wouldn't know where to start <g> [09:52] <omega_> yup [09:52] steveb (ste...@21...) joined #gstreamer. [09:52] <omega_> what about taking a really big fpga and spatially partit= ioning it to get the same effects as multiple contexts? [09:53] <omega_> and when do you think these multi-context chips might hi= t the market? [09:53] <taaz> there's tradeoffs... as you increase logic you need to inc= rease routing resources. so you can pack more logic/mem into a certain a= rea with multiple contexts, but then you are sharing routing so you can r= un as much logic at once... tradeoffs galore [09:54] <omega_> right [09:54] Nick change: ajbusy -> ajmitch [09:54] <ajmitch> back [09:54] <taaz> i have no clue if these will ever be marketable [09:54] <omega_> if you're successful they will be [09:54] <omega_> I need to make sure I go to OLS, we should talk about th= is stuff more [09:55] Action: omega_ needs a fpga 101 course <g> [09:55] <omega_> luckily there are many talks that relate to rr's work, s= o I might be able to convince them to send me and maybe a few other guys [09:56] <taaz> you should decide before they fill up [09:56] <omega_> mwc: I'm building my local copy of CVS HEAD to see if I = can get it to work myself ;-( [09:56] <omega_> taaz: yeah, will decide by the end of the week [09:56] <omega_> rather, I should just sign up, since I'll go anyway.... [09:57] <taaz> omega_: issue with fpgas as coprocessors is that in the ge= neral case you always seem to end up optimizing for a specific architectu= re. same sort of problems VLIW has.. applications (at the bit level at l= east) are not compatible at all across different archs [09:58] <omega_> taaz: true, but one of the manuf's is gonna figure this = out first, and have a significant lead [09:58] <omega_> micros are the same way, and intel/x86 came out on top (= for better or worse) [09:59] <taaz> well.. you could argue that you could just use IA32 or wha= tever and let the processor itself detect dataflow and map onto its own p= rivate hidden configurable logic [10:00] <omega_> Crusoe [10:01] <taaz> i think you would need a different ISA for it to be practi= cal [10:01] <omega_> quite [10:01] <omega_> something more like alpha or ppc [10:01] <taaz> which i think people have researched [10:01] <omega_> definitely risc though [10:01] <omega_> unless it's like crusoe, with heavy software optimizing = cisc to risc [10:02] <omega_> personally, I think that silicon-threads are one of the = major futures [10:02] <taaz> err.. i think it needs to be bit oriented. ie, you can ad= dress and do ops on single bits [10:02] <omega_> like MAJC [10:02] <taaz> or ranges of bits [10:02] Action: ajmitch is browsgins newsgroups for g400 cards [10:02] <omega_> hmm, ok, you're coming at it from the fpga point of view [10:02] <taaz> which could be optimzed to single gates if possible and so= on [10:02] <omega_> with GA's, of course ;-) [10:03] <ajmitch> funny when you see people asking for MCSE stuff being d= irected to places like rute.sourceforge.net [10:03] <omega_> doh [10:03] <omega_> don't you mean rude.sourceforge.net ? [10:03] <ajmitch> umm, no ;) [10:03] <omega_> well, it is ;-) [10:03] <omega_> MCSE =3D Minesweeper Consultant and Solitaire Expert [10:04] <taaz> i'm not taking fpga point of view ;) i'm taking abstract = 'what am i trying to do' viewpoint. just need more expressive power in t= he ISA and languages to do all this [10:04] <omega_> yeah, ok [10:04] <omega_> bit thinking about individual bits *is* taking it from t= he hardware pov, whereas ISA's take it from at least the byte-oriented op= eration pov [10:04] <omega_> at least, all common ISA's [10:05] <omega_> er, s/^bit/but/ <g> [10:06] <taaz> but what if i only need 3 bits? no need to waste all thos= e other resources... if you want it to optimize properly would be nice to= just say you only need 3 bits for an op and it could map 2-1/3 of those = ops to one 8 bit op [10:06] <taaz> exploiting parallelizm automagically and all that [10:06] <omega_> and you think that compilers are gonna do this when they= can't even outdo =02me=02 right now? [10:07] Action: ajmitch wonders why he does not see a pile of commits to = gstreamer... ;) [10:07] Action: ajmitch is impatient ;) [10:07] <omega_> ajmitch: settling <g> [10:07] <ajmitch> omega_: still having fun with your asm? [10:07] <taaz> ajmitch: sumbit patches and commit volume will go up [10:07] <ajmitch> omega_: you just wanting fixes for 0.2.0 at the moment = i guess? ;) [10:07] <omega_> ajmitch: yup [10:08] <ajmitch> ok, my gramofile plugin can wait then ;) [10:08] <omega_> yeah, probably should [10:09] <ajmitch> omega_: ok, what should i attempt to look at? [10:09] <taaz> omega_: yeah, these are difficult problems to solve in gen= eral cases. which is why fpgas and similar technology are usually used f= or specialized things [10:09] <omega_> ajmitch: ? [10:10] <taaz> omega_: if your reorder thing has a hope of improving you = should organize a distributed effort to attack it [10:10] <omega_> heh, I'm going after small chunks [10:10] <ajmitch> omega_: what needs fixed? ;) [10:10] <omega_> I had only 6 instructions that resulted in 10 possible r= eorderings [10:11] <omega_> now, when I try to reorder my 80ish instruction yuv2rgb = routine, then I just need a good assembler interface and a day or two [10:11] <omega_> taaz: or were you planning on attacking *everything* thi= s way? <g> [10:12] <taaz> you just wait till i reveal my GA that works on the gstrea= mer code AST until an event system pops out and all bugs are fixed [10:13] <omega_> mu [10:13] Action: omega_ wants to get the event system done this century [10:13] <ajmitch> hmm, so you don't need my help before 0.2.0? that's gre= at! ;) [10:14] <omega_> um, no, we need to you keep testing [10:14] <ajmitch> ah, testing... ;) [10:14] Action: ajmitch tries to play with ganso_terminator.mpg [10:15] <ajmitch> nope, still crashes [10:15] <taaz> being slightly more serious... it would be interesting to = use GP to create a codec ;) just proved a much of low level blocks with = standard interfaces and plot them together with control parameters and do= in->enc->dec->out and measure difference and encoded size and stuff unti= l you get somethign you want [10:15] <ajmitch> i guess the caps nego stuff isn't resolved yet [10:15] <taaz> s/proved/povide/ [10:15] <taaz> or provide [10:16] <taaz> ok, i'm not typing too clearly ;) time for bed [10:16] <omega_> taaz: with or without a predefined bitstream format? [10:16] <taaz> omega_: i have no idea [10:17] <omega_> or are you thinking of just GA optimizing the known enco= der? [10:17] <omega_> cause that could be quite useful in motion prediction [10:17] <taaz> just something i thought would be interesting... see if it= starts to come up with patented algorithms [10:17] <omega_> but you're quickly getting into the field of machine-vis= ion there [10:17] <omega_> doh [10:17] <omega_> if so, then is tha GA itself illegal? <g> [10:18] <taaz> that's what Koza did... started making circuits to create = amplifiers with known response. circuits would come up that were patente= d and supposedly optimal [10:19] <taaz> s/up/out [10:19] <omega_> heh [10:19] <taaz> usually expired patents from way back... but still its int= eresting that this comes out of raw bit soup [10:19] <omega_> yup [10:20] <taaz> circuits are easier than video codecs i think... easier t= o mess with network topology [10:21] <omega_> yup [10:21] <omega_> but it would be interesting [10:21] <omega_> we know we can generate a "perfect" bitstream with enoug= h work [10:21] <taaz> oh really? [10:22] <omega_> so we train the GA by comparing total quality/bits to th= at perfect stream [10:22] <taaz> entropy measurement or something? [10:22] <omega_> no, just exhaustive search [10:22] <ajmitch> how can i see what is taking the CPU time when playing = back a file like an .ogg? [10:22] <omega_> ajmitch: you don't, right now [10:22] <omega_> I realized that we need a mechanism for tracking that [10:22] <ajmitch> just wondering whether it is the fault of vorbisdec (wh= ich is my responsibility, apparantly) [10:23] <omega_> libvorbis is not optimized in the least [10:23] <omega_> so it wouldn't surprised me [10:23] <ajmitch> yeah, but xmms has a lower cpu usage, and it uses the s= ame lib, afaik [10:23] <ajmitch> i want to get this more optimised than xmms ;) [10:24] <omega_> taaz: the idea is to find the absolute best match, but w= eight that against the cost of the motion vector relative to the final en= coded dct coeffs [10:24] <omega_> ajmitch: dunno [10:24] <omega_> we'll have to look at it once we have the instrumentatio= n for it [10:25] <ajmitch> xmms shows separate threads tho with different cpu usag= es [10:25] <omega_> yup, one is decoder thread, one is gui thread [10:25] <omega_> afaik [10:25] <omega_> rather, more than that [10:25] <omega_> reader, decoder, writer, gui [10:26] <taaz> omega_: seems like a stream format where everything is var= iable would allow for maximum codec flexability and also fun with GAs to = find optimal use for a particular stream [10:27] Action: mwc must go home. Will try gstmediadisplay tomorrow... [10:27] <ajmitch> cya mwc [10:27] <omega_> mwc: yeah, I'll try to be ready for you then <g> [10:27] <mwc> OK bye! [10:27] mwc (ma...@ly...) left irc: [x]chat [10:27] <omega_> taaz: you mean like a bare minimum 'language' for the GA= to describe what it 'hears' with ? [10:28] <taaz> i have no idea ;) [10:28] <omega_> I've thought about that a little, I don't see how you co= uld do it without very specific audio knowledge, which would quickly impo= se higher syntax requirements [10:29] <omega_> question is, how do you *decode* that? [10:29] <taaz> i haven't really thought much beyond this about it. i rea= lly don't know enough about audio/video codecs to do things intelligently [10:29] <omega_> you have to have some minimum language to start from [10:29] <omega_> have you seen the thing NASA did that taught aliens all = about us starting from 1 + 1 =3D 2 ? [10:30] <taaz> yeah [10:30] <omega_> you have to do that, somehow [10:32] <omega_> start with a minimum language, then actually send langua= ge description inline with the data, to construct higher-level syntax ove= r time [10:32] <omega_> with enough time and data set, you could eventually redu= ce that to a single bitstream that uses the optimal coding, and ditch the= pre-learning stuff [10:32] <ajmitch> hehe, and i thought gstreamer was a little excessive wi= th cpu time - try using nautilus ;) [10:32] <omega_> but a significant part of the cost function has to be en= code time [10:33] <omega_> ajmitch: I did ;-( [10:33] <taaz> omega_: yeah... but i've always thought for some things it= would be very cost (as in $$) effective to be loose with the enocde time [10:34] <omega_> taaz: yeah, but you realize that the vast majority of en= coding these days is done live? [10:34] <taaz> like for instance dvd movies... a studio might take a week= with a supercomputer to master a dvd if they could fit 50% more stuff on= it [10:34] <omega_> taaz: yes, but that's very very rare [10:34] <omega_> and yes, you'd think they'd do that, but no such softwar= e exists afaict [10:35] <taaz> well, it would be nice to allow the impossible ;) [10:35] <omega_> yup [10:35] <taaz> this css stuff is absurd [10:35] <omega_> yup [10:35] sienap (sy...@ip...) joined #gstreamer. [10:35] <sienap> Hi all [10:36] <taaz> grr... i don't know how best to use it [10:36] <taaz> the auth stuff is ok legally? [10:36] <omega_> taaz: ? [10:36] <sienap> hej taaz/omega [10:36] <omega_> yo [10:37] <taaz> libcss has the auth code too but if you need libcss to tel= l if you need other parts of libcss that's just silly [10:37] <ajmitch> hey sienap [10:37] <taaz> as far as keeping crypto code out unless you need it [10:37] <sienap> ajmitch! [10:37] <omega_> oh, no, you can tell just by looking at the stream from = the filesystem if you need decss or not [10:37] <taaz> oh? [10:37] <omega_> check to see if it parses as a valid mpeg2 stream, if no= t, start decrypting [10:38] <omega_> beware that css doesn't always make it look screwed [10:38] <omega_> you can sometimes see useful bits of video in mpeg2dec i= f you play an encrypted stream [10:38] <omega_> I should do a bit-level diff between encrypted and not [10:39] <taaz> err... i'm looking at this ioctl that tells you if its enc= rypted or not [10:39] <omega_> yeah, you can do that to [10:39] <omega_> er, too [10:39] <taaz> but it would be nice to libraryize all the dvd code [10:39] <omega_> yup [10:39] <omega_> that's what libcss is supposed to do, right? [10:39] <taaz> which means 2 libs [10:40] <omega_> yup [10:40] <omega_> libnav and libcss [10:40] <taaz> one just to have checks if you need the one with the crypt= o [10:40] <omega_> and libudf maybe [10:40] <taaz> and runtime link in the crypto one if its available so no = linking issues [10:40] <taaz> stupid mess [10:40] <omega_> right [10:41] <omega_> of course, strictly speaking, even with the DMCA, libcss= is 100% legal [10:41] <taaz> oh? [10:42] <omega_> yes, it falls under the reverse-engineering for the purp= oses of interoperability clause [10:42] <omega_> no matter what the RIAA claims, it does [10:42] <taaz> do you have a pointer to a judge decision transcripts that= agree with this? [10:43] <omega_> no, because no judge has figured that out yet [10:43] <taaz> well, i fall under that category of people that cant affor= d to deal with a court case [10:43] <omega_> true [10:44] <omega_> though I wonder why the RIAA hasn't gone after oms? [10:44] <taaz> ahh... you have found our secret... we made it too hard fo= r them to get running ;) [10:44] <omega_> point [10:44] <omega_> no, I think it's because they know they can't win that [10:45] <omega_> what are they going to do, go to a judge and say "these = people are being evil and communist and illegal because they're distribut= ing a player that lets people play the DVD's they bought!" [10:45] <omega_> no, they go after "DeCSS" because it carries connotation= s of being used for piracy [10:46] <omega_> and they might be able to win that one, if they can impo= rt enough wool to pull over the judges' eyes [10:47] <taaz> piracy arguements are dumb. pirates just copy dvd bits wh= olesale, regardless of encryption or not [10:47] <omega_> tell that to the judge [10:47] <taaz> they did [10:47] <omega_> it seems to me that they didn't tell him enough [10:47] <omega_> I really should go over the court transcripts and call u= p Robin and suggest a few things [10:47] <omega_> they're not pushing the points that are so obvious if th= ey would just explain them [10:48] <omega_> like the fact that decryption has *nothing* to do with b= ulk piracy [10:48] <taaz> pirating is an issue... lots of people ask in #livid and e= lsewhere 'how can rip my dvd to divx'? and programs like drip are out the= re (and on sf no less...).. i'm sure the winders world is rampent with cr= appy rippers of all sorts [10:49] <omega_> yup [10:49] <omega_> but that's not what RIAA tells the judge [10:49] <ajmitch> someone will use gstreamer to make a ripper, i'm sure o= f it [10:49] <omega_> they not-so-subtly imply that decss allows bulk piracy [10:49] <omega_> which is done by way of pressing, not decrypting [10:49] <omega_> ajmitch: dvdsrc ! mpeg2dec ! 3ivxenc ! disksink [10:50] <ajmitch> yeah, like that ;) [10:50] <taaz> but that's lame! i like the dvd features. [10:50] <omega_> mu [10:50] <omega_> they should be done with a real codec family using xml [10:51] <omega_> did the opendvd stuff ever get anywhere near usefulness? [10:51] <taaz> oh.. vektor had a really great reason users should be able= to rip dvds. he's working on improving the quality ;) [10:51] <omega_> pfff [10:51] <omega_> improving over what's already on the disc? [10:51] <taaz> doing deinterlacing and pulldown (or whatever that is) opt= imization for the target format [10:52] <omega_> yeah [10:52] <omega_> well [10:52] <omega_> 1080p48 is the long-term goal [10:52] <taaz> he works in video industry now so i'll just trust he knows= what he's talking about ;) [10:52] <omega_> at which point all the video will be both sourced and vi= ewed in the same format [10:53] <taaz> what is 1080p48? [10:53] <omega_> 1920x1080, progressive, 48 frames/sec [10:54] <omega_> 1080p60 is one of the atsc specs, I think, but people ar= e realizing that 60fps is more than people need, and a lot of wasted band= width [10:54] <taaz> make sure xiph people target that format [10:54] <omega_> not to mention not being in line with film... [10:54] <omega_> yup [10:54] <omega_> they're already targetting the audio half, which is arbi= trary uncoupled channels used to encode ambisonics [10:55] <omega_> no one does ambisonics with it yet because it's still ju= st 2-channel and isn't flexible enough to handle more just yet [10:55] <omega_> but the format is, so... [10:55] <omega_> thing is, I think tarkin has a chance, because MPEG carr= ies so many of these old things with it [10:56] <ajmitch> problem is, how to build a codec to get around the pate= nts? [10:56] <taaz> if they can get throught the patent mine field [10:56] <omega_> and I'm of the opinion that if we target something like = 1080p48 in a few years, general-purpose computing hardware will be up to = that spec and beyond by then [10:56] <omega_> right. [10:56] <omega_> I think the only way that's gonna happen is if there's p= atent reform in the meantime [10:56] <ajmitch> damn i hate patents on things like that [10:56] <omega_> if the EU gets on with it, the US might get smacked arou= nd hard enough to fix it [10:57] <taaz> ok, 5am. birds are chirping outside my window. time for = bed. [10:57] Nick change: taaz -> taazzzz [10:57] <omega_> yeah, 2am here, I should sleep [10:57] <ajmitch> hehe [10:57] <omega_> l8r [10:57] <ajmitch> 9pm here ;) [10:58] omega_ (om...@om...) left irc: [x]chat [11:07] steveb (ste...@21...) left irc: Read error = to steveb[212186169160.chello.com]: Connection reset by peer [11:42] sienap (sy...@ip...) left irc: Ping timeout for = sienap[ipc379c0d7.dial.wxs.nl] [12:09] thomas (th...@21...) joined #gstreamer. [12:11] <ajmitch> hi thomas [12:13] <thomas> hi ! [12:13] <thomas> I wonder why I'm not getting my password hash mail from = sourceforge ;( [12:13] <thomas> I *really* need to keep working on gstreamer [12:14] Action: ajmitch is going to try & compile gstreamer using a gcc 3= .0 snapshot [12:16] <thomas> no practical reason for that I suppose ? :- [12:16] <thomas> ) [12:17] <ajmitch> nope, none whatsoever ;) [12:17] <ajmitch> apart from just testing in different ways ;) [12:29] Action: ajmitch is getting a *lot* more warnings now on compilati= on [12:29] <ajmitch> well, for some stuff anyway, like the mpeg1 encoder [12:29] <ajmitch> the rest is ok [12:43] Nick change: ajmitch -> ajzzzz [12:58] richardb (ri...@ix...) joined #gstreamer. [13:15] <richardb> Woo. Anyone awake? [13:25] <thomas> somewhat [14:03] <thomas> richardb: does osssrc even work ? [14:06] <richardb> Never tried it... [14:37] Action: thomas really wonders why osssrc insists on using 44101 K= Hz when asking for 44100, and only at that frequency [14:47] <richardb> thomas: it may be the hardware (or driver) doing somet= hing screwy. [14:49] <thomas> hmm, that would be weird. I never encountered something= like this before then. [15:05] <thomas> hmmm, it's the ioctl call that's screwing things up. [15:13] <richardb> Hmm, why are the arts checks commented out in configur= e.in, I wonder. [15:13] <richardb> (line 528) [15:14] Action: richardb greps [15:15] <richardb> and the XMMS check [15:16] <thomas> I know arts used a weird thing to check which failed on = my machine causing configure to stop... [15:16] <thomas> ... I think wtay changed that because of that [15:16] <richardb> Hmm. [15:17] <richardb> The commented out bit just uses AC_CHECK_LIB to check [15:17] <richardb> That might fail, and might not be a sufficient check, = but should cause configure to stop [15:18] <richardb> It definitely failed while running configure, not make= ? [15:18] Action: richardb wishes people would put comments in explaining w= hy they comment things out. [15:18] <thomas> richardb: I'm not 100% sure, but almost [15:19] <richardb> (Not that I'm innocent of such behaviour, ofc0 [15:19] <thomas> the oss plugin is a mess ! [15:19] Action: richardb sigh [15:19] Action: thomas wishes people would include pointers to docs where= they based themselves upon [15:23] <richardb> thomas: I've a plan for merging the configure.ac and c= onfigure.in, but I'd like someone to confirm that it's reasonably sane. [15:24] <richardb> Do you have a reasonable understanding of how autoconf= works? [15:24] <richardb> (Doesn't need to be highly detailed understanding, jus= t reasonable) [15:26] <thomas> richardb: hmm, I don't understand it too much [15:26] <thomas> richardb: but I've tried hacking stuff in before [15:27] <richardb> Hmm. [15:27] <thomas> richardb: and I understand enough of it to see that all = of the custom detects in the current thing we have are totally different = ;) [15:27] <richardb> Indeed: that mess needs sorting out. [15:27] <thomas> richardb: maybe explaining what you want to do will help= you think of stuff you didn't think of before [15:28] <richardb> But it'll be lots more work to do it in configure.in a= nd configure.ac. [15:28] <richardb> So I want to merge those somehow first. [15:28] <richardb> The diffs between them are only 58 lines (unified diff= ). [15:29] <richardb> I'm thinking of putting the appropriate lines to gener= ate the configure.ac into the configure.in in special comments... [15:29] <richardb> ... and having a perl script to generate the configure= .ac from the configure.in and its comment. [15:29] <richardb> Then I have the two sources together. [15:30] <richardb> Does that sound too horrible? [15:31] <thomas> richardb: no, actually that sounds sane and clean [15:31] <thomas> richardb: anything that makes stuff better protected aga= inst developer laziness is ok IMO [15:31] <richardb> Good. It's probably not as clean as it sounds, but it= should be cleaner than the current situation... ;-) [15:31] <thomas> richardb: as long as when you test it, you make sure tha= t the files generated match the ones they should match...=20 [15:32] <richardb> Yes, definitely. [15:32] <thomas> ... then it looks ok ! [15:33] Action: richardb codes. ;-) [15:48] dobey (dobey@141.154.95.104) joined #gstreamer. [15:49] <dobey> bleh [15:49] <thomas> dobey: still down ? [15:50] <dobey> thomas: feeling muchos better today [15:50] <dobey> hadess just isn't here [15:51] <thomas> sorry, you'll have to make do with us ;) [15:51] <dobey> heh, none of you have macs ;-P [15:52] <thomas> not anymore, no *snif* [15:53] <dobey> i need to know where the hell he got 2.4.4 from [15:53] <dobey> blah, i need a terminal that doesn't suck too [15:56] <thomas> dobey: isn't it included in the new yellowdog ? [15:57] <dobey> eww [15:57] <dobey> *smack* [15:57] <dobey> i need a tarball [15:57] <dobey> 2.2.19 sucks ass [15:57] <thomas> well, that's also a nice feature to have in a kernel, bu= t I usually turn that off due to my hemorroids [15:58] <dobey> uh [15:58] <dobey> no [15:58] <dobey> it couldn't power my hard drive back up properly [15:58] <thomas> sorry ;) [15:58] <dobey> and the RTC doesn't work [16:20] <richardb> thomas: done my configure changes. [16:21] <richardb> Can you cvs update and check it works for you ... ;-0 [16:21] <richardb> Hmm: mailing lists are either broken or being _very_ s= low. [16:30] <thomas> richardb: sorry, I'm back. I'll check. [16:32] <thomas> richardb: what's the difference anyway between configure= .in and configure.ac ? [16:32] <richardb> I sent an explanatory email... [16:33] <richardb> But sourceforge doesn't seem to be in a mail deliverin= g mood. [16:33] <richardb> :( [16:33] <richardb> configure.ac is for autoconf version 2.50 [16:33] <richardb> configure.in is for autoconf version 2.13, or earlier [16:33] <richardb> Some macros obseleted, some new macros, basically. [16:33] <thomas> make started, looks good for now [16:34] <richardb> Good [16:46] <thomas> make went ok, congratulations [17:57] <thomas> =20 [18:31] steveb-away (st...@no...) left irc: Ping timeout fo= r steveb-away[node1ee04.a2000.nl] [18:31] steveb-away (st...@no...) joined #gstreamer. [18:36] Nick change: steveb-away -> steveb [18:36] marcf (ma...@be...) left irc: Read error to marcf[becomm.com= ]: Connection reset by peer [18:37] thomas (th...@21...) left irc: [x]chat [18:39] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee04.a2000.nl] [18:40] <richardb> who# [18:40] richardb (ri...@ix...) left irc: Read error to ric= hardb[ixion.tartarus.org]: EOF from client [18:40] richardb_ (ri...@ix...) joined #gstreamer. [18:41] Nick change: richardb_ -> richardb [18:42] Nick change: taazzzz -> taaz [18:42] steveb (st...@no...) joined #gstreamer. [18:44] <dobey> whee [18:58] Zeenix (pro...@ho...) joined #gstreamer. [18:58] heroines (root@205.229.113.88) joined #gstreamer. [18:59] <Zeenix> hi wtay [19:01] <Zeenix> any body home ? [19:02] <dobey> no, i'm at work [19:03] Nick change: wtay-zZz -> wtay [19:03] <wtay> yo [19:03] <Zeenix> hi wtay [19:04] <wtay> hi [19:04] <Zeenix> wtay: can you send me the source code of that app you ma= de for me yesterday ? [19:04] <wtay> Zeenix: heh, I deleted it [19:05] <Zeenix> wtay: i have gstreamer( & the required rpms for gstreame= r rpm ) installed, what i need to use gsm ? [19:06] <wtay> Zeenix: you'll need the cvs version for gsm [19:06] <wtay> and libgsm of course [19:07] <Zeenix> from where ? [19:08] <wtay> no idea, I use debian. you should try rpmfind.net and look= for libgsm [19:09] ChiefHighwater (pa...@te...) joined #gstreamer. [19:09] <Zeenix> wtay: how'll i learn to use it ? [19:10] <wtay> Zeenix: you should read the Gtk manual first and then the = gstreamer manual/API docs [19:10] <wtay> hello ChiefHighwater [19:11] <Zeenix> i searched for the gsm in the gstreamer manual, but fail= ed to find any thing about gsm=20 [19:12] <wtay> Zeenix: gsm is just another plugin, just like mpg123, you = handle it in the same way [19:12] Uraeus (csc...@c2...) joined #gstreamer. [19:13] <Uraeus> hi [19:13] <wtay> hi Uraeus [19:13] <dobey> hey Uraeus [19:13] <Uraeus> hi wtay; I just updated the roadmap with the latest chan= ges from Erik [19:14] <wtay> Uraeus: good, will read soon [19:14] <Zeenix> wtay: can gsm tackle 5Khz/short/mono ? [19:15] <wtay> Zeenix: yes [19:15] <Zeenix> 5Khz/short/mono =3D=3D 10KB/s ?=20 [19:16] <Zeenix> is it so ? [19:16] <taaz> Uraeus: you dont have dvdsrc on the roadmap [19:16] <wtay> yes [19:16] <Zeenix> wtay: was that 'yes' for me ? [19:17] <dobey> yes [19:17] <wtay> Zeenix: yes :) [19:17] <wtay> dobey: not realy <g> [19:17] <dobey> heh [19:17] <Uraeus> taaz: hmm, your right, what is the status and who is the= maintainer? [19:17] <dobey> ? [19:18] <Zeenix> wtay: when i'll have my app successfully completted, i'l= l add your name in credit list at the top [19:18] <dobey> heh [19:18] <wtay> Zeenix: that's not needed, really :) [19:18] <dobey> Uraeus: would encompass count as a front end? [19:18] <dobey> heh [19:19] <taaz> Uraeus: well, basic test failed for me. i'll need to get = it working for OMS but that doesnt mean i'm volunteering to be labeled it= s maintainer ;) [19:19] <Zeenix> when i run gstmediaplay, i get the following message on = the xterm: [19:20] <Zeenix> Xlib: extension XVideo missing on display :0.0 [19:20] <Zeenix> & [19:20] <dobey> yeah [19:20] <dobey> so do i [19:20] <dobey> heh [19:20] <Uraeus> dobey: well, for encompass to fit in the frontend catego= ry you need to make gtkhtml support plugins and make a plugin :) [19:20] <Zeenix> Xlib: extension XINERAMA missing on display :0.0 [19:21] <dobey> Uraeus: gtkhtml supports plugins, bonobo-media has some c= ontrols, and the htmlview i'm writing will be able to load them shortly [19:21] <dobey> heh [19:21] <Uraeus> taaz: sorry, I missed that..you said you want to be the = dvdsrc maintainer? [19:21] <wtay> brb [19:21] <dobey> Uraeus: plus i'm gonna make a neato radio toolbar! [19:21] <dobey> heh [19:21] <Zeenix> any clue ? [19:22] <taaz> Uraeus: I'm trying to avoid being held responsible for DVD= things since I live in US and dont want to be sued [19:22] <Uraeus> dobey: I think I still have the code for the iCast panel= applet if you want it [19:22] <dobey> icast panel applet? [19:22] <dobey> damn [19:22] <Uraeus> dobey: yes [19:22] <dobey> i need a keyboard? [19:22] <dobey> err [19:22] <dobey> -? [19:22] <dobey> Uraeus: what's it do? [19:22] <Uraeus> no a mouse [19:23] <Uraeus> dobey: it lists the iCast streaming mp3 channels [19:23] <dobey> oh [19:23] <Uraeus> and plays the one you choose in xmms [19:23] <dobey> bleh [19:24] <dobey> no, i need a keyboard, it took me like 5 hits to make the= l appear [19:24] <Uraeus> dobey: I have a spare one with norwegian keys if you wan= t to type =F8=E6=E5 easy :) [19:25] <Zeenix> ? [19:25] <dobey> Uraeus: is it for a powerbook? [19:25] <Uraeus> Zeenix: are you using XFree? [19:25] <Uraeus> dobey: no standard ps/2 for pc :( [19:25] <dobey> Uraeus: i need a titanium kb ;-P [19:26] <Zeenix> Uraeus: yes i am using XFree [19:27] <Zeenix> do you mean XFree86 ? [19:27] <Uraeus> yes [19:27] <wtay> back [19:27] <Uraeus> Zeenix: well do you have Xinerama support congfigured? [19:27] <wtay> Zeenix: >=3D 4.0.1? [19:28] <Zeenix> wtay: XFree86 4.0.1, you mean ? [19:28] <wtay> Zeenix: yes [19:30] <Uraeus> wtay: I thought I should add a LICENSE file to each of t= he plugins, should I send em to you or can I get CVS access? [19:30] <wtay> Uraeus: are you going to behave yourself with this new pow= er? <g> [19:31] Zeenix (pro...@ho...) left irc: Ping timeout= for Zeenix[host-91.netzone.net.pk] [19:32] <Uraeus> wtay: no promises made :) [19:32] Zeenix (pro...@ho...) joined #gstreamer. [19:33] <wtay> Uraeus: I'll add you then :) [19:34] <Uraeus> great [19:34] <wtay> Uraeus: I added you. you'll need to checkout a new cvs cop= y thouhg, with your own userid [19:34] <Uraeus> ok [19:35] <wtay> I gotta take a bath [19:35] Nick change: wtay -> wtay-away [19:36] Zeenix (pro...@ho...) left irc:=20 [19:38] Zeenix (pro...@ho...) joined #gstreamer. [19:39] <dobey> Uraeus: heh, wait till you see the applet i'm gonna write [19:39] <dobey> you'll like it ;-P [19:39] Zeenix (pro...@ho...) left irc:=20 [19:41] Zeenix (pro...@ho...) joined #gstreamer. [19:42] Uraeus (csc...@c2...) got netsplit. [19:42] big_T (th...@fy...) got netsplit. [19:42] heroines (root@205.229.113.88) got netsplit. [19:42] steveb (st...@no...) got netsplit. [19:43] <ChiefHighwater> ok, so who netsplit? [19:43] <ChiefHighwater> must be me [19:44] big_T (th...@fy...) returned to #gstreamer. [19:44] steveb (st...@no...) returned to #gstreamer. [19:44] heroines (root@205.229.113.88) returned to #gstreamer. [19:44] Uraeus (csc...@c2...) returned to #gstreamer. [19:44] Zeenix (pro...@ho...) left irc: Ping timeout= for Zeenix[host-87.netzone.net.pk] [19:46] Uraeus (csc...@c2...) got netsplit. [19:46] big_T (th...@fy...) got netsplit. [19:46] heroines (root@205.229.113.88) got netsplit. [19:46] steveb (st...@no...) got netsplit. [19:46] <dobey> eek! [19:49] <ChiefHighwater> another one? [19:49] steveb (st...@no...) returned to #gstreamer. [19:49] Uraeus (csc...@c2...) returned to #gstreamer. [19:49] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee04.a2000.nl] [19:49] <dobey> glah [19:50] <ChiefHighwater> they must be reconfiguring at openprojects [19:51] big_T (th...@fy...) returned to #gstreamer. [19:51] heroines (root@205.229.113.88) returned to #gstreamer. [19:51] <dobey> oh well [19:52] <dobey> Uraeus: heh, so would encompass count? [19:52] <Uraeus> dobey: show me the code and it counts :) [19:53] <dobey> Uraeus: heh [19:53] steveb (st...@no...) joined #gstreamer. [19:57] <steveb> i have a cracker of a bug here [19:59] <steveb> every *second* call to my ladspa chain func is given a r= eference to the upstream element instead of the ladspa element [20:03] Nick change: wtay-away -> wtay [20:04] <wtay> steveb: check your chain function setup wrt the padss [20:05] <wtay> who's having problems with gstmediaplayer? [20:05] <Uraeus> wtay: my first plan was to add a LICENSE file to each pl= ugin clearly stating its License, but i realise this will be wrong [20:06] <wtay> Uraeus: why? [20:06] <dobey> hrmm [20:07] <dobey> wtay: i probably would if i had the time to patch automak= e, and get it built [20:07] <Uraeus> wtay: cause the plugin code is usually LGPL, but when we= link it to something GPL like the cdparania the binary is GPL, so I thin= k maybe I instead should add info saying that our plugins are LGPL but th= is and this plugin will be GPL due to the underlying library [20:08] <dobey> heh [20:08] <dobey> that's the great thing with bonobo [20:08] <wtay> Uraeus: makes sense [20:08] <dobey> i can make the htmlview LGPL, and the shell GPL [20:08] <dobey> heh [20:11] <Uraeus> I mailed RMS in regard to the debate on Gnotices about i= t being illegal for Sun etc. to ship GNOME due to their libc license not = being GPL compatible, he said that it was correct that it is illegal, but= that he will fix it in GPL 2.1 [20:12] <dobey> heh [20:13] <Uraeus> wonder how he will solve it, since he probably will not = want to open a loophole in the GPL for people to make 'GPL' software on t= op of proprietary libs [20:15] <dobey> heh [20:15] <Uraeus> which would mean that people can just take this or that = GPL app, and add functionality in a library instead of the app itself [20:15] <dobey> there's already a loophole like that [20:15] <Uraeus> dobey: the component system one you mean? [20:15] <dobey> Uraeus: well, yeah, the asp crap [20:17] <Uraeus> hmm, well the asp loophole is small, since it contains i= tself to one entity [20:17] <dobey> well [20:18] <dobey> i don't really care [20:18] <dobey> heh [20:18] <Uraeus> anyone know what this mean: gnapster: error while loadin= g shared libraries: gnapster: undefined symbol: stat [20:18] <dobey> yeah [20:18] <dobey> you are running rh71? [20:18] <Uraeus> yes [20:19] <Uraeus> and? [20:19] <dobey> and gnapster is broke there [20:19] <dobey> and update as soon as it's fixed [20:19] <dobey> ;-) [20:19] <dobey> and my headache is returning to haunt me :-/ [20:19] <Uraeus> dobey: well you are at the source, when will it be fixed= ?=20 [20:20] <dobey> Uraeus: it's not my job, i have no idea, all i know is it= 's known and i think someone is working on it as we speak [20:20] <dobey> my job right now is fucking mandrake [20:21] Action: Uraeus decides to test lopster [20:21] <dobey> heh [20:21] <dobey> that was ugly [20:22] <Uraeus> wtay: gstreamer fails to compile for me [20:22] <Uraeus> passthrough.c:21:39: ../../../include/gstaudio.h: No suc= h file or directory [20:23] <wtay> Uraeus: I mailed it to thomas allready [20:23] <dobey> heh [20:23] <Uraeus> the bug or the solution? [20:23] <wtay> Uraeus: he forgot to commit that file [20:24] <Uraeus> ah [20:24] <wtay> I'll commit a fix to not compile that plugin [20:25] <Uraeus> np, I can wait [20:25] <wtay> ok [20:25] <wtay> Uraeus: you can remove passthrough from plugins/filters/Ma= kefile.am if you want [20:25] <Uraeus> wtay: is the 0.2.0 release planed this week or next? [20:26] <dobey> whee [20:26] <Uraeus> dobey ? [20:26] <dobey> Uraeus: gnucash is building [20:27] <wtay> Uraeus: this friday is the goal [20:28] <Uraeus> wtay: ok, do you want to me write a release announcement= or is someone already on that? [20:28] <Uraeus> dobey: lopster seems really nice to bad it is GTK+ only = or I would suggest Ximian switching [20:29] <dobey> Uraeus: well, heh, i think i am getting control of gnome-= napster anyway [20:30] <Uraeus> dobey: well if you want to hacke a napster client I thin= k gnomifying lopster is better choice [20:30] <dobey> Uraeus: i don't like it's ui [20:30] <dobey> heh [20:31] <Uraeus> dobey: but functionilitywise it is kicking gnapster and = gnome-napster ass [20:31] <dobey> uh [20:31] <dobey> yeah [20:31] <dobey> so? [20:31] <dobey> i haven't hacked it yet [20:31] <dobey> :-P [20:31] <Uraeus> heh :) [20:33] <Uraeus> dobey: ok, but since I want a top-notch GNOME napster cl= ient, please just cutn'paste non-gui code from Lopster ok :) [20:33] <dobey> haha [20:34] <dobey> how about i just make a libgnapster or something? [20:34] <Uraeus> hmm, actually I think... [truncated message content] |
From: <wim...@ch...> - 2001-06-01 04:32:17
|
[06:45] Nick change: ajmitch -> aj_busy [07:24] aj_busy (aj...@p2...) left irc: Ping timeout = for aj_busy[p20-max2.dun.ihug.co.nz] [07:47] _gst_newt_ joined #gstreamer. [07:50] aj_busy (aj...@p1...) joined #gstreamer. [08:07] steveb (st...@no...) joined #gstreamer. [08:34] aj_busy (aj...@p1...) left irc: Ping timeout = for aj_busy[p10-max7.dun.ihug.co.nz] [08:35] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee02.a2000.nl] [08:36] steveb (st...@no...) joined #gstreamer. [08:40] steveb (st...@no...) left irc: [x]chat [08:49] Nick change: taaz -> taazzzz [09:00] aj_busy (aj...@p2...) joined #gstreamer. [09:02] mwc (ma...@ly...) left irc: [x]chat [09:32] walken (fo...@c1...) joined #gstreame= r. [10:40] thomas (th...@21...) joined #gstreamer. [11:03] Nick change: aj_busy -> ajmitch [11:03] <ajmitch> hey [11:06] <walken> hi [11:06] <walken> I'm not staying long [11:06] <ajmitch> k [11:07] <walken> well I'm leaving actually :) [11:07] <walken> cya :) [11:07] walken (fo...@c1...) left irc: l8r [11:50] ajmitch (aj...@p2...) left irc: Ping timeout = for ajmitch[p20-max3.dun.ihug.co.nz] [11:51] ajmitch (aj...@p2...) joined #gstreamer. [12:33] thomas (th...@21...) left irc: Ping ti= meout for thomas[212-100-172-175.adsl.easynet.be] [12:45] thomas (th...@21...) joined #gstreamer. [13:00] <thomas> hmmm... gstmediaplay still crashes whatever you try to p= lay ? [13:02] BBB (BB...@uc...) joined #gstreamer. [13:07] <ajmitch> hi thomas [13:07] <ajmitch> thomas: what was that? [13:07] <thomas> ajmitch: well, gstmediaplay still doesn't work anymore f= or me [13:07] <thomas> it just crashes whatever I try to play [13:08] <thomas> I was wondering if that is normal or not ;) [13:08] <ajmitch> thomas: since when? [13:08] <thomas> ajmitch three days [13:09] <BBB> hmm...... [13:09] <BBB> who added the check for automake? [13:09] <richardb> me [13:09] <BBB> I have two automakes: automake and automake_gst (patched) [13:10] <richardb> What's the check saying? [13:10] <BBB> kind of annoying because it tries to patch the old one... := ) [13:10] <BBB> "unable to check whether patch worked or not" [13:10] <BBB> :) [13:10] <richardb> It should just be warning, waiting a bit, and then con= tinuting [13:10] <BBB> I did change automake into automake_gst everywhere in autog= en.sh [13:10] <ajmitch> ok, gstmediaplay has been working for for me all day [13:11] <richardb> It would probably be easier for you if you just called= automake_gst automake, and put it earlier in your path. [13:11] <richardb> The patch should never cause a problem... [13:11] <BBB> richardb: I had an old ./configure which works too - so I c= an still build it :) [13:11] <BBB> I'll probably install the new automake [13:11] <richardb> automake also gets called from makefiles, so you might= need to change it there... [13:11] <BBB> hmm...aha [13:11] <BBB> that makes sense [13:12] <BBB> I'll go and install a new automake [13:12] <richardb> automake 1.4g or whatever has other problems... [13:12] <richardb> being a development version. [13:12] <richardb> I've not actually managed to compile with it. [13:12] <BBB> hm... [13:12] <BBB> 1.4f? [13:12] <richardb> ... though I'm getting there. [13:12] <richardb> 1.4g is cvs, 1.4f is release [13:12] <BBB> ok.... [13:13] <BBB> I'll try 1.4f then [13:13] <richardb> let me know if you get it to work. [13:13] <BBB> but first, let's play and see if gstmediaplay works already= :) [13:13] <BBB> <g> [13:13] <BBB> gstreamer won't compile ;) [13:13] <ajmitch> thomas: nope, gstmediaplay still works for me [13:13] <richardb> 1.4e onwards have the patch to fix the problem we had,= but they may fail for other reasons. [13:13] <thomas> ajmitch: hm, i'll check some more then [13:14] <ajmitch> thomas: where's it crashing? [13:15] <BBB> gstinfo.c:456: `Dl_info' undeclared (first use in this func= tion) [13:15] <BBB> where is the struct DL_info declared? [13:15] <BBB> it refuses to compile : [13:15] <BBB> :) [13:16] <thomas> ajmitch: as soon as I select something for play and it s= tarts playing, I get a gnome dialog window showing the seg fault. no cor= e though. [13:16] <ajmitch> BBB: gst/gstinfo.c [13:16] <ajmitch> thomas: run from within gdb then ;) [13:17] <ajmitch> BBB: oh, that's where it needs it... [13:17] <thomas> BBB: looks like a typo, no ? [13:18] <thomas> 0x40e36a4d in gst_static_autoplug_to_render (autoplug=3D= 0x81069d8,=20 [13:18] <thomas> srccaps=3D0x80772d8, target=3D0x809f8c0, args=3D0xbf= e17b78) [13:18] <thomas> at gststaticautoplugrender.c:279 [13:18] <thomas> 279 pad =3D GST_PAD_REALIZE (gst_element_get_pad= _list (targetelement)->data); [13:18] <thomas> ajmitch: that's where it crashes [13:18] <thomas> ajmitch: what are you playing with it ? [13:18] <ajmitch> thomas: mpeg1, mp3, ogg [13:18] <ajmitch> thomas: i think i got a crash like that when i removed = xvideosink... [13:20] <thomas> hmm, I still have xvideosink [13:23] <thomas> i'll try on a different computer [13:23] ajmitch (aj...@p2...) left irc: Ping timeout = for ajmitch[p24-max5.dun.ihug.co.nz] [13:23] Action: BBB wonders what Gl_info is and where to find it [13:24] ajmitch (aj...@p3...) joined #gstreamer. [13:24] <ajmitch> crappy isp [13:26] <BBB> :) [13:27] <ajmitch> thomas: all i can recommend at the moment is that you e= nter it in the SF bug tracker (that's what i was told to do for my crashe= s ;) ) [13:29] <BBB> does windows support FILE *...; thingies? [13:30] <BBB> or are they windoze only [13:30] <BBB> uhm [13:30] <BBB> unix only [13:30] <ajmitch> umm, i think it should [13:30] <BBB> nice [13:31] <ajmitch> i dunno who would want to develop for windoze, but i th= ink it does [13:31] Action: BBB might have to make a pdf-to-txt thingy for windows fo= r a company [13:31] <BBB> so I need to know how to open and close files etc. in windo= ws :) [13:31] <ajmitch> sick... [13:31] <BBB> :) [13:32] <BBB> I'd prefer to call it a job if you don't mind [13:32] Action: ajmitch must sleep now [13:32] <ajmitch> night all [13:32] Nick change: ajmitch -> ajzzzz [13:33] <richardb> BBB: getting anywhere with automake 1.4f? [13:33] <BBB> richardb: first, I'd like to compile gstreamer at all [13:33] <BBB> then, I'll install a new automake [13:33] <BBB> goodnight ajzzzz [13:34] <richardb> BBB: fair enough. ;-) [13:34] <thomas> night aj [13:34] <BBB> shall I submit a bug report to gst-devel@? [13:37] <richardb> what for? [13:38] <richardb> But in general, yes. Or use sourceforges bug tracker. [13:38] <BBB> for the fact that it won't compile [13:38] <BBB> doh [13:38] <BBB> :) [13:38] <BBB> [rbultje@tux gstreamer]$ cat /usr/include/*.h | grep Dl_inf= o [13:38] <BBB> } Dl_info; [13:38] <BBB> extern int dladdr (const void *__address, Dl_info *__info) = __THROW; [13:38] <BBB> [rbultje@tux gstreamer]$=20 [13:38] <BBB> hmm..... [13:38] <BBB> now to find out which file it is :) [13:39] <richardb> grep Dl_info /usr/include/*.h [13:39] <BBB> dlfcn.h [13:39] <BBB> that's the file [13:39] <richardb> Yup, on my system too. [13:40] <BBB> #include <dlfcn.h> [13:40] <BBB> it's there [13:40] <richardb> Hmm. What compiler and options are you using? [13:41] <BBB> stupid thing [13:41] Action: BBB kicks compiler [13:41] <richardb> Perhaps you could paste the compile line here? [13:41] <BBB> I just copypasted the struct to the .c file [13:41] <BBB> I mailed it to gst-deel [13:41] <BBB> so you should get the command in a few seconds [13:41] <richardb> Ah, good. [13:41] <BBB> I just copypasted the struct to my gstinfo.c [13:41] <BBB> and now it works [13:42] <richardb> Yeah, I suspect the problem is that _USE_GNU isn't def= ined for some reason. [13:42] <richardb> __USE_GNU, I mean. [13:42] <BBB> that was my wild guess too [13:42] steveb (ste...@21...) joined #gstreamer. [13:42] <BBB> only based on the fact that there was an ifdef there ;) [13:43] <richardb> Whats your system? [13:43] <BBB> used to be RH-70 [13:43] <BBB> is now some self-made system [13:43] <richardb> Intel? [13:43] <BBB> yes [13:43] <BBB> intel p-2 400 [13:44] <richardb> Hmm. Try putting #define _GNU_SOURCE in gstinfo.c bef= ore the #include <dlfcn.h> [13:44] <BBB> ok [13:45] <BBB> yes [13:45] <BBB> works too [13:46] <BBB> how about: [13:46] <BBB> #ifndef _GNU_SOURCE [13:46] <BBB> #define _GNU_SOURCE [13:46] <BBB> #endif [13:47] <richardb> Not a solution: that would be just a hack to get it to= work. [13:48] <richardb> Um, my compile line has -D_GNU_SOURCE in it. [13:48] <richardb> I wonder where that comes from... ;-) [13:49] <BBB> my compile line doesn't [13:49] <BBB> -I/usr/include/glib-1.2 -I/usr/lib/glib/include -D_REENTRAN= T [13:49] <BBB> -I/usr/include/gtk-1.2 -I/usr/include/glib-1.2 -I/usr/lib/g= lib/include [13:49] <BBB> -I/usr/X11R6/include -I/usr/include/gnome-xml=A0 -O6 -Wall = -c gstinfo.c [13:49] <BBB> gcc -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/include/glib-1.2 [13:49] <BBB> -I/usr/lib/glib/include -D_REENTRANT -I/usr/include/gtk-1.2 [13:49] <BBB> -I/usr/include/glib-1.2 -I/usr/lib/glib/include -I/usr/X11R= 6/include [13:49] <BBB> -I/usr/include/gnome-xml -O6 -Wall -Wp,-MD,.deps/gstinfo.pp= -c gstinfo.c [13:49] <richardb> Ah. What's in your Makefile.am [13:49] <BBB> in gst/ or in ./ [13:49] <BBB> ? [13:49] <richardb> Where CFLAGS are defined. [13:49] <richardb> In gst/ [13:50] <richardb> You should have [13:50] <richardb> CFLAGS =3D $(LIBGST_CFLAGS) -D_GNU_SOURCE [13:50] <BBB> CFLAGS =3D $(LIBGST_CFLAGS) -D_GNU_SOURCE [13:50] <BBB> same here [13:50] <BBB> weird [13:50] Action: BBB kicks automake [13:50] <richardb> Then, your makefiles are not updated. [13:50] <richardb> Have you run autogen.sh recently? [13:50] <BBB> no, :) [13:50] <richardb> If not, try again. [13:51] <richardb> Remember to give any arguments you want passing to con= figure. [13:51] <richardb> That should fix it. [13:52] <BBB> ok [13:52] <BBB> are you guys interested in automake warnings? :) [13:52] <thomas> hmmm... is it normal that pipes set up with gstreamer-la= unch do not end anymore but spin idly when done ? [13:53] <richardb> BBB: yeah, I probably am. [13:53] <richardb> I'd like to ensure none happen... [13:53] <BBB> ok, let's mail them too [13:53] <BBB> plugins/Makefile.am:105: GNOMEVFS_SUBDS multiply defined in= condition [13:53] <BBB> plugins/Makefile.am:107: GNOMEVFS_SUBDS multiply defined in= condition [13:53] <BBB> plugins/Makefile.am:110: bad macro name `=3D=3D=3D=3D=3D=3D= ' [13:53] <BBB> plugins/Makefile.am:112: SDL_SUBDS multiply defined in cond= ition [13:53] <BBB> plugins/Makefile.am:114: SDL_SUBDS multiply defined in cond= ition [13:53] <BBB> plugins/Makefile.am:157: bad macro name `=3D=3D=3D=3D=3D=3D= ' [13:53] <BBB> plugins/Makefile.am:169: bad macro name `=3D=3D=3D=3D=3D=3D= ' [13:53] <BBB> plugins/Makefile.am:148: required directory plugins/<<<<<<<= does not exist [13:53] <BBB> plugins/Makefile.am:148: required directory plugins/Makefil= e.am does not exist [13:53] <BBB> there ya go [13:53] <BBB> plugins/filters/Makefile.am:13: required directory plugins/= filters/level does not exist [13:54] <richardb> You have conflicts in your Makefile.am [13:54] <BBB> this is your newest just-checked-out-CVS [13:54] <BBB> testsuite/Makefile.am:1: required directory testsuite/refco= unting does not exist [13:54] <richardb> Did you alter them yourself ever? [13:54] <BBB> no [13:54] <BBB> let me do a total fresh checkout [13:54] <richardb> Odd. Do cvs -q update -Pd and see what it says. [13:55] <richardb> You probably want a .cvsrc file containing "update -Pd= " [13:56] <BBB> rather not - I have mjpegtools where I have a lots of thing= s that I do want to be changed [13:56] <BBB> they are not the same as the CVS [13:56] <BBB> it's kind of "debugged"code [13:56] <BBB> lots of comments and printfs that are useless for others [13:57] <richardb> Okay - well, remember to to -Pd with gstreamer, or it'= ll often break the build... ;-) [13:58] <richardb> Can't see how -Pd would cause problems for you though;= it just causes new directories added to the CVS to get checked out, and = old empty ones to be ignored. [13:59] <BBB> hmm..... [13:59] <BBB> ok [13:59] <BBB> I'll just update like that manually [14:04] Action: BBB will bbl (including reports and bugs and crashes) [14:04] <BBB> I gotta go to city hall quickly [14:04] Action: BBB is away: brb [14:27] hadess (ha...@pc...) joined #gst= reamer. [14:34] <thomas> someone here who can answer an audio-related question ? [14:34] <richardb> Thats a little bit of a vague request... ;-) I might = be able to. [14:35] <thomas> Well, I'm starting on a noisegate plugin ... [14:35] <thomas> ... first of all it makes a difference if you want to us= e this kind of thing real-time or not. [14:35] <richardb> Ah: dsp. I can't really help there. [14:36] <thomas> richardb: well, I was wondering what I should do about b= uffering [14:36] <thomas> richardb: I would like to somehow have the plugin use th= e standard gstreamer queue element [14:40] <hadess> hey guys [14:40] <thomas> hi haddess [15:05] <thomas> hadess: can a plugin send signals to the main applicatio= n somehow ? [15:13] <hadess> sure it can [15:13] <hadess> take a look at the gnomevfssrc, it sends an "eos-hack" s= ignal [15:15] <thomas> great, thanks [15:42] dobey (do...@dr...) joined #gstreamer. [15:48] <thomas> hadess: do you also have an example how the app receives= the signal ? [15:48] <thomas> hi dobey [15:49] <dobey> hey [15:53] <BBB> hey guys this is pretty cool, gstmediaplay works for me [16:06] <richardb> Anyone tried building docs lately? [16:09] <dobey> no [16:14] Action: richardb goes away for a couple of hours [16:14] Nick change: richardb -> richardb-away [16:27] <BBB> [rbultje@tux gstreamer]$ make docs [16:27] <BBB> make: getcwd: : No such file or directory [16:27] <BBB> make: *** No rule to make target `docs'. Stop. [16:27] <BBB> [rbultje@tux gstreamer]$=20 [16:27] <BBB> :) [16:32] Nick change: taazzzz -> taaz [16:33] Action: taaz pokes richardb-away=20 [16:35] <BBB> http://ronald.bitfreak.net/gstreamer-docs [16:35] <BBB> my logs when building the docs [16:53] steveb (ste...@21...) left irc: quit has st= eveb, yes [17:11] hadess (ha...@pc...) left irc: m= ooooh! [17:22] Nick change: taaz -> taaz-away [17:25] _gst_newt_ joined #gstreamer. [17:34] wtay-zZz (wi...@ca...) left irc: Ping = timeout for wtay-zZz[cable-195-162-214-58.upc.chello.be] [17:39] wtay-zZz (wi...@ca...) joined #gstream= er. [18:04] omega_ (om...@bo...) joined #gstreame= r. [18:04] Nick change: omega_ -> omega_rr [18:04] <dobey> whee [18:08] <richardb-away> Evening omega_ [18:08] steveb (st...@no...) joined #gstreamer. [18:08] Nick change: taaz-away -> taaz [18:08] Action: taaz pokes richardb-away again [18:08] Nick change: richardb-away -> richardb [18:08] <richardb> Hello [18:08] <taaz> richardb: rumors are true! some of us (me) have doc build= issues=20 [18:09] Action: richardb grins [18:09] <richardb> What goes wrong? [18:09] <taaz> from a (i think) clean checkout i'd like to run 'make dist= ' [18:09] <taaz> so i run autogen.sh then make dist [18:10] <taaz> but the docs fail cause it needs libgst.la at some point a= nd doesnt even try to build it [18:10] <richardb> Ah. [18:10] <richardb> That sounds like just a dependency problem. [18:10] <taaz> i guess... but it seems more messy than that [18:10] <taaz> well... actually, i couldnt really tell why it needed that [18:11] <taaz> why is make dist building docs at all? [18:12] <richardb> taaz: with --disable-docs-build (which is the default) [18:12] <richardb> It shouldn't be, but I can see how it might be: [18:12] <richardb> docs are turned off by putting docs into SUBDIRS condi= tionally. [18:13] <richardb> But make dist uses DIST_SUBDIRS, because even subdirs = which are configured off should go into dists. [18:13] <richardb> So I need to add a little more logic to that... [18:14] <richardb> It needs libgst.la because the gnome doc stuff needs t= o build a little program to scan for symbols [18:14] <richardb> (I don't quite understand that bit of the process, but= that's not important right now) [18:15] <taaz> why is it doing that on a make dist though? seems like som= ething that should just happen during a actual build. [18:16] <richardb> Hmm, not sure. [18:17] <richardb> I think its because the generated docs are in EXTRA_DI= ST or whatever, so are needed to make the dist [18:17] <richardb> So it tries to generate them. [18:17] <BBB> richardb: my log is at http://ronald.bitfreak.net/gstreamer= -docs [18:18] <richardb> BBB: thanks [18:18] <BBB> my doc-compile errors [18:18] Action: BBB is gone again :) [18:19] <richardb> BBB: those look like errors due to having the wrong do= cbook DTD installed... [18:19] <richardb> BBB: Gnome (and us) use a modified DTD for some reason= , and the standard DTD won't work with our docs. [18:20] <richardb> Fixing that is a matter of ensuring the correct packag= es are installed. [18:20] <richardb> Unfortunately, I don't know what the correct packages = are. [18:20] <richardb> taaz: I think I now understand your whole problem... ;= -) [18:20] <richardb> Just not sure what the solution is. [18:21] Action: richardb gets a cup of tea to help thought process. [18:32] <taaz> so why are generated docs in EXTRA_DIST? why not just hav= e the source (since that's the whole point, imho) in the EXTRA_DIST and t= he n build the docs later [18:33] Zeenix (pro...@ho...) joined #gstreamer. [18:36] <richardb> Thats a possible approach. [18:36] <richardb> I don't think the things ever really been thought abou= t... ;-) [18:36] <richardb> It's nice to have the generated docs easily available,= given how hard they are to generate... [18:37] <richardb> But that's not really an excuse. (We should make a se= parate docs package available anyway) [18:38] <richardb> Hmm: my attempt at running make dist on a fresh checko= ut failed in the arts plugin... [18:39] <Zeenix> when will be the new version of gstreamer launched ? [18:39] <richardb> Zeenix: soon. [18:39] <richardb> Tomorrow is the target, I believe. [18:41] thomas (th...@21...) left irc: [x]chat [18:42] <Zeenix> Will it be having visible( to the user ) changes ? [18:45] <Zeenix> any body home ? [18:45] <richardb> taaz: Hmm, can't build dists without having mcopidl in= stalled, since its needed to generate plugins/arts/gst_artsio.cc [18:45] <richardb> Zeenix: probably. Many new plugins, actually works, e= tc. [18:46] <Zeenix> richardb: looks like, i am the only one talking to you &= yourself [18:47] <richardb> Zeenix: others are around, I imagine they are just a b= it busy too. [18:47] Zeenix (pro...@ho...) left irc:=20 [18:52] <taaz> richardb: yup [18:52] <taaz> that idl file needs to be added to EXTRA_DIST for that dir= . i should commit that patch [18:54] <richardb> taaz: yup, that's needed. [18:55] <richardb> It's still a requirement to have that stuff installed = to make a dist though. [18:55] <richardb> Which is a shame, but unavoidable, I think. [18:56] <richardb> It'll be solved when the plugins are split off. [18:56] <taaz> hmm? developers are the people doing 'make dist' and they= -should- have everything installed [18:56] <taaz> its not like end users need to ever run that [19:00] <richardb> I could extend that argument to "developers should hav= e everything installed to build docs" [19:00] <richardb> Which would solve the docs problem: you just need to h= ave docs configured on to build a dist [19:01] <richardb> Hmm. It should still work if they're configured off, = though. [19:01] <taaz> bah ;) [19:02] <richardb> Why should a dist contain that idl generated .cc file,= actually? [19:02] <richardb> It only gets used if arts is installed. [19:02] <richardb> In which case mcopidl should be present. [19:02] <richardb> In which case the file can be generated automatically. [19:02] Action: taaz has no clue [19:03] Action: richardb tries to remember how to tell automake that a so= urce is not expected to be present in a dist. [19:04] <taaz> well... its probably a GPL violation to not include the pr= efered source form if we consider the tarball sufficient to satisfy rest = of GPL requirements for having source available. [19:04] marcf (ma...@be...) left irc: Read error to marcf[becomm.com= ]: Connection reset by peer [19:05] <richardb> The idl file is surely the preferred form? GIven that= the .cc is automatically generated. [19:06] <taaz> prefered form imho is what the developers would edit to ma= ke changes, not machine generated code [19:06] Action: omega_rr hacks on an embedded system [19:07] <richardb> taaz: yes. And you shouldn't be editing a generated f= ile, because your changes wilkl get lost next time its generated... [19:08] <richardb> First line of it is: [19:08] <richardb> /* this file was generated by the MCOP idl compiler -= DO NOT EDIT */ [19:08] Action: richardb grins [19:10] <taaz> yeah i think we both understand this ;) its just a questi= on if the .idl should go in the EXTRA_DIST instaed of (or in addition to)= the generated kde crap [19:11] <taaz> our fearless leader omega_rr will guide us [19:11] Nick change: taaz -> taaz-lunch [19:15] <richardb> Actually, I suggest that we simply change gst_artsio.c= c to gst_artsio.idl in libgst_arts_la_SOURCES [19:15] Nick change: wtay-zZz -> wtay [19:15] <wtay> hello [19:16] <richardb> As we would for a .y or .l or any other source file. = And as the automake docs suggest. ;-) [19:16] <richardb> wtay: morning. ;-) [19:17] <wtay> hmm, a guy on -devel wonders where the plugins are... [19:17] <wtay> I'll reply [19:18] <richardb> Odd. [19:19] <omega_rr> we =02need=02 to write a getting-started guide for the= se ppl [19:20] <wtay> yes, I'll start on that right now [19:20] <omega_rr> first section: who should be using gstreamer right now= ... [19:20] <omega_rr> answer: developers only! [19:21] <wtay> second: is it a media player... [19:21] <wtay> ? [19:21] <wtay> answer: no! <g> [19:22] <wtay> plain HTML will do? [19:25] <richardb> wtay: as a starting point, definitely. [19:25] <richardb> plain text might even be better... [19:26] <wtay> I'll do the html version then and use something to create = plain text from it when needed [19:26] <richardb> Anyone know if its possible to have arts development s= tuff installed without having mcopidl installed? [19:27] <richardb> (On typical distributions?) [19:28] <wtay> I don't think so, mcop is really tied to arts [19:28] ChiefHighwater (pa...@te...) joined #gstreamer. [19:28] <wtay> yo [19:29] <ChiefHighwater> ello [19:30] <richardb> wtay: Good. [19:37] <richardb> omega_rr: Is it desirable to have generated docs prese= nt in distributions? [19:38] <richardb> If so, docs tools must be installed on any machine tha= t runs make dist (and I'll put assertions in the Makefile.ams to check th= is0 [19:38] <richardb> If not, I'll disable it. ;-) [19:41] <omega_rr> um... I'd say it's not a requirement [19:47] <ChiefHighwater> omega_rr:on our webpage..should the newsflash be= updated to something like "Testing, Testing: 1, 2, 3" [19:47] <omega_rr> why? [19:47] <omega_rr> once we have pre1 out maybe [19:47] <omega_rr> which should be any time, I'd think. wtay? [19:47] <ChiefHighwater> is our biggest news the fact we have a new maill= ist? [19:47] <omega_rr> atm, yes ;-) [19:48] <ChiefHighwater> pffft [19:48] <wtay> omega_rr: writing some kind of FAQ now... [19:48] <omega_rr> wtay: ok, call it gstreamer.net/getting-started/ [19:48] <wtay> I can start a new make dist and upload it to gst.net if yo= u want [19:48] <ChiefHighwater> please cc the samba folks with a how to 8-] [19:49] <ChiefHighwater> omega_rr:how to write a faq that is [19:59] <wtay> http://gstreamer.net/getting-started/ [20:00] <omega_rr> things like why you shouldn't make install from cvs [20:00] <omega_rr> what the differences are between autogen and configure= d trees [20:00] <omega_rr> basically a first-level FAQ ;-( [20:03] thomas (th...@ad...) joined #gstreamer. [20:04] <thomas> hi [20:04] <wtay> hi [20:04] <thomas> anyone care to validate my thought on an app I'm wanting= to write tomorrow ? [20:05] <omega_rr> we can try ;-) [20:05] <wtay> yeah [20:05] <thomas> It's basically pretty easy. What it does is listen to a= udio-in, when it detects sufficient level it should write audio to a file= , when it's not, it should pause, when it's loud enough again it should w= rite to the next file [20:06] <thomas> so I made a cutter plugin which detects level and sends = a stop or start signal when appropriate [20:06] <omega_rr> yeah, should be fairly easy, but you'll probably have = to write a filter [20:06] <thomas> so the main app constructs a bin ossrc-cutter-disksink [20:06] <omega_rr> thomas: another option is for that level-check to simp= ly not write buffers at all unless it's loud enough [20:06] <omega_rr> that will have the effect of simply not writing tot he= file, you don't have to state-change at all [20:07] <thomas> omega_rr: but I do need to write to a new file for each = transition [20:07] <thomas> so I could either do a multisink or have it handled by t= he app [20:07] <omega_rr> oh, ok [20:07] <omega_rr> yeah, fire a signal when it starts and stops, only wri= te buffers when it's live, and have the app reset the filename on the dis= ksink [20:07] <thomas> my point is : now that incsched is done, I suppose it's = ok to pause the bin, remove the disksink, and put in a new disksink or a = fakesink in place and go on ? [20:08] <omega_rr> yup, should work no problems [20:08] <thomas> ok, that should make a good example as well. [20:08] <omega_rr> you will probably have to update disksink to handle cl= osing one file and starting a new one while PAUSED [20:08] <omega_rr> my guess is that it won't like that right now, like di= sksrc didn't until recently [20:08] Action: omega_rr writes a kernel driver [20:08] <thomas> omega_rr: I was thinking about destroying the disksink e= lement altogether and making a new one [20:09] <omega_rr> thomas: don't bother [20:09] <thomas> I'll try both ;) [20:09] <omega_rr> it should be capable of starting a new file, it just i= s not set up just right atm, easily fixable too [20:09] <omega_rr> so you just need to PAUSE that element, even [20:09] <omega_rr> just the disksink, set filename, unpause, and go [20:09] <thomas> well since I made disksink, I should be responsible for = fixing it as well ;) [20:09] <thomas> I'll try that then [20:10] <thomas> one other thing: the eos problem is a known thing, right= ? [20:10] <thomas> what's the suggested way around it at the moment ? [20:10] <omega_rr> um, don't hit EOS ;-( [20:10] <omega_rr> wtay: status on a hack? [20:10] <wtay> omega_rr: didn't look at it yet [20:11] <thomas> I mean: if I use gstreamer-launch, it spins idly at the = end, that's my eos problem ;) [20:11] <omega_rr> thomas: yup [20:11] <omega_rr> I'm working on non-gstreamer stuff at work right now, = but soon after 0.2.0 I intend to get the event system built [20:11] <thomas> ok [20:12] <thomas> One other thing: I tried gstplay both at work and at hom= e and on my laptop, all three of them segfaulted as soon as playing shoul= d start, no core dump [20:12] <omega_rr> wtay: you think pre1 would be ready? [20:12] <thomas> what can I do to check what went wrong ? [20:12] <omega_rr> thomas: run it in gdb: libtool gdb gstmediaplay [20:12] <omega_rr> r mediafile [20:12] <omega_rr> and/or --gst-mask=3D-1 [20:12] <wtay> omega_rr: yes, I think so [20:13] <omega_rr> wtay: ok, you want to build one? [20:13] <omega_rr> we'll put this out on -devel [20:13] <wtay> omega_rr: ok [20:13] <omega_rr> hmmm, I just got a new chair at work [20:13] <wtay> reload the getting-started from time to time and comment p= lz :) [20:13] <omega_rr> it's labeled as "Ergo-Optional Seating" [20:13] Action: omega_rr is confused [20:13] <richardb> Hmm: just tried a make install to see what happens whe= n plugin-srcdir is on... [20:14] <omega_rr> wtay: yup [20:14] <omega_rr> btw, we have a finalized logo now, lemme look at it an= d send it around [20:14] <richardb> ... it does the install fully, then fails with an erro= r message. :( [20:14] <richardb> (The idiot check error message I put in) [20:14] <omega_rr> richardb: yeah, lemme look at that [20:14] <wtay> make dist started [20:14] <omega_rr> richardb: which check? [20:14] <richardb> Read toplevel Makefile.am, at the bottom. [20:15] <richardb> install-exec-hook runs _after_ install, sadly., [20:15] Action: omega_rr left his 10bt dongle at home, so he's stuck with= viewcvs atm [20:15] <omega_rr> richardb: there should be a pre-hook [20:15] <omega_rr> in fact, I'd put that warning in at configure time too [20:15] Action: richardb is reading docs [20:15] <omega_rr> just like gtk+ does for 1.3 [20:15] <richardb> Good plan. [20:16] <omega_rr> problem with plugin-srcdir is that it should be plugin= -builddir ;-) [20:16] <omega_rr> make distcheck would need to actually have some make c= heck targets that do something to notice that [20:16] <wtay> make dist succeeded [20:16] <omega_rr> and distcheck doesn't do plugin-builddir either ;-( [20:16] <wtay> trying to build it now [20:16] <omega_rr> wtay: distcheck? <g> [20:16] <wtay> doing make distcheck then :) [20:17] <omega_rr> dnl Set location of uninstalled plugin directory [20:17] <omega_rr> PLUGINS_SRCDIR=3D`pwd`/$srcdir [20:17] <wtay> spotted an error in include/Makefile.am... fixing [20:17] <omega_rr> should be the builddir, but the `pwd` trick is really = lame [20:20] <richardb> distcheck does make dist then unpacks the dist and bui= lds [20:20] <richardb> So it won't have plugin-srcdir set. [20:21] <richardb> So that shouldn't be a problem. [20:21] Action: richardb hasn't tried make distcheck yet. ;-) [20:21] <omega_rr> richardb: but then make check won't work [20:21] <richardb> Ah. [20:22] <omega_rr> so we might want to come up with a way for make check = to set an env var that gst_init prepends the builddir to the plugin path [20:22] <omega_rr> that's cleanest I think [20:22] <richardb> Agreed. [20:22] <richardb> You typed it faster than me. :) [20:22] <omega_rr> hehehe [20:22] <omega_rr> brb [20:27] <richardb> wtay: Have you committed to include/Makefile.am yet/ [20:28] <wtay> richardb: nope... [20:28] <wtay> sec.. [20:29] <wtay> commited [20:33] <richardb> idiottest should now run first. :) [20:34] <wtay> what's idiottest? <g> [20:34] <richardb> You pass. :) [20:34] <wtay> richardb: who wouldn't? <g> [20:37] <wtay> codectest.o: In function `main': [20:37] <wtay> codectest.o(.text+0x18): undefined reference to `gst_init' [20:37] <wtay> codectest.o(.text+0x25): undefined reference to `gst_libra= ry_load' [20:37] <wtay> codectest.o(.text+0xe9): undefined reference to `ICOpen' [20:37] <wtay> codectest.o(.text+0x11e): undefined reference to `ICSendMe= ssage' [20:37] <wtay> codectest.o(.text+0x144): undefined reference to `ICClose' [20:37] <wtay> collect2: ld returned 1 exit status [20:37] <wtay> :( [20:38] <wtay> I'm going to remove it... [20:38] <wtay> or wait, I might be able to fix it... [20:40] <wtay> can someone try to do make check in the plugins/avi/ dir a= nd see if the codectest builds [20:40] <wtay> ? [20:42] <richardb> trying [20:44] Action: omega_rr is up on the EVM, will disappear for a while [20:48] <richardb> codectest-codectest.o: In function `main': [20:48] <richardb> codectest-codectest.o: In function `main': [20:48] <richardb> /data/home/richard/Working/gstreamer/gstreamer/plugin= s/avi/codectest.c:15: undefined reference to `gst_init' [20:48] <richardb> /data/home/richard/Working/gstreamer/gstreamer/plugin= s/avi/codectest.c:17: undefined reference to `gst_library_load' [20:49] <wtay> richardb: I added $(GST_LIBS) to the libs so that shouldn'= t happen, can you verify it's correct? [20:49] <richardb> Oh: that was before update, sorry. [20:49] <dobey> bbl [20:49] dobey (do...@dr...) left irc: eh [20:49] <wtay> richardb: right :) [20:49] Action: richardb retries [20:50] Uraeus (csc...@c2...) joined #gstreamer. [20:50] <wtay> hi [20:50] <Uraeus> hello [20:50] <richardb> wtay:Works [20:50] <wtay> Uraeus: gstreamer.net/getting-started/ [20:50] <wtay> richardb: ok [20:50] <wtay> Uraeus: I need *lots* of suggestions :) [20:51] <Uraeus> wtay: what about calling it a FAQ? :) [20:52] Action: Uraeus tries to awake after a 10 minute coach-nap having = taken 2 hours [20:53] <ChiefHighwater> hehe Uraeus:i hate it when that happens [20:54] <wtay> brb [20:54] <Uraeus> ChiefHighwater: yeah, the troublesome thing is being mor= e tired afterwards than one was before [21:00] <Uraeus> wtay: so you when your are back [21:00] <Uraeus> s/so you/say yo/ [21:02] <ChiefHighwater> Uraeus: I always wake up 'foggy' and it takes fo= rever to shake it off [21:03] <wtay> yo <g> [21:03] Action: wtay always *is* foggy [21:04] marcf (ma...@be...) joined #gstreamer. [21:04] <wtay> damn my GFs 2.4.4 kernel is hit by the swap bug :( [21:04] <Uraeus> wtay: I looked on your sourceforge background page today= , with a background like that I am having trouble understanding why you h= aven't made a sid plugin yet <g> [21:04] <wtay> Uraeus: 'cause the lib sucks :) [21:04] <Uraeus> hehe [21:05] <wtay> I have a small collection of C64 tunes :) [21:05] <Uraeus> wtay: isn't there like 2 or 3 different implementations = available? [21:05] <wtay> yeah, but it's not really a streamable format [21:06] <Uraeus> neither is FLI right? [21:07] <Uraeus> wtay: what did you think of my test document (other than= Hadess being right about the media clips :) [21:07] <wtay> Uraeus: ? [21:08] <Uraeus> wtay: my mail? [21:08] <wtay> Uraeus: FLI is somewhat streamable [21:08] <wtay> Uraeus: checking [21:08] <richardb> taaz-lunch: if you're still around, I'd be interested = in knowing what make dist does for you in the docs tree now... [21:08] <Uraeus> hi richardb [21:08] <richardb> hi [21:08] <wtay> Uraeus: very good, I wonder where the cothread test is... [21:09] <Uraeus> richardb: I too had problems with the docs, the excact s= ame one the automatic build had [21:09] <richardb> I think that got fixed. [21:09] <richardb> Do you still have them? [21:09] <wtay> any comments on the FAQ? [21:09] leviathan (le...@p3...) joined #gstreamer. [21:09] <wtay> hi [21:09] <Uraeus> wtay: is lies in the compilation itself, if 20 people ma= il us saying it compiles then we 'prove' cothreads work fine [21:09] leviathan (le...@p3...) left irc: [x]chat [21:10] leviathan (le...@p3...) joined #gstreamer. [21:10] <Uraeus> richardb: haven't tested since, will do so now [21:10] Nick change: taaz-lunch -> taaz [21:10] <wtay> Uraeus: well, it actually has to run too... [21:10] <richardb> Uraeus: great. [21:10] <taaz> richardb: i'll test build now [21:11] <Uraeus> wtay: well, if those 20 say it compiles but nothing disp= lays then we have 'proof' that someting is wrong=20 [21:11] <Uraeus> wtay: I can't think of any other way for odinary people = to test it [21:12] <Uraeus> s/odinary/ordinary/ [21:12] <wtay> Uraeus: a simple test app would be better IMO [21:12] Parapraxis (Sp...@us...) joined #gs= treamer. [21:12] <wtay> it depends on what we want to test of course [21:12] <wtay> yo [21:14] <Uraeus> wtay, omega_rr: leviathan is developer of asd (the esd r= eplacement) [21:15] <wtay> leviathan: nice to meet you :-) [21:15] BBB (BB...@uc...) left irc: Read error to BBB[ucu-1= 05-116.ucu.uu.nl]: EOF from client [21:15] Action: wtay builds a new kernel [21:15] <leviathan> uraeus told me that there are some asd interested per= sons in this channel? [21:16] <leviathan> is that right? [21:16] <leviathan> somebody interested in joining development on asd? [21:16] <leviathan> that would be quite nice [21:17] <wtay> omega is most likely the most interested person [21:17] <wtay> he's gone for a moment though.. [21:17] <leviathan> (i am not fluent in irc, could you please tell me if = rr stands for anything?) [21:17] <Uraeus> he probably be in later, ow3n might also be interested i= n helping out if a GNOME 2.0 timetable is achievable [21:18] <wtay> leviathan: the company he works for (ridgerun) [21:19] <leviathan> it should be achievable, if someone joins development= . at the moment i am the only (really working on it) developer and theres= much to do [21:19] <leviathan> right at the moment i am trying to write some sample = conversion routines [21:19] <wtay> leviathan: what are the major things that need work? [21:20] <leviathan> it's quite a tricky task to get the work with good sp= eed [21:20] <wtay> leviathan: omega is your man if you want high speed MMX co= de :) [21:20] <leviathan> anybody knows some code i'd be able to steal? [21:20] <wtay> leviathan: I know of a good samplerate converter [21:20] <leviathan> sox has one [21:20] <leviathan> but it is rather slow and really bad implemented [21:21] <wtay> leviathan: lemme upload resample to gstreamer.net... [21:21] <leviathan> doesn't gstreamer have a some code for doing such con= versions? it shoudl be a rahter common task, i think? [21:21] <wtay> leviathan: we plan to use resample (with heavy MMX optimis= ations) [21:22] <wtay> s/we/I [21:22] <taaz> richardb: make dist failed :( [21:22] <wtay> taaz: I commited a patch [21:22] <wtay> leviathan: gstreamer.net/resample-1.6.tar.gz [21:22] <taaz> for what? [21:22] <wtay> leviathan: the core looks simple enough [21:23] <leviathan> the ideal conversion library would be able to do conv= ert between arbitrary bit depths, sample rates, byte orders and signee [21:23] <richardb> taaz: what went wrong/ [21:23] <wtay> taaz: testsuite/refcounting [21:23] <taaz> mv -f gstreamer-scan.o gstreamer-scan.lo [21:23] <leviathan> i'll have a look on it [21:23] <wtay> leviathan: at the expense of speed... [21:23] <taaz> then it tries to link -scan with the .o [21:23] <taaz> and fails cause it just moved it... ?? [21:23] <wtay> taaz: yes, I patched gtk-doc [21:24] <leviathan> wget can't find gstreamer.net/resample-1.6.tar.gz [21:24] <wtay> leviathan: oops, wrong dir... [21:25] <richardb> taaz: which dir is that in? [21:25] <wtay> leviathan: try again [21:25] <taaz> richardb: docs/gst i think [21:25] <leviathan> 170k. that's quite big [21:25] <wtay> leviathan: it has a lot of other bullshit in it [21:25] <taaz> wtay: wtf? i need to patch gtk-doc to make this build? [21:25] <wtay> leviathan: the core is pretty simple I think [21:25] <wtay> taaz: yes, it doesn't handle libtool very well [21:26] <richardb> taaz: I have a rm gstreamer-scan.o; ln -s gstr= eamer-scan.lo gstreamer-scan.o [21:26] <richardb> taaz: But not a mv [21:26] <leviathan> wtay: what of making a real library out of it? [21:26] <leviathan> wtay: i mean a shared library [21:26] <wtay> leviathan: yes please :) [21:26] <taaz> wtay: is the gtk-doc debian maintainer ignoring the bugs/p= atches i know you submitted already? (hint hint) [21:27] <leviathan> wtay: what licence? [21:27] <leviathan> wtay: i see no LICENCE file. am i blind? [21:27] <wtay> leviathan: GPL [21:27] <wtay> leviathan: in the README [21:28] <wtay> taaz: I'm happy after patching it. [21:28] <taaz> argh... i wish someone with a fast machine could fix this = for me... it takes minutes to retest this crap [21:28] <leviathan> wtay: that's not right. can't find the word GPL there= . in fact a "grep GPL *" in the directory just prints out one line [21:29] <wtay> leviathan: look at the COPYING section in the README [21:29] <leviathan> wtay: OK, you convinced me [21:30] <wtay> leviathan: it's just a suggestion, I didn't find anything = better yet [21:31] <wtay> leviathan: resamplesubs.c has the code [21:31] <leviathan> wtay: you said, that you were trying to adapt resampl= e to gstreamer, haven't you? What do you think of making a library out of= it? [21:31] <wtay> leviathan: I didn't start on that yet.. I think a library = would be great [21:31] <wtay> leviathan: it does need a bit of work 'cause the resample = code works on floats [21:31] <leviathan> wtay: the code looks like the code of a windows progr= ammer ;-)=20 [21:32] Sopwith (so...@na...) joined #gstreamer. [21:32] <wtay> hi [21:32] <wtay> leviathan: yeah [21:32] <leviathan> wtay: hmmh. i need fast code, real time capable [21:32] <wtay> leviathan: the core is int only, and it looks like it can = be MMXified=20 [21:33] <wtay> leviathan: but it does have some interesting algorithms an= d filters... [21:33] <leviathan> wtay: at the moment i am programming a rather simple = resampler. doesn't make any further caluclations, just takes the data and= duplicates it or cuts it [21:34] <wtay> leviathan: that's what's in gstreamer too now, very simple [21:34] <Uraeus> hi sopwith [21:34] <Sopwith> ullo [21:34] <leviathan> wtay: a funny coincidence is, that stephan westerfeld= , the arts-developer studies a the same university here in hamburg, as i = do [21:35] <leviathan> wtay: he told me that he want's to do all his streami= ng with fp, too [21:35] <wtay> leviathan: well... [21:35] <wtay> leviathan: too bad the HW doesn't handle it... [21:36] <leviathan> wtay: strange idea, but he argued that that would be = a great benefit, when different filters apply on the data sequentially [21:36] <Uraeus> Sopwith: in case you didn't realize leviathan is the asd= developer :) [21:37] <wtay> leviathan: when the overhead of switching between filters = is low enough, yes [21:37] Nick change: omega_rr -> omega_lunch [21:37] <omega_lunch> brb [21:37] <richardb> taaz: I can't see where a mv would happen anywhere in = the docs directory. [21:37] <taaz> wtay, richardb: what should i do about this. i still cant= build. if its debian tools problem then =02please=02 submit bug reports= with patches so the tools get fixed. [21:37] <richardb> taaz: if you could give me some context that would hel= p lots... [21:38] <richardb> Mail me the output from the end of make dist would be = great. [21:38] <richardb> Last 1000 lines or so... ;-) [21:38] <leviathan> wtay: a guy called brian craft volunteered to do some= mmx based resampling routines [21:38] ChiefHighwater (pa...@te...) left irc: Read error to = ChiefHighwater[temple-baptist.com]: Connection reset by peer [21:38] <taaz> richardb: that rm ln thing happens. i think its the -scan= obj cmd that's failing [21:38] <wtay> leviathan: cool, me wants [21:38] <leviathan> wtay: he is registered as co-developer for asd asd so= urceforge [21:39] <leviathan> s/asd/at/ [21:39] <leviathan> wtay: but i heard nothing from him for about two mont= h now [21:39] <wtay> leviathan: hmm [21:39] <leviathan> wtay: he sayd he had very limited time [21:39] <richardb> taaz: Could you give me the output of "make scanobj" i= n docs/gst/ [21:40] <Sopwith> mikmod has some possibly useful mixing routines, FWIW. = can even do dolby surround if needed [21:40] <leviathan> wtay: don't now if he did work on it. but i think not [21:40] <richardb> taaz: it may be helpful if you run make in <topdir>/gs= t first. [21:40] <wtay> leviathan: too bad :( [21:40] <leviathan> wtay: did you have a look on the routines of arts? [21:40] <richardb> taaz: ri...@ta... [21:41] <wtay> leviathan: vagely... [21:41] <leviathan> wtay: i beleive that thay have such a code too [21:41] <wtay> leviathan: yup [21:41] ChiefHighwater (pa...@te...) joined #gstreamer. [21:42] <wtay> leviathan: If I do a 44100->441 sample rate conversion I w= ant *exactly* 44100 samples in and 441 samples out [21:42] <leviathan> wtay: i'd have a quick look, when there would be a cv= sweb for artsd-cvs. has kde one? [21:42] <wtay> leviathan: no idea... [21:43] Nick change: ChiefHighwater -> CHW_lunch [21:43] <taaz> richardb: sent [21:43] <wtay> richardb: uh? does make distcheck fail when a check_PROGRA= MS fails? [21:43] <taaz> richardb: but wtay says its gtk-doc/libtool issue that he = feels isnt important enough to file a bug report on ;) [21:44] <Uraeus> leviathan: don't think they have one [21:44] <leviathan> wtay: and the other things asd has to learn (besides = resampling) is: more complete protocol, a way for clients to get a extra-= low-latency-stream without mixing, many opitimizations for latency [21:44] <Uraeus> leviathan: unless it is on sourceforge [21:44] <richardb> wtay: I believe so. [21:44] <leviathan> wtay: sample caching has not yet made its move from a= sd3 to asd4 [21:44] <richardb> taaz: reading, but ah, right. [21:45] <leviathan> wtay: the xmms-plugin has do make the way too [21:45] <wtay> leviathan: you plan on doing DMA? [21:45] <leviathan> wtay: the current esd compatibility protocol needs so= me improvements too [21:45] <leviathan> wtay: and testing ist needed, much testing [21:46] <leviathan> wtay: and more docs. a the moment i did very limited = comments. only on some header-files [21:46] <Uraeus> leviathan: well if we have a gstreamer plugin we could t= est asd :) [21:46] <leviathan> wtay: dma? how does this work with oss? i did it year= s ago under DOS with direct hw-access, but under unix? [21:47] <wtay> leviathan: you can mmap the buffer IIRC [21:47] <leviathan> wtay: i expect that i mark a buffer asd dma-buffer an= d write my data to it [21:48] <wtay> leviathan: you write directly into the mmaped buffer I thi= nk [21:48] <leviathan> wtay: i don't know if this fits well into my current = design [21:48] <leviathan> wtay: it would be best if the data isn't copied more = than needed [21:49] <wtay> leviathan: maybe you should, John Carmack would thank you = very much :) [21:49] <leviathan> wtay: a the moment asd does only a read() and a write= (). no data is ever copied. [21:50] <leviathan> wtay: for dma i'd have to copy the data into the dma-= buffer instead of calling a write() [21:50] <leviathan> wtay: is this more efficient? copy+dma than write()? [21:50] <wtay> leviathan: you could do that in the write() code too if th= e audio device is local [21:50] <richardb> taaz: I shall do my best to disable docs build when do= ing make dist. [21:50] <richardb> taaz: Not sure why its working for me though. I run d= ebian... [21:51] <taaz> richardb: isn't this going to bite me when i try to really= build the docs? [21:51] <wtay> leviathan: if the app could write straight into the DMA bu= ffer that would be fastest of course. [21:51] <wtay> leviathan: that's what we are going to do at some point wi= th our bufferpools [21:51] <taaz> richardb: really? latest unstable debs? [21:51] <leviathan> wtay: as i noticed, oss has some nice ioctls for gett= ing current playing position, when using dma [21:51] <richardb> taaz: yes, it will bite then. :( [21:52] <Sopwith> leviathan: The ioctls work the same for non-DMA, though [21:52] <wtay> leviathan: yup, you'll need those [21:52] Nick change: omega_lunch -> omega_eating [21:52] <wtay> leviathan: why don't you implement something like X?=20 [21:52] <richardb> taaz: aaaaaaaah. I have a custom gtkdoc in my bin dir= ectory. Forgot about that. [21:52] <leviathan> wtay: X? [21:52] <taaz> richardb: hold up a sec... i figured it was a debian issue= but maybe its just me. i have cvs source, run autogen.sh, then run 'ma= ke dist' and it fails. [21:52] Action: richardb goes to look where they came from. [21:52] <wtay> leviathan: like, request a buffer from the asd lib and let= the app write into it [21:52] <leviathan> wtay: shared memory access to the asdoutput buffer? [21:52] <wtay> leviathan: yes [21:53] <Sopwith> wtay: You can't really do that, I don't think [21:53] Action: taaz smacks wtay and richardb for not getting the maintai= ner to fix the bugs [21:53] <Sopwith> can you? :) [21:53] <leviathan> wtay: you mean additionaly to the normal streaming ac= cess [21:53] <wtay> leviathan: yup [21:53] <leviathan> wtay: it should not be so complicated [21:53] <wtay> Sopwith: why not? [21:54] <leviathan> wtay: hmm... how does shm work under unix? [21:54] <taaz> speaking of that... has anyone asked the automake debian m= aintainer if that patch could go in the debian version? [21:54] <leviathan> wtay: do i request shared memory and get a new memory= area? [21:54] <wtay> leviathan: man shmget [21:54] <richardb> taaz: this is on a system where I don't have root and = the package installed is old. [21:54] <Sopwith> wtay: How would you get the mmap to go into the shm are= a? [21:54] <leviathan> wtay: or do i mark some existing area as shared=DF [21:54] <richardb> On my home machine, I have gtk-doc-tools 0.4-0.2 [21:55] <wtay> Sopwith: you *request* a buffer, the lib decides on shm or= dma [21:55] <wtay> Sopwith: like how Xv *should*"be implemented [21:55] <Sopwith> wtay: So you are essentially "use shm for a transport, = or use DMA for hardware access" [21:55] <Sopwith> you can't do both, was what I was asking [21:55] <wtay> Sopwith: ok [21:56] <leviathan> wtay: mmap has a flag MAP_SHARED for sharing a mmaped= file [21:56] <Sopwith> leviathan: Yes, but you can't do multiple opens on many= sound devices [21:56] <Sopwith> so MAP_SHARED won't do any good to get it straight into= the sound device [21:56] <wtay> the DMA should be the optimized case, you'll need a fallba= ck too [21:57] <Sopwith> well, I will leave you to get it right :) [21:57] <leviathan> wtay: yes. that's clear. not every client is local [21:57] <omega_eating> btw, what are we discussing right now? /me looks a= round [21:57] <wtay> leviathan: omega_eating has written something about his id= eas for a sound server, very interesting IMO, I'm just quoting him now :-= ) [21:57] <leviathan> Sopwith: a local client could get a pointer to the dm= a-buffer, which was opened by the asd-server, via shm [21:57] <Uraeus> omega_eating: leviathan is the asd author [21:58] <omega_eating> ok, I was guessing that... what else? <g> [21:58] <Sopwith> leviathan: No, I don't think it could [21:58] <Uraeus> omega_eating: thats it :) [21:58] <leviathan> Sopwith: why not? [21:58] <Sopwith> leviathan: How are you going to get the mmap'd memory t= o be used for the shm segment? [21:58] <omega_eating> leviathan: I don't think you can shm mmaps [21:58] <omega_eating> you request mmap, you request shm, you can't merge= them afaik [21:58] <wtay> you don't need to shm them, do you? [21:58] <leviathan> Sopwith: what is MAP_SHARED for than? [21:58] <omega_eating> leviathan: multi-open, for devices that support it [21:58] <leviathan> from the manpage: [21:59] <leviathan> Share this mapping with all other processes [21:59] <leviathan> that map this object [21:59] <Sopwith> leviathan: Yes, but MAP_SHARED is not sysv shm, and MAP= _SHARED will not work [21:59] <omega_eating> files implicitely support it, but the device (soun= d driver) has to a) support multi-open, and b) handle multiple mmaps [21:59] <Sopwith> what omega said :) [21:59] <leviathan> omega_eating: no soundcards i have can be opened simu= ltaneasly twice [22:00] <omega_eating> my maestro2e can be opened once for read and once = for write [22:00] <omega_eating> my sblive can be opened by 32 clients, hardware mi= xer [22:00] <omega_eating> but most cards can't [22:00] <leviathan> omega_eating: yes. my sb128 can do open for read and = for write simultanesaly by two processes to [22:00] <omega_eating> ok [22:00] thomas (th...@ad...) left irc: Read err= or to thomas[adsl-64110.turboline.skynet.be]: EOF from client [22:01] Nick change: Uraeus -> Ura_brb [22:01] <omega_eating> specifically, I mean most cards can't handle many = writers [22:01] <omega_eating> I'd hope drivers that can't handle two-client full= -duplex are getting rarer [22:01] <leviathan> all i can say: that in fact i never worked with neith= er mmap nor shm. so i don't know what works, and what not [22:01] <omega_eating> I havne't used it either ;-) [22:01] <omega_eating> but I've read up while planning to use them <g> [22:02] <leviathan> i think it would be a great feature [22:02] <leviathan> so i'll have a look on it [22:02] <leviathan> however [22:03] <omega_eating> ok, this is what I was really asking: what feature= ? <g> [22:03] thomas (th...@ad...) joined #gstreamer. [22:03] <leviathan> if shm and mmap would fit together: how would a clien= t - which has not a file handle to asd - be able to get the current playi= ng position? [22:03] <richardb> taaz: what version of gtk-doc-tools do you have? [22:03] <wtay> Sopwith: can I ask you for a feature request in glib? [22:04] <thomas> wtay: what's the difference between osssrc/sink and ossg= st ? [22:04] <omega_eating> ossgst is like pipefilter [22:04] <wtay> thomas: ossgst hijacks /dev/dsp [22:04] <Sopwith> wtay: You can but it won't do much good :) [22:04] <leviathan> he would need that for using mmaped oss-devices [22:04] <omega_eating> wtay: the RTLD_GLOBAL thing? [22:05] <wtay> yup [22:05] Nick change: omega_eating -> omega_rr [22:06] <thomas> wtay: ok, so ossgst takes stuff away from oss output and= puts into gstreamer then ? [22:06] Nick change: ajzzzz -> ajmitch [22:06] <wtay> thomas: yes [22:06] <wtay> thomas: for *old* apps [22:06] <omega_rr> thomas: yeah, you give it a commandline and it forks t= hat off, hijacking that program's use of /dev/dsp and sending the audio d= ata out a pad [22:07] <omega_rr> implemented in response to l-a-d's comments about ReWi= re [22:07] <wtay> thomas: all very experimantal of course [22:07] <thomas> of course ;) [22:07] <ajmitch> morning [22:07] <thomas> hi aj [22:07] <omega_rr> wtay: as if that disclaimer is needed for ossgst speci= fically? it's a global disclaimer <g> [22:07] <wtay> what about me FTPing a new tarball to gstreamer.net? [22:07] <omega_rr> wtay: it's reasonably sane? [22:08] <omega_rr> wondering where best to put it [22:08] <thomas> hmmm... can the xml functions in gstreamer save partial = pipelines as well ? [22:08] <wtay> omega_rr: make diskcheck fails on the check_PROGRAMS [22:08] <omega_rr> thomas: yup [22:08] <wtay> omega_rr: will try to build it first [22:08] <omega_rr> wtay: hrm, that should be a hint, but it can slide for= pre1 <g> [22:08] <thomas> cool... that should take care of writing code to handle = different kinds of input/output bits ;) [22:08] <leviathan> where cann i find docs about shm? neither do i have a= manpages, nor has libc any docs bout it [22:08] <omega_rr> thomas: yup, or the autoplugger can do it [22:08] <omega_rr> leviathan: man -k shm gives nothing? [22:09] <leviathan> omega_rr: ok. it does. did not know -k yet. thanks [22:09] <omega_rr> wtay: created gstreamer.net/releases/0.2.0/pre/ [22:09] <wtay> omega_rr: lemme build and run it first :-) [22:09] <omega_rr> leviathan: grep is such a wonderful tool ;-) [22:09] <omega_rr> wtay: detauls [22:09] <omega_rr> er, details [22:11] <wtay> uploading the tarball... not sure it works... [22:12] <omega_rr> getting pre0 into that dir [22:12] Nick change: leviathan -> lev_coding [22:12] Nick change: omega_rr -> omega_ [22:13] Topic changed on #gstreamer by ChanServ!s@ChanServ: Try out 0.2.0= -pre1: http://gstreamer.net/releases/0.2.0/pre/gstreamer-0.2.0-pre1.tar.g= z [22:14] <omega_> er, as soon as the whole tarball is there, at least..... [22:14] Action: omega_ drafts -devel mail [22:14] <ajmitch> yay, something new to break ;) [22:14] Action: omega_ takes the sledgehammer away from ajmitch [22:15] <ajmitch> you want us to test it to death again? [22:15] <wtay> it should be there now [22:15] <wtay> 2.4MB [22:18] Action: richardb goes for food [22:18] Nick change: richardb -> richardb-away [22:18] <omega_> tarbzll is only 1.8MB [22:19] <omega_> mail sent [22:19] <wtay> beware, I didn't verify it yet [22:19] <omega_> ok, that's what -devel is for <g> [22:19] <wtay> plugins and core passed without problems though [22:19] <omega_> (at least I didn't say "that's what users are for -M$" <= g>) [22:20] <wtay> man.. gstreamer is a huge beast :) [22:21] <ajmitch> what QA procedures you have ;) [22:21] <omega_> no, it's not a beast, beast is a beast. <g> [22:21] <omega_> beast/bse, that is [22:21] <wtay> heh [22:21] Action: ajmitch runs cvs update [22:22] <wtay> ajmitch: what are you saying? *you* are our QA guy <g> [22:22] <omega_> why did it jump 70KB between pre0 and pre1? [22:22] Action: omega_ gives the sledgehammer back to ajmitch [22:22] <wtay> ok, builds fine here [22:22] Action: omega_ backs away [22:22] Action: wtay does make install [22:22] <ajmitch> i'll tell you whether it builds fine in a few hours ;) [22:23] <wtay> ajmitch: we need *detailed* bug reports :-) [22:23] <wtay> installs fine... [22:23] <ajmitch> wtay: *how* detailed? all output from running with --gs= t-mask=3D-1 & a full backtrace? [22:23] <omega_> wtay: we need to add an env-var to prepend the plugin se= arch path [22:24] <omega_> ajmitch: um, no ;-) [22:24] <wtay> omega_: isn't there one allready? [22:24] <richardb-away> omega_: for making make check work? [22:24] <ajmitch> damn ;) [22:24] <omega_> ajmitch: let's put it this way: if you break the irc dai= ly log record, you'll be hunted down <g> [22:24] <omega_> richardb-away: yeah, among other things [22:24] <omega_> wtay: don't think so [22:24] <richardb-away> GST_PLUGIN_PATH [22:25] <ajmitch> omega_: hehe [22:25] Action: ajmitch warms up gdb [22:25] <wtay> omega_: gst.c line 213 [22:26] <omega_> guess so [22:26] <wtay> ok, everyhting works here :-) [22:28] Action: ajmitch needs a faster computer.... [22:34] <thomas> pre1 is compiling here... you know, it does show a lot o= f warnings throughout [22:38] <thomas> 2 of 4 tests failed [22:38] <thomas> when doing make chec [22:38] <thomas> k [22:38] <thomas> problem or not ? [22:39] <wtay> thomas: problem, don't do that [22:39] <thomas> wtay: ok, then my make worked ;) [22:39] <wtay> thomas: ah, does it work? [22:40] <thomas> well, depends, what do I need to check ? ;) [22:41] <wtay> thomas: -launch and gstmediaplay [22:42] <thomas> the mixer example works ;) [22:42] <thomas> yay, AlienSong plays ! [22:43] <wtay> cool [22:43] <thomas> ugh, that was ugly. Somebody should turn off fullscreen= option until it actually works instead of locking up X !!! [22:44] <wtay> ship it! [22:45] <thomas> ok, off to bed. hope to write a good audio cutter tomor= row [22:45] Action: thomas thanks the gods for having make uninstall in gstre= amer [22:46] thomas (th...@ad...) left irc: Client E= xiting [22:46] heroines (root@205.229.113.88) joined #gstreamer. [22:47] <ajmitch> bbl, breakfast time ;) [22:47] <wtay> hi=20 [22:47] <heroines> Here's the story on Gstreamer porting to gobject. [22:47] Action: taaz was away [22:48] <taaz> richardb-away: i have same gtk-doc-tools version you have [22:48] <wtay> heroines: ah? [22:48] <heroines> After hacking on the source code for a while we decide= d to stick to gtk. [22:48] <wtay> heroines: uh? [22:48] <heroines> But link just the gtk objects that are needed for basi= c gstreamer functionality. [22:49] <heroines> This amounts to about 6 gtk objects. [22:49] <wtay> heroines: but you'll still need X [22:50] <heroines> We took out the GtkWidget and gtk_main dependancies fr= om gstreamer so it now compiles without X. [22:50] <heroines> The next step is to test it. [22:50] <wtay> hmm [22:50] <omega_> heroines: I gave that a try twice and never had any luck [22:50] <Sopwith> heroines: What's wrong with GObject directly? [22:51] <heroines> The problems with GObject arise from the scope of the = Gtk dependancies. [22:51] <omega_> ? [22:51] <Sopwith> As in "it should be possible to switch" not "GObject is= perfect" :) [22:52] <Sopwith> unless you are using gtk widgets everywhere, it shouldn= 't be hard [22:52] <omega_> Sopwith: we only use gtkobject as the base, so nothing d= epends on X except that base object [22:52] <heroines> Things like the GtkArg, gtk_main, GtkObject flags, wou= ld require big changes to the gstreamer architecture. [22:52] <heroines> It could be done within a year but not within a month. [22:52] <Sopwith> heroines: There are equivalents in glib. You need to lo= ok a little harder [22:52] <omega_> heroines: no, they're str... [truncated message content] |
From: <wim...@ch...> - 2001-06-02 04:28:57
|
[06:34] arik (ar...@sd...) left #gstreamer. [07:02] mwc (ma...@ab...) left irc: Ping timeout for mwc[abu.ntu.edu.au] [07:21] _gst_newt_ joined #gstreamer. [08:14] plazma (pl...@cc...) left irc: Client Exiting [08:35] steveb (st...@no...) joined #gstreamer. [08:59] mwc (ma...@ab...) joined #gstreamer. [09:40] mwc (ma...@ab...) left irc: Read error to mwc[abu.ntu.edu.au]: EOF from client [10:21] Nick change: taaz -> taazzzz [10:50] walken (fo...@c1...) left irc: l8r [11:09] thomas (th...@21...) joined #gstreamer. [11:15] leviathan (le...@p3...) joined #gstreamer. [11:30] <thomas> hi everyone [11:53] Nick change: ajmitch -> ajzzzz [13:15] leviathan (le...@p3...) left irc: [x]chat [14:00] _gst_newt_ joined #gstreamer. [15:48] big_T (th...@fy...) joined #gstreamer. [16:07] <thomas> anyone know what you need to set to the lame plugin in a gstreamer-launch command line ? [16:11] dobey (do...@dr...) joined #gstreamer. [16:19] <thomas> yo dobey [16:19] <thomas> you wouldn't happen to know how to do gstreamer-launch with lame without it segfaulting ? [16:19] <dobey> yo [16:19] <dobey> no [16:20] <dobey> :-/ [16:20] <dobey> i happen to know there's going to be an encompass release sunday though [16:20] <thomas> is that your browser ? [16:21] <dobey> and i happen to know, that it may include support for bonobo-media [16:21] <dobey> yes [16:21] <dobey> ;-) [16:25] <dobey> and i happen to also know [16:25] <dobey> that it will support ghelp: [16:25] <dobey> which means i have a lot of hacking to do tonight/tomorrow [16:47] <dobey> hrmm [16:47] Action: dobey misses hadess [17:03] Uraeus (csc...@c2...) joined #gstreamer. [17:04] <dobey> Uraeus: !!!!!!!1 [17:04] <Uraeus> dobey! [17:04] <dobey> dude [17:04] <dobey> heh [17:04] <Uraeus> dobey: cool news, Creatures 3 for Linux is GNOME based :) [17:04] <dobey> what's that? [17:05] <Uraeus> http://ds.creatures.net/ [17:05] <dobey> i have cooler news [17:05] <Uraeus> oko [17:05] <dobey> oh [17:05] <dobey> "that" [17:05] <Uraeus> ? [17:05] <dobey> encompass 0.5.0, sunday [17:06] <Uraeus> hehe, cool, gotten any usefull patches from your new developers for it? [17:09] <dobey> no [17:09] <dobey> heh [17:10] <Uraeus> dobey: the french guy didn't mail back after you sent him some fix suggestions for his code? [17:10] <dobey> Uraeus: i didn't get a chance to mail him back yet [17:11] <dobey> Uraeus: and this release won't have proxy support anyway [17:13] <Uraeus> dobey: remember they depend upon you, if you want them onboard you have to take care of 'em ;) [17:13] <dobey> yeah [17:13] <dobey> <- really really really really busy :-/ [17:16] <Uraeus> dobey <- really really really really overworked probably ? [17:16] <dobey> heh [17:16] <dobey> nah [17:16] <dobey> mandrake is just a bitch [17:20] <Uraeus> heh, never used mandrake, but nice to see mandrake hackers active on Nautilus [17:21] <dobey> heh [17:21] <dobey> 1 of them [17:22] <Uraeus> no 2 I think [17:22] <dobey> oh [17:23] <dobey> i only know of fcrozat doing that [17:23] thomas (th...@21...) left irc: [x]chat [17:46] wtay-zZz (wi...@ca...) left irc: Client Exiting [17:50] arnd (ar...@pr...) joined #gstreamer. [17:51] steveb (st...@no...) joined #gstreamer. [17:56] arnd (ar...@pr...) left irc: using sirc version 2.211+4KSIRC/1.1 [17:58] <Uraeus> hi |
From: <wim...@ch...> - 2001-06-03 04:40:15
|
[06:34] arik (ar...@sd...) left #gstreamer. [07:02] mwc (ma...@ab...) left irc: Ping timeout for mwc[abu.n= tu.edu.au] [07:21] _gst_newt_ joined #gstreamer. [08:14] plazma (pl...@cc...) left irc: Client = Exiting [08:35] steveb (st...@no...) joined #gstreamer. [08:59] mwc (ma...@ab...) joined #gstreamer. [09:40] mwc (ma...@ab...) left irc: Read error to mwc[abu.ntu.= edu.au]: EOF from client [10:21] Nick change: taaz -> taazzzz [10:50] walken (fo...@c1...) left irc: l8r [11:09] thomas (th...@21...) joined #gstreamer. [11:15] leviathan (le...@p3...) joined #gstreamer. [11:30] <thomas> hi everyone [11:53] Nick change: ajmitch -> ajzzzz [13:15] leviathan (le...@p3...) left irc: [x]chat [14:00] _gst_newt_ joined #gstreamer. [15:48] big_T (th...@fy...) joined #gstreamer. [16:07] <thomas> anyone know what you need to set to the lame plugin in a= gstreamer-launch command line ? [16:11] dobey (do...@dr...) joined #gstreamer. [16:19] <thomas> yo dobey [16:19] <thomas> you wouldn't happen to know how to do gstreamer-launch w= ith lame without it segfaulting ? [16:19] <dobey> yo [16:19] <dobey> no [16:20] <dobey> :-/ [16:20] <dobey> i happen to know there's going to be an encompass release= sunday though [16:20] <thomas> is that your browser ? [16:21] <dobey> and i happen to know, that it may include support for bon= obo-media [16:21] <dobey> yes [16:21] <dobey> ;-) [16:25] <dobey> and i happen to also know [16:25] <dobey> that it will support ghelp: [16:25] <dobey> which means i have a lot of hacking to do tonight/tomorro= w [16:47] <dobey> hrmm [16:47] Action: dobey misses hadess [17:03] Uraeus (csc...@c2...) joined #gstreamer. [17:04] <dobey> Uraeus: !!!!!!!1 [17:04] <Uraeus> dobey! [17:04] <dobey> dude [17:04] <dobey> heh [17:04] <Uraeus> dobey: cool news, Creatures 3 for Linux is GNOME based := ) [17:04] <dobey> what's that? [17:05] <Uraeus> http://ds.creatures.net/ [17:05] <dobey> i have cooler news [17:05] <Uraeus> oko [17:05] <dobey> oh [17:05] <dobey> "that" [17:05] <Uraeus> ? [17:05] <dobey> encompass 0.5.0, sunday [17:06] <Uraeus> hehe, cool, gotten any usefull patches from your new dev= elopers for it? [17:09] <dobey> no [17:09] <dobey> heh [17:10] <Uraeus> dobey: the french guy didn't mail back after you sent hi= m some fix suggestions for his code? [17:10] <dobey> Uraeus: i didn't get a chance to mail him back yet [17:11] <dobey> Uraeus: and this release won't have proxy support anyway [17:13] <Uraeus> dobey: remember they depend upon you, if you want them o= nboard you have to take care of 'em ;) [17:13] <dobey> yeah [17:13] <dobey> <- really really really really busy :-/ [17:16] <Uraeus> dobey <- really really really really overworked probably= ? [17:16] <dobey> heh [17:16] <dobey> nah [17:16] <dobey> mandrake is just a bitch [17:20] <Uraeus> heh, never used mandrake, but nice to see mandrake hacke= rs active on Nautilus [17:21] <dobey> heh [17:21] <dobey> 1 of them [17:22] <Uraeus> no 2 I think [17:22] <dobey> oh [17:23] <dobey> i only know of fcrozat doing that [17:23] thomas (th...@21...) left irc: [x]chat [17:46] wtay-zZz (wi...@ca...) left irc: Clien= t Exiting [17:50] arnd (ar...@pr...) joined #gstreamer. [17:51] steveb (st...@no...) joined #gstreamer. [17:56] arnd (ar...@pr...) left irc: using =02sirc=02 ve= rsion 2.211+=034KSIRC=03/1.1 [17:58] <Uraeus> hi [13:03] _gst_newt_ joined #gstreamer. [13:04] <ajmitch> the newt is back... [13:05] <steveb> gotta pay the rent [13:05] Nick change: steveb -> steveb-away [13:16] Nick change: wtay-zZz -> wtay [13:17] <wtay> 'morning [13:17] <ajmitch> morning wtay [13:17] <ajmitch> (even tho it's evening here still ;) ) [13:18] <wtay> well, it's afternoon here actually :) [13:18] makki (yv...@ma...) joined #gstreamer. [13:18] <wtay> yo [13:19] <makki> io [13:19] <ajmitch> hi makki [13:22] <ajmitch> wtay: any new issues with gstreamer that have come up f= or us to fix? ;) [13:24] <wtay> ajmitch: ? [13:24] <ajmitch> anything i can do to gstreamer tonight? [13:25] <ajmitch> before i turn my attention to a cool looking game calle= d xtux? [13:27] <makki> is there a working divx codec for gstreamer ? [13:27] <wtay> ajmitch: does everything work? [13:27] <wtay> makki: working on that now [13:27] <wtay> makki: divx video should play fine in yesterdays CVS [13:27] <makki> avifile based ? [13:28] <wtay> nope, own implementation [13:28] <ajmitch> wtay: no, am recompiling now, but will test that infamo= us ganso_terminator.mpg (does it work for you?) [13:28] <wtay> ajmitch: where can I get it? [13:30] <ajmitch> wtay: gstreamer.net/media/ganso_terminator.mpg [13:30] <wtay> ajmitch: ok [13:30] <ajmitch> omega_ uploaded it after i sent it to him ;) [13:31] <wtay> hmm, I need to chmod it first... [13:32] <steveb-away> anyone know about /usr/lib/vfsConf.sh and why I don= 't have it? (using gnome-vfs ximian rpm) [13:32] Nick change: steveb-away -> steveb [13:32] <wtay> steveb: you have -devel too? [13:32] <steveb> no [13:33] <ajmitch> it's in libgnome-vfs-dev on debian ;) [13:34] <ajmitch> i'm guessing gnome-vfs-devel for ximian rpms [13:34] <steveb> -devel fixed it [13:34] <steveb> ta [13:34] <ajmitch> it's been awhile since i had to put up with an rpm-base= d system ;) [13:35] <steveb> i need to upgrade, but have nothing to do backups to [13:35] <steveb> grr - damn laptops [13:35] <wtay> grr, only omega can read ganso_terminator and I cannot chm= od it [13:36] <ajmitch> heh [13:36] <ajmitch> want me to dcc it to you? [13:36] <ajmitch> (from my 33.6k modem) [13:36] <wtay> ajmitch: just mail it to me [13:36] <wtay> ajmitch: or try dcc [13:37] <ajmitch> -rw------- 1 omegahac gstreame 591193 May 31 15:10= ganso_terminator.mpg [13:37] <ajmitch> doh, i see what you mean [13:38] <wtay> 2mins.. not bad [13:38] <ajmitch> funny how SF is setup - i'm not a member of your projec= t yet i can still look around the gstreamer dir [13:39] <steveb> is --enable-plugin-srcdir on by default in 0.2.0 tarball= ? [13:41] <wtay> steveb: nope [13:41] <steveb> good [13:41] <ajmitch> wtay: i suspect that it will play fine for you [13:42] <ajmitch> wtay: i think it's more to do with what my system can s= upport :( [13:42] <wtay> ajmitch: it does play fine [13:42] <ajmitch> damn [13:43] <ajmitch> well, we need to stop it crashing on crappy systems ;) [13:44] <wtay> yes [13:46] <steveb> is there some way to make the XvQueryExtension test fail= if there is no shared version of Xv? [13:48] <wtay> I don't think so [13:49] <steveb> oh well :) [14:02] <ajmitch> wtay: think it's worthwhile having a configure option t= o disable tests & examples? [14:03] <wtay> yeah, why not.. [14:03] <ajmitch> ok, cos i seem to waste half the compile time on those = ;) [14:03] <ajmitch> (and i have a damned slow machine for compiling) [14:03] Action: steveb waits for tests and examples to build [14:09] <steveb> yay, just played AlianSong for the first time with gstpl= ay [14:10] <steveb> i guess all i needed was Hermes [14:13] <ajmitch> what should i call the configure option? [14:13] <wtay> --disable-tests and --disable-examples? [14:14] <ajmitch> ok [14:17] <steveb> should thx.vob work with gstmediaplay? [14:17] <wtay> yes [14:17] <steveb> does it need any external libs? [14:17] <wtay> mpeg2dec is needed [14:17] <ajmitch> if sound, then some ac3 libs, if video, mpeg2dec [14:18] <ajmitch> or what wtay says (he knows more than me ;) ) [14:18] <wtay> nope, ac3 libs are not needed :) [14:18] <ajmitch> ah, that is good ;) [14:18] <wtay> for some definitions of good <g> [14:19] <steveb> hmm, mpeg2dec no such element [14:19] <ajmitch> wtay: if i send in patches, i should remove all the com= ments i add about my stuff being ugly hacks? [14:20] Action: steveb wgets [14:20] <wtay> ajmitch: if they are, you shouldn't remove them :-) [14:20] <ajmitch> hehe [14:20] <ajmitch> yes, i should leave in all the //FIXME comments ;) [14:20] <wtay> just make sure we can verify that you didn't add a backdoo= r :) [14:21] <ajmitch> oh, so i can't run it thru cobfusc? ;) [14:22] <wtay> ehm.. no :) [14:23] <ajmitch> richardb mustn't be away too far, he's committing chang= es ;) [14:29] Action: ajmitch starts to get a little paranoid when he sees the = Netcraft commentary saying that quite a lot of ecommerce sites running on= IIS are possibly compromised already [14:38] Action: ajmitch stops doing things the stupid way & edits the top= level Makefile.am instead... [14:56] makki (yv...@ma...) left irc: Ping timeout for makk= i[makki.wait-a-bit.com] [15:00] <ajmitch> anyone still around? [15:00] <wtay> yup [15:01] <ajmitch> wtay: ok, --disable-tests stops test/, tests/, and test= suite/ from being built=20 [15:01] <wtay> hacking on avi decoder (reworking it a 3th time) [15:01] <wtay> ajmitch: cool [15:01] <ajmitch> --disable-examples stops examples/ from being built [15:02] <ajmitch> shall i send you patch to configure.base & Makefile.am? [15:02] <wtay> ajmitch: shall I give you cvs write access? [15:02] <wtay> or is that too dangerous? :-) [15:03] <ajmitch> you can if you want, since i'll probably want to commit= plugins at some stage [15:03] <wtay> ok, sec... [15:04] <wtay> ajmitch is the SF name? [15:04] <ajmitch> am testing that patch doesn't break normal build first.= .. ;) [15:04] <ajmitch> yup [15:05] <wtay> ajmitch: I added you [15:06] <ajmitch> wtay: ok, i guess i have to check gstreamer out again n= ow ;) [15:06] <wtay> yup, follow the instructions on gstreamer.net [15:09] ajmitch_ (aj...@p3...) joined #gstreamer. [15:09] ajmitch (aj...@p6...) left irc: Ping timeout f= or ajmitch[p6-max5.dun.ihug.co.nz] [15:09] <ajmitch_> oh how i love my ISP [15:09] <ajmitch_> wtay: say anything important to me in the last couple = of minutes? ;) [15:09] Nick change: ajmitch_ -> ajmitch [15:10] <wtay> naah, I would never do that :) [15:10] <ajmitch> hehe [15:14] <ajmitch> damn automake is picky on syntax - i am running thru an= d having to add a space after an =3D, etc ;) [15:16] <ajmitch> wtay: there's no ChangeLog to maintain? [15:17] <wtay> ajmitch: nope [15:17] <wtay> cvs commit log is out changelog [15:17] <wtay> s/out/our [15:18] <ajmitch> ok [15:18] <ajmitch> yeah, the cvs commit mailing list is quite helpful, btw= ;) [15:47] makki (yv...@ma...) joined #gstreamer. [15:54] <steveb> hmm, gstplay.c doesn't compile - parse error in gst_play= _init [15:54] <ajmitch> not good... [15:55] <ajmitch> damnt his computer is slow when running autogen.sh ;) [15:56] <ajmitch> steveb: ok, i have no probs compiling it [15:56] <ajmitch> paste the error? [16:01] xie (xie@61.135.49.30) joined #gstreamer. [16:01] <xie> Hello [16:01] <steveb> ajmitch: never mind - me stupid [16:01] <ajmitch> hi xie [16:01] <ajmitch> steveb: hehe, what was it? [16:02] <steveb> a local hack which didn't like the latest cvs update [16:03] <ajmitch> ah, one of them ;) [16:06] <ajmitch> hmm, what is happening with the volume plugin? is it be= ing dropped in 0.2.0 in favour of volenv only? [16:06] <ajmitch> xie: how can we help, btw? ;) [16:07] <xie> sorry..I just listening..;-) [16:07] <xie> bye [16:07] <ajmitch> xie: quite a few ppl asleep or away at the moment i thi= nk ;) [16:07] xie (xie@61.135.49.30) left #gstreamer. [16:07] <ajmitch> hmm [16:12] Action: ajmitch sleep now [16:12] <ajmitch> night all [16:12] Nick change: ajmitch -> ajzzzz [16:22] dobey (do...@ip...) joined #gstreamer. [16:38] <wtay> yo [16:38] Nick change: wtay -> wtay-away [16:48] greg_ (gr...@ho...) joined #gstreamer. [16:59] greg_ (gr...@ho...) left irc: [x]chat [16:59] hadess (ha...@pc...) joined #gst= reamer. [17:00] <hadess> yo gang [17:01] <dobey> hey slut [17:01] <hadess> hey dobish [17:01] <dobey> dude [17:02] <ajzzzz> damn automake, i curse the day i ever touched it... [17:02] Action: ajzzzz had to get up & check the build [17:03] <dobey> hadess: you have flash on yer imac? [17:03] <ajzzzz> hi hadess, dobey [17:03] <hadess> dobey: nope [17:04] <dobey> hmm [17:06] <dobey> http://primates.ximian.com/~dobey/encompass-ghelp.png [17:07] <hadess> cool [17:07] Action: ajzzzz wonders what he broke with trying to disable build= ing of tests [17:18] Nick change: hadess -> hds-afk [17:26] <makki> yves@portapunch:~/src/gstreamer$ sh autogen.sh=20 [17:26] <makki> Checking for patched automake... not found. [17:26] <makki> Detected automake version 1.4 (or near) without patch. [17:26] <makki> Your version of automake needs a patch applied in order t= o operate correctly. [17:26] <makki> euh, great... :-) [17:26] <dobey> README [17:26] <hds-afk> wicked stuff [17:27] Nick change: hds-afk -> hadess [17:27] <ajzzzz> dammit, i can just about go to sleep after i commit my n= ice fixup (to my mistake, of course...) [17:31] <makki> yeah, but the automake is trashing my system... [17:31] <makki> 160M memory [17:33] <makki> 12272 yves 9 0 212M 138M 129M D 1.6 74.2 = 0:25 automake [17:33] <hadess> makki: that's normal [17:33] Action: dobey smacks ghelp: [17:33] <makki> it's the first time i see this behaviour [17:33] netnoise (ni...@rm...) joined #gstreamer. [17:34] <hadess> makki: it's the first time you have makefile.am's with s= o much conditionals... [17:35] <ajzzzz> hadess: plus i added a couple more conditionals tonight = ;) [17:35] <hadess> ajzzzz: right ;) [17:36] <ajzzzz> hadess: you now have the option to --disable-tests and -= -disable-examples [17:36] netnoise (ni...@rm...) left irc: Read error = to netnoise[rm2-85.dialup.tiscalinet.it]: Connection reset by peer [17:36] <hadess> ajzzzz: neat [17:36] Nick change: wtay-away -> wtay [17:36] <wtay> hi again [17:37] <wtay> incsched is f*cking cool :-) [17:37] <hadess> hey wtay [17:37] <ajzzzz> wtay: umm, i know a bit more about autoconf now, after m= aking a few mistakes ;) [17:37] <wtay> ajzzzz: are you the "funny CVS log messages" guy now? :-) [17:37] Action: ajzzzz lets wtay hit ajzzzz a few times with the cluestic= k... [17:37] <ajzzzz> wtay: umm, if you want ;) [17:38] netnoise (ni...@rm...) joined #gstreamer. [17:38] makki (yv...@ma...) left irc: Ping timeout for makk= i[makki.wait-a-bit.com] [17:38] <hadess> wtay: i am a pierced hacker now ;P [17:38] <wtay> hadess: show us! [17:39] <dobey> hah [17:39] <hadess> wtay: accessing my webcam freezes the puter, i'm 'piling= a new kernel [17:39] <ajzzzz> anyway, night all, i gotta get up in about 4 hours [17:39] <wtay> ajzzzz: cya [17:39] <dobey> heh [17:40] Action: ajzzzz sleeps [17:40] <hadess> cya ajzzzz [17:41] <hadess> wtay: do you know which files contains the version numbe= r for the kernel ? [17:43] <hadess> found it... [17:46] Nick change: taazzzz -> taaz [17:47] netnoise (ni...@rm...) left irc: Ping timeou= t for netnoise[rm2-85.dialup.tiscalinet.it] [17:47] <wtay> yo taaz [17:47] <taaz> hey [17:47] <hadess> hey taaz [17:47] <taaz> time to play with matrox g400 tool [17:51] makki (yv...@ma...) joined #gstreamer. [17:51] <makki> grmbl [17:51] <makki> that automake ate all my resources [17:52] Action: taaz is not amused... a fish just swam across my desktop.= .. [17:53] <wtay> taaz: heh you too? :) [17:53] <makki> hadess: still here ? [17:53] <hadess> yup [17:55] <taaz> seriously... i don't think any of my running processes sho= uld be putting fish on my desktop. [17:56] <makki> hadess: that automake ate all my resources by the way... = something is definitely wrong [17:57] <hadess> taaz: heh, that's the panel ;) [17:57] <hadess> makki: you need at least 128 megs of ram, and as much of= swap [17:57] <taaz> hadess: i see no "disable random fish" option [17:57] <hadess> taaz: that's an easter egg [17:57] <taaz> makki: didn't the autogen.sh test detect you have an old a= utomake? [17:58] <taaz> err... not old... unpatched [17:58] <makki> taaz: it detected an old makefile and patched it yes [17:58] <hadess> makki: old automake you mean [17:59] <makki> hadess: i have 200M of ram and 500M of swap [17:59] <makki> an old automake yes [17:59] <makki> sorry [18:02] <hadess> makki: i have 320 megs of ram and 128 of swap, and my sy= stem doesn't die, even unpatched [18:03] <makki> my automake is 1.4-p2-1 and dates 2001-05-24 (from Chang= eLog) [18:05] <hadess> same here [18:05] <hadess> just close all your big apps, and run the autogen.sh aft= er [18:07] <makki> oh, the autogen doesn't use the patched version ? [18:07] <taaz> jeez... just patch automake and then there's no problem. = what's so hard about this? [18:08] <makki> taaz: it wasn't clear to me if it did the automake-patchi= ng automagically... [18:08] <hadess> makki: it doesn't [18:08] <makki> hadess: yes, i see that now [18:19] hadess (ha...@pc...) left irc: r= eboot [18:24] hadess (ha...@pc...) joined #gst= reamer. [18:25] <hadess> i guess that the usb drivers don't like my webcam too mu= ch [18:25] Apoc (ep...@pp...) joined #gstreamer. [18:26] <hadess> "usb-ohci.c: OHCI Unrecoverable Error, controller usb-01= :18.0 disabled" [18:26] <Apoc> Yo=20 [18:27] <hadess> yo Apoc [18:27] <wtay> yo [18:29] <Apoc> I have a problem with the pre1 release ... [18:29] <wtay> Apoc: ? [18:30] <Apoc> A program which run fine with 0.1.1, coredump with the pre= 1 .. [18:30] <wtay> which one? [18:31] <Apoc> One of mine with cdparanoia plugin ... [18:31] <Apoc> ** CRITICAL **: file gstelement.c: line 778 (gst_element_s= et_state): assertion `element->sched !=3D NULL' failed=20 [18:31] <wtay> Apoc: you need a pipeline as the top element [18:31] <wtay> a bin doesn't have a scheduler [18:32] <Apoc> That's a change from the 0.1.1 ?? [18:32] <wtay> yup [18:33] <wtay> major scheduler changes [18:33] <Apoc> Ok thanks ... I'm going to try :) [18:33] <wtay> ok, check back here if that didn't work [18:34] <Apoc> wtay : Thanks ... [18:36] <Apoc> wtay : The helloworld example doesn't work too :( [18:37] <wtay> hmm, it uses a bin too... [18:38] <wtay> Apoc: fixed in CVS [18:40] <Apoc> wtay : thanks ... [18:53] <hadess> wtay: seems the ohci driver doesn't like my webcam :/ [18:56] <wtay> hadess: v4lsrc? [18:57] <hadess> wtay: nope, every tool i tried [18:57] <wtay> uhm [18:57] <hadess> "usb-ohci.c: OHCI Unrecoverable Error, controller usb-01= :18.0 disabled" [18:58] <wtay> wow [18:58] <wtay> that sucks [18:58] <wtay> it used to work, no? [18:58] Zeenix (pro...@ho...) joined #gstreamer. [18:58] <hadess> yep, i guess if i switched back to an older kernel... [18:58] <Zeenix> hi [18:58] <hadess> hey [18:59] <Zeenix> i istalled the libgsmxxx.rpm, but there isnt any 'man gs= m' [19:00] <wtay> yo [19:00] <wtay> Zeenix: there are docs on the site [19:01] <Zeenix> where ? [19:02] <wtay> http://kbs.cs.tu-berlin.de/~jutta/toast.html [19:05] <Zeenix> wtay: if you take 50,000 shorts from /dev/dsp, using AFM= T_SE16_LE , & direct it to a file, what size you expect the ouput file ? [19:06] <wtay> 2500 bytes [19:06] <wtay> er 10000 bytes sorry [19:07] <Zeenix> but i get a 80,000 bytes file [19:07] <Zeenix> err 180KB file [19:08] <wtay> it should be 100K [19:08] <wtay> 50000 shorts is 100000 bytes [19:09] <wtay> make sure you don't use staero [19:09] <wtay> er stereo [19:09] <Zeenix> no i use mono [19:10] <Zeenix> i'll see if something isnt wrong in my code [19:10] <wtay> then it should be exectly 100,000 bytes [19:11] <Zeenix> wtay: i tried 'toast sample', it did compressed it to 10= % of original but at much slower speed than you told me that day [19:12] <wtay> Zeenix: you don't have a 1.2GHz PC I guess :-) [19:13] <Zeenix> a Cirex-333Mhz with 64MB RAM [19:13] <Zeenix> it actually works like a 250MHz [19:13] <wtay> that's about 8 times slower I think [19:15] <Zeenix> wtay: why you said, you have make an app. to test it if = you knew about the toast command [19:15] <Zeenix> ? [19:15] <wtay> Zeenix: lazyness :) [19:16] <wtay> and I just knew about the URL because I did a google searc= h [19:18] <Apoc> wtay : Thanks it works :) [19:18] <wtay> Apoc: ;-) [19:19] Action: hadess installs HM&M3 [19:19] <wtay> HM&M3? [19:20] <dobey> ok [19:20] <dobey> who wants to help me do some string munging? [19:21] <wtay> dobey: I can try... [19:21] <dobey> so, ok [19:22] <dobey> hrmm.. [19:22] <hadess> wtay: heroes of might and magic 3 [19:23] <dobey> ok, i need to go back to the second / from the end of the= string... [19:23] <wtay> hmm [19:23] Nick change: hadess -> hds-busy [19:23] <dobey> i think [19:24] <dobey> hds-busy: it works on ppc?! [19:25] <hds-busy> dobey: sure does [19:25] <Zeenix> wtay: failed to find anything on http://kbs.cs.tu-berli= n.de/~jutta/toast.html [19:26] <hds-busy> dobey: railroad tycoon does as well [19:26] <dobey> hrmm [19:26] <dobey> neat [19:26] <dobey> i wish UT/q3 did [19:26] <dobey> or something [19:28] <dobey> damnit! [19:28] <dobey> file:/usr/share/gnome/help/panel/C/figures/launcher_prope= rties_basic.png [19:28] <dobey> file:/usr/share/gnome/help/panel/C/figures/figures/launch= er_properties_advanced.png [19:28] <dobey> i keep having that happent [19:29] <dobey> (passed url is ./figures/...) [19:29] <Zeenix> wtay: where will i find 'man gsm' [19:29] <dobey> rpm -ql libgsm|grep man [19:30] <Zeenix> dobey: it has three man page: toast, untoast & another o= ne [19:31] <dobey> Zeenix: then you type man <those> [19:31] <dobey> man toast [19:31] <dobey> man untoast [19:31] <dobey> man other one [19:31] <Zeenix> dobey: i have done that [19:31] <wtay> lemme check [19:32] <Zeenix> but at those manual pages, i see gsm(3) in see also sect= ion [19:32] <dobey> i don't even have man strchr so don't ask me [19:32] <dobey> man 3 gsm [19:32] <wtay> http://kbs.cs.tu-berlin.de/~jutta/gsm/gsm-3.html [19:32] <dobey> or man gsm 3 [19:32] <Zeenix> but there is no man gsm [19:33] <wtay> Zeenix: look at the URL, it has everything [19:33] <Zeenix> when i do rpm -qpl libgsmxxx.rpm, i dont see any man. fo= rm gsm in the list [19:33] <wtay> well, it is the man [19:34] <wtay> vi plugins/gsm/gsmdec.c is also sortof a man :-) [19:35] netnoise (ni...@rm...) joined #gstreamer. [19:35] <Zeenix> which rpm istalls that [19:35] <wtay> Zeenix: http://kbs.cs.tu-berlin.de/~jutta/gsm/gsm-3.html [19:36] <wtay> Zeenix: that's man page you're looking for [19:38] <Zeenix> wtay: thanks [19:39] <Zeenix> wtay: do i need somethin else to use gsm with gstreamer = ? [19:41] <wtay> Zeenix: nope [19:41] netnoise (ni...@rm...) left #gstreamer. [19:47] <Zeenix> wtay: for 8Mhz/short/mono, is 'toast -l sample' correct = ? [19:47] <Apoc> Bye=20 [19:47] <wtay> Apoc: cya [19:47] <wtay> Zeenix: I think so.. [19:48] Apoc (ep...@pp...) left #gstreamer (Cli= ent Exiting). [19:48] <Zeenix> wtay: would you mind to check it plz [19:49] <wtay> -l (linear) 8 kHz, 16 bit signed linear encod= ing in [19:49] <wtay> host byte order with 13 significant bits (fi= le suf=AD [19:49] <wtay> fix .l) [19:51] <Zeenix> i read that man, but i asked you, as i can be sure=20 [19:51] <wtay> I think it's ok [19:52] <Zeenix> host byte order has something to do with endian prob. ? [19:52] <wtay> yup [19:52] <Zeenix> does it mean that it solves that prob. for you [19:52] <dobey> how would one get the string before a strrchr()? [19:52] <wtay> most likely yes [19:53] <wtay> dobey: use memcpy [19:53] <dobey> ehh [19:54] <wtay> dobey: or strncpy [19:54] <dobey> ? [19:54] <wtay> dobey: to copy the bytes out of the original string? [19:54] <dobey> if i do strrchr(url, '/'); it returns everything after an= d including the last / in the url [19:54] <dobey> i need everything before the last / in the url [19:55] <wtay> dobey: yup, a pointer in the orig string where the / start= s [19:55] <wtay> so strncpy (orig, found-orig) gives you the stuff before i= t [19:55] <dobey> hrmm [19:55] <wtay> no nice, I know... [19:56] <wtay> s/no/not [19:57] <dobey> uhm [19:58] <dobey> strncpy() needs more than 2 args [19:58] <wtay> hmm, yeah, you need to alloc a space for the new string [19:59] Action: dobey sees what glib can do for him [20:01] <wtay> dobey: glib has nice g_strdup_printf [20:01] <dobey> yes [20:01] omega_breakfast (om...@om...) joined #gstreamer. [20:02] <wtay> yo [20:02] <omega_breakfast> yo [20:02] Zeenix (pro...@ho...) left irc: Ping timeou= t for Zeenix[host-106.netzone.net.pk] [20:05] Zeenix (pro...@ho...) joined #gstreamer. [20:05] <omega_breakfast> wtay: copying the padtemplate from the pad to i= ts ghostpad simply writes the padtemplate pointer over with its own conte= nts [20:05] <Zeenix> wtay: sorry got disconnected for a while [20:05] <wtay> omega_breakfast: yes [20:05] <omega_breakfast> GST_PAD_PADTEMPLATE is GST_PAD_REALIZE(pad)->pa= dtemplate iirc [20:06] <Zeenix> wtay: would you like to repeat your last message(s) for = me [20:06] <wtay> omega_breakfast: the padtemplate is in GstPad [20:06] <omega_breakfast> wtay: ? [20:06] <wtay> Zeenix: I didn't say anything [20:06] <omega_breakfast> reading [20:07] <omega_breakfast> ummm, ok [20:07] <wtay> omega_breakfast: I wasn't sure I could change the semantic= s of GST_PAD_X (using a realize) [20:07] <omega_breakfast> is there a reason for it to be per-[ghost]pad ? [20:07] <omega_breakfast> a ghostpad should be invisible to the system, f= or the majority of purposes [20:07] <wtay> omega_breakfast: I wasn't sure of that either [20:07] <omega_breakfast> which should include padtemplates, imo [20:07] <Zeenix> wtay: did you got the last question i asked you ? [20:08] <omega_breakfast> only thing it has is a different name and priva= te pointer [20:08] <wtay> Zeenix: no [20:08] <wtay> omega_breakfast: ok, then it should be moved to GstRPad [20:08] <omega_breakfast> wtay: yup [20:08] <omega_breakfast> I thought it was there originally.. [20:08] <wtay> so did I until it failed :) [20:09] <omega_breakfast> hmm [20:09] <Zeenix> wtay: i asked that what if i make a 1.2GHz system a requ= irement for my app. [20:09] <Zeenix> wtay: whats your opinion ? [20:09] <wtay> Zeenix: yeah, you should do that [20:09] <dobey> bleh [20:09] <omega_breakfast> Zeenix: which app? [20:11] <Zeenix> omega_breakfst: i am aiming to build an app. for Interne= t Voice Chat [20:11] <wtay> omega_breakfast: I'm having fun with AVI now :) incsched r= ules! [20:11] <Zeenix> wtay: i think when i'll complete my app, 1.2GHz would be= come very common [20:12] <omega_breakfast> 'internet phone' shouldn't require much power a= t all, unless you're doing video [20:12] <omega_breakfast> wtay: did you ever doubt? <g> [20:13] <Zeenix> omegaxxx: it would in our country [20:13] <omega_breakfast> Zeenix: huh? [20:13] <Zeenix> our net connection never goes any further than 3KB/s [20:14] <omega_breakfast> KB/s or Kb/s ? [20:14] <Zeenix> omegaxxx: why dont you keep a simple & nice nic for your= self [20:14] <omega_breakfast> btw, what country is that, anyway? .pk doesn't= ring any bells [20:14] <Zeenix> KB/s [20:14] <omega_breakfast> Zeenix: ? [20:15] <wtay> I got to go now... cya later [20:15] <Zeenix> its not my country, but i live in pakistan [20:15] <omega_breakfast> wtay: ugh [20:15] Nick change: wtay -> wtay-away [20:15] <omega_breakfast> ok [20:15] <omega_breakfast> so you want to do voice-only chat? [20:16] <Zeenix> omegaxx: thats the primary objective [20:17] <Zeenix> i'll think about video later, if it gets to any success [20:17] <omega_breakfast> ok. but why 1.2GHz? lame can do 24kbps on a 2= 00MHz afaict [20:17] <makki> i have a design-question [20:17] <omega_breakfast> vorbis is heavier, but that's because it's not = as mature [20:17] Nick change: omega_breakfast -> omega_ [20:17] <makki> why is there a need for /etc/gstreamer created by a scrip= t ? [20:18] <makki> shouldn't those go into ~/gstreamer or something ? [20:18] <omega_> makki: a script? the -register program creates it, othe= rwise startup of each application takes a long time, because it has to lo= ad all the plugins to find out what's available [20:25] _gst_newt_ joined #gstreamer. [20:27] steveb (st...@no...) got netsplit. [20:27] omega_ (om...@om...) got netsplit. [20:27] wtay-away (wi...@ca...) got netsplit. [20:27] thomas (th...@ad...) got netsplit. [20:27] dobey (do...@ip...) got netsplit. [20:27] big_T (th...@fy...) got netsplit. [20:27] <makki> omega_evm: is gstreamer-register responsible for creating= /etc/gstreamer/compreg.xml ? [20:29] <richardb-away> makki: /etc/gstreamer/compreg.xml is created by g= streamer-compprep [20:30] Nick change: richardb-away -> richardb-near [20:30] <richardb-near> Its used by gstreamer-complete. [20:30] <makki> richardb-near: yes, because gstreamer-complete segfaults = and strace tells me it segfailts on that file [20:33] wtay-away (wi...@ca...) returned to #g= streamer. [20:33] omega_ (om...@om...) returned to #gstreamer. [20:33] steveb (st...@no...) returned to #gstreamer. [20:33] <omega_> yo [20:34] dobey (do...@ip...) returned to #gstrea= mer. [20:34] <makki> re [20:35] thomas (th...@ad...) returned to #gstre= amer. [20:38] big_T (th...@fy...) got lost in the net-split. [20:46] <makki> omega_: seen my error on that chains thing in the schedul= er ? [20:46] <omega_> yeah, I must have disconnected before you got my reply [20:46] <omega_> -launch requires a commandline, that's what it does [20:47] <omega_> i.e. gstreamer-launch fakesrc ! identity ! fakesink [20:47] <omega_> I committed a fix to cvs that will catch the case where = there's no pipeline on the cmdline and print an error [20:47] <makki> aha, gonna check that out first :-) [20:49] <makki> omega_: could it be there is no Makefile.am in /component= s ? [20:50] <omega_> probably, that code is depracated afaik [20:55] <makki> mmh [20:55] <makki> config.status: error: cannot find input file: plugins/avi= /winaudio/Makefile.in [20:55] <dobey> uh [20:55] <dobey> yeah [20:55] <omega_> heh, wtay didn't finish his checkins ;-( [20:55] <makki> seems like it [20:55] <makki> that /winaudio is empty [20:56] Action: omega_ will thunk wtay for you when he comes back [20:57] <hadess> bad boy didn't finish checkins before leaving [20:57] Topic changed on #gstreamer by ChanServ!s@ChanServ: Try out 0.2.0= -pre1: http://gstreamer.net/releases/0.2.0/pre/gstreamer-0.2.0-pre1.tar.g= z [20:57] <makki> np [21:05] omega_evm (om...@bo...) left irc: Pin= g timeout for omega_evm[boisdslgw5poolb76.bois.uswest.net] [21:08] ChiefHighwater (pa...@su...) joined #gstre= amer. [21:21] dobey (do...@ip...) left irc: eh [21:21] dobey (do...@ip...) joined #gstreamer. [21:36] hadess (ha...@pc...) left irc: p= ub [21:41] walken (fo...@c1...) joined #gstreame= r. [21:41] <omega_> yo [21:41] Action: omega_ has a question [21:42] <walken> yop [21:42] Action: walken has an answer [21:42] <walken> but I dont know if it matches your question [21:42] <omega_> in the escape code handler, you do a [21:42] <omega_> val =3D (val + ~SBITS (val, 1)) | 1; [21:42] <walken> 42 ! [21:42] <omega_> that seems rather odd. is is, um, part of oddification? [21:42] <omega_> no, actually, -45 [21:42] <omega_> (at least for the flc in question) [21:43] <omega_> or is it -46? that's the question [21:43] <walken> well [21:43] <walken> its in mpeg1 right ? [21:43] <walken> so yeah its oddification [21:43] <omega_> yup [21:43] <omega_> ok, what bit is that? [21:43] <walken> oddification is odd :) [21:43] <omega_> even so ;-) [21:43] <walken> the sbits ? I guess its the sign [21:43] <omega_> hrm, ok, I'll analyse it some more [21:43] <omega_> I have my routine in place, but there are coeff differen= ces [21:43] <omega_> somehow it's overwriting the DC [21:44] <omega_> and it's slower, but that doesn't surprise me, it hasn't= been optimized at all beyond the tables that were autogenerated [21:45] greg_ (gr...@ho...) joined #gstreamer. [21:45] <omega_> yo [21:45] <omega_> and there seem to be other minor problems [21:45] <walken> omega. looking at the code - yes this is the oddificatio= n [21:45] <omega_> ok [21:45] <omega_> I removed it to match my coeff stuff [21:46] <omega_> heh, convenient [21:46] <omega_> my debug output has exactly 10 lines per block, so I jus= t /10 to get blocknum from lineno [21:48] <omega_> hrm, gweather claims light rain, I see blue sky overhead [21:53] <omega_> um [21:53] <omega_> run 0, level 4, pos now 1 [21:53] <omega_> run 1, level -2, pos now 3 [21:53] <omega_> run 1, level 1, pos now 5 [21:53] <omega_> run 2, level 1, pos now 8 [21:53] <omega_> run 2, level -1, pos now 11 [21:53] <omega_> run 0, level 1, pos now 12 [21:53] <omega_> run 0, level 1, pos now 13 [21:53] <omega_> run 3, level 1, pos now 17 [21:53] <omega_> run 1, level -1, pos now 19 [21:53] <omega_> run 5, level -1, pos now 25 [21:53] <omega_> have escped run,level of 38,-1, 28 bits long: 0x066ff8a [21:53] <omega_> run 38, level -1, pos now 64 [21:53] <omega_> warning, warning!! pos is 64 [21:54] <omega_> pos can't be 64, it goes from 0(dc) to 63 [21:54] <walken> hmmmmmmmmm [21:54] <walken> well if it gets too high you should exit the routine [21:54] <omega_> is this expected? [21:54] <walken> I *think* you're supposed to have a break code though [21:55] <omega_> seems like the encoder is broken then [21:55] <omega_> this was encoded with Ligos [21:55] <walken> I'm not 100% sure what the spec mandates, especially mpe= g1 [21:55] <walken> I think mpeg2 mandates we have a break code [21:56] <omega_> it seems like a horrible waste of bits to use an flc to = walk off the end of the coeff table just to close a block [21:57] <omega_> I'm gonna see if the same happens in your code [21:58] <omega_> in scan[i], does i range from 1 to 63 like in my code? [21:59] <omega_> um, I don't see an escaped code at all in yours ;-( [21:59] <omega_> would explain why the next block is fragged too [22:00] Action: ChiefHighwater frags a block with his railgun [22:01] <ChiefHighwater> hrmm, sounded more fun than that [22:01] Action: omega_ marks up walken's code again [22:06] <omega_> hmm, the rest of the block matches, but I'm getting a fa= lse vlc [22:06] <omega_> er, fls [22:06] <omega_> er, flc [22:06] <omega_> btw, I was dumping flcs wrong, but afaik that's fixed no= w [22:07] Action: ChiefHighwater wonders when omega_ is going to come heal = his router 8-] [22:07] <omega_> can you talk to the cbos? [22:07] <ChiefHighwater> no, but not sure about my port settings [22:08] <omega_> I know my laptop can talk to it, I can confirm port sett= ings on mine here before I try yours [22:08] <ChiefHighwater> 8-] [22:08] <omega_> walken: you're not getting an EOB either.... [22:08] <ChiefHighwater> bring your management cable too, just in case [22:08] <omega_> yup, if I can locate it [22:08] <ChiefHighwater> hehe [22:08] <ChiefHighwater> we have several at church 8-] [22:08] <omega_> yup [22:09] <walken> omega. hmmm dont know :) [22:09] <omega_> what code would escape on i >=3D 64 ? [22:09] <ChiefHighwater> i got us battery rechargers...show you when you = are here [22:09] <omega_> ah, found it [22:09] <omega_> ok [22:10] <walken> omega. also I dont have explicit code to handle the esca= pe. I handle it with the same code that handles buffer overflows [22:10] <walken> basically I make sure the escape forces an overflow [22:10] <omega_> run 1, level -1, pos now 19 [22:10] <omega_> run 5, level -1, pos now 25 [22:10] <omega_> illegal EOB! [22:10] <omega_> if (i >=3D 64) { [22:10] <omega_> fprintf(stderr,"illegal EOB!\n"); [22:10] <omega_> break; /* illegal, check needed to avoi= d buffer overflow */ [22:10] <omega_> } [22:10] <omega_> at top of escape code section, right after i+=3D [22:11] <omega_> hrm, I wonder if the ligos encoder was scrambling for wa= ys to pad the bitstream?? [22:14] <walken> omega. it sure looks weird. shouldnt happen too. [22:14] <walken> omega. I think the reference decoder will complain on th= is too [22:14] <omega_> ok [22:14] <omega_> unfortunately, I get a zeroed block right after that now= ;-( [22:14] <walken> hmmm [22:15] <walken> why dont you test on a dvd stream [22:15] <omega_> because that's mpeg2 [22:15] <omega_> my decoder is only mpeg1 right now [22:15] <walken> well mpeg2 is fine no ? [22:15] <walken> oh [22:15] <omega_> I don't have tables and such for mpeg2 [22:17] <omega_> your code seems to use part of that flc to construct the= next block [22:18] <walken> huh ? [22:18] <omega_> checking [22:19] <omega_> um, given original and current bufptrs, plus bits, what = will give me the total bits read? [22:20] <walken> hmmmm [22:21] <walken> I'd guess something like 8*(bufptr-buf)+bits or somethin= g [22:21] <omega_> +bits? [22:21] <omega_> I'll try it and derive based on match to my code [22:23] <omega_> looks like +bits-16 maybe [22:23] <omega_> yup [22:23] <walken> looks about right [22:23] Apoc_ (ep...@pp...) joined #gstreamer. [22:23] <omega_> yo [22:23] big_T (th...@fy...) joined #gstreamer. [22:24] <Apoc_> yo [22:28] <Apoc_> omega : I've made the cdparanoia plugin patch ... [22:28] <omega_> cool [22:29] <Apoc_> I send it to you or to the gstreamer-devel list ? [22:29] <omega_> the list [22:29] <omega_> along with a short description, of course [22:30] <Apoc_> Ok .. [22:30] <omega_> walken: you start the next block 18 bits sooner than I d= o [22:31] <walken> in the case where the escape code has an invalid run ? [22:31] <omega_> yup [22:31] <omega_> checking vlcs by hand [22:33] Nick change: dobey -> dobey-eat [22:34] Nick change: ajzzzz -> ajmitch [22:34] <ajmitch> morning all [22:34] <omega_> yo [22:35] <omega_> oooooh [22:35] <ajmitch> what's up? [22:35] <ajmitch> whaaaat? [22:35] <omega_> I think I found a bug in mpeg2dec [22:35] <ajmitch> hehe [22:35] <walken> argh [22:35] <omega_> muhahaha [22:36] <walken> now I'll have to kill you ;-) [22:36] <omega_> the valid coeffs total 52 bits [22:36] <omega_> after that is an invalid 20-bit flc [22:36] <omega_> (pos =3D=3D 64) [22:36] <ajmitch> omega_: i think it's a pain when gstmediaplay crashes f= or me but for noone else because of my hardware ;) [22:36] <ajmitch> makes it harder to debug... [22:36] <omega_> yet you start parsing the next block only 54 bits after = the start of the prev block [22:36] <omega_> 18 bits sooner [22:36] thomas (th...@ad...) left irc: Read err= or to thomas[adsl-63632.turboline.skynet.be]: EOF from client [22:36] <walken> well its after an invalid block right ? [22:37] <omega_> yes, but the codes are just fine [22:37] <omega_> you're flushing a 2-bit EOB code instead of a 20-bit flc= code [22:37] <walken> well the stream is invalid so I dont care as long as it = doesnt crash [22:37] <ajmitch> hehe [22:37] <omega_> hmmmm [22:37] <walken> basically [22:37] <omega_> invalid stream =3D=3D invalid codes, imo [22:37] <omega_> the codes are fine, so they should be flushed correctly [22:38] <omega_> it's kinda by accident that you manage to not wedge on t= he remaining bits and find an EOB sorta where you expected it [22:38] <walken> only if mpeg1 says this escape code is valid in this pos= ition [22:38] <walken> I know its not valid in mpeg2. I'm not 100% sure in mpeg= 1 though. [22:38] <omega_> do you have the mpeg1 video spec? [22:38] <walken> no :-/ [22:39] <omega_> whoops [22:39] <walken> systems neither [22:39] <walken> I have all of mpeg2, and I have mpeg1 audio, but thats i= t [22:39] <omega_> hmm [22:39] <ajmitch> ah well, cya, off to breakfast then to church... ;) [22:40] Nick change: ajmitch -> ajbusy [22:40] <omega_> oh, right, it's Sunday already for you [22:40] <ajbusy> yup [22:40] <omega_> somewhere I have a tool that will dump bits at any offse= t in a file [22:41] Nick change: dobey-eat -> dobey [22:47] <omega_> walken: data point though...: [22:47] <omega_> I've compared every last parsed detail of my decoder wit= h mpeg_play, and it's identical [22:48] <walken> hmmmmm [22:49] <walken> so you have an mpeg1 stream that mpeg_play plays fine, a= nd my decoder plays it fine too except that it goes thru unexpected paths= ? [22:49] <omega_> apparently [22:49] <walken> ok.... [22:49] <walken> what does it do ? read 2 bits instead of 20 and then ? [22:50] <omega_> your code reads 2 bits and proceeds to consider the rema= inder of the flc to be the codes for the next block [22:51] <walken> apparently the reference decoder will read the 20 bits b= ut it will fprintf(stderr,"DCT coeff index (i) out of bounds (int= ra)\n"); [22:52] <omega_> there's still some other bug in my code because the bloc= k numbers get out of sync [22:52] <omega_> I'm getting extra codes from somewhere ;-( [22:52] <walken> well I guess if I dont read enough bits and I get end of= block codes later that explains the desync [22:53] <omega_> oh, the extra codes are due to a lacking debug in your r= outine [22:54] <walken> well yeah [22:54] <walken> after an error I pretend its fine to do ANYTHING except = crashing [22:54] <omega_> heheh [22:55] <walken> (though it'd be best to detect as many errors as possibl= e) [22:55] <omega_> hrm, we're getting different run levels [22:55] <walken> OK going to the shower back in 10 [22:55] <omega_> ok [22:55] <Apoc_> omega : patch sended :) [22:55] <omega_> cool [22:58] <Apoc_> I have another problem with queue and CVS gstreamer [22:59] <omega_> oh? [23:00] <Apoc_> the sound is low ... it worked with 0.1.1 [23:00] <omega_> low?? [23:02] <Apoc_> I don't know how to say it in english :) ... the frequenc= y seems to be lower ... a 78 rpm disc at 33 :) [23:04] <omega_> oh, hmm, that's a capsnego problem [23:04] <omega_> write up a message to -devel directed at wtay [23:04] <omega_> he can debug it and fix it [23:04] <omega_> I don't understand capsnego well enough yet (still!) [23:05] <omega_> basically the sound card is being set to a lower sample = rate than the audio is [23:05] <walken> hehe [23:05] <walken> (back) [23:05] <omega_> I've explicitely set the default for osssink to 11025, m= ono. that flushes out these problems, whereas the normal defaults don't = necessarily show them [23:06] <omega_> walken: I'm getting some really strange differences here [23:06] <omega_> 15570c15570 [23:06] <omega_> < run 64, level -1, pos now 39 [23:06] <omega_> --- [23:06] <omega_> > run 32, level -1, pos now 39 [23:06] <omega_> yours,mine [23:07] <omega_> it's an flc, obviously [23:08] <walken> huh [23:08] <omega_> next 16 bits are 0x259e [23:08] <omega_> run 1, level -3, pos now 2 [23:08] <omega_> run 3, level -1, pos now 6 [23:08] <omega_> then the run 64 [23:08] <omega_> interesting thing is that I don't get an illegal EOB fro= m yours [23:09] <omega_> oh wait [23:09] <omega_> I think my debug is wrong [23:10] <walken> hmmmmm [23:10] <walken> I have an API question [23:10] <omega_> ok [23:10] <walken> I dont know how to do a good API [23:11] <walken> have 5 minutes ? [23:11] <omega_> um, I yeah, I have 2hrs <g> [23:11] <omega_> or more [23:11] <omega_> actually no, 2hrs [23:11] <omega_> which hopefully is enough to develop a good API <g> [23:11] <walken> hmmm [23:11] <walken> have to find some paper bacc k :) [23:12] <omega_> ? [23:12] <omega_> whoah [23:12] <omega_> run -33, level -1, pos now 39 [23:12] <omega_> and you really do place a coeff there [23:13] <omega_> er, afaict [23:13] <Apoc_> omega : it works with esdsink but not with osssink=20 [23:13] <walken> OK [23:13] <omega_> Apoc_: right, they have different defaults [23:13] <walken> so its about the ac3 decoder [23:13] <omega_> Apoc_: if you play a slower audio stream to esdsink, it'= ll sound fast, cause esdsink doesn't nego at all [23:13] <walken> ac3 can have different number of channels : 1, 2, 3, 2+1= , 3+1, 2+2, 3+2 [23:13] <omega_> yup [23:13] omega_evm (om...@bo...) joined #gstre= amer. [23:13] <omega_> um, go away omega_evm.... [23:14] Action: omega_ can't log into work to kill that xchat ;-( [23:14] <Apoc_> amoga : I set the frequency to the osssink but no change [23:14] <walken> you can also have different numbers of output channels := for example you play a 3+2 stream on a stereo output [23:14] <omega_> Apoc_: hmm, not sure setting the rate works well on osss= ink [23:14] <omega_> walken: right [23:14] <walken> I'm not sure how to specify what output you want when yo= u call ac3dec [23:14] <walken> I mean [23:15] <omega_> are there standard #defines or something for the various= channel mixes? [23:15] <walken> you may have 3+2 channel output but if the ac3 stream is= stereo you probably only need to get stereo output.... other channels wi= ll be 0.... [23:15] <omega_> right [23:16] <walken> no standard defines... I guess I could define them as nu= mber of front and rear channels [23:16] <Apoc_> omage : I tell that to wtay ?? [23:16] <omega_> Apoc_: yeah, email to the -devel list and ask what's up [23:16] <omega_> but you'll probably want to sync up with wtay here in ir= c later today (few hours maybe?) [23:16] <omega_> walken: does the channel count ever change mid-stream? [23:17] <walken> should not... [23:17] <omega_> ok, that makes thins a little bit easier [23:17] <walken> dont know if its right to rely on this though, since we = might have had an error in the first frame [23:17] <omega_> right [23:19] <walken> for stereo output it can be dolby compatible or not too [23:20] <omega_> good point [23:20] <omega_> do you have the code to 5.1 -> surround-stereo ? [23:20] <walken> yes [23:21] <omega_> cool [23:23] <omega_> hmmm [23:24] <omega_> thing is, the solution depends heavily on the rest of AP= I [23:27] steveb (st...@no...) left irc: [x]chat [23:27] <walken> well.... [23:27] <walken> I dont know [23:27] <walken> I see two solutions... [23:28] <walken> either you ask for a given config of output channels and= if its too many you get an error [23:28] <walken> like if you want 3+2 and the ac3 is 2+0 [23:28] <omega_> ok, you're talking about the semantics, not the API [23:28] <walken> or you get something that describes what the ac3lib did = (basically says you only got 2+0) [23:29] <omega_> I think it has to be stream-driven [23:29] <walken> or.... [23:29] <walken> explain ? [23:30] <omega_> ac3dec tells the app what kind of stream it is, and expe= cts the app to then decide on an output format [23:30] <omega_> with the default being stereo [23:31] <walken> if the app decides on something excessive its an error t= hen ? [23:31] <omega_> right [23:31] <walken> OK [23:31] <walken> seems quite doable [23:31] <omega_> if you can construct defins for each format, you can eas= ily have a matrix of what's allowed, or rather a matrix of downmix funcpt= rs [23:32] <omega_> but the high-level API is what really matters [23:32] <omega_> the key is the bitstream code, which is causing you to h= ave to do copy_chunk in mpeg2dec [23:33] <walken> well I dont need that in ac3dec [23:33] <omega_> how? [23:33] Nick change: wtay-away -> wtay [23:34] <wtay> =F9 [23:34] <omega_> Apoc_: wtay's here <g> [23:34] <wtay> yo [23:34] <walken> ac3_frame_length (uint8_t * buf) returns frame length [23:34] <walken> (you have to pass at least 5 bytes of buf currently) [23:34] <omega_> wtay: have you gotten mail from -devel in the last hour? [23:34] <omega_> hmmmm [23:35] <walken> frame =3D ac3_decode_frame (uint8_t * buf) decodes the f= rame [23:35] <omega_> with a private struct as well? [23:35] <wtay> omega_ I have received mail, yes [23:35] <walken> with buf being the right size as decoded previously [23:35] <omega_> decode_frame(priv,buf) [23:35] <omega_> wtay: hrm, I haven't [23:35] <walken> none yet... will need one though [23:35] <omega_> but I've verified the path through sf [23:35] <wtay> omega_ omega_evm messes up my tab completion... [23:35] <omega_> wtay: yeah, I can't do anything about it [23:35] <walken> ac3 has very little state between frames, but still a bi= t of this (mdct delay samples) [23:35] <omega_> can't ssh into rr [23:36] <omega_> walken: ok, then you need a per-stream private struct, d= efinitely [23:36] <walken> yes [23:36] <omega_> globals =3D=3D BAD <g> [23:36] <walken> so maybe first function could be extended to also return= channel config, sample rate and bit rate [23:36] <Apoc_> wtay : yo=20 [23:36] <omega_> hmmm [23:36] <walken> and second function would take desired output channel co= nfig as a parameter [23:37] <wtay> Apoc_: you had problems? [23:37] <omega_> first/second function? [23:37] <Apoc_> wtay : Yes another one ;-) [23:37] <walken> ac3_frame_length/ac3_decode_frame [23:37] <omega_> hmmm [23:37] <Apoc_> cdparanoia + queue +osssink [23:37] <Apoc_> osssink in a thread [23:38] <wtay> Apoc_: ok [23:38] <Apoc_> the sound seems to be at a lower sample rate [23:38] <wtay> Apoc_: CVS I assume? [23:38] <omega_> walken: I'd be tempted to use a callback [23:38] <Apoc_> no problem with esdsink [23:38] <Apoc_> wtay : yes CVS version [23:38] <omega_> wtay: esdsink works because the default for that is 44.1= /stereo [23:38] <omega_> osssink is set to 11.025 to catch these problems [23:38] <wtay> yup, cdparanoia doesn't even set caps [23:39] <walken> I dont like callbacks much :) [23:39] <omega_> walken: yeah, but they're very efficient in cases like t= his [23:39] <wtay> f1x0r1ng [23:39] <Apoc_> wtay : I set the frequency properties of osssink element = to 44100 ... non change [23:39] <omega_> walken: lemme write up a possible API, sec.... [23:39] <wtay> Apoc_: nope, cdparanoia has to set the rate (but it doesn'= t) [23:40] <omega_> Apoc_: if you remove the queue, does it start working pr= operly? [23:40] <Apoc_> omega_ : yes=20 [23:40] <omega_> wtay: so it's a capsnego problem, like I thought <g> [23:40] <wtay> hmm, cdparanoia *does* have caps... [23:40] <wtay> dmn [23:40] <walken> also what if you have mono stream and stereo output [23:41] <wtay> damn even [23:41] <walken> ac3lib could say its invalid, and you'd need to ask mono= output [23:41] <omega_> walken: if the default isn't changed, you have to duplic= ate the channels [23:41] <omega_> imo *any* ac3 should be able to play out stereo [23:41] <omega_> because otherwise you have to force every app to worry a= bout it [23:41] Kuroyi (ri...@ub...) joined #gstream= er. [23:41] <walken> ok so ac3lib has to do the memcpy [23:41] <omega_> and why force them to worry about it when stereo is the = standard? [23:41] <walken> I guess its OK [23:41] <Apoc_> I just sended a mail to the list ... [23:42] <walken> but if you want 4-channel output you'll have to worry ab= out this though [23:42] <omega_> of course [23:42] <walken> since if the stream was stereo you wont get 4-channel [23:43] <omega_> the idea is that I should be able to load an ac3 and pla= y it out a stereo sound card without having to worry about the format at = all [23:43] <omega_> if I have anything other than stereo, I'd then start to = care and consider changing the up/downmix config [23:44] <walken> OK [23:44] <walken> sounds acceptable [23:44] <Apoc_> wtay : I use the cdparanoia plugin with my patch ... but = I don't modify the *core* of the plugin=20 [23:46] <wtay> Apoc_: it's a problem with caps nego proxy, which I intend= to fix sometime later [23:46] <Apoc_> wtay : Ok ... [23:47] <wtay> Apoc_: I'll make a temporary fix now to cdparanoia to circ= umvent the problem [23:47] <walken> omega. another way would be to let libac3 handle stuff. [23:47] <omega_> meaning? [23:48] <walken> like you tell you have 3+2 output, the stream is 2+0, li= bac3 decodes and tells you oh BTW I gave you 2+0 output [23:49] <omega_> I thought we were discussing the libac3 API ? [23:49] <walken> maybe this is simplest as you dont have to figure out wh= at you should ask given the stream and your output config [23:49] <walken> well yes we are [23:50] <omega_> btw, is the output interleaved or not? [23:50] <omega_> or is that another option? [23:50] <walken> I just said maybe libac3 should figure out what output y= ou really want instead of making you ask for it [23:50] <omega_> how would libac3 know what you want? [23:50] <walken> output is non interleaved and in floating point format [23:50] <omega_> hmm, ok [23:50] <walken> its what libac3 uses internally, it allows for >16 bit o= utput, and it avoids the interleave question :) [23:51] <omega_> but can't go very fast, or be optimized with mmx [23:51] <walken> well actually it lets you do the conversion in libao and= mmx optimize it [23:51] <walken> if it was in libac3 the mmx/floating point stuff wouldnt= mix very well [23:51] <omega_> howso? the majority of the work is done in ac3dec, and = that's where you're doing floats [23:51] <omega_> why can't ac3dec use all integer? [23:52] <walken> errr, hard to do... no say if it would be faster either [23:52] <wtay> Apoc_: fix in CVS [23:52] <walken> we would probably need 40-bit integers at least [23:52] <omega_> walken: definitely faster, float is slow [23:52] <Apoc_> wtay : The patch I send to the list will work ?? [23:52] <omega_> walken: if you survey where the decimal point is through= out the code, you can easily do it with 32bit [23:52] <wtay> Apoc_: I'll merge it now.. [23:53] <omega_> by definition, since you're using 32bit floats right now [23:53] <walken> well they can change in magnitude quite a lot [23:53] <walken> by 24 bits.... [23:53] <omega_> that's what shifts are for <g> [23:53] <walken> coeffs start with 16 bits mantissa and exponent between = 0 and 24 [23:53] <omega_> do you have a good roadmap of the current code? [23:54] <walken> then they go thru an imdct (looks like a fft) [23:54] <walken> errr, the standard :) [23:54] <walken> current code is quite straightforward [23:54] <omega_> hmm, ok, can you send me a tarball of your current worki= ng copy? [23:55] <omega_> I'll see if I can understand it [23:55] <walken> om...@om... ? [23:55] <omega_> om...@te... [23:57] <omega_> got it [23:57] <omega_> ideally I'd like to refactor it to use libcodec <g> [23:57] <omega_> parse.c:191: warning: implicit declaration of function `= memset' [23:58] <walken> oh yes. I thinkits fixed now... [23:58] <omega_> segfault [23:58] <wtay> Apoc_: you patch is merged in CVS [23:58] <walken> make sure your stream starts with 0b77 [23:58] <walken> (yes, it sucks) [23:59] <omega_> hrm, I have 48 bytes of noise at the front of mine [23:59] <Apoc_> wtay : thanks ... [23:59] <omega_> 1 frames decoded in 0.00 seconds (0.00 fps) [23:59] <walken> now change the #if in libao/output_linux.c [00:00] --- Sun Jun 3 2001 [00:00] <walken> (its in debug mode... outputs output.pcm) [00:00] <walken> should work then. I swear ! [00:00] <omega_> #if LIBAO_OSS ? [00:01] <walken> no #if 0 [00:01] <omega_> there is no if 0 [00:02] <walken> #if 1 then [00:02] <omega_> ok [00:03] <omega_> wtay: ac3dec in gst is very broken, afaict [00:03] <omega_> walken: works [00:04] <wtay> omega_: I was looking at using libac3 instead [00:04] <omega_> yeah, that's what I'm writing an API for right now <g> [00:04] <wtay> omega_: it has that weird plugin_output_audio crap again [00:04] <omega_> libao? yeah, libao should go away imo [00:05] <wtay> yes please [00:06] <wtay> is libac3 from CVS in a reasonably working state? [00:06] <omega_> it hasn't changed at all [00:06] <wtay> OMS CVS I mean [00:06] <omega_> I just ran libac3 from walken's homedir, and with caveat= s, it works [00:06] <walken> wtay. libao's going :) [00:06] <walken> I want to get rid of it too [00:07] <omega_> libao is useful for sinks, but not much more [00:07] <omega_> and imo the code in libao would be better in the appropr= iate gst sink elements [00:07] <walken> well the ac3dec libao is going I meant [00:07] <walken> I dont know if you have a different libao in gst [00:07] <omega_> we don't use libao [00:11] <omega_> walken: api sent, it's not complete, but should give you= an idea of how I approach this [00:12] <omega_> xfree 4.1.0 is out [00:13] <walken> yaaaaay [00:14] <walken> hmmm [00:14] <walken> have a web pointer ? [00:15] <omega_> no, I just saw it on de...@xf... [00:15] <omega_> public later today [00:15] <walken> is devel a public list ? [00:15] <walken> its not mentionned on their site [00:15] <omega_> ok, not out yet. tagging cvs now. will be tarballs soo= n [00:15] <omega_> no [00:15] <walken> damn [00:16] <walken> how did you get on it :) [00:16] <omega_> by talking to Dirk <g> [00:16] <walken> grrrr [00:16] <walken> I'm always left out on stuff [00:16] <omega_> I intend to use it eventually to replace Xv with somethi= ng that doesn't suck [00:16] <omega_> when I have time [00:16] <omega_> that's what I intended to do when I got on the list, but= I didn't have any clue at that point [00:17] <omega_> hmmm, compile failure on 4.1.0 tag, looks like it needs = some work still <g> [00:17] <omega_> hadess wouldn't be happy, it's a compile bug on ppclinux= .... [00:18] <wtay> what's new? anything interesting? [00:18] <omega_> dunno [00:19] Action: omega_ hopes it finally has complete xv support for rage = chips, I'm sick of going to get the gatos driver [00:19] <omega_> btw walken: there's a guy in spain wanting to run a TV s= tation off a PC [00:19] <omega_> sounds like he's gonna base in on gstreamer [00:20] <omega_> but he has some interesting problems to overcome [00:20] <omega_> his source material is all non-live, and can be mpeg2 [00:20] <omega_> dxr3 would be ideal with its svideo output... [00:20] <omega_> but he needs arbitrary graphics overlay, live [00:20] <omega_> so I'm thinking that he needs to do software decode with= fullcolor overlay, on an ati board [00:21] <walken> hmmmm [00:21] <walken> probably yeah [00:21] <omega_> how well do you think mpeg2dec would mix with some advan= ced form of ati driver that does really well at interlace handling? [00:22] <omega_> since it's gonna be a PAL signal, the computer should do= everything it can to closely match the interlacing and such, not try to = be progressive and work back [00:22] <omega_> though they video-edit the stuff, that's gonna be progre= ssive unless they move to a complete solution that implicitly handles int= erlace [00:22] <omega_> (betacam -> matrox digisuite -> mepg2 -> pc -> broadcast= ) [00:23] <omega_> right now it's betacam -> digisuite -> beta -> broadcast [00:23] Nick change: richardb-near -> richardb [00:23] <richardb> Evening all. [00:23] <omega_> yo [00:23] <wtay> hi [00:23] <richardb> How're we doing with progress towards release? [00:24] <richardb> I'm seeing a lot of patches being made... ;-) [00:24] <omega_> dunno, getting closer afaict [00:24] <omega_> we should do a pre2 maybe? [00:24] <richardb> Yeah, soon. [00:24] <omega_> it'd be nice if we could release this weekend [00:24] <richardb> I was looking at the arts configure problem someone re= ported. [00:24] <wtay> I need some work work on avi... [00:24] <richardb> Can't see any way to get the right paths. [00:25] <wtay> s/work/more/ [00:25] <omega_> richardb: great [00:25] <omega_> wtay: good, one case where the lacking 'g' works <g> [00:26] <wtay> a coincidence :) [00:29] <richardb> Hmm: when writing a configure check to find paths for = arts, would it be a sane idea to try using the output of: artsc-config --= cflags | sed 's/artsc/arts/g' [00:29] Action: richardb grins manically [00:31] <wtay> it would work on my system :-) [00:31] <richardb> Worth a try then. ;-) [00:31] <wtay> better than what's in CVS now=20 [00:32] <Apoc_> Just a question : Is it possible to listen to a wav file = and to encode it to ogg in the same time ? With tee elements ?? [00:33] <wtay> Apoc_: I doubt it... [00:33] <wtay> Apoc_: tee doesn't do capsnego at all [00:33] <richardb> Anyone know why the XMMS check is commented out in con= figure.base? [00:34] <wtay> richardb: no idea, I noticed that too.. [00:35] <Apoc_> wtay : Another element maybe ??? [00:36] <wtay> Apoc_: disksrc ! vorbisenc ! ... [truncated message content] |
From: <wim...@ch...> - 2001-06-05 04:57:09
|
[06:41] mwc (mal...@ab...) joined #gstreamer. [07:14] chillywilly (bau...@d6...) joined #gstreamer. [07:14] <chillywilly> omega_evm: you here? [07:16] derek (de...@cp...) joined #gstreamer. [07:16] Nick change: taaz -> taazzzz [07:16] <chillywilly> sorta dead here at te moment [07:19] Nick change: ajmitch -> ajfood [07:36] omega_ (om...@om...) joined #gstreamer. [07:37] <chillywilly> yo [07:37] <omega_> yo [07:37] <chillywilly> wassup [07:37] <omega_> not much [07:37] <chillywilly> my bud here [07:37] <chillywilly> has some questions for you [07:37] <chillywilly> about mixing and software you can use on GNU/Linux [07:37] <walken> honk [07:38] <omega_> chillywilly: current software or planned software [07:38] <omega_> walken: ? [07:38] <derek> omega_ i here you are the audiophile :) [07:38] <walken> hi [07:38] <omega_> derek: sorta <g> [07:38] <chillywilly> omega_: no, what he could use (if anything) [07:39] <omega_> what kind of use? [07:39] <derek> ok [07:39] <derek> two things [07:39] <derek> the first is looking for studio quality mixing appliction [07:39] <omega_> live or post? [07:39] <derek> searched sourceforge and foudn a few that might be nice [07:40] <derek> omega_: thats why its two part :) [07:40] <omega_> heh [07:40] <derek> first part is post [07:40] <derek> btw: im not a huge audio guy [07:40] <derek> but trying to get some musicians off microsoft [07:40] <omega_> heh, good idea <g> [07:41] <omega_> MS has really bad latency guaranteeds, and performance in general [07:41] <derek> they keep complaining how expensive their 'habit' is [07:41] <derek> so if i can find something decent on linux they will switch [07:41] <omega_> then stop paying the M$ tax <g> [07:41] <derek> then second part of the question [07:41] <derek> is i run sound/video for church services [07:41] <omega_> me too [07:41] <derek> i would like to pipe right off our mixing board [07:41] <omega_> just finished mixing/recording a choir favorites concert [07:41] <derek> on top my laptop [07:42] <derek> and then post the mp3's to the net [07:42] <omega_> right [07:42] <derek> or stream live [07:42] <derek> etc etc etc [07:42] <omega_> on $0 budget, of course <G> [07:42] <derek> no i have money [07:42] <omega_> ah, ok, that helps [07:42] <derek> but prefer 0 budget [07:42] <derek> i wont pay for software [07:43] <derek> i will pay for hardware [07:43] <omega_> right [07:43] <derek> if i find good software [07:43] <omega_> what kind of mixer, how many channels in use on an avg sunday? [07:43] <derek> ive been known to donate [07:43] <derek> either money or time [07:44] <derek> um crud i threw away the manual (copy i had at home) [07:44] <derek> it has 32 channels i believe [07:44] <derek> use normally [07:44] <derek> 10-12 channels for band [07:44] <derek> usually [07:44] <omega_> ok [07:44] <omega_> we have 24 channels, use usually ~10 on avg sunday [07:45] <omega_> I've got some high-end hardware with the goal of eventually having the computer do all the live and record mixing [07:45] <omega_> but the problem is the lack of a mixing surface that controls the computer [07:46] <derek> yip [07:46] <derek> i would love to go 100% digital [07:46] <omega_> but am having problems with the driver, only get peak levels of 1/13th the range [07:46] <omega_> which is insane.... [07:46] <derek> and ditch the mixing board [07:46] <derek> anyhow the package i grabbed [07:46] <derek> was protux [07:46] <derek> or something [07:47] <omega_> searching [07:47] <omega_> on sf? [07:47] <derek> yeah [07:47] <derek> if you know something better im game [07:47] <derek> i am fairly new to the audio scene in general [07:48] <derek> i got asked to do at church on saturday nights [07:48] <derek> as im the 'geek' [07:48] <omega_> heh [07:48] <derek> which equated to 'must do anything electronic' [07:48] <omega_> problem is I've never done post anything [07:48] <omega_> derek: yeah, beware, or you get sucked into all kinds of things <g> [07:48] <derek> no kidding :) [07:48] <chillywilly> hehe [07:48] <derek> but i have wanted to get the things digital for some time [07:48] <omega_> my list of major projects to be done is way too long [07:49] <derek> so figured it was good way to get that done [07:49] <derek> as when i miss a service [07:49] <derek> having it handed to me on really crappy sounding tape [07:49] <derek> is just sick [07:49] <omega_> yeah, tapes suck [07:49] <omega_> I was forced to record tonights concert to tape because I couldn't get the computers to cooperate [07:49] <omega_> and I managed to get it sounding OK, but it's still flat and icky [07:50] <walken> omega. do you have 24-bit output on your computers ? [07:50] <walken> or even better, 24-bit input ? <g> [07:50] <derek> well for me big ouchie on tapes isnt only the quality but the media [07:50] <omega_> the machine in question, when it works, has theoretical 24-bit I/O, but I only have 18 bits of A/D/A [07:50] <derek> to me tape player is like LP [07:50] <derek> hard to find somewhere to play it:) [07:50] <omega_> hehehe [07:51] <omega_> yeah, I have no doubt that I'm gonna be asked for CDs of this concert [07:51] <omega_> I just need to find a way to get it captured off tape somehow, with w2k if all else fails [07:51] <derek> ok so no good tips for post? [07:51] <walken> I'm deaf enough to not have sound quality issues :) [07:51] <walken> lol [07:51] <derek> what do i need to stream? [07:51] <omega_> walken: tape vs. CD? if you can't tell that difference, you need to go get your ears transplanted... [07:51] Action: derek is wondering if i coudl do it now [07:51] <omega_> derek: just to stream to mp3? a sound card and gstreamer [07:51] <walken> well yeah I can tell that one :) [07:52] <omega_> walken: good <g> [07:52] <derek> ? [07:52] <walken> 16bit vs 24, I dont think so :) [07:52] <derek> omega_: ok on the mixing board [07:52] <omega_> derek: gstreamer-launch osssrc ! lame ! disksink location=service.mp3 [07:52] <derek> i woudl run from an output to what on my computer? [07:52] <walken> ac3 vs mp3... oh wait the ac3 has nicer code :) [07:52] <omega_> derek: soundcard line input [07:52] <derek> ok [07:52] <derek> will have to get an adapter or something [07:52] <derek> omega_: where are you located physically? [07:52] <omega_> but the problem is that most sound cards are going to introduce a lot of RFI [07:53] <omega_> derek: Portland, OR [07:53] <derek> ah cool [07:53] <derek> have you seen [07:53] <derek> http://www.geeks4christ.com [07:53] <derek> i think micah is going to start a chapter in portland this month [07:53] <omega_> I think I ran across it at one point, lemme look [07:53] <derek> you should really talk to him [07:53] <derek> it would be cool [07:53] <derek> perhaps we could do a virtual one [07:54] <walken> hahahaha [07:54] <derek> anyhow shouldnt go on that tangent in this channel [07:54] <walken> today's garfield comic is great [07:54] <omega_> derek: tangent? what tangent? <g> [07:54] <walken> http://garfield.ucomics.com/garfield/gaview.htm [07:54] <derek> ok on the board->soundcard [07:54] <omega_> walken: heheeh [07:54] <chillywilly> derek: heh, this place is about as free flowing as #gnuenterprise ;) [07:54] <derek> where can i get an adapter? (Radio shack) [07:54] <omega_> derek: depends on the outputs [07:55] <omega_> you do stereo -> tape now? [07:55] <derek> yes [07:55] <omega_> it's rca to rca for that? [07:55] <omega_> from aux busses or something? [07:55] <derek> i planned on maybe just stealing those [07:55] <derek> i think its some gross plug [07:55] <derek> to rca [07:55] <omega_> yeah, just grab those, tee htem off (RS has neat piggy-back rca cables) [07:55] <derek> i dont know what they call the gross plug :) [07:55] <omega_> hmm [07:55] <omega_> not 1/4" [07:55] <omega_> ? [07:55] <derek> um yeah [07:56] <derek> probably bout 1/4" [07:56] <derek> like OLD headphones [07:56] <omega_> right [07:56] <derek> but dont know what thats called [07:56] <omega_> ok, somehow get from that to a standard dual-rca to stereo 1/8" [07:56] <derek> i call call it old head phones :) [07:56] <omega_> 1/4" is either TS or TRS [07:56] <omega_> Tip-sleeve, or tip-ring-sleeve [07:56] <omega_> you've probably got a pair of TS coming out, to L/R tape [07:57] <derek> yip [07:57] <derek> i think so [07:57] <derek> if they have t's at radio shack [07:57] <derek> would be sweet [07:57] <omega_> just piggyback off the RCA on the tape dec, gender-bend to a dual-rca to 1/8" stereo [07:57] <derek> as i can just go off of the tape out [07:57] <derek> and then can make tape [07:57] <omega_> yup [07:57] <derek> at same time streaming [07:57] Nick change: ajfood -> ajmitch [07:57] <derek> so expirementation doesnt 'harm' anything [07:57] <omega_> then you'll want a decent sound card. I haven't tried line in on the SBLive, but that's a good place to start [07:58] <omega_> derek: only thing you could do wrong is to make a loop somewhere [07:58] <derek> for now [07:58] <omega_> which I have never managed to pull off <g> [07:58] <derek> it will have to be crappy sound card [07:58] <omega_> hmmm [07:58] <derek> on my laptop [07:58] <omega_> oh, mucho noise [07:58] <omega_> but if it's for streaming, it's probably OK [07:58] <derek> just to 'test' to confirm i can get things working [07:58] <omega_> trick is that the hard-drive tends to make the most noise [07:58] <derek> btw: probably cant be much worse than tapes :) [07:59] <omega_> so if you can stream it, and have some other computer write it to its hard drive.... <g> [07:59] <omega_> derek: tape noise is constant, drive noise is bursty [07:59] <derek> for now i would be happy [07:59] <derek> if i could go from board to disk [07:59] <derek> in mp3 [07:59] <derek> and have it be audible [07:59] <omega_> yup [07:59] <derek> from there [07:59] <derek> i would work on making better [08:00] Action: derek should really be doing work [08:00] <derek> but this is more fun [08:00] <omega_> yeah, so you can start there and it should be pretty easy [08:00] <omega_> gstreamer can do in its sleep [08:00] <ajmitch> sounds like a fun discussion... [08:00] <omega_> the commandline I quoted above is all you need, really [08:00] <derek> so how much processor/ram does gstreamer eat up? [08:00] <omega_> with some options to lame to set the bitrate and such [08:01] <omega_> gstreamer doesn't eat hardly anything, it's the various plugins that do [08:01] <ajmitch> derek: much cpu for mp3 encoding, i guess ;) [08:01] <derek> i have pIII 700mhz/128MB RAM in laptop [08:01] <omega_> plenty [08:01] <ajmitch> i think that will be good enough... [08:01] <omega_> lame is what you're gonna be running, it tends to suck up the CPU, but isn't gonna touch that 700... [08:01] <chillywilly> derek: what card is in that beast anyway? [08:01] <derek> maestro i think [08:02] <omega_> 2e or 3? [08:02] <omega_> I have a 2e [08:02] <derek> 3 [08:02] <derek> i think [08:02] <derek> is there way to tell? [08:02] <omega_> lsmod maybe [08:02] <derek> by probing something [08:02] <ajmitch> derek: you want mp3 or ogg? [08:02] <omega_> or lspci [08:02] <omega_> yeah, lspci says 2E for me [08:02] <derek> just says maestro [08:02] <omega_> 00:08.0 Multimedia audio controller: ESS Technology ES1978 Maestro 2E (rev 10) [08:02] <mwc> omega_: Just of of interest do you use oss or alsa for your 2e? [08:02] <derek> lspci bombs for me [08:03] <omega_> oss for now [08:03] <omega_> derek: bombs?? [08:03] <derek> dneighbo@latitude:~/downloads/sound/protux0.10.0-src$ lspci [08:03] <derek> pcilib: proc_read: tried to read 64 bytes at 64, but got only 0 [08:03] <derek> lspci: Unable to read cardbus bridge extension data. [08:03] <derek> dneighbo@latitude:~/downloads/sound/protux0.10.0-src$ [08:03] <omega_> eeeek [08:03] <omega_> cat /proc/pci then <g> [08:03] <chillywilly> there ya go [08:03] <chillywilly> do it the "real" man's way [08:04] <ajmitch> chillywilly: what, morse code on the PCI bus? [08:04] <derek> omega_: helps if im root :) [08:04] <derek> 00:08.0 Multimedia audio controller: ESS Technology ES1978 Maestro 2E (rev 10) [08:04] <omega_> chillywilly: no, the "real" man's way is cat /proc/bus/pci/devices [08:04] <derek> we have same card :) [08:04] <omega_> derek: ok [08:04] <omega_> derek: /me has an Insprion 7500 [08:04] <omega_> er, Inspiron [08:04] Action: derek has latitude CPx [08:04] <derek> :) [08:04] <omega_> yup [08:05] <chillywilly> Dell supports GNU/Linux right? [08:05] <omega_> sorta [08:05] <chillywilly> heh [08:05] <derek> yip [08:05] <derek> have a few models [08:05] <ajmitch> masta: what church is this for? [08:05] <derek> you can get with linux installed [08:05] <omega_> derek: sorta <g> [08:05] <derek> but if you cal lthem for support [08:05] <omega_> ->linuxcare [08:05] Action: mwc has a latitude as well [08:05] <derek> they dont know doink about linux [08:05] <omega_> mwc: you know alsa? [08:06] Action: chillywilly has a gateway laptop [08:06] <derek> i like the layout of the dell [08:06] <derek> the performance of the dell [08:06] Action: omega_ wants an 8000 [08:06] Action: ajmitch doesn't have a laptop yet :( [08:06] <derek> but the quality is crud [08:06] <mwc> omega_: tried it once but did not inhale [08:06] <derek> i have had for about 1 year [08:06] <omega_> mwc: heh [08:06] <derek> had keyboard replaced twice [08:06] Action: chillywilly uses ALSA [08:06] <omega_> derek: I've had both left and rigth hinge mounts of my LCD snap [08:06] <derek> fan/cooling unit relpaced once [08:06] <derek> hinge snap (have service call out now) [08:07] <omega_> neat [08:07] <derek> and had mouse button snap off [08:07] Action: chillywilly has to replace his fan and the hard drive [08:07] <mwc> I find I have to have the PC speaker input turned down or BIG BAD NOISE [08:07] <derek> now i will admit [08:07] <derek> im on my computer about 18hours a day [08:07] <derek> 6+ days a week [08:07] <chillywilly> mwc: big bad beeeeep? [08:07] <derek> and i transport it to and from work daily [08:07] <omega_> they replaced the left-hand mount but not the right mount... you'd think that if the left mount is broken, the right mount will be stressed to the point of breaking [08:07] <omega_> derek: same here [08:07] <derek> and use on public transportation so i do beat the snot out it [08:07] <mwc> Nah big bad screach screach bingy bingy blong screach [08:08] <omega_> derek: used to do public transport, but now I work at home half the time and fly to work the other half [08:08] <chillywilly> mwc: eeek [08:08] <derek> omega_: i did that for a while :) [08:08] <derek> 2 days week flew from phoenix to san jose [08:08] <derek> and rest worked from home [08:08] <omega_> oddly enough, it takes only a little longer to fly to Boise than it took to get across town on the light-rail [08:08] <derek> omega_: no kidding [08:08] <derek> same for sanjose [08:08] <omega_> about 1:45 door to door [08:09] <derek> i had one day away from family [08:09] <derek> but took me 2hr flight [08:09] <derek> but twice in two days [08:09] <derek> so 4 hour total [08:09] <omega_> pdx<>boi is only 1hr [08:09] <derek> going to office was 1 hour each way or 2 hour a day [08:09] <derek> so 4hr in 2 days [08:09] <derek> and then have the 10sec commute [08:09] <omega_> heh [08:10] <derek> the rest of the week made up for it :) [08:10] <omega_> these days I go there for a week at a time every 2-3 weeks [08:11] <derek> thats cool [08:11] <derek> they get you an apartment there? [08:11] <omega_> no, a hotel for now [08:11] <derek> in boise [08:11] <omega_> the CEO comes from San Jose, he talked about getting an appt [08:11] <derek> probably could get you a studio [08:11] <derek> for same price [08:11] <omega_> but the hotel is 2 blocks from the new offices [08:11] <derek> :) [08:11] <derek> and then if they have others in [08:11] <omega_> an appt would be a lot further, and I don't drive.... [08:11] <derek> they coudl 'reuse' it [08:12] <omega_> yup [08:12] <derek> mind me asking whats in boise :) [08:12] <omega_> have to be 2br though, because the CEO and I probably overlap schedules randomly [08:12] <omega_> in boise? um.... [08:12] <derek> not known for its technology and all :) [08:12] <omega_> oh? micron + HP [08:12] <omega_> and others [08:12] <derek> ah [08:12] <derek> no plans to move there? [08:12] <omega_> plus a small startup called RidgeRun, at least ;-) [08:13] <omega_> no, they tried again last week to get me to move.... but I'm tied to pdx [08:13] <derek> boise would be nice calm place to be [08:13] <derek> omega_: you single still? [08:13] <walken> how tied ? [08:13] <derek> if you dont mind me asking? [08:13] <omega_> walken: friends, family, etc. [08:13] <omega_> derek: yup [08:13] <derek> cause if you are, then i wouldnt wnat to move to boise either :) [08:13] <omega_> derek: quite [08:14] <derek> not to say there is problem with girls in boise [08:14] <derek> but your chances are much better in pdx :) [08:14] <omega_> derek: wouldn't know, but probably <g> [08:14] Action: derek is going to go grab gstreamer [08:15] <derek> cant find rca to audio in converter [08:15] <chillywilly> another satisfied customer [08:15] <chillywilly> :P [08:15] <derek> hey [08:15] <derek> wait [08:15] <omega_> derek: rca -> 1/8" ? [08:15] <derek> can it stream from a mic? [08:15] <omega_> yeah, but don't <g> [08:15] <derek> why? [08:15] <chillywilly> that'll be a $100 finders fee (yeah right) [08:15] <omega_> not any mic you have [08:15] <derek> i just want to 'test' [08:15] <derek> to see if it works [08:15] <omega_> oh, yeah [08:16] <omega_> if you have a cheap internet-phone mic or something [08:16] <derek> yip i do [08:16] <derek> a free one actually :) [08:16] <omega_> just set the mixer and give it a go [08:16] Action: derek is grabbing it now [08:16] <omega_> derek: as if it's worth anything anyway ;-) [08:16] <derek> will you be for another 30 min [08:16] <omega_> yup [08:16] <derek> if i run into issues [08:16] <derek> warning [08:16] <derek> IM CURSED [08:16] <chillywilly> loll [08:16] <omega_> hehehe [08:16] <derek> when it comes to software [08:16] Action: chillywilly can attest to that [08:16] <derek> chillywilly: can vouch [08:16] <omega_> just get CVS [08:17] <derek> if your program has ANY bug ANYWHERE [08:17] <omega_> it's been cleaned up a lot recently [08:17] <derek> my hardware will find it [08:17] <derek> i promise [08:17] <chillywilly> :P [08:17] <walken> I have some of that too [08:17] <derek> omega_ for now not gonna grab cvs [08:17] <derek> oh momma [08:18] <omega_> derek: wouldn't recommend it [08:18] <derek> this does video too? [08:18] <omega_> 0.1.1 is very not guaranteed to work for this.... [08:18] <walken> hmmmmm [08:18] Action: derek just bought a jvc digital video recorder [08:18] <walken> 100M free [08:18] <derek> ANY chance i could get it to work with this? [08:18] <walken> need to clean up a little :) [08:18] <ajmitch> derek: warning - i've touched this project, you sure you want to trust it? ;) [08:18] <omega_> 0.1.1? yeah, maybe, but I wouldn't bother [08:18] <omega_> I can send you a CVS tarball if you want [08:18] <derek> i can slurp cvs [08:18] <derek> or can try at least [08:18] <omega_> ok [08:18] <chillywilly> ajmitch: where's the GNUe code that you prmised eh? [08:19] Action: omega_ watches ajmitch's back as he runs away [08:19] <ajmitch> chillywilly: umm, sitting out back... ;) [08:19] <chillywilly> see masta I told you he was cheating us [08:19] <chillywilly> :P [08:19] <ajmitch> thx omega_ ;) [08:19] <derek> you support pserver [08:19] <derek> anon cvs? [08:19] <omega_> sf does, yeah [08:19] <ajmitch> it's SF [08:19] <derek> this is pretty cool [08:19] <omega_> http://sourceforge.net/cvs/?group_id=1936 [08:19] <derek> have you guys gotten any press? [08:19] <ajmitch> a little [08:20] <derek> chillywilly: tells me you have dorked your license [08:20] <ajmitch> iirc [08:20] <derek> but i will forgive you [08:20] <chillywilly> derek: they have gotten mega press, they were even at the Gnome conference thingy [08:20] <chillywilly> I saw the pictures [08:20] <chillywilly> pretty cool [08:21] <ajmitch> chillywilly: you mean GUADEC? [08:21] <chillywilly> yes [08:21] <omega_> ChiefHighwater apparently made a version with names on it, I need to find that, but in the meantime there's http://gstreamer.net/guadec-crew.jpg [08:22] <derek> what is the module name [08:22] <derek> gstreamer? [08:22] <omega_> yup [08:22] Action: omega_ udpates CVS HEAD to get all the patches in the last 2 days [08:23] <derek> cvs now slurping [08:23] <chillywilly> yum [08:23] <omega_> derek: what version of automake do you have? [08:23] <ajmitch> omega_: it'll need patched, i know it [08:24] <omega_> right, but if he's got an unpatched copy, he can test the check in autogen... <g> [08:24] <omega_> then patch it and get it built [08:24] <chillywilly> oh no, not the infamous automake leak [08:24] <omega_> yup [08:24] <derek> omega_: which one are you? i see no names, but this group despite poor licensing choices cant be all bad, i mean two of them ahve futbol jerseys on [08:24] <omega_> bottom left [08:24] <ajmitch> chillywilly: if you got about 1GB swap you should be ok i think [08:25] <omega_> ajmitch: um, I think it's up to 2GB now ;-) [08:25] <ajmitch> omega_: oh, ok, someone added another plugin? ;) [08:25] <omega_> think so [08:25] <chillywilly> heh [08:25] <omega_> aforementioned CEO now has a new kid.... [08:25] Action: omega_ downloads the pictures [08:26] <chillywilly> well, does debian sid's automake have those problems ajmitch? [08:26] <derek> ok want me to run [08:26] <derek> autogen [08:26] <omega_> yeah [08:26] <derek> any special param? [08:26] <omega_> nope [08:26] <chillywilly> what a motely crew [08:26] <omega_> it should fail because you don't have a patched automake [08:26] <chillywilly> ;) [08:27] <ajmitch> chillywilly: i think it might be a little old, yeah [08:27] <chillywilly> omega_: you have the Gnome shirt right? [08:27] <omega_> chillywilly: which? [08:27] <omega_> yeah, /me on bottom left [08:28] <omega_> top row right to left is zaheer, wtay, Ow3n, Uraeus, richardb [08:28] <ajmitch> omega_: it was bad in here today, i had to use my (very) limited knowledge of dutch to try & follow the conversation... [08:28] <omega_> bottom right is hadess [08:28] <omega_> ajmitch: uh oh, the rot13 filter didn't cut it? [08:29] <derek> hmmm [08:29] <derek> this work on win32? [08:29] <omega_> derek: nope [08:29] <omega_> could be ported, with some effort, but none here is gonna bother.... [08:29] <omega_> couple people have expressed interest, but didn't hear anything further [08:30] <omega_> shouldn't be too hard though [08:30] <ajmitch> derek: you'd want to use it on win32? [08:30] <chillywilly> you have your own GTK+ thread object? [08:30] <derek> no [08:30] <derek> i just say win32/ dir [08:30] <omega_> chillywilly: thread object? [08:30] <derek> i did get the error [08:30] <derek> about automake [08:30] <omega_> derek: ok, patch as per in README [08:30] <derek> so what need i do [08:30] Action: derek is too lazy to readme [08:30] <derek> ok will readme [08:30] <omega_> patch is in the README... [08:31] <chillywilly> omega_: you know a thread class [08:31] <chillywilly> GstThread???? [08:31] <omega_> no such thing as thread class, but we have a GStreamer element that happens to contain a thread, yeah [08:31] <ajmitch> omega_: good thing richardb changed it so it failed on old automake, rather than letting it run ;) [08:31] <omega_> ajmitch: quite... [08:32] <omega_> [01:51] <kenneth> wel, ik kan ook andere talen spreken [08:32] <omega_> eesh. yeah, um, what he said [08:33] <derek> so just edit my $PATH ? [08:33] <derek> would be course of action [08:33] <derek> or do you recommend [08:33] <omega_> derek: huh? /me hasn't read the README recently [08:33] <derek> grabbing new copy of automake [08:33] <derek> omega_: it says it created patched version [08:33] <omega_> oh, do you have root on the box? [08:33] <derek> and put in root dir [08:33] <omega_> if so, just patch it [08:33] <derek> how do i patch it [08:33] <omega_> it's a bugfix, so there's no reason not to, unless you prefer a pristine system [08:33] <derek> i dont know junk about automake [08:34] <omega_> I manually updated mine [08:34] <ajmitch> derek: automake is a perl script, edit it in emacs ;) [08:34] <omega_> just edit it and make it look like the +'s rather than the -'s [08:34] <chillywilly> eh? [08:34] <derek> where is it [08:34] <omega_> /usr/bin/automake [08:34] <derek> that was the question i SHOULD have asked :) [08:34] <chillywilly> patch -px < patchfile [08:34] <derek> chillywilly: yip thats what i plan to do :) [08:34] <omega_> chillywilly: there's more in the README than that.. [08:34] <omega_> but it should work, I think [08:35] <chillywilly> oh [08:35] <derek> ok will hand do [08:35] <chillywilly> sorry, don't listen to me [08:37] <derek> well actually didnt listen to anyone [08:37] <omega_> brb in couple minutes [08:37] <ajmitch> chillywilly: i never do... [08:37] <derek> i copied the patched version it made [08:37] <derek> and just copied over top fo my version [08:37] <derek> now its complaining wants newer libtool [08:38] <chillywilly> ajmitch: how did I know you were gonna chime in witha smart ass comment eh? [08:39] <ajmitch> chillywilly: i dunno, explain how [08:39] <derek> argh [08:39] <derek> someone know if rpms exist for libtool 1.3.5 [08:40] <ajmitch> rpms?? [08:40] <chillywilly> rpms are for weenies [08:40] <chillywilly> :P [08:40] <chillywilly> j/k [08:40] <chillywilly> ajmitch: this is on his laptop ya know [08:40] <chillywilly> (er, maybe) [08:41] <chillywilly> no debian on the laptop masat? [08:41] <chillywilly> masta [08:41] <derek> chillywilly: no kidding [08:42] <derek> no debian yet [08:42] <derek> im DYING for it though [08:42] <chillywilly> ;) [08:42] <derek> just have stuff on here [08:42] <derek> i need [08:42] <derek> that takes a bit to configure [08:42] <derek> so until im done with a project [08:42] <derek> and copy stuff over to another box [08:42] <derek> i cant switch [08:42] <derek> soon though [08:42] <derek> very soon i hope [08:42] <derek> i have debian box at home though :) [08:43] <ajmitch> good masta ;) [08:43] <derek> i could compile from source but would rather not [08:44] <chillywilly> and what is omega_ running? [08:44] <chillywilly> have you shown him the way ajmitch? [08:44] <derek> looks as though i will have to compile from source [08:45] <ajmitch> chillywilly: omega_ is running an unspeakable evil, iirc [08:45] <chillywilly> what's that? [08:45] <chillywilly> RedHat? [08:45] <ajmitch> RH, i think [08:45] <chillywilly> ugh [08:45] <ajmitch> i know [08:46] <ajmitch> at least it's not caldera or corel [08:46] <chillywilly> rm-techno-remix is feaky [08:46] <chillywilly> freaky [08:46] <derek> didnt i give that you you a long time ago [08:46] <derek> its still in my playlist :) [08:46] <chillywilly> maybe [08:46] <chillywilly> I think I may have gotten it off napster though [08:47] <chillywilly> derek: you know what the best program is though to run on your debian box? [08:47] <derek> configure.in:141: warning: AC_TRY_RUN called without default to allow cross compiling [08:47] Last message repeated 1 time(s). [08:47] <derek> configure.in: 56: required file `./ltconfig' not found [08:47] <derek> some warnings [08:48] <derek> chillywilly: whats that [08:48] <chillywilly> vrms [08:48] <chillywilly> ;) [08:48] <derek> checking whether ln -s works... yes [08:48] <derek> updating cache ./config.cache [08:48] <derek> ./ltconfig: ./ltconfig: No such file or directory [08:48] <derek> configure: error: libtool configure failed [08:48] <derek> configure failed [08:48] <derek> dneighbo@latitude:~/downloads/sound/gstreamer$ [08:48] <derek> chillywilly: yes its great :) [08:48] <ajmitch> hmm [08:48] <derek> btw omega_ i warned you was cursed [08:48] <chillywilly> libtoolize? [08:48] <ajmitch> [ajmitch @ ajmitch gstreamer-anon] libtool --version [08:48] <ajmitch> ltmain.sh (GNU libtool) 1.4 (1.920 2001/04/24 23:26:18) [08:49] <ajmitch> version 1.4 works great ;) [08:49] <chillywilly> stinking libraries [08:50] <chillywilly> we don't want to reuse code, I want to rewrite every program from scratch like the bad old days [08:50] <derek> dneighbo@latitude:~/downloads/sound/gstreamer$ libtool --version [08:50] <derek> ltmain.sh (GNU libtool) 1.4 (1.920 2001/04/24 23:26:18) [08:50] <derek> [6]+ Done emacs REQUIREMENTS [08:50] <derek> dneighbo@latitude:~/downloads/sound/gstreamer$ [08:50] <derek> ajmitch: i have same libtool [08:50] <derek> so now explain the problem :) [08:51] <chillywilly> type libtoolize in the top level once [08:51] <chillywilly> shit I need to go to bed [08:51] <ajmitch> chillywilly: nah, it's early [08:51] <ajmitch> chillywilly: getting dragged away? [08:52] <chillywilly> no [08:52] <chillywilly> just wanted to go to bed early [08:52] <derek> this probably has something to do with it [08:52] <derek> dneighbo@latitude:~/downloads/sound/gstreamer$ ./autogen.sh [08:52] <derek> Checking for patched automake... found. [08:52] <derek> I am going to run ./configure with no arguments - if you wish [08:52] <derek> to pass any to it, please specify them on the ./autogen.sh command line. [08:52] <derek> You should update your `aclocal.m4' by running aclocal. [08:52] <derek> configure.in:141: warning: AC_TRY_RUN called without default to allow cross compiling [08:52] Last message repeated 1 time(s). [08:52] <derek> configure.in: 56: required file `./ltconfig' not found [08:52] <ajmitch> chillywilly: shouldn't 'need' to run libtoolize, i think [08:52] <derek> bye chillywilly [08:52] <derek> omega_ you coming back soon? [08:53] <chillywilly> ajmitch: where do you think that script comes from? [08:53] <chillywilly> autogen.sh usually doe sit for you though [08:53] <derek> didnt help [08:53] <derek> i did libtoolize [08:53] <chillywilly> and? [08:53] <chillywilly> anything? [08:53] <ajmitch> chillywilly: libtoolize is called from autogen.sh [08:53] <ajmitch> # Generate configure.in and configure.ac [08:53] <ajmitch> sed <configure.base >configure.in '/^SUBSTFOR configure.ac:.*/d;s/^SUBSTFOR configure.in://g' [08:53] <ajmitch> sed <configure.base >configure.ac '/^SUBSTFOR configure.in:.*/d;s/^SUBSTFOR configure.ac://g' [08:53] <ajmitch> libtoolize --copy --force [08:54] <derek> i did that aj [08:54] <derek> libtoolize --force anyhow [08:54] <derek> i wiped cvstree [08:54] <chillywilly> just whip him masta [08:54] <derek> and reslurped [08:54] <derek> see if that will help [08:54] <ajmitch> hmm [08:54] <chillywilly> you and your broad band access [08:54] <chillywilly> makes me sick [08:54] <ajmitch> it works fine here [08:54] <chillywilly> :P [08:55] <derek> get EXACTLY the same thing on a fresh copy [08:55] <derek> no omega_ :( [08:55] <chillywilly> derek: you really are cursed [08:55] <derek> its almost bed time for me [08:55] <derek> chillywilly: i wasnt saying it to be funny [08:55] <ajmitch> lemme try on a fresh copy i have checked out [08:56] <chillywilly> derek: I know, I am just in awe of it [08:56] <ajmitch> is working fine on fresh copy [08:57] <derek> chillywilly: whats scary [08:57] <derek> i run REALLY standard stuff [08:57] <derek> i mean dell laptop NO modifications [08:57] <derek> RH 6.2 (stable old beast) [08:57] <derek> install almost everything from packages when i can [08:57] <derek> etc etc [08:57] <chillywilly> heh [08:58] <chillywilly> poor masta [08:58] <chillywilly> my system is very stable even though it is debian unstable :P [08:58] <derek> i owuld blame it on redhat [08:58] <derek> but my curse extends to debian as well :) [08:58] <chillywilly> :) [08:59] <derek> im an equal opprotunity jinx [08:59] <derek> did omega_ say if he was coming back? [08:59] <chillywilly> dunno [08:59] <derek> ajmitch: how is that test going? [08:59] <ajmitch> test? [09:00] <derek> you were gonna try [09:00] <derek> latest cvs [09:00] <derek> btw: where does ./ltconfig [09:00] <derek> come from [09:01] <ajmitch> libtoolize --copy --force [09:01] <derek> as thats what its complaining about [09:01] <derek> same thing [09:01] <ajmitch> i think [09:01] <derek> dneighbo@latitude:~/downloads/sound/gstreamer$ libtoolize --copy --force [09:01] <derek> You should update your `aclocal.m4' by running aclocal. [09:01] <derek> dneighbo@latitude:~/downloads/sound/gstreamer$ aclocal [09:01] <derek> dneighbo@latitude:~/downloads/sound/gstreamer$ ./autogen.sh [09:01] <derek> Checking for patched automake... found. [09:01] <derek> I am going to run ./configure with no arguments - if you wish [09:01] <derek> to pass any to it, please specify them on the ./autogen.sh command line. [09:01] <derek> You should update your `aclocal.m4' by running aclocal. [09:01] <derek> configure.in:141: warning: AC_TRY_RUN called without default to allow cross compiling [09:01] Last message repeated 1 time(s). [09:01] <derek> configure.in: 56: required file `./ltconfig' not found [09:01] <derek> so you will BELIEVE ME [09:02] <chillywilly> copy t by hand [09:02] <derek> copy what by hand? [09:02] <chillywilly> to heck with it [09:02] <chillywilly> ltconfig [09:02] <ajmitch> i just rm'ed ltconfig, ran libtoolize, and it didn't reload it [09:02] <chillywilly> heh [09:02] <chillywilly> how screwed up is that? [09:03] <derek> ok [09:03] <derek> better question [09:03] <derek> why dont i have an ltconfig? [09:03] <derek> doesnt it come out of cvs? [09:03] <ajmitch> nope [09:03] <ajmitch> it's generated, and in .cvsignore [09:03] <chillywilly> it is auto genrated [09:03] <chillywilly> jinx [09:03] <omega_> hrm, Ebert and Roeper just smashed "Pearl Harbor" to bits.... [09:04] <chillywilly> man [09:04] <chillywilly> it's bed time for real [09:04] <chillywilly> night night [09:04] <omega_> chillywilly: l8r [09:04] chillywilly (bau...@d6...) left irc: Philosophers and plow men, each must know his part, to sow a new mentality closer to the heart... [09:04] <derek> well i copied from /usr/share/libtool/ltconfig [09:04] <derek> and now autogen works [09:05] <derek> but it does get those two warnings [09:05] <derek> hi omega_ [09:05] <derek> you have missed my cursing :) [09:05] <omega_> heh [09:05] <derek> if you have a buffer [09:05] <derek> you may want to read back some [09:05] <derek> to see if there is anything that is buggy in gstreamer [09:05] <derek> or if its solely a libtool problem [09:06] ChiefHighwater (pa...@su...) joined #gstreamer. [09:06] Action: derek is now actually compiling gstreamer (yeah) [09:06] <ajmitch> hey ChiefHighwater [09:07] <ChiefHighwater> Ello ajmitch 8-] [09:07] <ajmitch> hmm, ltconfig isn't autogenned on my system, but it's not complaining [09:08] <derek> well kiss my bootie then ajmitch ;) [09:08] <ajmitch> this is on a fresh checkout of gstreamer [09:08] <omega_> derek: still have the ltconfig problem? [09:08] <ajmitch> maybe ltconfig is needed by older versions of some tools? [09:08] <derek> well yes and no [09:08] <derek> the problem is there [09:08] <omega_> ChiefHighwater: derek is after the same stuff we're trying to get sonic to do [09:09] <derek> i hand copied it from /usr/share/libtool [09:09] <ChiefHighwater> kewl [09:09] <derek> to the root [09:09] <derek> and it let things slide [09:09] <derek> but gets cross compile warnings [09:09] <derek> i ignored the warnings and am compiling anyhow [09:09] <omega_> yeah, the cross warnings are annoying [09:09] <ajmitch> derek: hmm, ltconfig is designed for creating a system-specific libtool, apparantly (am looking at source) [09:10] <derek> omega_: what is sonic [09:10] <derek> and sounds as though i should be interested in it [09:10] <ChiefHighwater> 8-] [09:10] <omega_> derek: machine name [09:11] <ajmitch> ChiefHighwater: how are u this evening/morning? ;) [09:11] <omega_> it's supposed to be the mixing machine, but it's minus a mobo currently [09:11] <ajmitch> hehe [09:11] <omega_> ajmitch: 12:11am [09:11] <omega_> so, um, morning? <g> [09:11] <ajmitch> yeah, thought it was morning there ;) [09:11] Action: ChiefHighwater is tired [09:11] Action: ChiefHighwater conducted a big conct this evening [09:11] <ajmitch> cool ;) [09:11] Action: ChiefHighwater can't even type [09:12] <ChiefHighwater> concert [09:12] <ajmitch> i figured ;) [09:12] <omega_> looks like CHW needs to look it up in the dictionary too <g> [09:12] <derek> omega_ if you get stuff out on mp3 [09:12] <derek> let me know [09:12] <omega_> ok [09:12] <derek> i would be interested in hearing it [09:12] <omega_> ChiefHighwater: any requests for CD's yet? <g> [09:12] <ajmitch> derek: good thing GNUe is BugFree (TM), eh? ;) [09:12] <ChiefHighwater> hehe [09:13] <omega_> ChiefHighwater: can try to do a record via w2k of the tape, get it to CD and mp3 [09:13] <ChiefHighwater> um, we could..yeah [09:13] <derek> man gstreamer takes a bit to compile :) [09:13] <derek> good thing its night time out [09:13] <omega_> derek: yup. takes 15+min on my P3/500 [09:14] <ajmitch> derek: should run autogen.sh with --disable-tests --disable-examples to speed it up (if you don't want the funky demo apps) [09:14] <derek> not complaining [09:14] <derek> just noting [09:14] <ajmitch> derek: i complained about it ;) [09:15] <omega_> ajmitch even submitted the patch he just mentioned [09:15] <derek> see for me [09:15] <derek> my complaint would be that there werent binaries [09:15] <derek> and that to get new features have to get cvs [09:15] <derek> and would say [09:15] <derek> really you need to release more often :) [09:15] <ajmitch> omega_: that's why i was allowed to complain about it - because it was one of the few things i could fix ;) [09:16] <ajmitch> derek: when was last GNUe release? ; [09:16] <derek> but i cant gnue to release either so im not about to complain to someone else :) [09:16] <omega_> derek: yeah, need to release more, an 0.2.0 is due soon [09:16] <omega_> binaries come with that release (debs, even, if taaz gets in gear) [09:16] <derek> its a no win situation [09:16] <derek> if gstreamer is like gnue [09:16] <derek> its so young [09:16] <derek> that features are going like mad [09:17] <derek> and so you constantly tell people get cvs [09:17] <derek> the optimum would be to release about every 2 weeks [09:17] <omega_> derek: yup. gonna stay that way for a bit more, then hopefully settle down soon [09:17] <derek> but it takes away from coding time to relleas [09:17] <derek> so releases happen every month at best [09:17] <derek> and more like every 6 weeks [09:17] <omega_> yup, we haven't added much new stuff in the last few days tryig to get the release happening [09:17] <ajmitch> yeah, to release you have to freeze features to try & get a semi-stable release [09:18] <omega_> the main thing is that the perious of instability here have been a bit long, so a release would be a waste of time and counterproductive [09:18] <omega_> I hope that speeds up now that the major stuff is donw [09:18] <omega_> er, done [09:18] <ajmitch> omega_: 0.3.0 will be glib 1.3.x/2.0 dependant? [09:18] <omega_> ajmitch: yeah, something along those lines [09:18] <omega_> we're still trying to figure out how to proceed there [09:19] <ajmitch> k [09:19] <derek> dependencies are evil :) [09:19] <derek> but necessary (sigh) [09:19] <ajmitch> i got glib from cvs so i'm happy ;) [09:22] <derek> curious [09:22] <ajmitch> omega_: i get an incredible amount of warnings when compiling some of the mpeg1 plugins (using gcc 3.0 snapshot) [09:22] <derek> how much are the plugins changing? [09:22] <omega_> derek: depends on which ones <g> [09:22] <derek> im wondering if you would be better off [09:22] <derek> releasing binaries of the plugins [09:22] <derek> then in cvs [09:22] <omega_> moving them out of the main package? that's the plan [09:22] <derek> have option to not compile them [09:23] <derek> which would speed things up greatly [09:23] <omega_> binaries aren't viable separately until we have a fully stable API [09:23] <derek> ok [09:23] <ajmitch> derek: i've added that option to my local copy here ;) [09:23] <omega_> we're still changing things that affect every plugin [09:23] <omega_> ajmitch: howso? [09:23] <derek> gotcha [09:24] <derek> bad thing about free software [09:24] <derek> waiting for it to compile :) [09:24] <derek> im anxious to use this thing :) [09:24] <ajmitch> omega_: same way as i added the option to disable test & examples from building [09:24] <derek> im like kid waiting for dessert [09:24] <ajmitch> dunno if it works yet, haven't tried it ;) [09:25] <derek> my luck [09:25] <derek> i will get it compiled [09:25] <derek> you all will leave [09:26] <derek> and i wont be able to get ti to work :) [09:26] <derek> remember the curse [09:26] <omega_> heh [09:27] <ajmitch> hehe [09:27] <mwc> should v4lsrc be working? If so, how do I use it? [09:27] <ajmitch> omega_: i don't think there's much point me committing the --disable-plugins patch, is there? ;) [09:28] <omega_> ajmitch: um, no <G> [09:28] <omega_> mwc: I don't think it's fully functional right now [09:28] <ajmitch> btw, here are some warnings i said about: [09:28] <ajmitch> system_encode.c:140:3: warning: pasting "(" and ""pick_streams: %lld, %lld\n"" does not give a valid preprocessing token [09:28] <ajmitch> system_encode.c:201:3: warning: pasting "(" and ""system_encode::multiplex: update mta\n"" does not give a valid preprocessing token [09:28] <ajmitch> system_encode.c:238:3: warning: pasting "(" and ""system_encode::multiplex: data_rate %u, video_rate: %u, audio_rate: %u\n"" does not give a valid preprocessing token [09:28] <omega_> but you can try -launch v4lsrc ! xvideosink [09:28] <omega_> ajmitch: hmm, lemme look [09:28] <mwc> omega_: That's what I tried. Lots of assertion failures. Oh well. [09:29] <omega_> hmm, check with wtay, he wrote v4lsrc [09:29] <omega_> ajmitch: do you get a lot of these building libgst.la ? [09:29] <ajmitch> omega_: not that i've seen, but i son't pay much attention... [09:30] <omega_> check, because that's GST_DEBUG there.... [09:30] <ajmitch> ok, will check cron results... [09:30] <derek> do i need to do a make install? [09:30] <derek> its done [09:31] <derek> its done [09:31] <omega_> no, don't install [09:31] <omega_> just run it in place [09:31] steveb (st...@no...) joined #gstreamer. [09:31] <omega_> there's a reason for that<g> [09:31] <omega_> yo [09:31] <steveb> yo [09:31] <ajmitch> derek: i got told off for installing it [09:31] <ajmitch> hey steveb [09:31] <steveb> hi [09:31] <derek> well i asked :) [09:31] <ajmitch> derek: meet a kiwi who's not called andrew ;) [09:31] <derek> ok i have it now what [09:31] <omega_> derek: go to tools [09:31] <derek> steveb: are you in counseling? [09:31] <omega_> make sure you can write to /etc/gstreamer [09:32] <omega_> run ./gstreamer-register [09:32] <derek> steveb: as it must be emotionally traumatizing to be NZ'er not named andrew [09:32] <omega_> derek: good point <g> [09:33] <derek> ok [09:33] <derek> registered [09:33] <mwc> ajmitch: we have two kiwi's here - one is called andrew the other is not! [09:33] <ajmitch> omega_: get none of those warnings with libgst.la, btw [09:33] <derek> now what [09:33] <ajmitch> mwc: only kiwis derek has met are called andrew ;) [09:34] <omega_> derek: now try: [09:34] <steveb> the ones that aren't called andrew are called steve [09:34] <omega_> ./gstreamer-launch fakesrc ! fakesink [09:34] <mwc> ajmitch: if derek comes here I will introduce him to Mark. (and an Andrew) [09:35] <derek> omega_: its scrolling stuff like nuts [09:35] <derek> same thing over and over [09:35] <omega_> derek: Ctrl-C [09:35] <derek> fakesrc: ******* (fakesrc0:src)> [09:35] <derek> fakesink: ******* (fakesink0:sink)< (0 bytes) [09:35] <derek> i did [09:36] <omega_> now, do you have libmad or libvorbis installed? [09:36] <derek> um beats me [09:36] <derek> how do i tell [09:36] <omega_> ./gstreamer-inspect | grep mad [09:36] <omega_> or grep voris [09:36] <derek> i thought i grabbed libogg the other day [09:36] <omega_> er, vorbis [09:36] <omega_> derek: if you think dependencies are evil, read REQUIREMENTS [09:36] Action: omega_ expects never to hear from derek again <g> [09:36] <derek> [root@latitude tools]# ./gstreamer-inspect | grep mad [09:36] <derek> INFO ( 3770:-1) Initializing GStreamer Core Library [09:36] <derek> INFO ( 3770:-1) CPU features: (0c040882) MMX SSE [09:36] <derek> [root@latitude tools]# [09:36] <omega_> well, no mad then [09:37] <derek> [root@latitude tools]# ./gstreamer-inspect | grep vorbis [09:37] <derek> INFO ( 3788:-1) Initializing GStreamer Core Library [09:37] <derek> INFO ( 3788:-1) CPU features: (0c040882) MMX SSE [09:37] <derek> [root@latitude tools]# [09:37] <ajmitch> oops ;) [09:37] <omega_> yeah. the majority of the interesting plugins simply wrap existing libraries (as it should be) [09:37] <omega_> so you have to get those libraries before configuring and building gstreamer [09:37] <omega_> so pick the ones you want, get them, and then rebuild <g> [09:38] <derek> yikes [09:38] <derek> you have to rebuild [09:38] <omega_> hehehe [09:38] <derek> thats painful [09:39] <omega_> well, yeah, it's a compile-time dependency [09:39] <omega_> each plugin has to link against dependent libs, that's not unusual [09:39] <derek> definitely need to get those plugins separated otu [09:39] <derek> out [09:39] <omega_> yup [09:39] <omega_> you don't have to make clean, if that's what you mean [09:39] <derek> ok tell me what i should need and i will go get [09:39] <ajmitch> rebuilding isn't that painful [09:39] <derek> i just installed libogg [09:39] <omega_> but because config.h is used so much, it tends to have about the same effect [09:39] <omega_> and libvorbis? [09:39] <derek> (this is when i HATE not running debian) [09:40] <omega_> what are you running? [09:40] <ajmitch> omega_: pure evil [09:40] <derek> rh6.2 on this laptop (and hating it) [09:40] <omega_> ajmitch: never heard of that distro [09:40] <omega_> derek: can you upgrade to at least 7.0 or something? [09:40] <derek> um if i do anything [09:40] <derek> it will be switch to debian :) [09:40] <omega_> he [09:40] <omega_> um, well... [09:40] <omega_> lemme check something [09:41] <derek> only reason this is still RedHurt is i have stuff i cant eliminate yet [09:41] <omega_> well, to do what you're after, you need at least mad and lame [09:41] <omega_> or libvorbis if you want to do .ogg's instead of .mp3's [09:41] <ajmitch> omega_: yet another person who recognises the superiority of debian ;) [09:41] <omega_> ajmitch: hehe [09:42] <derek> ok [09:42] <derek> question [09:42] <derek> is there way to tell on red hurt if those things are installed? [09:43] <derek> whats in the requirements file? [09:43] <omega_> rpm -qa | grep xyz [09:43] <ajmitch> rpm -qa |grep lame [09:43] <derek> otay [09:43] Action: omega_ needs to call Dell tomorrow, his new laptop drive is making really nasty whining noises [09:43] <ajmitch> omega_: not good, that the one you code on? [09:44] <omega_> ajmitch: yup [09:44] <omega_> it's been doing this to varying degrees for a couple weeks [09:44] <ajmitch> omega_: too close to 0.2.0 release to have a laptop die! [09:44] <omega_> ajmitch: um, it's called CVS <g> [09:44] <omega_> but yeah, that'd suck [09:44] <steveb> should the refcount on a buffer be inced for the duration of a chain function? [09:44] <ajmitch> omega_; i meant, we cannot afford to have you away from keyboard [09:45] <omega_> steveb: no, only if you need it beyond that point [09:45] <omega_> ajmitch: I have other machines I can use [09:45] <steveb> ok [09:45] <omega_> steveb: the default refcount of 1 is passed from element to element [09:46] <omega_> only if it's held up for some reason (like the autoplugcache) does the refcount go up [09:46] <steveb> and if the buffer is replaced with a different one, it just needs to be unreffed? [09:47] <omega_> yup [09:48] <derek> omega_: i have lame [09:48] <omega_> derek: and liblame ? [09:48] <derek> but i dont see liblame [09:48] <derek> are they diff [09:48] <omega_> ok, the usual lame packages really are lame.... (pun intended) [09:48] <omega_> your best bet is to get the latest lame from sourceforge [09:48] <omega_> build and install that [09:49] <omega_> you'll get better mp3s too [09:49] <derek> okok [09:49] <derek> you realize [09:50] <derek> this is more compiling i have done on stuff im not working on [09:50] Nick change: ChiefHighwater -> CHWZZzz [09:50] <derek> in like 8 mos [09:51] <omega_> heh [09:51] <ajmitch> derek: keeps you in practice [09:51] <omega_> well, problem is a lot of things like lame don't have good packages at all [09:51] <omega_> the best package I've ever seen is the one I made myself [09:51] <omega_> for lame, that is [09:51] <steveb> omega_: so what is a meaningful benchmark test for ladspa? [09:51] <derek> omega_: yes this does require [09:51] <derek> that if i use this thing [09:51] <derek> i will probably start contributing [09:51] <derek> i mean after all just getting it running [09:52] <omega_> steveb: bench? um, does it run? [09:52] <derek> is a big time investment ;) [09:52] <omega_> derek: yeah, the entry curve is a little steep <g> [09:52] <steveb> mono _chain and _ get does atm [09:52] <omega_> steveb: cool. um. freeverb come in mono? [09:52] <steveb> on my box that seems to be all it comes in [09:53] <omega_> hmm, well, try it <g> [09:53] <derek> omega_: i got oggvorbis nightly [09:53] <derek> and it creates 5 dirs [09:53] <derek> which one do i need? [09:53] <derek> just vorbis? [09:53] <omega_> no, vorbis depends on ogg afaik [09:55] <derek> argh [09:55] <steveb> oh, my freeverb is stereo [09:55] <ajmitch> derek: you sound like you are having much fun ;) [09:56] <derek> hmmm [09:56] <derek> mad just failed to compile [09:56] <derek> make[2]: Leaving directory `/home/dneighbo/downloads/mad-0.13.0b/libmad' [09:56] <derek> Making all in winamp [09:56] <derek> make: read.c:892: read_makefile: Assertion `*p2 != '\0'' failed. [09:56] <derek> make[1]: *** [all-recursive] Error 1 [09:56] <derek> make[1]: Leaving directory `/home/dneighbo/downloads/mad-0.13.0b' [09:56] <derek> make: *** [all-recursive-am] Error 2 [09:56] <derek> dneighbo@latitude:~/downloads/mad-0.13.0b$ [09:56] <omega_> derek: welcome to the world where package management hasn't caught up with CVS HEAD yet [09:56] <omega_> derek: try 0.12.3 [09:56] <omega_> or get gstreamer.net/releases/0.2.0/support/mad-0.13.0b-2.i386.rpm and try that [09:57] <omega_> it was probably build on rh7.1 though.... [10:00] <derek> same issue with .12 [10:00] <derek> i think its cause its failing on making some crap for winamp [10:01] <omega_> winamp?? [10:01] <omega_> thought you were running rh6.1, not rh95 <g> [10:01] <derek> btw: that support rpm stuff is way to go [10:01] <omega_> er, um, win95 <g> [10:01] <derek> enlightenment used to do that [10:01] <derek> as it had litney of dependencies [10:02] <derek> and yip rpm was mad in rh7.x [10:03] <derek> guess its sign to go to bed :( [10:03] <derek> maybe some other time [10:03] <omega_> yeah, bed is good [10:03] <omega_> come back tomorrow evening, try again <g> [10:04] <derek> for vorbis [10:04] <derek> i only need libvorbis? [10:04] <derek> or will i need lame too? [10:04] <ajmitch> night then derek... [10:05] <ajmitch> lame is an mp3 encoderm not needed for vorbis [10:05] <ajmitch> i don't have lame [10:05] <omega_> derek: lame encodes mp3, vorbis enc/dec's oggs [10:05] <omega_> mad decodes mp3 [10:05] <derek> ok [10:05] <omega_> and REQUIREMENTS has the rest ;-) [10:05] <derek> im trying [10:05] <derek> libvorbis [10:05] <derek> if i get it to compile i might try this [10:06] <derek> but i probably dont have necessary junk to play an .ogg file in xmms :) [10:06] <derek> or will gstreamer play it back for me? [10:06] <derek> its bed time [10:06] <derek> libvorbis failed [10:06] <derek> it wants ogg :) [10:06] <omega_> if you have libvorbis and configure gst with it, yes [10:06] <omega_> derek: yeah, it requires ogg [10:06] <omega_> not sure what for though [10:07] <ajmitch> derek: use debian & grab the debian packages ;) [10:07] <derek> ajmitch: NO KIDDING [10:07] <derek> thats why this is KILLING ME [10:07] <derek> omega_ you in here often? [10:07] <omega_> derek: all the time [10:08] <derek> someday when i get all dependencies installed i will come back [10:08] <derek> er 'requirements' [10:08] <derek> :) [10:08] <omega_> you only need the ones you're interested in using [10:08] <ajmitch> to think i build this beast nightly... [10:09] <ajmitch> plugins seem to be added at the rate of 1 a week at the moment ;) [10:09] <derek> funny thin [10:09] <derek> g [10:09] <derek> is that libvorbis says it wants ogg [10:09] <derek> but [10:09] <derek> [root@latitude gstreamer]# rpm -qa |grep ogg [10:09] <derek> libogg-1.0beta4-ximian.3 [10:09] <derek> [root@latitude gstreamer]# [10:09] <ajmitch> maybe i shoudl try & get gnue to build for once [10:09] <derek> i had just grabbed this [10:09] <ajmitch> you need the -devel package i'd say [10:11] <derek> me also wonders why they have the libogg [10:11] <derek> but not libvorbis [10:12] hadess (ha...@pc...) joined #gstreamer. [10:13] <derek> crud [10:13] <hadess> hey derek [10:13] <derek> i might have oggvorbis installed now [10:13] <derek> hi hadess [10:13] <ajmitch> hi hadess [10:13] <hadess> hey ajmitch [10:14] <ajmitch> hadess: watch out, he's seen your pic from guadec ;) [10:14] <derek> omega_: if so what do i need to do to recompile gstreamer? [10:14] <derek> just rerun ./autogen.sh ? [10:14] <omega_> derek: yeah, then make [10:14] <omega_> and hope that it doesn't take too long [10:14] <omega_> in the meantime, I must sleep [10:14] <walken> omega. did you progress on the dct parsing stuff ? [10:14] <omega_> walken: not today [10:14] <walken> ok [10:14] <omega_> tomorrow sometime [10:14] <derek> ok [10:15] <derek> omega_ then do i rerun that register [10:15] <omega_> derek: then once it's all built, do ./gstreamer-launch osssrc ! lame bitrate=128 ! disksink location=output.mp3 [10:15] <omega_> yes, after -register [10:15] <hadess> hey omega_ [10:15] <omega_> yo hadess [10:15] <derek> i thought i didnt need lame? [10:15] <omega_> derek: if you want to encode mp3s you do [10:15] <omega_> else use vorbisenc [10:15] <derek> what was libogg for? [10:16] <hadess> ajmitch: why is the fact that he's seen pictures of me any bad for me <g> [10:16] <derek> er libvorbis? [10:16] <omega_> dep for vorbis [10:16] <derek> will that give me vorbisenc? [10:16] <derek> or no [10:16] <omega_> yes [10:16] <derek> so waht is command line? [10:16] <ajmitch> hadess: helps him know that you're not the nice, kind, friendly person you appear to be on IRC ;) [10:17] <omega_> just replace lame with vorbisenc [10:17] <hadess> ajmitch: lol [10:17] <derek> ./gstreamer-launch osssrc ! vorbisenc bitrate128 ! disksink location=file.ogg [10:17] <omega_> and ditch the bitrate=128 because vorbisenc doesn't have that yet [10:17] <derek> ok [10:17] <ajmitch> derek: when do you need all this working by? [10:17] <derek> ./gstreamer-launch osssrc ! vorbisenc ! disksink location=file.ogg [10:17] <omega_> atm there's still only the default quality for vorbis, which is ~128 I think [10:17] <omega_> then to play back, disksrc location=file.ogg ! vorbisdec ! osssink [10:17] <omega_> just turn the pipeline around, basically [10:18] <derek> cool [10:18] <derek> i hope it works [10:18] <omega_> it should [10:19] <derek> do i need to do that register thing? [10:19] <omega_> but I must sleep now [10:19] <omega_> after rebuild yet [10:19] <omega_> er, yes [10:19] <derek> ./gstreamer-register or whatever [10:19] <derek> ok cool [10:19] <derek> night [10:19] <derek> i need sleep too [10:19] <omega_> l8r all [10:19] <ajmitch> night [10:19] <steveb> l8er [10:19] omega_ (om...@om...) left irc: killall -SLEEP omega_ [10:28] makki (yv...@ma...) joined #gstreamer. [10:30] <ajmitch> hi makki [10:46] Nick change: ajmitch -> ajbusy [11:00] <makki> hello [11:16] Nick change: mwc -> mwczzz [11:40] walken (fo...@c1...) left irc: l8r [11:52] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee4c.a2000.nl] [11:53] steveb (st...@no...) joined #gstreamer. [12:09] makki (yv...@ma...) left irc: Ping timeout for makki[makki.wait-a-bit.com] [12:15] Nick change: hadess -> hds-busy [12:22] thomas (th...@ad...) joined #gstreamer. [12:25] hds-busy (ha...@pc...) left irc: Ping timeout for hds-busy[pc2-guil2-0-cust150.gui.cable.ntl.com] [12:28] richardb (ri...@ix...) joined #gstreamer. [12:29] hds-busy (ha...@pc...) joined #gstreamer. [12:54] Nick change: ajbusy -> ajmitch [13:19] <richardb> Hurray! [13:19] <richardb> Just got docs/gst to compile without patched tools... [13:21] <ajmitch> yay! ;) [13:22] <richardb> Are you one of the people who's been having problems with that? [13:25] <richardb> Fix committed, anyway. [13:28] <ajmitch> no, i always compiled with --disable-docs-build [13:41] <thomas> anyone use abcde ? [13:48] <richardb> thomas: no, but I might in future. Is it good? [13:56] <thomas> richardb: it looks pretty good, but I'm trying to get freedb in instead of cddb [13:56] <thomas> richardb: I need a one-stop rip shop for my gbox :) [14:20] richardb (ri...@ix...) left irc: Work [14:25] Nick change: wtay-zZz -> wtay [14:25] <wtay> yo [14:26] <steveb> to [14:26] <steveb> er, yo [14:29] ajmitch (aj...@p3...) left irc: Ping timeout for ajmitch[p32-max3.dun.ihug.co.nz] [14:30] ajmitch (aj...@p2...) joined #gstreamer. [14:30] <wtay> thomas: I'm looking at lame now.. [14:30] <ajmitch> yo [14:31] <wtay> yo aj [14:31] Action: ajmitch knows it is time to go to bed when wtay gets up ;) [14:32] <hds-busy> hey wtay [14:32] Nick change: hds-busy -> hadess [14:33] <ajmitch> night all [14:33] <hadess> night dude [14:33] <wtay> ajmitch: cya [14:33] Nick change: ajmitch -> ajzzzz [14:33] <wtay> hey hadess [14:33] <hadess> wtay: you don't work today ? [14:33] <wtay> nope, a holiday [14:34] <hadess> what holiday is it ? [14:34] <ajzzzz> hehe, it was holiday here too [14:34] <hadess> it isn't here... [14:34] <wtay> pinksteren? [14:34] <wtay> the monday after [14:34] <ajzzzz> queen's birthday here (any excuse for a holiday...) [14:35] sienap (sy...@ip...) joined #gstreamer. [14:35] <steveb> pink day here [14:35] <hadess> pentecost [14:35] <sienap> hadess! [14:35] <sienap> steveb! [14:35] <sienap> paripraxi has been on lately ? [14:35] <steveb> sienap! [14:35] <wtay> yo sienap [14:35] <sienap> hej wtay! [14:35] <hadess> hey sienap [14:35] <ajzzzz> ah, you celebrate pentecost with a holiday? [14:35] <sienap> let's checkout pre2 :) [14:36] <ajzzzz> hey sienap [14:36] <sienap> hej ajzzzzzzzzzzzzzzzzz ;) [14:36] <hadess> ajzzzz: they do, we don't [14:36] <sienap> but has paripraxi been here lately ? :) [14:36] <wtay> sienap: yes [14:36] <ajzzzz> sienap: hey, i'm just going to bed now... ;) [14:36] <sienap> he [14:36] <sienap> ok [14:36] <sienap> :) [14:36] <sienap> hope he'll be on tonight ;) [14:37] <sienap> ajzzz i just got out of it :) [14:37] <sienap> Mon Jun 4 14:41:03 CEST 2001 [14:37] <sienap> :) [14:37] <ajzzzz> ;) [14:37] <ajzzzz> Tue Jun 5 00:36:22 NZST 2001 [14:37] <sienap> my vacation somehow just begun :) [14:37] Action: ajzzzz has exams starting on saturday [14:37] <ajzzzz> so i'll be studying hard ;) [14:37] <sienap> hehe :) [14:38] <ajzzzz> anyway, night all (again) [14:38] <hadess> on saturday i get a new lappy, that'll be more interesting than exams :P [14:38] Action: ajzzzz is away: sleep... [14:38] <sienap> night dude [14:38] <wtay> cya [14:40] <sienap> hadess :) [14:41] <steveb> i get these alsa entries with -inspect: [14:41] <steveb> alsasrc: ESS AudioDrive ES1688(0/0)-src: ESS AudioDrive ES1688 at 0x220, irq 5, dma 1 [14:41] <steveb> alsasrc: alsasrc: ESS AudioDrive ES1688 at 0x220, irq 5, dma 1 [14:41] <steveb> alsasink: ESS AudioDrive ES1688(0/0)-sink: ESS AudioDrive ES1688 at 0x220, irq 5, dma 1 [14:41] <steveb> alsasink: alsasink: ESS AudioDrive ES1688 at 0x220, irq 5, dma 1 [14:41] <wtay> cool [14:42] <steveb> those nasty names are crashing gst-compprep [14:42] <wtay> uh? [14:44] <steveb> it looks like it is actually registering an element with the short name "ESS AudioDriv... [truncated message content] |
From: <wim...@ch...> - 2001-06-07 04:29:23
|
[06:31] wtay-sleep (wi...@ca...) joined #gstreamer. [06:32] <taaz> heh... LAD suggestion they need a wiki like gstreamers ;) [06:35] wtay-sleep (wi...@ca...) got netsplit. [06:36] wtay-sleep (wi...@ca...) returned to #gstreamer. [06:39] Action: ajmitch is back (gone 00:36:43) [06:44] walken (fo...@c1...) joined #gstreamer. [07:16] walken (fo...@c1...) left irc: l8r [07:21] ajmitch (aj...@p5...) left irc: Ping timeout for ajmitch[p53-max11.dun.ihug.co.nz] [07:25] ajmitch (aj...@p5...) joined #gstreamer. [07:45] Nick change: taaz -> taazzzz [08:09] derek (de...@cp...) left irc: Ping timeout for derek[cpe-24-221-112-50.az.sprintbbd.net] [08:25] <CHW_sicko> gnite all 8-] [08:25] <omega_gst> um [08:25] <omega_gst> question [08:26] <CHW_sicko> k [08:26] <omega_gst> is that trip to SJ the weekend of the 7th? [08:26] <CHW_sicko> nope..next weekend [08:26] <CHW_sicko> second saturday [08:26] <omega_gst> ok, good [08:26] <CHW_sicko> no wedding that weekend 8-] [08:26] <omega_gst> right [08:27] <omega_gst> l8r [08:27] <CHW_sicko> l8rz 8-] [08:27] CHW_sicko (pa...@su...) left irc: [08:28] <ajmitch> heh, just saw your mail on the number of gstreamer commits [08:28] <ajmitch> and this is when it's in a semi-freeze for release? ;) [08:30] <omega_gst> yup <g. [08:31] <ajmitch> think we can get gstreamer cvs count in the next gnome summary? ;) [08:31] <omega_gst> prob not [08:31] <ajmitch> think it'll be part of gnome 2.0? [08:31] <omega_gst> prob so [08:31] <ajmitch> kewl [08:32] <ajmitch> lots of users coming here with problems.... [08:32] <omega_gst> yipee [08:32] <ajmitch> ppl like me saying 'wait for the great omega_ to arrive...' [08:32] <omega_gst> if that happens, we may need to move to #gst-dev and leave #gstreamer for the users [08:33] <omega_gst> so we can at least keep the conversations somewhat separate [08:33] <ajmitch> yes [08:33] <ajmitch> i'd like to try & help with the core at some stage [08:34] <ajmitch> keep #gst-dev invite-only or similar? [08:34] <omega_gst> sorta, yeah [08:34] <ajmitch> cos i know that discussions can get pretty technical here for users [08:34] <omega_gst> more likely just 'private' in that we don't tell people unless they need to know, and don't mention it publicly [08:35] <ajmitch> k [08:35] <omega_gst> right, and interleaving user and developer discussions can lead to some pretty incomprehensible logs [08:35] <ajmitch> yup [08:35] mwc (mal...@ab...) joined #gstreamer. [08:35] <ajmitch> since they sorta deal with different issues [08:36] <ajmitch> plus if gstplay becomes popular then we'd have other sorts of users in here [08:36] <omega_gst> right [08:37] <ajmitch> stuff like 'd00d, y0r software suz cos i can't play this divx pr0n on my p100!!!' [08:46] wtay-sleep (wi...@ca...) got netsplit. [08:47] wtay-sleep (wi...@ca...) returned to #gstreamer. [08:50] derek (de...@cp...) joined #gstreamer. [09:22] steveb (ste...@21...) joined #gstreamer. [09:23] <steveb> yo [10:01] omega_gst (om...@om...) left irc: killall -SLEEP omega_gst [10:08] steveb (ste...@21...) left irc: quit has steveb, yes [10:52] Nick change: ajmitch -> ajzzzz [10:52] Action: ajzzzz is away: sleep... [11:08] ajzzzz (aj...@p5...) left irc: Ping timeout for ajzzzz[p53-max11.dun.ihug.co.nz] [11:08] ajzzzz (aj...@p5...) joined #gstreamer. [11:43] thomas (th...@21...) joined #gstreamer. [11:43] ajzzzz (aj...@p5...) left irc: Ping timeout for ajzzzz[p5-max2.dun.ihug.co.nz] [11:44] ajzzzz (aj...@p2...) joined #gstreamer. [12:09] thomas (th...@21...) left irc: [x]chat [12:09] thomas (th...@21...) joined #gstreamer. [12:09] <thomas> [13:44] big_T (th...@fy...) joined #gstreamer. [14:34] thomas (th...@21...) left irc: [x]chat [14:46] thomas (th...@21...) joined #gstreamer. [15:25] dobey (do...@dr...) joined #gstreamer. [15:51] ajzzzz (aj...@p2...) left irc: Ping timeout for ajzzzz[p20-max2.dun.ihug.co.nz] [16:15] steveb (ste...@21...) joined #gstreamer. [16:15] <steveb> thomas: i saw that :) [16:15] <thomas> steveb: yeah, good idea or not ? ;) [16:15] <thomas> care to help ? [16:16] <steveb> yep, good idea [16:16] <thomas> I'm starting it out easy; I'm still learning how this dynamical library stuff works [16:16] <thomas> it only has one function, one to calculate the real-time length of a buffer for audio processing [16:16] <thomas> since I need that in a few plugins [16:17] <thomas> but it's all uphill from there ;) [16:17] <thomas> let me commit the other stuff using it first, it seems to work and I want to test that so as to not break the compilation like last time ;) [16:17] <steveb> i started something like this a while ago - i was thinking of putting common code snippets like IIR/FIR filters and stuff [16:17] <thomas> steveb: yeah, I'm also looking at extracting some sox stuff to do resampling [16:18] ajzzzz (aj...@p5...) joined #gstreamer. [16:18] <thomas> also, I'd like to move the RMS routines I've written to that library as well [16:19] <steveb> cool [16:20] <steveb> i might do many of the snippets as macros - just for speed [16:21] <thomas> yeah - I suppose the buffer length function could be done with a macro as well, but I'll worry about that later ;) [16:21] <steveb> although if you were thinking of putting common caps/padtemplates in this, a real api would be better than a bunch of #defines [16:22] <thomas> steveb: wtay doesn't like the common caps either, I'm still not sure how I should do that [16:22] <thomas> I really think plugin writers should try to adhere to some common caps structures; they should only worry about getting the basic DSP routines right [16:22] <thomas> so if you have any suggestions for that ... [16:23] <steveb> i was just saying do it with functions, not defines. figure out an api which will make it easy for most people to create caps for common audio formats [16:24] <thomas> hmmm... what's the benefit in using functions for that ? [16:24] <steveb> with caps for example, they will almost always want to set the rate and channels [16:25] <thomas> steveb: I agree on channels, but a lot of filter plugins will not enforce any rate specification other than what is considered a sensible limit [16:25] <steveb> and padtemplates channels would be useful too [16:26] <steveb> are you confusing caps and padtemplates? if you're talking about padtemplates i agree [16:29] <thomas> steveb: yeah, I was talking about padtemplates [16:29] <thomas> for the caps, most filter-type plugins will proxy most of them [16:30] <steveb> unless the developer actually cares about the rate, it should just be left out of padtemplates [16:31] <thomas> steveb: maybe that's an option too... I don't know what the consequences are of not specifying template stuff [16:31] <thomas> hmmm, anyone know why gtk-config is not installed in the path when you do make install when compiling from source ??? [16:31] <steveb> no :) [16:43] <dobey> thomas: uhm, you did --prefix=/something? [16:51] <thomas> dobey: I found the error, I didn't have /usr/local/bin in my root path variable [16:51] <thomas> thanks for the hint [16:51] <dobey> heh [16:54] steveb (ste...@21...) left irc: quit has steveb, yes [16:58] Nick change: thomas -> thomas-ontheroof [17:18] Nick change: taazzzz -> taaz [17:23] Nick change: derek -> dnWork [17:27] Nick change: thomas-ontheroof -> thomas [17:40] wtay-sleep (wi...@ca...) left irc: Read error to wtay-sleep[cable-195-162-214-58.upc.chello.be]: EOF from client [17:47] sienap (sy...@ip...) joined #gstreamer. [17:47] <dobey> hi sienap [17:47] <sienap> hej [17:47] <sienap> damn again ! parapaxi :( [17:47] <dobey> dude [17:47] <sienap> d0p3y ;) [17:48] <dobey> go install neon 0.15.0 and encompass 0.4.2 and be happy [17:49] <sienap> i've got galeon ;) [17:49] <dobey> bah [17:50] <sienap> he? [17:52] Action: dobey poops on galeon [17:54] <sienap> ? [17:54] <sienap> explain [17:55] <dobey> heh [17:55] <dobey> obviously i don't like it [17:55] <dobey> ;-) [17:56] <sienap> yeah i got that far [17:56] <sienap> but i think you judge more on a personal problem then a quality problem [17:56] <dobey> both [17:56] <sienap> however i still don't see encompass playing flash movies [17:56] <dobey> uh [17:56] <dobey> whatever [17:56] <dobey> i can't view flash anyway [17:57] Action: dobey kicks macromedia in the pants. [17:57] <sienap> dobey why you can't ? [17:57] <dobey> sienap: because macromedia is too lame to make flash for linux/ppc [17:57] <sienap> haha [17:57] <sienap> cool dudes :) [17:57] <sienap> >:) [17:57] <dobey> :-( [17:57] <sienap> just kidding dopey [17:58] <dobey> dude [17:58] <dobey> learn to use Tab-completion, or where the 'b' key is [17:59] <sienap> he i am using tab completion dopey :) [17:59] <sienap> my bitchx just has a natural language plugin >:) [18:00] <dobey> :-( [18:02] wtay (wi...@ca...) joined #gstreamer. [18:02] <wtay> yo [18:03] <sienap> wtay! [18:04] <wtay> hey sienap [18:04] sienap (sy...@ip...) left #gstreamer. [18:04] <dobey> heh [18:04] <wtay> hi d0hbie [18:04] <wtay> <g> [18:05] <dobey> :-/ [18:12] <thomas> woah, where did everyone come from all of a sudden ? ;) [18:12] <thomas> wtay: why are the cvs messages so late ? [18:13] <wtay> seems like SF is screwed up again [18:13] <thomas> dobey: you should put the </A> tag right after the <IMG's on your encompass site ;) [18:13] <thomas> dobey: I'll try it on a new machine [18:13] <dobey> thomas: ? [18:13] <thomas> dobey: the whitespace from the newline causes a real small clickable underscore next to the images [18:14] <thomas> dobey:the sourceforge/gimp/wc3 ones [18:14] Action: dobey blames html [18:14] <thomas> wtay: I added a gstaudio library today [18:14] <wtay> yeah, I could see that [18:15] <thomas> dobey: no, html isn't to blame. just put the </A> right behind the > of the <IMG tag [18:15] <thomas> I downloaded a fresh anonymous cvs copy to make sure it compiles ok this time ;) [18:15] <dobey> thomas: yes, html is to blame, if i can't have cleanly formatted html [18:16] <thomas> dobey: it's a standard. every html browser is *forced* to treat whitespace as whitespace and not drop it. hence the clickable space, no ? [18:16] <thomas> dobey: you don't have to change it on my account, you know, it just looks better for a web browser project ;) [18:16] <dobey> thomas: no, they treat whitespace as one space [18:16] <dobey> ;-) [18:17] <dobey> maybe i should just make it xml [18:17] <thomas> wtay: I tried installing gstreamer on a machine that once had gstreamer-0.1.1 [18:17] <thomas> now when I run -register, it only looks in /usr/local/gstreamer [18:17] <thomas> what can I do to make it check in the current src tree ? [18:18] <wtay> rebuild with --enable-plugin-srcdir [18:19] Nick change: dobey -> dobey-eat [18:21] steveb (st...@no...) joined #gstreamer. [18:21] <thomas> wtay: I've started building an apache toolbox-like gstreamer downloader/installer for gstreamer, it's dependencies packages and plugin source libraries... [18:21] <thomas> ... is that a good idea for new users or not ? [18:21] <wtay> whohoo, col [18:21] <wtay> cool even [18:21] <thomas> wtay: what, the toolbox or something else ? [18:22] <wtay> the toolbox [18:22] <thomas> glad you're enthusiastic about it ;) [18:22] <thomas> right now it downloads and installs gtk, glib, libxml and lame [18:22] <thomas> so that proves it works ;) [18:22] <wtay> nice [18:22] <wtay> gotta eat now [18:22] Nick change: wtay -> wtay-eating [18:23] <thomas> wtay-eating: ok, later [18:36] thomas (th...@21...) left irc: [x]chat [19:03] Nick change: wtay-eating -> wtay- [19:03] Nick change: wtay- -> wtay [19:04] Kuroyi_ (ri...@ub...) left irc: Ping timeout for Kuroyi_[ubr-35.82.223.wmelbourne.cfl.rr.com] [19:05] <steveb> so, anyone see the flamebait on l-a-d? [19:06] <wtay> steveb: I still have 30 LAD-mails unread :-) [19:06] <steveb> "Lately someone posted [19:06] <steveb> twice in our featurelist, that glame is going into a totally wrong [19:06] <steveb> direction and we should join forces with gstreamer."... [19:06] <steveb> "This guy probably [19:07] <steveb> just didn't have one look in our codebase to understand that we're [19:07] <steveb> not a streaming media application, but an audio editor that allows [19:07] <steveb> to setup filter networks. " [19:09] CHW_sicko (pa...@su...) joined #gstreamer. [19:09] Nick change: CHW_sicko -> ChiefHighwater [19:10] <wtay> feeling better? :) [19:10] Nick change: dobey-eat -> dobey [19:11] <ChiefHighwater> somewhat, yes. thank you [19:29] Uraeus (csc...@c2...) joined #gstreamer. [19:29] <Uraeus> hi [19:29] <dobey> hi [19:30] <wtay> hi [19:32] <Uraeus> hi wtay [19:45] <wtay> oh man, noatum really sucks... [19:46] <Uraeus> wtay: that is common knowledge :) [19:46] <wtay> I suppose... sad [19:46] <wtay> brb [19:46] Nick change: wtay -> wtay-bath [19:52] Nick change: dobey -> dobey-brb [19:56] Nick change: ajzzzz -> ajmitch [19:56] <ajmitch> morning [19:58] Nick change: dobey-brb -> dobey [19:58] <Uraeus> morning ajmitch [19:59] <Uraeus> if hadess where here he would be proud, I am downloading the latest windowmaker to try out :) [19:59] <dobey> heh [19:59] <dobey> hey aj [20:00] <ajmitch> hi dobey [20:00] Action: ajmitch wonders why he got so damned early [20:00] <dobey> hrmm [20:05] omega_ (om...@om...) joined #gstreamer. [20:06] Nick change: richardb-asleep -> richardb-out [20:06] <ajmitch> hey omega_ [20:07] <omega_> yo [20:09] <Uraeus> hi omega_ [20:10] thomas (th...@ad...) joined #gstreamer. [20:10] <thomas> hi all [20:11] <ChiefHighwater> ello [20:11] <Uraeus> hi thomas [20:13] <steveb> anyone have an opinion on hand coded bindings vs SWIG? (for python) [20:14] <omega_> swig is better, if it works, because it can be trivially kept up to date [20:14] <thomas> anyone know of a graphical boot loader/init displayer other than aurora ? [20:14] <steveb> and the performance overhead? [20:14] <omega_> steveb: dunno [20:14] <omega_> I would expect fairly low [20:15] Nick change: wtay-bath -> wtay [20:16] <wtay> yo [20:16] <omega_> I'd like to try to get 0.2.0 done today, if possible [20:16] <ajmitch> today?? [20:16] <wtay> no prob [20:16] <steveb> ooo [20:16] <omega_> any reason we can't? [20:17] <wtay> nope [20:17] <wtay> we just do make dist <g> [20:17] Action: wtay goes to check the makefiles of the newly added files [20:17] <ajmitch> it's stable enough at the moment? [20:18] <omega_> ajmitch: dunno, you tell me ;-) [20:18] <wtay> I think it is for a 0.2.0 release :) [20:18] <Uraeus> same here [20:19] <ajmitch> i guess we'll let you unleash it then ;) [20:19] <Uraeus> release early, release often is a popular quote :) [20:20] <wtay> does anybody have a smtp server for me (my ISPs is down ) [20:20] <omega_> omegacs.net [20:20] <wtay> checking... [20:20] <ajmitch> why don't you just ask for an open relay to spam thru? ;) [20:21] <omega_> oh, wait, right, it's not open [20:21] <omega_> and I had zero luck opening temple-baptist.com to ChiefHighwater at home yesterday [20:21] <wtay> heh, I noticed :) [20:21] <omega_> lemme give it a try on this machine [20:22] <wtay> ajmitch: you have an open relay server I can use to spam my enemies? [20:22] <ajmitch> wtay: sadly no, mine's not open [20:22] <thomas> anyone have a clean machine without gtk/glib to test my gstreamer toolbox on ? ;) [20:23] <dobey> uh [20:23] <omega_> wtay: try again [20:23] <dobey> "without gllib"? [20:23] <dobey> pam requires glib nowadays [20:23] <omega_> bingo [20:23] <wtay> omega_: yup, it's gone..thx [20:23] <Uraeus> wtay: since you use chello smtp.chello.no might work [20:23] <wtay> Uraeus: you sure? [20:24] <Uraeus> 'might work' == not sure :) [20:24] <thomas> dobey: it's possible to run linux without glib installed you know... anyways, it can be with glib but without gtk as well, just to test some stuff [20:26] <dobey> thomas: oh well, i can't right now anyway [20:28] Parapraxis (Sp...@us...) joined #gstreamer. [20:28] <wtay> hi [20:28] <Uraeus> hello Parapraxis [20:35] _gst_newt_ joined #gstreamer. [20:35] Parapraxis (Sp...@us...) joined #gstreamer. [20:36] Nick change: thomas -> thomas-eats [20:38] <ajmitch_> heh [20:40] Action: wtay is preparing pre3 [20:41] <ajmitch_> why pre3? [20:41] Nick change: ajmitch_ -> ajmitch [20:41] <taaz> 16 maxcothreads [20:41] <wtay> after pre2 comes pre3 :) [20:41] <taaz> ? [20:41] <wtay> yup, until lame is fixed [20:41] Action: taaz thinks gstreamer should handle 6000 cothreads... [20:42] <omega_> if taaz can find a way to make that happen generically, more power to him [20:42] <taaz> bah... i was just about to checkin debian stuff.. can pre3 wait a few mins? [20:42] <omega_> yeah [20:42] <wtay> sure [20:42] <wtay> I spotted a few more bugs.. so.. [20:43] <taaz> ok 2 questions: 1) adding configure.{ac,in} to .cvsignore ok now i guess? 2) adding *.o *.lo to docs/gst/Makefile.am clean-local rule ok? [20:44] <omega_> taaz: 1) no, 2) yes [20:44] <omega_> taaz: 1) oh, right ,yes, it is [20:44] Uraeus (csc...@c2...) left irc: Read error to Uraeus[c224s9h5.upc.chello.no]: Connection reset by peer [20:46] Action: omega_ starts on the slides for tomorrows PLUG talk [20:46] <omega_> (hey, at least I'm starting the day before, unlike GUADEC where I started the morning of...) [20:46] <wtay> taaz: you or me? [20:46] <wtay> taaz: you :) [20:48] Nick change: thomas-eats -> thomas-nofood [20:49] <thomas-nofood> omega_: any chance you could point me in the right direction re: my problem ? [20:49] <omega_> anyone know the proper name of the BeOS media framework? [20:49] <omega_> thomas-nofood: oh, right. um... checking [20:50] <omega_> gst_object_destroy (GST_OBJECT (audiosink)); [20:50] <omega_> don't do that [20:50] <omega_> esp just before connecting an element to its pad [20:50] <thomas-nofood> ok, why ? [20:50] <omega_> um, hrm [20:51] <thomas-nofood> the don't do that I mean, is that function broken ? [20:51] Action: omega_ is more confused now [20:51] Action: omega_ reads the code some more [20:52] <thomas-nofood> omega_: are you saying I should create a new object with a new pointer name different than audiosink ? [20:52] <omega_> no, I missed some of that code [20:52] <omega_> but you will run into the problem that we don't free cothreads right now [20:52] <omega_> which means this work for about 5 iterations with wtay's latest patch [20:52] <thomas-nofood> uh, ok. why aren't cothreads freed ? to difficult right now ? [20:53] <omega_> we don't have support for it, it's not difficult, but it hasn't been a prio [20:53] <thomas-nofood> because I tried with just resetting the disksink location argument to either /dev/null or /tmp/some.mp3 ... [20:53] <thomas-nofood> ... but of course if you do that in the middle of a buffer then lame doesn't know it should write a new header [20:53] Uraeus (csc...@c2...) joined #gstreamer. [20:53] <ajmitch> wb Uraeus [20:53] <omega_> hrm, right. need to add flush support for lame [20:53] <thomas-nofood> ... meaning that for now I would be better of doing this in raw audio instead of mp3 for now ? [20:53] <omega_> yeah, for now ;-( [20:54] <thomas-nofood> ok, so basically for now it makes no sense to take this approach in making a cutter app ? [20:54] <thomas-nofood> just add flush to lame maybe and stick to location ? [20:55] <wtay> omega_: did I screw something up? [20:55] <omega_> wtay: no [20:55] <wtay> ah [20:58] <steveb> anyone help me with why i'm getting this? error in loading shared libraries: /home/steveb/Projects/gstreamer/./plugins/mpeg2/mpeg2dec/.libs/libmpeg2dec.so: undefined symbol: mm_accel [20:58] <omega_> thomas-nofood: um. question: where does 'it in' and 'it out' come from? [20:59] <wtay> steveb: what version of libmpeg2? [20:59] <thomas-nofood> omega_: it in is just before the iterate call, meaning it's going into the iterate function... [20:59] <thomas-nofood> and it out is the other way around [20:59] <thomas-nofood> just to check if it's making iterations, because I've been noticing some weird stuff lately [20:59] <steveb> wtay: erm, how do i tell? [20:59] <thomas-nofood> the problem with the cutter is that after the disconnect/reconnect, it doesn't leave iterate anymore [20:59] <wtay> steveb: where did you get it? [21:00] <thomas-nofood> and today I noticed with another implementation that, every ten iterates, it does a double one [21:00] <thomas-nofood> not sure if that is my fault or not yet [21:01] <dobey> thomas-nofood: homeruns.com [21:02] <steveb> wtay: http://www.linuxvideo.org/mpeg2dec/ 0.2.0 I think (installed it on Saturday) [21:03] <thomas-nofood> dobey: thanks ;) do they do one-hour cross-atlantic delivery as well ? [21:03] <dobey> i dunnow [21:03] <dobey> peapod.com maybe? [21:04] Action: taaz had computer console stolen for a few mins... [21:04] <wtay> steveb: nm /usr/local/lib/libmpeg2dec.a | grep mm [21:06] <steveb> wtay: 00000000 T mm_accel [21:06] <omega_> whoops [21:07] <wtay> steveb: same here... are you sure there isn't some other libmpg2dec somewhere? [21:07] <taaz> should autoplugtest be in the gst/autoplug dir or test(s)/? [21:08] <ajmitch> 0.2.0 still gonna be out today? ;) [21:08] <wtay> ajmitch: are you a non-believer or what? <g> [21:08] <taaz> just wondering if i should add the bin to .cvsignore there or will it be moved anyway? [21:09] <ajmitch> wtay: heh, me? ;) [21:10] <ajmitch> wtay: i guess 0.2.0 doesn't have to have the stability, etc of a 1.0 release ;) [21:10] <omega_> ajmitch: kinda by definition... [21:11] <ajmitch> ok, 0.2.0 today, 1.0 tomorrow then ;) [21:14] Action: ajmitch watches another conflict arise on plugins/filters/Makefile.am.... damned plugin i made ;) [21:16] <wtay> eek, red-carpet crashes.. [21:16] <ajmitch> heh [21:16] <ajmitch> why would you use redcarpet? [21:16] <wtay> eye candy [21:17] <dobey> it gets you girls. [21:17] <wtay> dobey: it does? [21:17] <dobey> i dunno, it hasn't worked if it's supposed to [21:18] <wtay> I thought so <g> [21:18] <steveb> wtay: this would do it (building mpeg2dec0.2.0) cpu_accel.c: In function `mm_accel': [21:18] <steveb> cpu_accel.c:105: fixed or forbidden register 3 (bx) was spilled for class BREG. [21:18] <steveb> This may be due to a compiler bug or to impossible asm [21:18] <steveb> statements or clauses. [21:19] <omega_> steveb: neat, I've seen that before [21:19] <steveb> is there a fix [21:19] <steveb> ? [21:19] <omega_> checking [21:20] <wtay> omega_: omegacs.net does not work? [21:20] <omega_> ? [21:20] <wtay> cannot parse 'B' error? [21:20] <omega_> wtay: huh? [21:20] <wtay> omega_: dunno [21:21] maYam (ma...@ca...) joined #gstreamer. [21:21] <wtay> ah [21:21] <maYam> grrr [21:21] <ajmitch> hi maYam [21:21] <maYam> out of my way everybody! [21:21] <ajmitch> uh oh [21:22] Action: maYam walks straight up to omega [21:22] <wtay> oh dear [21:22] <maYam> hehe sorry ;) [21:22] <wtay> my fault :( [21:22] <maYam> hi ajmitch! [21:22] <Uraeus> hi maYam [21:22] <omega_> maYam: ? [21:22] <maYam> hey Uraeus [21:23] <maYam> omega_: nevermind.. i'm complaining about my mail trouble ;) [21:23] <ajmitch> hehe [21:23] <ajmitch> bbl [21:23] Nick change: ajmitch -> ajbusy [21:25] <taaz> blah... i think debs are out-of-date already ;) got some audio lib and cutter lib... what a hassle to keep up [21:26] <omega_> taaz: need to autogen the deb files from some plugin list, then we can do the same for the spec file [21:26] <Uraeus> anyone heard of this file: glibconfig.h ? [21:26] <omega_> yeah, that's the basis for glib machine-dependent stuff [21:26] <taaz> should that plugins/avi/README thing be in EXTRA_DIST? [21:26] <omega_> if it exists, yes [21:27] <Uraeus> omega_: where is it usualle placed in the filesystem? [21:27] <omega_> /usr/lib/glib afaik [21:27] <taaz> glib-config tells you where it is [21:29] Nick change: taaz -> taaz-away [21:31] <omega_> wow: http://richardberg.net/Postings/stickfight.swf [21:31] <omega_> need flash for that, and a reasonably fast machine [21:33] <wtay> ick [21:36] <Uraeus> heh, nice but not as nice as my new wm dockapp which is a bubblemonitor with a small duck swimming on top [21:36] Action: Uraeus remembers why he used to love windowmaker so much [21:37] <maYam> hahah [21:44] <Uraeus> hmm, should I use the usefull worldclock dockapp or the eyecandy clock with small fishes swimming around......eyecandy wins again [21:45] <wtay> hmmm.... fish... [21:45] <omega_> gstreamer.net/slides/plug/ [21:45] <omega_> very in progress, suggestions welcome [21:46] <maYam> fish = eyecandy for Uraeus.. ;) [21:46] <Uraeus> maYam: hey! they are cute little yellow fish :) [21:47] <maYam> yellow even.. need i say more? ;) [21:47] <Uraeus> besides anyone claiming fish aren't eyecandy has never owned a tropical aquarium [21:47] <omega_> Uraeus: any chance we can get gstreamer commit counts listed in the gnome summary? <g> [21:47] <maYam> that is a good point, Uraeus [21:47] <omega_> there's a newchannel chopper circling overhead. that's usually a bad sign [21:47] <omega_> er, newschannel [21:47] <ChiefHighwater> yup [21:48] <Uraeus> omega_: heh, yes moving gstreamer into GNOME CVS [21:48] <Uraeus> omega_: besides the reason we beat everyone is cause we do a lot of small commits [21:49] <maYam> omega_: yes, you! get your hands where i can see 'em and move away from the computer.. slowly! [21:49] <maYam> they're coming to get you, omega [21:49] <maYam> with a helicopter, you lucky bastard! [21:49] <wtay> omega_: is the 0.2.0 release going to be on television :-) [21:50] <omega_> ChiefHighwater: no, the mass of cars parking in the area tells me that it's likely it's probably a parade [21:50] <omega_> wtay: um [21:50] <wtay> s/video chroam key/video chroma key/ [21:50] <wtay> slide 2 [21:50] <ChiefHighwater> tis the season for parades [21:50] <omega_> whoops [21:50] <omega_> http://www.rosefestival.org/junior_parade.htm [21:50] <omega_> yup. that's what the chopper is for [21:51] <omega_> starts about 5 blocks from here [21:51] <wtay> hollywood? [21:51] <ChiefHighwater> that'll do it [21:52] <omega_> whoah. cars parking left and right. that's the first time I've seen that [21:52] <ChiefHighwater> wtay:in portland, all the neighborhoods have names [21:52] <wtay> ChiefHighwater: ah [21:52] <wtay> I thought that name was allready taken... <g> [21:53] <ChiefHighwater> hehe [21:53] <ChiefHighwater> there's a theater near his home with that name...somewhat of a landmark...especially in the good ole days [21:54] <omega_> good thing it didn't burn down with the building next to it a few years ago.... [21:54] <omega_> eesh. I count 8 unknown vehicles just line-of-sight from my window [21:55] <omega_> only 3 known [22:01] _gst_newt_0 joined #gstreamer. [22:01] Action: omega_ would love to see the reaction of the webmasters when they see that poor page getting hammered all at the same time by many people from around the world [22:01] cschalle_ (csc...@c2...) joined #gstreamer. [22:01] maYam (ma...@ca...) got netsplit. [22:01] richardb-out (ri...@ix...) got netsplit. [22:01] _gst_newt_ (My...@ca...) got netsplit. [22:01] Uraeus (csc...@c2...) got netsplit. [22:01] Nick change: _gst_newt_0 -> _gst_newt_ [22:01] Possible future nick collision: _gst_newt_ [22:01] cschalle_ (csc...@c2...) left #gstreamer. [22:01] Nick change: ajbusy -> ajmitch [22:02] <omega_> um. is the log gonna get fragged by that confusion? [22:02] Uraeus (csc...@c2...) returned to #gstreamer. [22:02] <Uraeus> hmm [22:02] <wtay> omega_: dunno, I suppose this is not the first time this happens [22:02] richardb-out (ri...@ix...) returned to #gstreamer. [22:02] <ChiefHighwater> wtay:nor the last [22:02] <ChiefHighwater> not even this first time today [22:02] <omega_> wtay: never seen _gst_newt_0 before [22:03] <Uraeus> omega_: you are aware that making free software that can't be screenshootet is the linux eqivalent of devil-worship? [22:03] <omega_> mu [22:03] <Uraeus> http://www.linuxrising.com/files/screenshot.jpg [22:03] Apoc (ep...@pp...) joined #gstreamer. [22:03] <omega_> hehehehe [22:04] <ajmitch> heh [22:04] <omega_> ENOTMYFAULT [22:04] <Uraeus> maYam_: what do you think of my fish clock -> http://www.linuxrising.com/files/screenshot.jpg [22:04] <Apoc> Yo [22:04] <Uraeus> hi Apoc [22:04] Action: omega_ thinks "fish", "barrel", "gun" [22:05] <ajmitch> omega_: violent person.... ;) [22:05] <Uraeus> omega_: you have to fix, not release even a 0.0.0.0.1alpha can be releases if screenshots can't be made [22:05] <maYam_> i see the duck! [22:05] <ChiefHighwater> ajmitch:comes from hanging around ChiefHighwater [22:05] <maYam_> ooh and the fish!! [22:05] <maYam_> *aaaaah*!! [22:06] Action: ajmitch wonders at the sanity of some of these people... [22:06] <omega_> Uraeus: if you can find a way to screenshot that, be my guest.... [22:06] <wtay> yummy... fish.. [22:06] <maYam_> yes, i complained about the screenshot prob to wtay [22:06] <wtay> I can fix it [22:06] <Uraeus> omega_: well, if gstmediaplay took and held the last image played when you press pause, that would do it [22:06] <omega_> is there maybe a way for that xwindow to pick up a different event and redraw itself? [22:06] <Uraeus> wtay: cool [22:06] <omega_> Uraeus: no, that's not it at all [22:07] <wtay> I don't think there's an event for a screenshoot [22:07] <omega_> it's Xv doing its thing, the actual framebuffer really does look like that [22:07] <wtay> Xv is an overlay, meaning a blue image where the video card draws the image (in HW) [22:08] <omega_> wtay: you'd have to have gstmediaplay support for it, because it'd have to stick a colorspace converter in there and paint the background directly before the screenshot [22:08] <maYam_> (nothing that wtay can't do when put under pressure :p) [22:08] <Uraeus> maYam_: don't let him go to sleep until we can make gstmediaplay screenshots :) [22:08] <dobey> maYam_: !!!!!!!!!!!!!!1 [22:08] <wtay> omega_: xvideosink can be told not to use Xv images using an arg [22:09] <maYam_> Uraeus: no prob ;) [22:09] <maYam_> dobeeeeey! :-D [22:09] <omega_> wtay: right, but will it then cause a renego and stick the colorspace in there, then redraw the picture, all in time for the screengrab? [22:09] <dobey> ;-) [22:09] <Uraeus> hmm, the cool thing amount my clock fish is that they get scared when the mouse pointer moves towards them :) [22:09] <wtay> omega_: don't push it :-) [22:10] <steveb> 'night all [22:10] <omega_> l8r [22:10] <wtay> sleep well [22:10] <Uraeus> night steveb [22:10] <maYam_> yes, in fact i can't get enough of your screenshot, Uraeus, it's so.. neo-art-deco.. [22:10] <omega_> omegacs.net/~omega/pictures/dscn4782.jpg <- taken last night at 1:25am [22:10] <ajmitch> hehe [22:10] steveb (st...@no...) left irc: [x]chat [22:10] <wtay> You don't have permission to access /~omega/pictures/dscn4782.jpg on this server. [22:10] <ChiefHighwater> Forbidden [22:10] <omega_> mu [22:10] <dobey> heh [22:10] <dobey> meh [22:11] <ChiefHighwater> there we go [22:11] <omega_> chmod 711 ~ [22:11] <Uraeus> maYam_: I live and breath for useless eyecandy stuff, that is why I like GStreamer so much :) [22:11] Action: Uraeus runs [22:11] <dobey> Uraeus: vaporware rules! [22:11] <ajmitch> Uraeus: heh, no matter where you run, we will fins you... [22:12] maYam (ma...@ca...) got lost in the net-split. [22:12] <omega_> if GStreamer was just vaporware, I'd be getting pretty high inhaling it all... [22:12] <maYam_> heheh [22:12] <wtay> omega_: what's that? [22:12] <omega_> oh wait.... <g> [22:12] <omega_> wtay: moon, trees, clouds [22:12] <Uraeus> 'relase late, release never' the new slogan [22:12] <omega_> long exposure [22:13] <ajmitch> prolonged exposure to forbidden substances? ;) [22:13] <wtay> hmm [22:13] Action: omega_ fills that directory up with some more pictures, slowly [22:14] <omega_> and be warned, that website is really really really REALLY old [22:15] Action: ajmitch wonders why a forbidden error still pops up... [22:15] <omega_> no index, fixing [22:15] Action: omega_ gets annoyed at the constant chopper noise circling his head [22:15] <ajmitch> omega_: got a decent rifle? [22:16] <wtay> Uraeus: I'll add a --noxv option to gstmediaplay [22:16] <omega_> ChiefHighwater: ? [22:16] <Uraeus> wtay: thnx [22:16] <ChiefHighwater> no, t's a .22 [22:16] <ajmitch> wtay: hmm? i can play mpeg2 videos now without Xv ;) [22:16] <omega_> ChiefHighwater: it's channel 8, so they're fair game, no chance Dean is up there <g> [22:16] <ChiefHighwater> a squirrel gun [22:16] <ChiefHighwater> hehe [22:17] <Uraeus> think I need to get hadess to make us a windowmaker dockapp output plugin, playing videos in a 64x64 square is just what I need to survive [22:17] <omega_> mu [22:17] <ajmitch> hehe [22:17] <ChiefHighwater> although, bet they'd run when they hear..tink, tink, tink...under their chopper [22:17] Action: omega_ drops CD 1 of 4 (soon to be 6) into his drive to start looking for good pics [22:17] Nick change: taaz-away -> taaz [22:18] <wtay> gstreamer-lauch shopper_missile ! enhanced_aiming ! shoppersink [22:18] <omega_> ChiefHighwater: they'd better, let's try it! [22:18] <maYam_> yes playing videos, that would be nice.. wtay, where's my video? :p [22:18] <omega_> wtay: chopper... <g> [22:18] <omega_> ok, index now fixed [22:18] <maYam_> a shopper's nice too! [22:18] <ChiefHighwater> show em ecola 8-] [22:18] <wtay> I guess I'm still hungry :-) [22:18] <omega_> dscn0010.jpg [22:19] Action: Uraeus gets back to editing Jim Gettys interview [22:19] <ChiefHighwater> there we go...who wanted eye candy? [22:20] <Uraeus> me me me [22:20] <ChiefHighwater> http://omegacs.net/~omega/pictures/dscn0010.jpg [22:20] Action: ChiefHighwater watches omega_'s bandwidth evaporate [22:21] Action: omega_ is gonna regret this [22:21] <ajmitch> ChiefHighwater: what's so great about it? ;) [22:21] <maYam_> i see.. i see.. a gray sky...? [22:21] <ChiefHighwater> ajmitch:you sure you're looking at the right image? [22:21] <ChiefHighwater> umm....scroll [22:21] <maYam_> aha! ;) [22:21] <omega_> yeah, it's, um, 1600x1200 [22:22] <ajmitch> http://omegacs.net/~omega/pictures/dscn0010.jpg [22:22] <maYam_> still downloading.. heheh [22:22] <maYam_> ooh nice! would make a peaceful wallpaper [22:23] <omega_> didn't you see it on my laptop at GUADEC ? [22:23] <ChiefHighwater> is the background shot on like half the computers i see now [22:23] <omega_> or was it so covered by gnome-terminals that you couldn't? <g> [22:23] <maYam_> indd! ;) [22:23] Action: ajmitch sees a fairly avg coastline pic... [22:23] <Uraeus> hmm, someone elses holiday pictures as my wallpaper, no way :) [22:24] <maYam_> i have a self made wallpaper oh yes [22:24] <ajmitch> heh, should i share my background image? ;) [22:24] <omega_> dscn1810 is CHW [22:24] <ChiefHighwater> uhoh [22:24] <omega_> overexposed and with redeye though ;-( [22:24] <ajmitch> mine's a bit dark, i think it was taken late in the day... [22:25] <maYam_> here's a small version of my wallpaper http://users.chello.be/ws35703/pictures/homepage/egowar.jpg [22:26] Action: ajmitch should not tell people to connect to the webserver on his machine.... [22:26] <omega_> whoops. guess that blender came batteries included.... [22:27] <maYam_> ajmitch: you're too shy to show us your wallpaper? hmm..? [22:27] <maYam_> probably a 'dirty' pic ;) [22:28] <dobey> hahah [22:28] <ChiefHighwater> just mail it to omega_ [22:28] <ChiefHighwater> 8-] [22:29] <ajmitch> maYam_: nah, it's 2360*1590, a little big for a 33.6k modem ;) [22:29] <ajmitch> ChiefHighwater: i don't think a picture of a lake is dirty ;) [22:29] <maYam_> that's not an excuse, i scaled my image! [22:29] <omega_> unless there's dead fish and scum and stuff on the lake surface <g> [22:30] <ChiefHighwater> ajmitch:i didn't mean to suggest it was dirty, i was suggesting let omega_ use his bandwidth to show it [22:30] <ajmitch> heh [22:34] <ajmitch> http://ajmitch.dhis.org/pics/ [22:34] <wtay> http://users.chello.be/ws35703/maYam02.html <<-- wtay is using this background :) [22:37] <omega_> many more pics uploaded [22:37] <ajmitch> wtay: btw, the other lake pic is my current background, but Lake.JPG looks better i think [22:37] <wtay> ajmitch: it was too slow to download :) [22:38] <ajmitch> wtay: hehe [22:38] <ajmitch> wtay: be patient ;) [22:38] <taaz> you all can actually see your background? i have windows covering every piece of background... [22:38] <ajmitch> taaz: i see mine when i login ;) [22:39] <taaz> ajmitch: i don't bother to logout anymore ;) [22:39] <ajmitch> taaz: oh, and when i turn on transparancy in Eterm ;) [22:39] <ajmitch> taaz: i don't voluntarily logout often either ;) [22:39] <maYam_> 1% of Lake.JPG downloaded [22:40] <maYam_> but i'm a patient person.. ehem [22:40] <omega_> 83% and climbing [22:40] <omega_> done [22:40] <ajmitch> maYam_: hehe, wtay must be using all my bandwidth ;) [22:40] <omega_> it's rather pixelated.... [22:40] <ajmitch> omega_: which? [22:40] <omega_> Lake [22:40] <maYam_> lol omega [22:41] <ajmitch> omega_: wasn't taken with a digital camera sorry, it was scanned ;) [22:42] <maYam_> oh what omega wasn't joking? [22:43] <maYam_> ah you're all such perfectionists! [22:43] <ajmitch> hehe [22:43] <ajmitch> maYam_: did it finish downloading? [22:43] <maYam_> yes it did! [22:43] <maYam_> on to the next one ;) [22:43] <ajmitch> maYam_: do you think it looks pixellated? ;) [22:43] <maYam_> beautiful though.. [22:43] <dobey> YES POKEY!!! [22:43] <maYam_> no not at all! [22:44] <ajmitch> dobey: eh? [22:44] <maYam_> i refuse to complain about quality when i've just spent 10 minutes downloading [22:44] <dobey> heh [22:44] <dobey> ajmitch: www.yellow5.com/pokey/ [22:45] Action: ajmitch sees all these different ip addresses in his apache logs... [22:45] <ChiefHighwater> 8-] [22:46] <maYam_> 41%.. [22:46] <ChiefHighwater> 19 [22:46] <maYam_> 00:02:13 remaining [22:46] <maYam_> i might consider taking a bath in the mean time [22:47] <ChiefHighwater> 30% [22:47] <ajmitch> hehe [22:47] <maYam_> 57% [22:47] <ajmitch> might as well, my modem wasn't designed to server all these users at once ;) [22:47] <ChiefHighwater> 40% [22:48] <ChiefHighwater> 70% [22:49] <ChiefHighwater> maYam_ must be done, my bandwidth just went up 8-] [22:49] <ajmitch> hehe [22:49] <ajmitch> ::ffff:128.173.52.12 - - [07/Jun/2001:08:47:33 +1200] "GET /pics/Lake.JPG HTTP/1.0" 200 329200 "http://ajmitch.dhis.org/pics/" "Mozilla/4.77 [en] (WinNT; U)" [22:49] <ajmitch> ::ffff:195.162.214.58 - - [07/Jun/2001:08:48:11 +1200] "GET /pics/Mountains.JPG HTTP/1.1" 200 353678 "http://ajmitch.dhis.org/pics/" "Mozilla/5.0 (X11; U; Linux 2.4.0 i686; en-US; Galeon) Gecko/20010421" [22:49] <ajmitch> ::ffff:128.173.52.12 - - [07/Jun/2001:08:48:17 +1200] "GET /pics/Mountains.JPG HTTP/1.0" 200 15758 "http://ajmitch.dhis.org/pics/" "Mozilla/4.77 [en] (WinNT; U)" [22:49] <ajmitch> ::ffff:63.105.18.226 - - [07/Jun/2001:08:48:40 +1200] "GET /pics/Lake.JPG HTTP/1.0" 200 329200 "http://ajmitch.dhis.org/pics/" "Mozilla/4.72 [en] (Win98; I)" [22:50] <ajmitch> 4 more pics requested in a minute... my poor modem.... :) [22:50] <dobey> heh [22:52] <ajmitch> hmm [22:53] <ajmitch> everybody got their sample of NZ scenery? ;) [22:55] Action: ajmitch will bbl [22:56] Nick change: ajmitch -> ajbusy [22:56] <maYam_> yes i'm totally relaxed now ;) [22:59] Action: wtay starts 10 wget threads to ajmitch's machine [22:59] ajbusy (aj...@p5...) left irc: Ping timeout for ajbusy[p54-max3.dun.ihug.co.nz] [22:59] <wtay> hehe [22:59] <omega_> whoops [22:59] <wtay> hmm, it was only a joke [23:00] <Apoc> 'night all [23:00] <omega_> more pics up, again [23:00] <wtay> cya [23:01] Apoc (ep...@pp...) left #gstreamer (Client Exiting). [23:02] Action: omega_ gets hammered [23:03] <wtay> ...a flower... [23:03] <wtay> .. and angel.. sorta [23:03] <omega_> hrm, so that's what I just created: an image->word association game? [23:05] <wtay> any picture of you doing you audio stuff? [23:05] <wtay> s/you au/your au/ [23:05] <ChiefHighwater> i think omega_ needs to lable some of his pics 8-] [23:05] <ChiefHighwater> somehow dscnxxxx just doesn't do it for me [23:05] <wtay> I seee..... a bidge... [23:05] <omega_> yeah, that's another project [23:06] <omega_> a what?? [23:06] <wtay> http://omegacs.net/~omega/pictures/dscn2891.jpg [23:06] <wtay> a bridge? [23:06] <omega_> yup [23:06] <wtay> oops type [23:06] <wtay> typo [23:06] <omega_> somehow I got the wrong picture in there, I'll have to get the one I intended off the other Cd [23:07] <omega_> dscn4712, one can see hadess on the far right, with gleblanc next to him [23:08] <omega_> wtay and maYam_ are behind the two guys standing up [23:09] <omega_> as seen in 4713 [23:11] <ChiefHighwater> 4673 needs rotating [23:11] <omega_> just did [23:11] <ChiefHighwater> ahh, much better [23:13] <omega_> I just noticed something about http://gstreamer.net/guadec-pics/dscn0001.jpg [23:13] <omega_> stw and tim janik happen to be in the picture too <g> [23:13] <omega_> and maYam_ needed some sleep I think [23:13] <maYam_> oh thanks [23:14] <maYam_> i was wide awake, i always look like that [23:15] <omega_> temple-baptist.com/~omega/images/4xBHG-3c.jpg [23:15] <omega_> ChiefHighwater can explain what they are [23:16] <thomas-nofood> damn, my gstreamer toolbox script failed because active ftp doesn't work behind my firewall ;( [23:16] <ChiefHighwater> they are bore hole geophones [23:16] <ChiefHighwater> made by ChiefHighwater [23:17] <ChiefHighwater> they plug into your seismograph [23:17] <omega_> put 'em in the ground, hit the ground really hard, and watch the squigles [23:17] <ChiefHighwater> here: http://www.georadar.com/geostuff.htm [23:18] <ChiefHighwater> i build all the stuff on that page now [23:18] <omega_> dscn4537.jpg is the reason I had a borrowed laptop at GUADEC [23:18] dobey (do...@dr...) left irc: home [23:18] <omega_> part 1 of 2, one could say [23:18] <taaz> omega_: what are those gst logos in that last dir? (yes i'm snooping around ;) [23:19] <omega_> which dir? [23:19] <omega_> taaz: oh, rally logos [23:19] <omega_> rally.temple-baptist.com [23:20] <omega_> and http://temple-baptist.com/~omega/images/lbabel.jpg is a POV image I did years ago [23:20] <taaz> i mean logo6 and gstlogo1 [23:20] <omega_> stuff maYam_ did I think [23:22] <omega_> ChiefHighwater: did you make those BHG's, or is that an older pic? [23:22] Action: taaz wishes he had a digital camera [23:24] <taaz> omega_: was i talking with you a while ago about portable flash<->hd storage or was that someone else? [23:24] <omega_> um, I think it was someone else [23:24] <omega_> er, no, must'v been me [23:24] <taaz> ok, nevermind then ;) [23:24] <omega_> you mean a box with a hd and a cf reader, spins up to empty cf and then down again? [23:25] <taaz> yeah [23:25] <omega_> if you can make one that's just the shell with no HD for $100, you'll be rich [23:26] <omega_> tarbzll of pre3 available now too [23:27] <taaz> there is one company that sells hd-less ones... its more than $100 though [23:27] <omega_> which? [23:27] Action: taaz looks for the link... [23:29] <thomas-nofood> night all [23:29] thomas-nofood (th...@ad...) left irc: Client Exiting [23:30] <taaz> its $199.95: http://www.simacorp.com/cp150.html http://www.steves-digicams.com/2001_reviews/imagebank.html [23:31] <taaz> there are other options but i think this is the only one that you can get without a hd [23:33] <taaz> i built some debs, totally untested ;) [23:33] <omega_> so here's my question. what's with all the buttons, and the LCD? I want exactly one button: transfer [23:33] <omega_> and I don't need a smartmedia reader [23:33] <omega_> it should be trivial to make one of these for <<$100 [23:34] <taaz> yes it should [23:34] <taaz> i could do it ;) [23:34] <omega_> what would you use as the main chip? [23:34] <omega_> fpga or an embedded cpu? [23:35] <taaz> microcontroller of some sort [23:35] <taaz> fpga isn't the right tool for this job [23:35] <omega_> yup [23:35] <omega_> a tiny embedded arm would be ideal [23:36] <omega_> it needs: IDE, CF, USB, and a GPIO input pin [23:36] <taaz> and flash with uClinux and minimal filesys and usb(or whatever) driver [23:36] <taaz> and flash driver [23:36] <taaz> err compactflash [23:37] <taaz> anyway... anyone listening could make some cash from this ;) [23:37] <omega_> quite [23:37] <taaz> so where should debs go? i recall arguing about this a while ago... [23:37] <omega_> ucsimm is $300 now, way out of the range [23:37] <omega_> pre debs should go where the pre release is, as per the topic [23:38] <taaz> do you remember my arguements? ;) [23:38] <omega_> nope [23:39] <taaz> i want to setup and apt accessable location. just one place so when new debs come out it will just upgrade them automagically [23:39] <wtay> omega_: you still have the video of the presentation? [23:39] <taaz> s/and/an [23:39] <taaz> versioned directories dont work for this [23:39] <omega_> wtay: which? guadec? [23:39] <wtay> omega_: yup [23:40] <taaz> symlinks and stuff will work i guess, but that still kinda sucks to maintain [23:40] <omega_> taaz: hrm, ok, then we'll put them in the versioned dirs and symlink into a central location [23:40] <omega_> why does that suck? [23:40] <omega_> we can have a script do it, if it's that annoying <g> [23:40] <taaz> um... i dunno... how do you want it setup? [23:40] <omega_> rather have the real files in the versioned dirs [23:40] <omega_> much easier to have a page that lists the released binary fies [23:43] <taaz> there's some extra files too... perhaps a x.y.z/debian/ dir with the files and releases/debian or just gst.net/debian that links to the latest dir? [23:43] <omega_> ok [23:44] <Uraeus> omega_, wtay, taaz: want to read what JG has to say about sound servers? [23:44] <taaz> JG? [23:44] <Uraeus> Jim Gettys [23:44] <omega_> Uraeus: yup [23:44] <Uraeus> http://www.linuxrising.com/files/jimgettysinterview.html [23:44] <omega_> and maybe we should have an rpm/ dir per release as well, if there are going to be as many rpms as debs [23:45] <taaz> just use sf's download archive for tarballs? [23:45] <taaz> Uraeus, that page is busted [23:45] <omega_> yeah, it doesn't render anything [23:46] <omega_> taaz: hrm, oh, yeah [23:46] <omega_> um, how could we integrate sf's release mechanism into dpkg stuff? [23:46] <taaz> no clue [23:46] <taaz> probably too hard to do [23:46] <Uraeus> try again now [23:47] <omega_> what is the key file that dpkg uses to find packages on a site? [23:47] <omega_> Uraeus: no go [23:47] <taaz> and what's the point anyway? seems permanent tarball archiving would be enough... but i dunno... [23:47] <omega_> the idea of the sf releases is that they go on a really well-connected http server [23:48] <omega_> geographically mirrored, no less [23:49] <taaz> well... as far as debs go i see the main distribution point being the debian machines... use sf for unstable CVS builds or maybe not at all [23:49] <Uraeus> ok, try again now : http://www.linuxrising.com/files/jimgettysinterview.html [23:49] <omega_> ok, that didn't suck as bad <g> [23:50] <Uraeus> omega_: what browser do you use? [23:50] <omega_> N$ 4.75 [23:50] <Uraeus> weird [23:50] <omega_> reading interview now [23:50] <taaz> i've got netescape 4.77 on winnt here... yeah nt, got a problem with that? huh? [23:51] <Uraeus> <g> [23:51] <omega_> /kickban taaz [23:51] <omega_> ;-) [23:52] <wtay> uh? did I readme winnt? [23:52] <wtay> s/readme/read [23:53] <taaz> yeah. for codewarrior usage... i do rest of work on linux boxes ;) [23:53] <wtay> ah [23:54] <wtay> no cursing in this channel :) [23:54] <Uraeus> taaz: isn't codewarrior availble on linux? [23:55] Action: wtay is tired waiting for gst to compile on his PIII [23:55] <omega_> configure and make of pre3 on my laptop took 23m52s [23:55] <Uraeus> wtay: read the interview meanwhile then [23:55] <taaz> Uraeus: maybe... its embedded ppc crosscompiler stuff though... so i'm not sure [23:56] <Uraeus> taaz: I remember reading some comments from John Carmack when trying out a early CW release for Linux [23:56] <omega_> so, one of the things I've been thinking about recently is the fact that gtk is huge [23:57] <Uraeus> .... [23:57] <omega_> if he's pushing X as a std for handhelds, there still needs to be a widget set that isn't massive [23:58] <omega_> IMO gtk and its deps could be half their current size or less, if someone put the work into it [23:59] <Uraeus> by the way of code optmimizations or by removing widgets etc? [23:59] <taaz> omega_: let us know when you're done ;) [23:59] <omega_> any and all of the above [23:59] <omega_> taaz: there's a chance we may be trying that at rr, if no other small widget set can be found [00:00] --- Thu Jun 7 2001 [00:00] Action: Uraeus adds GTK+ to the roadmap [00:00] <Uraeus> :) [00:00] <taaz> how big is big? how small do you want it? [00:00] <omega_> cvs glib is 250K, gobject is 120K [00:01] <omega_> atk is 58K [00:02] <taaz> with all the asserts and g_return_if_fail stuff removed? [00:02] ChiefHighwater (pa...@su...) left irc: Ping timeout for ChiefHighwater[sub18-226.member.dsl-only.net] [00:02] <omega_> theoretically, yes [00:02] <Uraeus> omega_: did you add the wiki comments about arts? [00:02] <omega_> pango is 140 + 20 + 28 + 90 + 70 K [00:02] <omega_> Uraeus: no [00:02] <omega_> what comments? [00:02] <Uraeus> hmm, arts blah blah blah [00:03] ChiefHighwater (pa...@su...) joined #gstreamer. [00:04] <taaz> so you're at 776k... what's the target? [00:04] <omega_> at the very least there should be the option to remove the atk and pango deps from gtk, for embedded uses [00:04] <omega_> taaz: ideally about .5MB total, gtk included [00:05] <taaz> total meaning + gstreamer too? [00:05] <taaz> or linux + gst + gtk* > 500k? [00:05] <taaz> ;) [00:05] <omega_> no, the goal for a minimal gstreamer is 50K, as defined by my boss, which I think we can reach with things like xml support disabled, etc. [00:06] <omega_> no, glib + gdk + gtk should take 500K [00:06] <omega_> linux 500K [00:06] <omega_> support libs 1M [00:06] <omega_> (these numbers are completely off the top of my head, btw) [00:06] <omega_> gst + some plugins 500K [00:06] <omega_> apps 500K [00:06] <omega_> that totals 3M [00:07] <omega_> assume our target is 4MB and try to fit that [00:07] <omega_> (I don't think our target is that small, since flash is so cheap these days, but for the sake of argument....) [00:07] <Uraeus> omega_: found the sentence on wiki: No. But one could be written using GStreamer. Please note that despite the popular belief, aRts is a complete media server (audio, video) with components, schemes and plugins (MCOP) just like GStreamer. [00:07] <omega_> ah, what node? [00:08] <Uraeus> http://www.gstreamer.net/wiki/GstFaq [00:09] <omega_> 66.37.66.32 [00:09] <taaz> no, that wasnt me ;) [00:09] <omega_> swva.net [00:10] <taaz> yeah, that is me last editing it... but i'm not the person who wrote that part [00:10] <omega_> yeah, I know, it was someone else, dunno who because the history mechanism is lame [00:10] Action: omega_ responds [00:11] <Uraeus> prolly stw going over our site and adding stuff as 'needed' <g> [00:11] <omega_> no, not stw, he knows better than this [00:11] <omega_> I'm putting stw's own comments in there.... [00:11] <Uraeus> njaard? [00:12] <omega_> doubt it [00:12] <taaz> the revision history is there, its just the print that sucks. newer versions fix that. i've just not upgraded yet [00:12] <omega_> the fact that it's not actual version-controlled is what sucks [00:13] <omega_> taaz: I get 1,710 + 280 + 45 + 45 K for gtk+gdk [00:14] <taaz> good luck trimming it [00:14] <omega_> 2.8M [00:15] <omega_> a lot of the widgets could be turned off [00:15] <omega_> by size, in hex: [00:15] <omega_> 00005278 T gtk_range_get_value [00:15] <omega_> 000057c0 T gtk_clist_undo_selection [00:15] <omega_> 000058fc T draw_spin_entry_shadow [00:15] <omega_> 00005dd4 T gtk_text_view_place_cursor_onscreen [00:15] <omega_> 00006084 T gtk_notebook_new [00:15] <omega_> 00006350 T gtk_ctree_get_type [00:15] <omega_> 00009450 T gtk_tree_view_get_type [00:15] <omega_> 0000b9e4 T gtk_text_forward_delete [00:15] <omega_> those are ridiculously huge functions [00:16] <taaz> put refactoring hat on and get to work [00:16] <omega_> something's very wrong here. gtk_notebook_new: [00:16] <omega_> return GTK_WIDGET (gtk_type_new (gtk_notebook_get_type ())); [00:17] <omega_> there's no way that's 25KB [00:17] <taaz> isnt that the location not the size? [00:17] <omega_> no, not with --size-sort [00:17] <omega_> at least, theoretically [00:18] <taaz> 0x6350-0x6084 = 716 bytes... is that so bad? [00:18] <omega_> according to man nm, those are sizes, not offsets, and 716 bytes *is* bad for something that calls a function [00:19] <omega_> nm on gtknotebook.lo itself tells an entirely different story [00:19] <taaz> stripped? [00:20] <omega_> that says 66 bytes, which is still big, but probably includes debug and other stuff [00:21] Action: taaz should run that graphviz stuff on gstreamer code again [00:21] <omega_> ok, I commented on GstFaq [00:22] <omega_> gtknotebook alone is 35K [00:22] <omega_> wtay: btw, pre3 built fine [00:23] <taaz> isn't there some tool that scans apps/modules/etc and removes all symbols that arnt referenced? [00:23] <omega_> I dunno, I could use one.... [00:23] <Uraeus> omega_: agree with JG in regard to the soundserver? [00:23] <omega_> for the most part [00:23] <omega_> oh, and csl 0.1.2 was just released, speaking of.... [00:27] <maYam_> later guys! [00:27] maYam_ (ma...@ca...) left #gstreamer (Client Exiting). [00:27] <wtay> omega_: cool [00:28] <wtay> it built fine on maYams untainted pc too [00:28] <wtay> ship it [00:29] <omega_> well... need to set up release notes, and all that [00:29] <Uraeus> yup, ship it - as for Slashdot I think I have an angle that will get us posted [00:29] <wtay> changelog is in docs/random/NOTES.0.2.0 [00:29] <ChiefHighwater> hey wtay: is maYam pronounced "my am" or "may hem" or "ma'am [00:29] <wtay> ChiefHighwater: ma'am [00:29] <ChiefHighwater> txs [00:30] <wtay> ChiefHighwater: but with the "ma" like in "may" [00:30] <ChiefHighwater> like from Texas [00:30] <ChiefHighwater> a drawl [00:30] <wtay> ChiefHighwater: I wouldn't know :) [00:31] <wtay> ChiefHighwater: I think it also is the work for some eastern godess.. [00:31] <wtay> s/work/word [00:31] <ChiefHighwater> hrmm [00:31] <Uraeus> hehe [00:32] Sopwith (so...@na...) joined #gstreamer. [00:32] <wtay> yo [00:32] <Uraeus> hi sopwith [00:32] <omega_> yo Sopwith [00:32] <Sopwith> ullo [00:32] <omega_> Sopwith: any luck converting gst to gobject? <g> [00:33] <Sopwith> not with the previously-unknown funky requirements :) [00:33] <omega_> heh [00:33] <Uraeus> Sopwith: early preview with soundserver info: http://www.linuxrising.com/files/jimgettysinterview.html [00:33] <Sopwith> hmm [00:33] <omega_> Sopwith: if it were possible to link against gobject 2.0 while still letting people write gtk 1.2 apps with gstreamer, I'd go for it, but that's not gonna happen ;-( [00:34] <Uraeus> omega_: why can't people write GTK+ 2.0 apps? before we are ready GTK+ 2.0 should be ready for public consuption anyway [00:35] <omega_> yes, but it's not yet.... [00:35] <Uraeus> neither is gstreamer to some extent [00:35] <Sopwith> well, it is pretty ready [00:35] <omega_> until 2.0 is out and developers are actually developing against it, we'd be kinda jumping the gun [00:35] <wtay> chicken.. egg .. [00:35] <Sopwith> Because gtk 2.0 seems not to break too much these days, I agree with uraeus to some extent, but I have also seen the darkness of chasing a moving target [00:36] <omega_> Sopwith: right [00:36] <Uraeus> chicken.. egg .. bacon.. [00:36] <omega_> now, one option is to keep a gobject-based branch alive until gtk 2.0 is official [00:36] <omega_> Uraeus: don't, you're making me hungry [00:36] <Sopwith> yes, but that is as bad as just doing a bunch of #ifdef's [00:37] <Sopwith> Apparently CVS is not great at merging conflicts between branches [00:37] <omega_> Sopwith: not really, not of this magnitude [00:37] <taaz> is a complete mapping from 2.0 calls to 1.2 calls possible? [00:37] <omega_> Sopwith: Tim says that baulig has something that might be relevant here [00:38] <Uraeus> need to sleep, Sopwith if you want to read the interview do it tonight cause I am removing it tommorow morning (CET) since it isn't published yet [00:38] Uraeus (csc...@c2...) left #gstreamer. [00:38] <Sopwith> omega_: Yes, I had heard of something like that [00:38] <omega_> I will email him [00:38] <Sopwith> uraeus: I read it :) [00:38] ajbusy (aj...@p2...) joined #gstreamer. [00:38] <ChiefHighwater> ello ajbusy [00:39] Nick change: ajbusy -> ajmitch [00:39] <ajmitch> ello ChiefHighwater 8-] [00:40] <ajmitch> what's happening in here today? ;) [00:40] <taaz> question: what are the downsides of just dropping gtk 1.2 support in gstreamer completely? its not like there is a huge userbase to support... and so what if people have to upgrade... [00:40] <ajmitch> taaz: drop it, i say ;) [00:41] <omega_> taaz: well, as long as there's sufficient support for people to get gtk 1.3 and friends, sure [00:41] <ajmitch> let the users fear the mighty coders of gstreamer.... ;) [00:41] <taaz> its not really that bad to say "hey, we are using the latest stuff, if you want to use our code, upgrade" [00:41] <omega_> rh7.1 ships with gtkbeta packages, but they haven't been updated at all afaict [00:41] <omega_> dunno about debian [00:41] <Sopwith> Well, there are a number of problems with requiring gtk 1.3/2.0 stuff: [00:41] <omega_> they have to coexist, of course [00:42] <Sopwith> 1. You are using a theoretically unstable platform for devel (even though most of the unstableness is gone) [00:42] <Sopwith> 2. The barrier to getting the deps set up is higher than with gtk 1.2, making it harder to get users and developers for now. [00:43] <omega_> yup [00:43] <Sopwith> OK, I guess two is a number. [00:43] <omega_> last I checked, yeah <g> [00:44] <ajmitch> 1 isn't too much of a problem, imho. it's not as if gstreamer is any more stable ;) #2 i guess has to be compard against the effort to maintain gtk 1.2 compatibility [00:44] <omega_> ajmitch: right [00:44] <omega_> except for that stability part <g> [00:44] <Sopwith> 1. is a problem because the effect is multiplicative... [00:45] <Sopwith> the main question to address #1 would be whether gtk 1.3 is close enough to final 2.0 to warrant using it [00:45] <ajmitch> Sopwith: yeah, it'd be a hassle trying to see if the bug was in gtk or gstreamer at time [00:45] <ajmitch> i should really attempt to get gtk installed from CVS [00:46] <Sopwith> well, that's not so hard, but having binary compat broken every day can suck, for example [00:46] <omega_> mail sent [00:46] <omega_> Sopwith: quite [00:46] <ajmitch> i recompile gstreamer everyday anyway... ;) [00:47] <omega_> it would royally suck having to keep track of which version of gstreamer (day/hour) works against which version (day/hour) of gtk [00:47] <Sopwith> well, the way to solve that is to just explicitly specify which gtk version people are supposed to develop against [00:48] <ajmitch> how far away is gtk 2.0 ,as a rough guess? [00:48] <Sopwith> dunno, weeks-months [00:48... [truncated message content] |
From: <wim...@ch...> - 2001-06-08 04:29:54
|
[06:48] arik (ar...@sp...) joined #gstreamer. [06:48] <arik> hi [06:49] <omega_> yo [06:49] Action: arik is in ohio [06:49] <omega_> we were wondering, that friend with the $1700 webcam, is the webcam the primary use, or does she use it as a video camera primarily? [06:49] <arik> haha [06:49] <arik> no, it is used for video camera and still pictures [06:49] <arik> she was just wanting to try it as a webcam [06:49] <omega_> good <g> [06:49] <arik> ;-) [06:50] <omega_> ohio? [06:50] <arik> yeah [06:50] <arik> dealing with some school bullshit for a few days [06:50] <omega_> neat [06:50] <arik> heh [06:50] <omega_> not gonna be back in the area by tomorrow 7pm? [06:50] <arik> yeah, hanging out with friends and whatnot [06:50] <arik> yes i will [06:50] <arik> oh no [06:50] <omega_> darn [06:50] <arik> i will be back in the bay area by 1:30pm friday [06:50] <omega_> bay? [06:51] <arik> oh [06:51] <arik> nm [06:51] <arik> was confused [06:51] <omega_> I thought you were in seattle? [06:51] <arik> i will be back in seattle tomorrow [06:51] <omega_> heh [06:51] <arik> then i head to sf on friday for the summer [06:51] <ajmitch> omega_: when's your presentation? [06:51] <omega_> Thursday 7pm [06:51] <arik> yeah i would come [06:51] <ajmitch> 2 hours by my clock ;) [06:51] <arik> but i can't get to portland [06:52] <omega_> too bad [06:52] <arik> it is rather [06:52] <ajmitch> omega_: where's 0.2.0? ;) [06:53] Action: ajmitch ducks [06:56] <arik> hmm [06:56] <arik> i should disable full screen tonight [06:57] <arik> as i don't have time to fix it before 0.2.0 [06:57] <omega_> yup, or fix it [06:57] <arik> heh [06:57] <arik> i won't have any real time to spend on gst for two more weeks [06:57] <arik> then you're all mine ;-) [06:57] <ajmitch> i won't for the next week or 2, but i'll be free for 3 weeks after that... [06:58] <arik> heh [06:58] <arik> i just have to finish up with netscape [06:58] Nick change: taaz-away -> taaz [06:59] <omega_> arik: you're fixing netscape? yaaaaay!!! [07:00] <arik> hehe [07:00] <arik> ;-P [07:00] <ajmitch> i have exams starting in 2 days... [07:00] <ajmitch> (as you can see by my diligent study) [07:01] <omega_> given all your nick changes, I would have expected an ajstudy or something by now [07:01] <ChiefHighwater> sleeping on your books doesn't count as studying [07:01] <ajmitch> hehe [07:01] <ajmitch> why do i always get hassled? ;) [07:02] <arik> heh [07:02] <arik> cause you deserve it? [07:02] <ajmitch> hey, that's not fair ;) [07:03] <arik> ;-) [07:03] <ChiefHighwater> here, I'll harass someone else [07:03] Action: ChiefHighwater slaps omega_ around a bit with a large trout [07:04] Action: omega_ grabs the nerf weaponry in the closet..... [07:04] <ChiefHighwater> omega_ is sitting nect to me in my office [07:04] <ChiefHighwater> omega_ is getting up [07:04] <ChiefHighwater> omega_ is grabbing a gun [07:04] <ChiefHighwater> nerf of course [07:04] <ajmitch> uh oh... [07:04] <ChiefHighwater> who accused him of being violent? [07:05] <ChiefHighwater> HE'S SHOOTING AT ME 8-] [07:05] Action: ChiefHighwater takes the gun [07:05] <omega_> ack!!! [07:05] Action: omega_ ducks [07:05] Action: ChiefHighwater is a lot bigger than omega_ [07:05] Action: omega_ runs from the room [07:05] <ajmitch> haha [07:05] Action: omega_ has one manually fired piece of ammo [07:06] Action: omega_ is hoarding it [07:06] <ajmitch> ChiefHighwater: yes, i've seen your photo remember? ;) [07:06] <taaz> tonight on FOX: "when virtual violence goes bad" [07:06] <arik> hah [07:06] Action: ChiefHighwater has 8 rlounds left [07:06] <ChiefHighwater> rounds [07:06] <ajmitch> to think this goes into the logs... [07:06] <ChiefHighwater> 7 [07:06] <ChiefHighwater> 8 [07:06] Action: omega_ protects his laptop from fire [07:07] <ChiefHighwater> you feel better now, ajmitch? [07:07] <ajmitch> now we know why omega_'s hard drive makes funny sounds ;) [07:07] Action: omega_ uses his laptop case as a shield [07:07] <omega_> hehehee [07:07] <arik> bbiab [07:07] Nick change: arik -> arik|afk [07:07] <omega_> noooo! [07:07] Action: omega_ just lost his ammo [07:10] Action: ChiefHighwater thinks lame is living up to its name [07:10] <ajmitch> hmm, and i thought that omega_ & ChiefHighwater were serious, mature, and respectable members of society.... ;) [07:10] <omega_> pfff [07:10] <ChiefHighwater> rofl [07:11] Action: taaz never thought so ;) [07:11] <omega_> thx [07:11] <ajmitch> ah well, the illusions of the young... [07:11] <ChiefHighwater> lol [07:12] <ChiefHighwater> we only act respectable when we have too 8-] [07:12] <omega_> which is when, exactly? [07:12] <ChiefHighwater> I dunno [07:12] <omega_> right [07:12] <ChiefHighwater> 8-] [07:12] Action: ChiefHighwater needs a webcam in his office for y'all 8-] [07:12] <ajmitch> you see, i act mature & respectable all the time ;) [07:12] <omega_> er [07:13] <omega_> ajmitch: pfff [07:14] <ajmitch> hmm, when have i acted otherwise? [07:14] <omega_> no comment [07:14] Action: ChiefHighwater looks away whistling to himself [07:15] Action: taaz scans the logs for evidence ;) [07:15] <omega_> I wasn't gonna be that mean <g> [07:15] <taaz> ok ok.. i'm too lazy... someone else do it [07:16] Action: ajmitch runs to delete logs [07:16] <taaz> s/lazy/intoxicated/ [07:16] <ajmitch> hehe [07:16] Action: omega_ prepares the other nerf weapon while chw is out of the room [07:17] <ajmitch> lol [07:17] Action: ajmitch should go now to eat some fod [07:17] <omega_> since he remembered to take his with him [07:17] <ajmitch> food, that is... ;) [07:17] <ajmitch> bbl [07:17] Nick change: ajmitch -> ajfood [07:17] <ajfood> ;) [07:17] Action: taaz ponders why deb building tools fail because of "unrepresentable differences" in source trees... hmm [07:17] Action: omega_ is getting shot at [07:18] Action: omega_ fires and runs out [07:18] Action: taaz decides to ponder over a bowl of icecream [07:19] Action: omega_ has the gun removed again [07:24] Action: taaz awaits GOBJECT checkins... [07:24] Action: taaz holds breath waiting... [07:28] <ChiefHighwater> omega_'s gun is broked 8-[ [07:31] <ChiefHighwater> we need new guns 8-] [07:32] Action: omega_ runs out of ammo and breath [07:33] walken (fo...@c1...) joined #gstreamer. [07:37] <omega_> what is taaz holding his breath for? [07:37] <taaz> gobject checkins... [07:37] <omega_> um [07:37] Action: taaz is turning blue ;) [07:38] <omega_> taaz: maybe you can colormatch the BHG's ChiefHighwater's got on his screen that are supposed to be gray.... [07:38] Nick change: arik|afk -> arik [07:38] <taaz> the what? [07:38] <ChiefHighwater> no, there supposed to be grey [07:38] <arik> back [07:39] <omega_> http://temple-baptist.com/~omega/images/4xBHG-3c.jpg [07:40] Action: taaz is looking at prices for flights to ottawa [07:40] Action: omega_ needs to do the same [07:40] <arik> heh [07:40] <arik> isn't ridgerunner sending you omega_? [07:41] <omega_> ridgerun? dunno yet [07:41] <omega_> will email tomorrow [07:41] <arik> run [07:41] <omega_> I'm going either way, imo [07:41] <arik> yeah [07:41] <arik> hehe [07:41] <arik> good ;-) [07:41] <arik> noone is using the changelog [07:42] <arik> which makes me mad [07:42] <walken> OLS meeting ? [07:42] <omega_> yeah [07:42] <walken> yeeaaaah [07:42] <arik> mmaaaaaaaaadd [07:44] <taaz> arik: cvs2cl ;) [07:44] <omega_> temple-baptist.com/events/choir-favorites-2001/ [07:44] <arik> ;-P [07:44] <taaz> walken: are you crashing at aarons place again? [07:45] <walken> I guess [07:45] <walken> I havent asked him yet :) [07:45] <walken> he knows I'm coming though [07:45] Action: arik prepares to disable full-screen [07:45] <arik> then if i fix it [07:45] <arik> i can turn it back on [07:45] <omega_> just grey it out [07:46] <arik> yes [07:46] <arik> that's what i'm doing ;-) [07:46] <taaz> heh... jeff and i are coming again... jeff is trying to get his new company to pay for expenses. which means i'll have another option for lodgings ;) [07:46] <walken> lol [07:46] <walken> I'm still coming on my own [07:46] <taaz> you should talk to the guy... he's in the bay area now [07:46] <walken> jeff ? [07:46] <walken> why not [07:47] <walken> you know what he's doing ? and where he lives [07:47] <taaz> yeah, he's been there for a while. working at Intransa. i'm in a ytalk session with him 24/7 ;) [07:47] <walken> lol [07:48] Nick change: ajfood -> ajmitch [07:48] <walken> is he your twin brother or something ? :) [07:48] <taaz> something ;) [07:48] <walken> lol [07:49] <walken> whats intransa ? [07:50] <taaz> a startup (of course). one of my friends was main force getting it going, a bunch of my friends and now even my roommate work there! i probably would be too if i could finish my damn masters thesis [07:50] <arik> he [07:50] <arik> h [07:51] <arik> what are they doing? [07:51] <walken> haha [07:51] <walken> like you cant tell :) [07:51] <walken> OK [07:51] <arik> heh [07:51] <omega_> taaz: any suggestions on who to fly? [07:51] <taaz> jeff lives in fremont btw [07:51] <walken> ok [07:52] <walken> reminds me when I met nathan [07:52] <walken> he wouldnt even tell me if he worked on software or hardware [07:52] Action: taaz never been to that area... has no idea what that means ;) [07:52] Action: arik can't wait to get back to san francisco [07:52] <walken> I know where fremont is :) [07:52] <taaz> i never signed nothing... i could tell you more than i'm supposed to ;) [07:52] <arik> haha ;-) [07:52] <ChiefHighwater> omega_: http://www.topixonline.com/cgi-bin/sgin0101.exe?T1=TE+258&UID=2001060621544565 [07:52] <arik> the basic thing that won [07:52] <arik> 't get you sued would be fine [07:54] <taaz> well.. the web site says "developing Internet-scale storage solutions that can be deployed in Internet-time" whatever that means ;) (*cough* network storage *cough*) [07:54] <arik> haha [07:54] <arik> ;-) [07:54] <arik> sounds interesting i suppose [07:54] <ajmitch> heh [07:54] <ajmitch> but with some innovative new technology that noone else has tried? [07:55] <arik> heh [07:55] <arik> ;-P [07:57] <taaz> walken: jeff is asking if you want a job ;) [07:57] <arik> haha [07:57] <arik> let's make a deal in #gstreamer [07:58] <omega_> great. just realized that OLS is right over the top of a vacation plan [07:58] <arik> ek [07:58] <omega_> which we can probably move around a bit, but it's messy [07:58] <arik> right [07:59] Action: taaz considers time vs cost of driving vs flying to OLS [07:59] Action: ajmitch swears merrily at gtk+ ;) [07:59] <arik> ooh [07:59] <arik> i'll join in on that [08:01] <ajmitch> am merely trying to compile the latest version from HEAD ;) [08:02] <arik> heh [08:02] <arik> brave sole [08:02] <ajmitch> ;) [08:02] <ajmitch> gdk-pixbuf/pixops/Makefile.am:8: variable `GLIB_LIBS' not defined [08:02] <ajmitch> configure.in:421: warning: AC_TRY_RUN called without default to allow cross compiling [08:02] <ajmitch> loading cache ./config.cache [08:02] <ajmitch> ./configure: line 637: syntax error near unexpected token `AM_INIT_AUTOMAKE($PACKAGE,' [08:02] <ajmitch> ./configure: line 637: `AM_INIT_AUTOMAKE($PACKAGE, $VERSION, no-define)' [08:02] <ajmitch> doesn't even finish configure ;) [08:02] johnix (thomas@61.152.140.57) joined #gstreamer. [08:02] <arik> man, stop _needs_ to get fixed in gstplay [08:02] Action: arik puts that on the top of his list [08:03] <ajmitch> arik: yep ;) [08:03] <johnix> Hi, [08:03] <arik> ;-) [08:03] <ajmitch> hi [08:04] <johnix> Is it normal that I get a "connexion refused" when trying access http://gstreamer.net ? [08:04] <ajmitch> nope... [08:04] <johnix> Actually, I get a "host unreachable" [08:05] <ajmitch> hmm, works here, i should check that it's not cached [08:06] <arik> done [08:06] <arik> full-screen is now grayed out [08:06] <omega_> mu. I'm gonna have to rebranch GOBJECT1 <g> [08:06] <arik> heh ;-) [08:07] <ajmitch> johnix: nope, i can access the site ok, it seems [08:08] <omega_> I get it here [08:08] <ajmitch> omega_: why rebranch? ;) [08:08] <omega_> j/k [08:08] <johnix> ajmitch: So, is it still there? [08:09] <ajmitch> johnix: yes [08:09] <johnix> ok. [08:09] <johnix> But I definitely cant. [08:09] <omega_> hmm [08:09] <omega_> can you ping it? [08:09] <arik> man i hate sawfish sometimes [08:09] <johnix> omega_: nope. Destination host unreachable. [08:09] <omega_> but it gets an IP address? [08:09] <ajmitch> ajmitch@gw:~$ ping gstreamer.net [08:09] <ajmitch> PING gstreamer.net (216.136.171.204): 56 data bytes [08:09] <ajmitch> 64 bytes from 216.136.171.204: icmp_seq=0 ttl=240 time=619.4 ms [08:09] <johnix> So it might be a downstream network problem... [08:09] <ajmitch> could be... [08:10] <omega_> traceroute it [08:11] <johnix> Oops... [08:11] <johnix> Traceroute's not installed. Brb [08:13] <ajmitch> omega_: so what bugs remain before 0.2.0 is worthy of our great praise & adoration? [08:13] <omega_> dunno, you tell me [08:13] <arik> i wish i had time to fix the stop button, oh and i wish i knew enough about gst to be able too ;-P [08:14] <ajmitch> omega_: i still got crash after playing video (haven't updated for maybe 6-12 hours tho) [08:14] <arik> hmm [08:14] <arik> with gstplay? [08:15] <omega_> that's the ending crash, we know about that [08:15] <ajmitch> arik: yep [08:15] <arik> i don't get hat [08:15] <arik> er that [08:15] <ajmitch> omega_: only started happening a few days ago for me [08:15] <ajmitch> (not sure when, time has little meaning to me ;) ) [08:15] <johnix> weird! [08:16] <arik> oh [08:16] <arik> yeah i do ;-( [08:16] <arik> damnit [08:16] <johnix> I get to 202.97.33.86 in 6 jumps [08:16] <johnix> but at step 2, I get to a host with a private (NAT) address!!! [08:16] <arik> hmm [08:17] <omega_> whoops [08:17] <johnix> 192.168.0.10 [08:17] <ajmitch> you on a lan? [08:17] <ajmitch> i get this...: [08:17] <ajmitch> traceroute to gstreamer.net (216.136.171.204), 30 hops max, 38 byte packets [08:17] <ajmitch> 1 ajmitch (192.168.0.12) 2.101 ms 0.264 ms 0.238 ms [08:17] <ajmitch> 2 max2.dun.ihug.co.nz (203.109.135.12) 132.620 ms 129.513 ms 139.783 ms [08:18] <ajmitch> until it reaches the right address [08:18] <omega_> anyone know of a samba irc channel? [08:18] <johnix> The funny part is that that very address was in my hosts.conf from a lan I've been on lately [08:18] <taaz> ajmitch: we need a gstreamer-compprep.1 man page if you want to help out ;) [08:19] <johnix> I'm not on a 192.168.* lan! [08:19] <ajmitch> taaz: hehe, i have no idea what it does ;) [08:19] <johnix> I get this: [08:19] <ajmitch> johnix: some isps might set it up that way, i dunno [08:19] <ajmitch> i've seen a wacky one here in nz that uses private ips on it's machines [08:19] <johnix> 1) 61.152.140.214 (thats my NAT box's public address) [08:20] <johnix> 2) 192.168.0.10 (this shouldnt be there, as we're jumping from public to private to public!) [08:20] <johnix> 3) 202.96.201.255 [08:20] omega_omicron (om...@te...) joined #gstreamer. [08:20] <omega_> grrr [08:21] omega_omicron (om...@te...) left irc: [x]chat [08:21] Nick change: arik -> arik|afk [08:21] <ajmitch> taaz: so umm, what does gstreamer-compprep do? ;) [08:21] <johnix> 4) 202.101.63.5 [08:21] <johnix> 5) 202.101.63.242 [08:21] omega_omicron (om...@te...) joined #gstreamer. [08:21] <omega_> stupid xchat [08:21] <johnix> 6) 202.97.33.86 [08:21] <taaz> ajmitch: ask omega_ [08:21] <omega_> it creates a registry used by the complete program [08:22] <taaz> cat omega_ > gstreamer-compprep.1 [08:22] <ajmitch> ;) [08:23] omega_omicron (om...@te...) left irc: [x]chat [08:24] <johnix> WOW! When I try to go to max2.dun.ihug.co.nz (your step 2), it takes me 16 jumps! [08:24] <ajmitch> johnix: no surprise [08:24] <johnix> The last 6 are in the 203.* domain [08:24] <ajmitch> takes me 17 jumps to get to gstreamer.net [08:25] <ajmitch> the 203.* range is partly US, mostly within NZ [08:26] <johnix> This is weird! [08:26] <johnix> Who should be blamed? [08:27] Action: ChiefHighwater , it's always my fault [08:27] <johnix> If I can get to max2.dun.ihug.co.nz, and you get from there to gstreamer.net, [08:28] <omega_> johnix: but if you're not going through that host on the way already, it's not relevant [08:28] <johnix> then why cant I get to gstreamer.net? [08:28] <omega_> someone's BGP tables are fragged [08:28] <omega_> nothing you can do but wait [08:29] <johnix> omega_: some 36 hours, right? [08:29] <omega_> that's how long it's been down? [08:29] <omega_> or unaccessible, that is? [08:29] <johnix> No, thats my guess of how long it should take to recover... [08:30] <omega_> no, it should take seconds after someone realizes the problem [08:30] <omega_> the trick would be to find the AS that the problem is in and email or call them [08:30] <johnix> Well, [08:30] <omega_> but I dunno how to find the AS# based on IP [08:30] <ajmitch> not like DNS where entries have to propagate thru the caches [08:30] <johnix> Whats the AS number? [08:31] <omega_> a number assigned to a network, used for BGP routing [08:31] <omega_> treats each network as a black box, handles highest-level routing decisions [08:32] <johnix> How does it differ from the subnet/mask? [08:32] <ajmitch> this is where IPv4 can be messy, right? ;) [08:32] <omega_> johnix: very many conceptual levels above subnet [08:32] <omega_> when you hear MAE-East and MAE-West, those are the people dealing with this stuff [08:33] <johnix> Ok. I dont know about those high-level routing mechanismes... [08:33] <johnix> ... but this might help: I get the same problem with heroinevirtual.com [08:33] <omega_> I took a class on it, so... <g> [08:33] <omega_> do you get stuck in the same place? [08:33] <johnix> (which is hosting the libmpeg3 project) [08:33] <omega_> I think they might be sf hosted as well [08:34] <omega_> you shouldn't be able to get to sourceforge.net at all [08:34] <johnix> it's IP is 202.97.33.86 [08:34] ChiefHighwater (pa...@te...) left irc: Read error to ChiefHighwater[temple-baptist.com]: Connection reset by peer [08:34] <omega_> ChiefHigwater touched his Windoze computer, it had to reboot to make the change take effect.... [08:34] <johnix> I can ping sourceforge (216.136.171.196) [08:35] <ajmitch> inetnum: 202.97.32.0 - 202.97.63.255 [08:35] <ajmitch> netname: CHINANET-BB [08:35] <ajmitch> descr: CHINANET backbone network [08:35] <ajmitch> descr: Data Communication Division [08:35] <ajmitch> descr: China Telecom [08:35] <ajmitch> wtf? [08:35] <ajmitch> johnix: where are you? [08:35] <johnix> In China, why? [08:35] <johnix> In Shanghai [08:35] <omega_> that would explain that <g> [08:35] <ajmitch> seems like a problem near your end then ;) [08:35] <johnix> Huh? China blacklisting gstreamer??? [08:35] <omega_> johnix: the reason you get a NAT address is because the country you are located in is behind a BIG filewall [08:36] <ajmitch> man, i'd love to see what they run that on... ;) [08:36] <omega_> ajmitch: cluster, many gateway points [08:36] <johnix> So, the 192.168. address is a "public" address for chinese ISPs? [08:36] <omega_> no, that's private [08:36] <omega_> oh, right in china, yeah [08:36] <ajmitch> omega_: yes, how to manage it all tho? [08:36] <johnix> (well, for THE chinese ISP...) [08:37] <omega_> johnix: what's your public IP ? [08:37] <johnix> I mean, its private to the world, but public to us? [08:37] <omega_> right [08:37] <johnix> 61.152.140.213 [08:37] <johnix> fabulous... [08:38] <omega_> neat. [08:38] Nick change: taaz -> taazzzz [08:38] <omega_> oooh, cool [08:38] <omega_> 18 202.97.33.73 164.989 ms 172.021 ms 172.156 ms [08:38] <omega_> 22 192.168.0.9 170.244 ms * 169.721 ms [08:38] <omega_> end. [08:39] <omega_> ajmitch: how did you get the AS info for that ip range? [08:39] <ajmitch> there will be ways arond it, such as tunnelling it... [08:39] <ajmitch> omega_: whois [08:39] <omega_> whois what? [08:40] <omega_> what's the netblk syntax [08:40] <ajmitch> omega_: whois 202.97.33.86 [08:40] <ajmitch> ah, for you i think it's whois 202...@ap... [08:40] <omega_> nope [08:40] <ajmitch> i'm in the asia-pacific region so i just specify ip [08:40] <ajmitch> ok, @whois.apnic.net, iirc [08:41] <omega_> there we go [08:41] <johnix> ajmitch: btw, is there any openproject.net server closer than europe? [08:41] <ajmitch> johnix: not sure... [08:42] <omega_> check with lilo [08:43] <johnix> Yup. gstreamer.net and heroinevirtual.com both get to 202.97.33.86 in 6 jumps and block from there... [08:44] <ajmitch> ok, so everything seems to stop at that firewall or something... [08:44] <ajmitch> omega_: care to donate some bandwidth for ip6-on-ip4 tunnelling? ;) [08:44] <johnix> ajmitch: do you think this is a firewall? [08:44] <ajmitch> johnix: not sure, could be [08:44] <johnix> or just a misconfigured router? [08:44] <ajmitch> could just be bad router [08:45] <omega_> you're getting stuck before you leave 207 ? [08:45] <omega_> er, 202.67 ? [08:45] <omega_> or whatever it is [08:45] <johnix> I dont get that far. I stop at 202.97.33.86 [08:46] <big_T> omega_, you alive? [08:46] <ajmitch> big_T: we hope so, we're talking to him ;) [08:46] <big_T> oh, didn't notice that ;) [08:47] ChiefHighwater (pa...@te...) joined #gstreamer. [08:47] <big_T> omega_, one question/suggestion: in GstPluginDesc there is an entry for a plugin_init-function, but shouldn't also be an entry for a plugin_free och plugin_deinit-function??? [08:47] <omega_> big_T: I'm trying to get this STUPID windows networking problem here resolved [08:47] arik|afk (ar...@sp...) left #gstreamer. [08:47] <big_T> omega_, thats ok [08:47] <omega_> big_T: probably [08:48] <omega_> we can't unload plugins yet [08:48] <big_T> nay, I know that - but I think it should be added anyway, an "unsupported" "feature" ;) [08:48] <ChiefHighwater> hehe [08:49] <big_T> then the plugins can pretend to handle it, and we poor plugin-developers doesn't have to do all the work later ;) [08:49] <johnix> Ok. So I gotta get to the owner of that dead-end bit bucket so I can get my work done... [08:50] <ajmitch> johnix: either that or use some proxy server that you can connect to ;) [08:50] <ajmitch> http://poisk.hypermart.net/proxy.htm [08:54] <johnix> ajmitch: haha! I'll have a look at it... [08:55] <omega_> johnix: beware that trying to circumvent the national gateway is very likely a crime [08:55] <ajmitch> could be... [08:55] Action: omega_ is not joking [08:55] <ajmitch> it isn't here, but this is NZ [08:55] <ajmitch> omega_: it'll still have to go thru the national gateway tho [08:57] <johnix> ajmitch: guess what... [08:57] <ajmitch> johnix: i'm guessing... [08:58] <ajmitch> that site is unavailable? [08:58] <johnix> omega_: yup. 'hope I dont get in trouble for dowloading GPL'ed software... [08:58] <omega_> I'm wondering if that could be why you can't get to parts of sourceforge.... [08:58] <johnix> ajmitch And you probably even know who's blocking it. [08:59] <johnix> omega_: Aha... [08:59] <johnix> man, that really s**** [08:59] <ajmitch> johnix: hypermart.net is a site designed for cheap/free business webhosting tho [08:59] <johnix> So I cant see half of the world from here??? [09:00] <omega_> ajmitch: and you think that matters? [09:00] <omega_> johnix: welcome to China [09:00] <ajmitch> could still be a misconfigured router, it might not be a big conspiracy ;) [09:00] <omega_> .net [09:00] <johnix> ajmitch: thats what I was about to say [09:00] <johnix> But still... [09:00] <big_T> ajmitch, ofcourse it's a conspiracy, haven't you been watching the X-files? [09:00] Action: johnix added an entry to /etc/hosts [09:00] <ajmitch> johnix: you're not from china? [09:01] <ajmitch> big_T: not lately [09:01] <johnix> big_T: if the X-files talk about that, theyll be censored out of China... [09:01] <johnix> ajmitch: I'm from France [09:01] <ajmitch> cool [09:01] <johnix> I'm just taking a training for 3 months in Shanghai... [09:01] <omega_> ouch [09:02] <johnix> so yes, you can insult the govt, I wont feel offensed [09:02] <ajmitch> cool ;) [09:02] <omega_> they might though [09:02] Action: ajmitch remembers what France did to NZ back in '85... [09:02] <johnix> (althrough they may put me in trouble for that) [09:03] <johnix> ajmitch: Dont talk about that. [09:03] <ajmitch> oh, you mean the Chinese govt... ;) [09:03] <ajmitch> johnix: heh ;) [09:03] <omega_> DOH! [09:03] <johnix> I wont feel offensed if you _insult_ the french govt either... [09:03] <ajmitch> ;) [09:04] <omega_> <ajmitch> you see, i act mature & respectable all the time ;) [09:04] <omega_> hmmmm? [09:04] <ajmitch> my evil twin.... [09:04] <ChiefHighwater> lol [09:05] Action: ChiefHighwater needs one of those [09:05] <omega_> hrm, and I guess ajmitch has a bridge for sale too... [09:05] <ajmitch> how'd you guess? ;) [09:05] <ajmitch> of course, the current owners probably don't want me to sell it, but oh well ;) [09:06] <ChiefHighwater> details, details [09:06] <johnix> Ok. I think I'll export display for mozilla from France, then ;) [09:06] <ajmitch> lol [09:06] <omega_> mu [09:06] <ajmitch> you must have *lots* of bandwidth ;) [09:06] <omega_> + lbx.... [09:07] <johnix> lets try it! [09:07] <omega_> better off just running squid over there [09:08] <johnix> Nop. No X forwarding... [09:09] <johnix> omega_: would you set up a squid server for me? [09:09] <omega_> how much use would it be getting? [09:09] <omega_> and is China going to lob an ICBM at me? [09:09] <johnix> Mine and that of my co-worker. [09:09] <johnix> I can take no liability for your house. [09:10] <johnix> we have a 512Kbps connexion [09:10] <johnix> and only one ethernet cable;) [09:10] <omega_> I'd give it a day first, see if it's a routing problem [09:10] <omega_> then we can try a proxy [09:10] <johnix> (that will change in a few days, however...) [09:11] <johnix> omega_: I have another suggestion. [09:11] <johnix> omega_: what about my setting up a mirror for the gstreamer website in France? [09:11] <johnix> (or in China) [09:11] <johnix> (but thats not before a few weeks) [09:12] <omega_> it's on sourceforge, not sure there's really a need [09:12] <omega_> easier ways for you to get ahold of it [09:12] <johnix> Oh! [09:12] <johnix> Wait, [09:12] <omega_> if they really don't want you to see it, they'll get annoyed if you mirror it <g> [09:12] <johnix> Is there any proxy server requiring less resources than squid? [09:12] <omega_> yeah, search freshmeat [09:13] <ajmitch> yeah, several tiny non-caching ones [09:13] <johnix> omega_: I dont think they're after free software. [09:13] <johnix> ajmitch: Ok. [09:13] <johnix> I've got a root access on that server in France. [09:13] <johnix> I'll get it throuhg. [09:14] <johnix> Thanks a lot for your help!!!! [09:14] <omega_> johnix: can you try something like http://phpsips.sourceforge.net/ (randomly chosen) [09:15] <omega_> try codecs.org [09:15] <johnix> wow! 12 pages on freshmeat for "proxy"! [09:16] <ChiefHighwater> johnix:in the meantime, did you have a question about GStreamer we could help you with? (we are somewhat familiar with the contents of the webpage 8-] ) [09:16] <johnix> omega_: dead_end [09:16] <omega_> for the phpsips one? [09:16] Action: ajmitch gets back to swearing at gtk+ [09:16] <johnix> omega_: dead_end [09:16] <johnix> for both of them [09:16] <omega_> and codecs.org? [09:16] <omega_> ok [09:16] <omega_> um. [09:16] <omega_> but you can get to sourceforge.net ? [09:17] <johnix> yup. [09:17] <johnix> About gstreamer: yes, a very important question! [09:17] <omega_> phpsips is on .201, gstreamer.net and codecs.org are .204, and sourceforge.net is .196 [09:17] <omega_> all in the same subnet [09:17] <omega_> no way that's a routing problem [09:17] <johnix> Funny [09:18] <omega_> your question? [09:18] <ajmitch> unless sf.net is cached [09:18] <johnix> Well, then what are they after? [09:18] <omega_> ajmitch: doubt it [09:18] <johnix> Ok. [09:18] <omega_> johnix: ask em <g> [09:18] <johnix> I'm in China for a training in a Free Software company [09:19] <omega_> hrm, can *they* get to these IP's ? [09:19] <johnix> They want me to put together a kind of library which transparently plays various animation formats [09:20] <omega_> that would be gstreamer <g> [09:20] <johnix> I heared that's what your doing [09:20] <johnix> I'd need to look at your website, but well... [09:20] <omega_> heh [09:21] <omega_> I'm guessing you can't DCC either... [09:21] <ajmitch> ask, and omega_ will bless you with knowledge... [09:21] <omega_> I could DCC you the slides from the website, maybe [09:22] <ChiefHighwater> johnix:or maybe email would be better? [09:24] <johnix> wait, I'm setting up my proxy... [09:24] <omega_> ip_masq_irc ? [09:24] <johnix> omega_: nop. I get the DCC offer, but cant start the transfer [09:24] <omega_> ok, not surprising [09:24] <omega_> email addr? [09:24] <johnix> omega_: no use [09:24] <omega_> ? [09:25] <johnix> let me set up a proxy [09:25] <omega_> huh? [09:25] <johnix> on that server in France! [09:25] <ChiefHighwater> brb [09:25] <ChiefHighwater> chw_away [09:37] <ChiefHighwater> back [09:37] <omega_> as for animation, there's a fli codec, for general video there's a lot more stuff in cvs [09:43] <johnix> omega_: how about adapting a mpeg-2 decoder onto it? [09:44] <omega_> already have one [09:44] <johnix> how well does it run? [09:44] <omega_> it's one of the most solid formats we support [09:44] <omega_> walken (in the channel) is the author of the codec we use [09:45] <omega_> taaz is rebuilding the LiViD DVD player on top of GStreamer, based around it [09:46] <big_T> there, now the multiple-factories in gstalsa is gone [09:46] <omega_> cool [09:46] <omega_> 0.9? ;-) [09:47] <ajmitch> mpeg2 is watchable but slow on my system (no Xv) [09:48] <big_T> omega_, not yet... almost understood the new API though [09:48] <big_T> I'll try to fix it before the 0.2-release, but cannot promise anything [09:48] <omega_> no, 0.2.0 is frozen except for fixes [09:48] <omega_> but it should thaw tomorrow after the release [09:48] <big_T> btw, when is the 0.2-release due? [09:48] <omega_> I have a talk tomorrow, it'd be cool to release it in the middle of the talk <g> [09:49] <big_T> omega_, cool [09:49] <big_T> the talk at the lug? [09:49] <omega_> yup [09:50] <johnix> nice! [09:50] <big_T> have you put the slides on gstreamer.net yet? [09:50] <omega_> haven't finished them yet <g> [09:51] <big_T> haha [09:51] <ajmitch> omega_: what tims is talk? [09:51] <omega_> 7pm pst [09:51] <ajmitch> s/tims/time [09:51] <ajmitch> k [09:51] <omega_> in about 18hrs [09:51] <ajmitch> what's that in GMT? ;) [09:52] <omega_> 1900 + 0800 [09:52] <ajmitch> k [09:52] <omega_> I'll be on IRC at that time, of course [09:52] <ajmitch> of course ;) [09:52] <omega_> else how would I release? <g> [09:53] <ajmitch> hehe [09:53] <omega_> may need you and others to handle the news submissions [09:53] <ajmitch> righto [09:53] <ajmitch> i don't have an exam until the next day, so plenty of time ;) [09:54] <ajmitch> so long as we have our sla^H^H^Hcoordinator Uraeus on hand to help ;) [09:55] <big_T> hehe [09:57] derek (de...@cp...) left irc: Ping timeout for derek[cpe-24-221-112-50.az.sprintbbd.net] [10:03] derek (de...@cp...) joined #gstreamer. [10:05] walken (fo...@c1...) left irc: l8r [10:07] <big_T> any chanse I can get any credits in the AUTHORS-file for the alsa-plugin before the release? :) [10:08] <omega_> sure, are you in cvs write list? [10:08] <ajmitch> credits matter that much? ;) [10:08] <omega_> if so, add yourself [10:08] <ajmitch> if not, whinge to omega_ and he will add you ;) [10:09] <big_T> omega_, no I haven't got cvs-write-permissions :) [10:09] <ajmitch> omega_: you should correct that ;) [10:09] <ChiefHighwater> big_T:I thought you were on the sf project list? [10:09] <johnix> ajmitch: credit is everything, I believe... [10:09] <big_T> ChiefHighwater, nope am not... [10:09] <ajmitch> soon gstreamer will have the biggest list of developers of all SF projects ;) [10:10] <big_T> ajmitch, <rant> in all the sf-movies(pun not intended) they use credits to buy cool stuff </rant> [10:10] <ajmitch> johnix: doesn't bother me too much if my name isn't attached to any gstreamer code [10:11] <ChiefHighwater> ajmitch: omega_ says you canb add big_T 8-] [10:11] <ajmitch> ChiefHighwater: i'm not a gstreamer admin [10:12] <ChiefHighwater> omega_ says you don't need to be..just dd to the authors file [10:12] <ajmitch> afaik, one has to have admin rights to the project to add a developer [10:12] <ajmitch> oh, i thought he meant add him to SF [10:14] <ajmitch> big_T: your name is already in there for gstreamer m4 macros i see ;) [10:14] <ChiefHighwater> and the alsa stuff? [10:14] <ajmitch> nope, adding that [10:14] <big_T> ajmitch, yepp... but want more... MORE... MOOOREEE!!! muhahhahaha ... oh, sorry for that ;) [10:14] <big_T> haha [10:14] <ajmitch> list it as 'ALSA plugins' ? [10:15] <ajmitch> or 'insane ramblings' [10:15] <big_T> both would do fine ;) [10:15] <big_T> ALSA-plugins should work [10:15] <ajmitch> k [10:16] <ChiefHighwater> it is 1:15 am... omega_ and ChiefHighwater are going home [10:16] <ajmitch> ok ;) [10:16] <ajmitch> night [10:16] <omega_> l8r [10:16] <omega_> release today sometime.... [10:16] <ChiefHighwater> l8rz [10:16] omega_ (om...@te...) left irc: killall -SLEEP omega_ [10:16] ChiefHighwater (pa...@te...) left irc: [10:18] <ajmitch> big_T: so what else needs done for the alsa plugin? [10:22] <ajmitch> hmm, i wonder if i can close bugs that wtay has fixed... [10:27] <big_T> ajmitch, I'm trying to get the ALSA 0.9.x-support working... almost understood the API now :) [10:28] <ajmitch> excellent! [10:28] <big_T> ajmitch, just removed the multiple-factories from alsasink, and will do the same with alsasrc after lunch [10:28] <ajmitch> ok [10:28] <ajmitch> going to dump it into HEAD after 0.2.0 is released? [10:28] <big_T> then I'll try to do something similar to the OSS-stuff, multiple cards are fun ;) [10:28] <ajmitch> nice ;) [10:29] <ajmitch> what were the other plans for the OSS plugins? [10:29] <big_T> ajmitch, I'll send a patch to wtay or some other poor victim when I'm done [10:29] <ajmitch> hehe [10:29] <ajmitch> i can test out some of your alsa stuff, i'm running 0.9 drivers [10:29] <big_T> ajmitch, I haven't seen any other plans about the OSS-plugins... [10:29] <big_T> ajmitch, okey... sounds nice, just need to code them first ;) [10:29] <ajmitch> on the mailing lists awhile ago, what was being discussed? [10:30] <big_T> ajmitch, the essence of it was that multiple-cards should be added, I think anyway [10:31] <big_T> there was also a suggestion that it should support multiple-opens and such, but I think a mixer-plugin should be used for that [10:31] <ajmitch> ok [10:31] <big_T> mix all streams together and send the result to the output-plugin, doing it otherwise is not a good sollution [10:32] <ajmitch> yes, i think a mixer plugin would be quite doable, whereas mutiliating the osssink to support it would be ugly ;) [10:33] <ajmitch> a mixer plugin would also be useful for audio processing as well [10:33] <big_T> yes, I agree with that... no need to more features than needed in a given plugin, it just introduces more bugs and more debuging(which noone I know really enjoys) [10:33] <big_T> ajmitch, haven't anyone written a mixer yet??? [10:33] <ajmitch> ugh, debugging is evil task ;) [10:33] <ajmitch> dunno, i think there probably is one somewhere ;) [10:34] <big_T> haha, too many plugins, too few with a clue ;) [10:34] <big_T> I think all the plugins should have a README-file with info about that particular plugin [10:34] <ajmitch> yes! [10:34] <ajmitch> but they should all at least have some info in the header... [10:35] <big_T> hmm... should perhaps send a mail about this to the mailinglist [10:35] <ajmitch> gstreamer-inspect gives a quick overview [10:36] <big_T> yes, I know... but I still think that a README-file should be present, with explanation of options, known bugs, limitations, features and perhaps an FAQ [10:37] <ajmitch> we have an adder plugin in plugins/filters [10:37] <ajmitch> yes, but who would want to write such stuff? hackers are lazy ;) [10:37] <big_T> yeah, I know :) [10:37] <big_T> bad hacker, baad hacker ;) [10:38] <ajmitch> hehe [10:38] <ajmitch> anyway, i will bbl, have to ring someone... [10:40] <big_T> okey [10:41] ajmitch (aj...@p2...) left irc: Ping timeout for ajmitch[p28-max2.dun.ihug.co.nz] [10:43] ajmitch (aj...@p5...) joined #gstreamer. [10:45] Action: johnix finally got his poor proxy working... [10:45] <ajmitch> excellent ;) [10:45] Action: ajmitch just got his job sorted out for the break between semesters... [10:46] <ajmitch> ah crap, have to ring someone else.... brb ;) [10:48] <johnix> how easy is it to script gstreamer? [10:49] ajmitch (aj...@p5...) left irc: Ping timeout for ajmitch[p56-max6.dun.ihug.co.nz] [10:56] johnix (thomas@61.152.140.57) left irc: Read error to johnix[61.152.140.57]: EOF from client [10:58] johnix (thomas@61.152.140.57) joined #gstreamer. [10:58] <johnix> re [11:03] <johnix> Can I get the 0.2 release tonight? [11:03] <johnix> is it the current CVS ? [11:09] johnix (thomas@61.152.140.57) left irc: Read error to johnix[61.152.140.57]: Connection reset by peer [11:11] ajmitch (aj...@p2...) joined #gstreamer. [11:11] <ajmitch> back... [11:41] Nick change: ajmitch -> ajzzzz [12:41] thomas (th...@21...) joined #gstreamer. [15:16] dobey (do...@dr...) joined #gstreamer. [15:29] dobey (do...@dr...) left irc: brb [15:35] dobey (do...@dr...) joined #gstreamer. [15:35] Nick change: richardb-out -> richardb [15:39] <richardb> Wooohoo! [15:39] <richardb> New computer arrives. [15:39] Action: richardb unpacks [15:39] <thomas> richardb: :) good luck [16:06] sienap (sy...@ip...) joined #gstreamer. [16:06] <sienap> wank-o-matics! [16:07] <dobey> uhm [16:08] <sienap> D0p3y!! [16:08] <sienap> How are u dude? [16:08] <dobey> ok [16:08] <sienap> he [16:09] <thomas> hi sienap [16:09] <thomas> what's up ? [16:09] <sienap> thomas great stuff (TM) [16:09] <sienap> :) [16:09] <thomas> like ? [16:10] <dobey> heh [16:10] <dobey> yeah [16:10] <dobey> like what? [16:10] <sienap> going to my gf in some hours ? :) [16:10] <thomas> you go man [16:10] <sienap> dobey ? [16:10] <dobey> heh [16:11] <sienap> so something happened lately ? [16:11] <thomas> sienap: I'm hacking the apachetoolbox to work with gstreamer [16:11] <thomas> sienap: care to test it for me ? I need feedback ;) [16:12] <sienap> thomas he ok [16:12] <thomas> sienap: ok, I'll package it and put it online [16:13] <sienap> great [16:14] <thomas> sienap: do you have a machine without stuff like gstreamer, libxml and stuff installed ? [16:14] <sienap> ehm [16:14] <sienap> ? [16:14] <sienap> why :) [16:14] <sienap> btw what is the thingy doing ? [16:14] <thomas> well, I want to see it work on a machine that already has the necessary deps and one that hasn't [16:15] <thomas> sienap: the idea is that it downloads, configures and installs libraries or other packages that plugins need [16:15] <thomas> so if you don't have lame, it gets it automatically for you and installs it from source [16:15] <thomas> then after that it compiles gstreamer [16:15] <thomas> right now it gets gtk+, glib, libxml, libtool, lame, mad, and something else [16:16] <thomas> I want to test it both on machines that already have stuff and machines that don't [16:16] <sienap> thomas WOOW :) [16:16] <sienap> you are making an auto installer for gstreamer ? [16:16] <sienap> netjes ;) [16:16] <thomas> apachetoolbox is a neat script to do that for apache [16:16] <sienap> but not sure if i have the time to test it all right now [16:16] <thomas> when it works, it's great ;) [16:16] <thomas> sienap: it's fire and forget [16:17] <sienap> ehm [16:17] <sienap> i have to go really son [16:17] <sienap> soon [16:17] <sienap> ;) [16:17] <sienap> however [16:17] <sienap> put it onlien [16:18] <thomas> sienap: get it from http://urgent.rug.ac.be/thomas/download [16:18] <thomas> it's GStreamertoolbox-0.0.2.tar.gz [16:19] <thomas> I haven't taken out all of the apche stuff out yet though [16:19] <thomas> run it as ./install.sh for interactive, or ./install.sh -f to just assume yes on all questions [16:20] <sienap> thomas he [16:20] <sienap> i think i'll check later [16:20] <sienap> really have no time right now [16:20] <sienap> i figured.. [16:20] <sienap> sorry [16:20] <sienap> i have to go actually [16:20] <thomas> ok... [16:20] <sienap> damn sorry! [16:20] <dobey> later [16:20] <sienap> i'll try mondag [16:20] <sienap> when i am back [16:20] <thomas> bye [16:20] <sienap> sorry thomas kan er niks aan doen [16:20] <sienap> greets dopey! [16:20] sienap (sy...@ip...) left #gstreamer. [16:23] dobey (do...@dr...) left irc: brb, gdm [16:52] derek (de...@cp...) left #gstreamer. [17:40] hadess (ha...@de...) joined #gstreamer. [17:42] <hadess> hey guys [17:43] <thomas> hi hadess [17:43] Nick change: thomas -> thomas-roof [18:29] Nick change: taazzzz -> taaz [18:30] <hadess> hey taaz [18:32] <taaz> yo [18:43] Zeenix (programmer@203.135.12.205) joined #gstreamer. [18:44] <Zeenix> wtay is again busy, my luck always sucks [18:53] thomas-roof (th...@21...) left irc: [x]chat [18:58] ChiefHighwater (pa...@te...) joined #gstreamer. [18:59] steveb (st...@no...) joined #gstreamer. [19:03] dobey (dobey@141.154.95.104) joined #gstreamer. [19:03] <Zeenix> hello dobey [19:10] Nick change: wtay-zZz -> wtay [19:10] <wtay> yo [19:10] hadess (ha...@de...) left irc: home [19:10] <ChiefHighwater> ello wtay 8-] [19:11] <Zeenix> hello wtay [19:12] <Zeenix> wtay: you seem to be quite busy these day, what are you upto [19:12] <wtay> Zeenix: I'm always busy :) [19:13] <Zeenix> wtay: would you like to give me a favour [19:13] <Zeenix> ? [19:13] <wtay> sure [19:13] <wtay> depends :) [19:14] <Zeenix> just watch how many millisec. it take's on you CPU for gsm encoding of a single sample i-e 160 shorts ? [19:15] <wtay> hmm, that needs a couple of minutes work... [19:15] <Zeenix> wtay: are you at it now ? [19:15] <wtay> yup [19:15] <wtay> creting a makefile :) [19:24] <Zeenix> am i still connected or not ? [19:24] <wtay> ./test [19:24] <wtay> 264362 cycles [19:24] <Zeenix> i asked you for millisecs. [19:25] <wtay> Zeenix: working on it... [19:30] <wtay> Zeenix: 0.2 msec here [19:30] <Zeenix> wtay: 0.2 msec or sec [19:30] <wtay> msec [19:31] <wtay> 1 frame [19:31] <Zeenix> wtay: did you used the ftime function & timeb struct ? [19:32] <wtay> Zeenix: nope, RDTS [19:32] <Zeenix> whats that ? [19:32] <wtay> measures the number of clock ticks [19:32] <wtay> number of cycles actually [19:33] <Zeenix> what man pages ? [19:33] <wtay> It's a CPU instruction [19:34] <Zeenix> wtay: have you ever used ftime function [19:34] <wtay> yes [19:34] <wtay> not usefull for this measurement [19:34] <Zeenix> wtay: why ? i used that for it [19:36] <wtay> it's not accurate enough if you want to measure nanosecond stuff [19:37] <Zeenix> i got 3 milliseconds out of it, does it mean 0.003 sec. ? [19:37] <wtay> yup [19:38] <wtay> my PC is most likely 10 times faster then yours [19:38] <Zeenix> wtay: 0.2 msec = ? sec [19:38] <wtay> 0.0002 sec [19:41] <Zeenix> wtay: i'll go on two weeks vacation on this monday. And then i'll start the real work on my proj [19:41] <wtay> it's actually a bit faster with 0.15 msec for 160 bytes after the caches are warm [19:42] <wtay> nic [19:42] <wtay> nice [19:42] <Zeenix> wtay: do you it gets fast after encoding the first one or two samples ? [19:43] <wtay> of course [19:43] <wtay> MMX optimisations could make this even faster [19:44] <Zeenix> my CPU has got MMX [19:44] <wtay> there are patches for libgsm to add MMX encoding [19:45] <Zeenix> wtay: where'll i get that [19:45] <wtay> dunno [19:46] <wtay> http://www.megsinet.com/~stabro/gsm_mmx.html [19:47] <wtay> it claims a two time speedup on compression [19:47] <Zeenix> wtay: if it really does what they say, thats the stuff then [19:48] <wtay> well, you won't notice the speed as it's quite fast already [19:49] <wtay> well.. you decide :) [19:52] <Zeenix> wtay: ok i must go home now [19:52] <wtay> ok, cya [19:54] Zeenix (programmer@203.135.12.205) left irc: [20:00] <steveb> a soundserver written in gstreamer would need to run for months, but an internal clock based on nanos would overflow in a day [20:00] <steveb> is this a problem? [20:00] <wtay> depends on how you handle the overflow :) [20:01] <steveb> fair enough [20:01] <wtay> I think it's easy to detect the overflow and do something about it [20:02] <steveb> wtay: are you planning to put your java bindings into CVS? [20:03] <wtay> steveb: not yet, I'm waiting for the cvs restructure, the current cvs hierarchy is messed up enough [20:03] <steveb> bindings/java ? [20:03] <wtay> yup [20:04] <wtay> I'm going to put it into a separate cvs module [20:04] <wtay> like all the plugins and helper libs [20:04] <steveb> good idea [20:05] <wtay> people can then check out whatever they need instead of the current +2.4MB [20:05] <steveb> good idea [20:07] <taaz> definately seperate modules for each languages bindings [20:08] <wtay> yes [20:15] <steveb> I don't have the 'complete' command - is there another way to set up gstreamer-complete? [20:16] <wtay> steveb: it's a command internal to bash [20:16] <steveb> well, not my bash :( [20:17] <wtay> what version? [20:19] <steveb> 1.14.7(1) [20:20] <wtay> ah, you need at least 2.0 I think [20:20] <steveb> oh [20:30] Nick change: taaz -> taaz-away [20:41] big_T (th...@fy...) left irc: Ping timeout for big_T[fysgr763.sn.umu.se] [20:59] taaz-away (dlehn@66.37.66.32) left irc: Ping timeout for taaz-away[66.37.66.32] [20:59] hds-afk (ha...@pc...) left irc: Ping timeout for hds-afk[pc2-guil2-0-cust150.gui.cable.ntl.com] [21:00] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee4c.a2000.nl] [21:01] taaz-away (dlehn@66.37.66.32) joined #gstreamer. [21:01] steveb (st...@no...) joined #gstreamer. [21:01] hds-afk (ha...@pc...) joined #gstreamer. [21:01] Nick change: ajzzzz -> ajmitch [21:04] omega_ (om...@om...) joined #gstreamer. [21:05] <wtay> yo [21:05] <omega_> yo [21:05] <ChiefHighwater> ello [21:05] <ajmitch> morning ;) [21:06] Kuroyi (ri...@ub...) left irc: Read error to Kuroyi[ubr-35.82.223.wmelbourne.cfl.rr.com]: EOF from client [21:06] Kuroyi (ri...@ub...) joined #gstreamer. [21:06] <ajmitch> seems like i'll miss the release today, i'll be out all day :( [21:06] <dobey> release? [21:06] <dobey> where? [21:07] taaz-away (dlehn@66.37.66.32) left irc: Ping timeout for taaz-away[66.37.66.32] [21:07] <omega_> I'm tempted to 'officially' release 0.2.0 in the middle of my talk tonight <g> [21:07] <wtay> hehe [21:07] <dobey> heh [21:07] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee4c.a2000.nl] [21:07] <wtay> pre3 is good enough for me [21:08] <omega_> ok, so given that HEAD == 0.2.0, what else do we need to get ready? [21:09] <wtay> release notes [21:09] <omega_> need announce, release notes, a list of places to submit, rpms and debs, and support packages [21:09] <ChiefHighwater> some plugin readme files like the mail mentioned? [21:10] taaz-away (dlehn@66.37.66.32) joined #gstreamer. [21:10] <omega_> reading [21:15] Nick change: ajmitch -> ajstudy [21:15] steveb (st...@no...) joined #gstreamer. [21:23] Action: ChiefHighwater thought omega_ read faster than that 8-] [21:27] Action: omega_ is discussing vacation vs. OLS [21:28] <dobey> meh [21:37] Nick change: taaz-away -> taaz [21:38] Action: taaz thinks OLS is a vacation [21:38] <omega_> heh [21:38] <omega_> ok, sent mail to boss asking if he thinks I should go [21:45] Nick change: hds-afk -> hadess [21:45] <hadess> guys [21:49] <omega_> ok, so, who's gonna do what to get the release stuff ready? [21:50] Action: ChiefHighwater is ordering the pizza 8-] [21:50] <wtay> hmm [21:50] <omega_> is ChiefHighwater providing release-party pizza to all of PLUG?? [21:51] <omega_> eesh [21:51] <ChiefHighwater> only if wtay is coming 8-] [21:51] <wtay> ChiefHighwater: ok :) [21:51] <omega_> wtay: you better get on that Concorde RSN.... [21:51] <ChiefHighwater> you come and I'll get you the best pizza you've ever had [21:52] <wtay> yeah, the fast one I'll be there in 2 hours :) [21:52] <omega_> only if you're on a scramjet, which last I heard wasn't going so well... [21:54] <ChiefHighwater> umm, I think I successfully distracted us from omega_'s question [21:54] <omega_> yeah... [21:54] <wtay> hehe [21:58] <ChiefHighwater> wtay:you missed all the fun last night...nerf wars and frenchmen trapped in china [21:59] <ChiefHighwater> at least we did fix Erik's gun, whew 8-] [21:59] <omega_> yeah, but we need to get another just like it, and more ammo [21:59] Nick change: hadess -> hds-afk [21:59] <ChiefHighwater> lots more ammo [21:59] <omega_> it's not fair pitting those two different guns against each other [21:59] <omega_> anyway..... ;-) [21:59] <omega_> so who's working on what? [21:59] <omega_> I have slides to do, among other things..... [21:59] <wtay> I'm still a bit busy... [22:00] <wtay> any comments on the NOTES? [22:01] <wtay> oooh, my deinterlacer is somewhar working... [22:03] <wtay> ok, release time :) [22:04] <ajstudy> hmm, release now? ;) [22:05] <wtay> it's about time, no? [22:05] <ajstudy> yeah, why not? ;) [22:06] <wtay> plugins.html needs an update [22:07] <omega_> let's make a list somewhere [22:07] Action: omega_ thinks we need a realtime whiteboard attached to irc [22:07] <omega_> zircon (the irc client written in tk) had something like that [22:07] <ajstudy> hmm [22:08] <ajstudy> has 0.2.0 been packaged by make dist, and tested that it compiles? [22:08] <omega_> pre3 has [22:08] <ajstudy> there have been makefile changes since then, right? best to test it... ;) [22:09] <omega_> yeah [22:09] <omega_> ok, I'm creating a text file on sf [22:10] <wtay> I'm creating an index.shtml and notice.shtml in releases/0.2.0 [22:11] <ajstudy> ok, what can i do until i actually leave to go study? ;) [22:11] <omega_> um, if you want to start writing some of the READMEs for plugins, that'd be useful [22:11] <ajstudy> ugh, doc writing? ;) [22:11] <omega_> yup [22:12] <ajstudy> i see some have already been written? [22:12] <omega_> not of quite the type that was suggested [22:12] <omega_> actual docs on what it takes in, puts out, what it does, and what the arguments are [22:12] <ajstudy> ok [22:12] <omega_> is pretty straightforward for most plugins [22:13] <ajstudy> like in plugins/oss/README? [22:13] <omega_> hrm, not really [22:13] <omega_> like a man page for the plugin [22:13] <ajstudy> hmm, ok [22:15] Action: ajstudy guesses something more like plugins/avi/README [22:15] <omega_> sorta, yeah [22:15] <omega_> but that's still more internals docs, really [22:15] <ajstudy> we want the docs to be sort of similar in the way they are structured, imho [22:15] <omega_> structure it like a man page, stuff like that goes in one of the sections towards the bottom [22:16] <ajstudy> k [22:17] <wtay> what was the code name again? [22:17] <omega_> um [22:17] <omega_> 'Free the Mallocs' was suggested, CHW had another idea [22:17] <ajstudy> oh? [22:17] <dobey> Code Name: Earbleed [22:17] <omega_> ? [22:18] <dobey> heh [22:18] <ChiefHighwater> 'Critical Mass' [22:18] <ajstudy> ChiefHighwater: ready to explode in your face? ;) [22:19] <omega_> no, just glowing Cerenkov blue.... ;-) [22:19] <taaz> if you add READMEs make sure to add them to Makefile.am EXTRA_DIST rule [22:20] <omega_> hrm, I wonder if Crayola would ever name a crayon 'Cerenkov blue' <g> [22:21] <ChiefHighwater> I erally think that this release will propell us into the mainstream of Linux Multimedia [22:21] <ChiefHighwater> as witnessed by the Gnome interest [22:21] <omega_> yup [22:21] <ajstudy> so long as it hits /. ;) [22:21] <omega_> heh [22:21] <taaz> oh jeez.. don't worry about the press [22:22] <taaz> if you really want the code to be used i'd start verifying the docs are all up to date [22:22] <taaz> and then add numerous annotated plugin examples [22:22] <omega_> yup. that and a test suite are gonna be the major non-feature pushes of post 0.2.0 [22:22] <ajstudy> taaz: you suggesting we do all this for 0.2.0, or for a later release? [22:23] <taaz> and add a bunch of -launch examples in a noticable place just cause that has good hack value [22:23] <omega_> yup. I can write up the text of the manpage for that, if someone can nroff-manify it for me <g> [22:23] <ajstudy> omega_: going to branch it to have a gtk 1.2-based branch (just for adding docs, stability, etc)? [22:24] <taaz> not to throw a wet blanket on things... but i think making a big deal out of a "release" is rather silly at this stage [22:24] <omega_> taaz: yeah, outside of this group. it's a big deal in this channel though <g> [22:24] <taaz> opensource + CVS world has latest code always available... [22:25] <ajstudy> something tells me that 'mute' and 'format' arguments for the alsasunk plugin are not to be used... [22:25] <omega_> the issue with a release is that we're gonna get a lot more people using the release than CVS, which means we have to cover ourselves to avoid getting pestered to death when it doesn't work [22:25] <omega_> ajstudy: prob not used yet [22:26] <ajstudy> switch (id) { [22:26] <ajstudy> case ARG_MUTE: [22:26] <ajstudy> ALSA_DEBUG_MSG("ARG_MUTE not supported"); [22:26] <ajstudy> break; [22:26] <ajstudy> case ARG_FORMAT: [22:26] <ajstudy> ALSA_DEBUG_MSG("ARG_FORMAT not supported"); [22:26] <ajstudy> break; [22:26] <omega_> that's a hint, yeah [22:26] <taaz> yeah, i suppose so. i think releases should be more frequent too... ie, add new plugin for FooBar, make a release [22:27] <omega_> taaz: once the core isn't changing the way it is now, yeah [22:27] <ajstudy> that is where we'd run into problems if we split plugins off... [22:27] <taaz> just my opinion... no reason not to have 0.2.42 before we hit 0.3 [22:27] <omega_> right [22:27] <omega_> I think that'll come after 0.3.0 though, because of the core changes left to go in [22:27] <taaz> problem is coordinating work in multiple places so a release doesnt have half-broken features along with new stuff [22:28] <omega_> exactly [22:28] <ajstudy> so 0.3.0 is to be the last in changes to the API ? [22:28] <omega_> no, but the last of the big ones for a while, I hope [22:28] <omega_> we should get some practice with the frequent releases on the plugin modules once we split them [22:28] <wtay> http://gstreamer.net/releases/0.2.0/ [22:28] <ajstudy> ok, because how well do you expect binary plugins built against 0.2.0 to last with 0.3.0? [22:29] <wtay> you can click through to the notes too [22:29] <omega_> so we know how to keep it in a releasable state all the time.... [22:29] Action: ajstudy takes a look [22:29] <taaz> ajstudy: that would be foolish... change what needs to be changed when it needs to be changed... [22:29] <wtay> should I list the new plugins in the notices too? [22:29] <omega_> yes [22:29] <ajstudy> taaz: you know me, i am foolish :) but we are going to have people writing binary plugins, which might not be updated with the core [22:29] <wtay> ok [22:29] <omega_> also, we need a longer list of people at the end now <g> [22:30] <wtay> I figured.. [22:30] <omega_> I expect we can start to have pseudo-binary-compatibility somewhere post 0.5.0 [22:30] <taaz> ajstudy: tough. kernel seems to handle it ok, i think this little project can too ;) [22:30] <omega_> taaz: kernel is a lot more mature than gst is [22:30] <wtay> I think we should hav a plugin counter :-) [22:30] <ajstudy> taaz: you've never tried the binary-only lucent modem drivers? ;) [22:30] <omega_> wtay: hehehehe [22:31] <omega_> wtay: and an app counter [22:31] <wtay> number of supported plugins: 000085 [22:31] <omega_> elements: 000173 [22:31] <ajstudy> wtay: i see 2 different names for 0.2.0 on there ;) [22:31] <wtay> yeah, provide a 6 digit number and stuff [22:31] <wtay> ajstudy: <g> [22:32] Action: taaz wonders about cvs2cl usage again... [22:33] leviathan (le...@p3...) joined #gstreamer. [22:33] <ajstudy> dammit, no time to figure out what to write in this README, commit it anyway for someone to fix it? [22:33] <omega_> which? [22:33] <omega_> do others and commit them in a group [22:33] <ajstudy> plugins/alsa/README [22:33] <omega_> with a log message listing the ones that need more TLC [22:33] <wtay> yo [22:33] <ajstudy> omega_: yeah, but i am meant to be going out to a friend's place about 5 min ago ;) [22:34] <omega_> ok, then start on another one quickly <g> [22:34] <omega_> ok, release-status in aforementioned directory [22:35] <omega_> oplock that file via this channel [22:35] steveb (st...@no...) left irc: [x]chat [22:35] <omega_> wtay: chmod -R g+w . [22:36] <wtay> ok, sec.. [22:36] <omega_> g+rwx that is [22:36] <omega_> as appropriate [22:37] <wtay> plugins added [22:37] <omega_> yeah, the list is almost as long as the entire 0.1.1 list <g> [22:39] <wtay> who wants his name added to the list? <g> [22:39] <omega_> check AUTHORS for a first pass [22:39] <omega_> then the roadmap [22:39] big_T (th...@fy...) joined #gstreamer. [22:39] <omega_> and/or cvs annotate of gst/ for core, vs. plugins [22:39] <omega_> yo [22:40] <omega_> slides are being incrementally updated [22:42] <omega_> automake 1.4-p3 is out, someone should test gstreamer against it [22:42] <omega_> * Deal with configure.ac as well as configure.in. [22:42] <ChiefHighwater> I think you should have some voice recognition software tonight broadcasting your talk onto irc [22:42] <omega_> might make things easier [22:42] <omega_> ChiefHighwater: I'll tape it, at least [22:43] <ajstudy> we gonna be able to listen in with gstmediaplay? ;) [22:43] <omega_> um, no, that'll be later [22:43] Action: ChiefHighwater knows what his job is tonight [22:43] <ChiefHighwater> zoom in/ zoom out/ pan [22:44] <omega_> I do hope to eventually be able to go to do a talk and have it broadcast live via my camera and laptop [22:44] <ChiefHighwater> "I wonder how far up his nose this lense can see?" [22:44] <omega_> but that's gonna require significant speed improvements to libdv [22:44] <ajstudy> yeah [22:44] <omega_> ChiefHighwater: do I need to put you in the back of the room? <g> [22:45] <ajstudy> not sure what to put in the readme, have just got the argument list at the moment [22:45] <ChiefHighwater> who me? [22:45] Action: ChiefHighwater puts on his halo [22:45] <ChiefHighwater> o8-] [22:45] <ajstudy> rofl [22:46] Action: ajstudy hands ChiefHighwater some cleaner for his tarnished halo ;) [22:46] <omega_> doh [22:47] <omega_> ok, the slides link is on the documentation page, please comment on it as I work on it [22:47] <omega_> like, propose slide titles and such [22:49] <ajstudy> omega_: what software you use for making these slides? [22:49] <omega_> see generate.pl and outline in the same dir [22:49] <ChiefHighwater> Stuff You Know, Stuff You Should Know, Stuff You Wish You Knew, If Only You Knew 8-] [22:49] <omega_> same I used to make the slides for GUADEC <g> [22:49] <omega_> ChiefHighwater: um [22:50] <ChiefHighwater> you were looking for something more...instructive? [22:50] <omega_> that was the idea, yeah... [22:52] <omega_> any ideas re slides? [22:52] <ajstudy> umm... [22:52] <omega_> like, just a sequence of slide titles would help a lot <g> [22:57] <ajstudy> ok, can i paste the tiny amount i wrote for the esdsink plugin? ;) [22:57] <omega_> just commit them all [22:58] <ajstudy> ok [22:58] <wtay> ok added the authors.. plz check [22:58] <omega_> spare <> at the bottom [22:58] <wtay> yup, free space for more :-) [22:59] <omega_> heheheh [22:59] <omega_> put a ... in there ;-) [22:59] <ajstudy> heh, i see my name is there ;) [22:59] <wtay> "you" (with a point... [truncated message content] |
From: <wim...@ch...> - 2001-06-09 04:29:53
|
[06:48] Nick change: ajstudy -> ajmitch [06:54] <ajmitch> hi all [07:00] Dekar (ar...@ip...) joined #gstreamer. [07:01] <ajmitch> hi Dekar [07:01] <Dekar> hello, ajmitch [07:03] <ajmitch> what's up? [07:04] walken (fo...@c1...) joined #gstreamer. [07:04] <Dekar> not much - just got back from the Portland Linux Users Group meeting - Erik just did a great talk :) [07:05] <ajmitch> ah, ok ;) [07:05] <ajmitch> dunno where everyone is now... ;) [07:05] <walken> yopyop [07:05] <ajmitch> hi walken [07:05] <Dekar> I think he just released 0.2.0, too [07:06] <Dekar> he was uploading tarballs over wavelan after the meeting... [07:06] <ajmitch> yes, it's released [07:35] Nick change: ajmitch -> ajfood [07:53] steveb (st...@no...) joined #gstreamer. [07:57] Nick change: ajfood -> ajmitch [07:58] <ajmitch> hi steveb [07:58] <steveb> hi [07:59] <ajmitch> haven't seen 0.2.0 in the news yet, what's happened? ;) [08:00] <Dekar> I'm pretty sure Erik was releasing it remotely over a poor net connection that kept dying... [08:00] <ajmitch> yes, but it's available from what i've seen [08:01] <Dekar> yeah [08:01] <ajmitch> http://sourceforge.net/projects/gstreamer [08:01] <Dekar> he might have decided to wait and announce it when he got home :) [08:01] <Dekar> he was thinking about that... [08:01] <ajmitch> the plan was for people to be in here at the time of release to let other know (eg freshmeat, etc) [08:01] <steveb> it is announced on the site [08:01] <ajmitch> yes [08:01] <ajmitch> haven [08:02] <ajmitch> haven't seen it on gnome lists yet ;) [08:03] <ajmitch> hmm, has grown from 500k for 0.0.9 to 2.4MB for 0.2.0 (tar.gz) [08:04] <steveb> that should shrink somewhat after the package reorp [08:04] <steveb> that should shrink somewhat after the package reorg [08:05] <ajmitch> yes [08:06] <ajmitch> download count on sourceforge is still 0 ;) [08:09] <johnix> ajmitch: I'm downloading it right now. [08:10] <ajmitch> yay, we have one victim!! ;) [08:10] <walken> lol [08:10] <johnix> What, is there a trojan inside? [08:10] Action: Dekar gives in and downloads a copy, to let it sit for a while [08:11] <ajmitch> no [08:11] <ajmitch> no trojan [08:11] <ajmitch> we just joke about users being victims in here (and in other places) ;) [08:11] <ajmitch> or volunteer coders being victims ;) [08:11] <Dekar> there. two copies. :) [08:12] <johnix> configuring... [08:12] <johnix> What's that xml-config he's complaining about? [08:12] <ajmitch> pardon? [08:13] <johnix> checking for xml-config... no [08:13] <johnix> configure: error: Couldn't find xml-config [08:13] <ajmitch> it wants xml-config? install libxml-dev (i think that is pkg name on debian) [08:13] <Dekar> I haven't ever built it, but you probably need libxml [08:13] <Dekar> yeah [08:13] <johnix> Ok. [08:13] <ajmitch> not just the main package but the devel package as well [08:18] <johnix> Yup, seems to be working no... [08:18] <johnix> err now [08:23] taaz (dlehn@66.37.66.32) left irc: Ping timeout for taaz[66.37.66.32] [08:23] Nick change: steveb -> steveb-away [08:24] omega_ (om...@om...) joined #gstreamer. [08:25] <johnix> yop [08:25] <omega_> yo [08:25] <walken> yo :) [08:25] <ajmitch> hey omega_ [08:25] <ajmitch> what happened to the irc party? ;) [08:25] taaz (dlehn@66.37.66.32) joined #gstreamer. [08:25] Action: omega_ will be very distracted spending the next 35min rapidly buying 7-9 round trip tickets to work for the next 4 mo [08:25] <ajmitch> i haven't seen 0.2.0 on /., freshmeat, or linuxtoday.com either ;) [08:25] <omega_> no, start submitting them [08:25] <ajmitch> plus Uraeus needs to annouce it the the gnome lists ;) [08:26] <omega_> had major tech problems with the projector, didn't use my laptop ;-( [08:26] <ajmitch> crap [08:26] <Dekar> hey omega_ :) [08:26] Action: Dekar was the brown-haired kid in the eazel t-shirt who kept bugging omega_ after his talk, btw [08:27] <ajmitch> hehe [08:27] Action: walken was the guy who wasnt there [08:27] <Dekar> walken: *grin* [08:27] Action: ajmitch couldn't swim that far in time [08:34] Nick change: taaz -> taazzzz [08:35] Nick change: big_Tzzz -> big_T [08:35] <ajmitch> hey big_T [08:35] <big_T> good morning [08:35] <ajmitch> evening ;) [08:35] <big_T> damn, acording to GWeather it's raining outside...:( [08:37] <big_T> ajmitch, is about 0840 here right now :) [08:38] <Dekar> night [08:38] <ajmitch> night Dekar [08:38] <Dekar> night ajmitch :) [08:38] Dekar (ar...@ip...) left irc: zzz [08:39] <ajmitch> what would 0.2.0 classify as, minor/major enhancements, or minor/major bugfixes? [08:40] <johnix> the configure script doesn't check for the berkeley DB libs [08:41] <johnix> And at compile time, it complains about not having "-ldb1" (I think that's it) [08:42] <ajmitch> in building what part? [08:44] <ajmitch> linking to libdb seems to be a result of what your computer reports (from a config script i guess), as db1 does not appear in the gstreamer source at all on here (including makefiles) [08:44] <ajmitch> i'll search for db2 or db3 [08:45] <omega_> grrr, is taazzzz asleep yet? [08:45] <ajmitch> probably ;) [08:46] <ajmitch> omega_: having problems, or you want debs? [08:46] <omega_> both, I had questions about his OLS plans [08:46] <ajmitch> omega_: what should 0.2.0 be classed as on FM? [08:46] <omega_> isn't it already? [08:46] <omega_> there's nothing that really fits ;-( [08:47] <ajmitch> hmm, major enhancements? ;) [08:47] Action: omega_ just bought 4 round trips to work at $103 each, vs. $189 normal fare [08:47] <ajmitch> or major bugfixes? [08:47] <ajmitch> nice price ;) [08:47] <omega_> oh, you mean the release type? [08:47] <omega_> um, dunno, you figure it out <g> [08:48] <ajmitch> yes, 'release focus' [08:48] <omega_> um [08:48] <ajmitch> ok, major bugfixes it is then, unless i hear further objections ;) [08:48] <omega_> major enhancements [08:48] <ajmitch> k [08:48] mwc (mal...@ab...) left irc: Ping timeout for mwc[abu.ntu.edu.au] [08:49] <walken> to me major bugfixes would only be appropriate for a brownbag release. or something. [08:49] <ajmitch> damn, how much of the changes to paste from the release notes? ;) [08:50] <ajmitch> i think i'll just paste the top paragraphs and leave off the list [08:50] <omega_> walken: talked with krasic some tonight, I wish he could make it to OLS, but he'll be in France on vacation (Tour de ...) [08:50] <omega_> we discussed vlc reading code.... [08:50] <walken> who's krasic ? [08:50] <omega_> Buck Krasic, coauthor of libdv [08:50] <walken> ok [08:50] <walken> did he had some other cool ideas ? [08:51] <omega_> not atm [08:51] <walken> ok [08:51] Action: walken had a day off [08:51] <omega_> he's got his head deep in his research, which doesn't cover that directly [08:51] <walken> picked my mom at the airport (she stays here 2 weeks) [08:51] Action: omega_ also found out the Kieth Packard lives in pdx.... [08:52] <walken> and wrote a long email to that other developper at work that annoys me so much [08:52] <omega_> heh [08:52] <walken> kieth ? whats he working on [08:52] <walken> I know that name only I dont remember what he does :) [08:52] <omega_> he does ati drivers for xfree [08:52] <omega_> works for suse [08:52] <walken> ok [08:52] <omega_> specifically, I want to know what he knows about getting idct+mcomp on the ati hardware to work, maybe [08:53] <walken> yes I think its on the X list that I see his name all the time [08:53] mwc (mal...@ab...) joined #gstreamer. [08:53] <omega_> yo [08:54] <ajmitch> omega_: hey, where are the tarballs on gstreamer.net? there are no links that i see (aprt from the release section on the SF page) [08:54] <omega_> the links are on the release page [08:54] <omega_> hrm [08:55] <omega_> ok, making it less convoluted [08:55] <ajmitch> cool [08:55] <omega_> um? [08:55] <omega_> someone removed the links [08:55] <ajmitch> yeah, that's why i [08:55] <ajmitch> am complaining ;) [08:55] <omega_> fixed [08:55] <ajmitch> which page? [08:55] <omega_> I managed to remove the space in A HREF somehow [08:55] <omega_> download.shtml [08:56] <omega_> will add links to he release page [08:56] <ajmitch> l [08:56] <ajmitch> k [08:56] Action: ajmitch hits keyboard [08:56] <ajmitch> ok, will submit to freshmeat now [08:59] <omega_> what lists should I send the announce to? [08:59] <ajmitch> umm, l-a-d, just to annoy them ;) [08:59] <omega_> yeah [08:59] <omega_> searching for address list from prev announces [09:00] <ajmitch> who should send to gnome lists? i see that stefan westerfield announced CSL 0.1.2 on a few lists [09:00] <omega_> hrm, sf is having issues with their shell server, it's slow [09:00] <ajmitch> probably cracked again ;) [09:00] <omega_> better not be [09:00] <johnix> yup, someone just came in to bother me. [09:01] <johnix> Here's what I get: [09:01] <ajmitch> yeah, i've logged in there today [09:01] <johnix> "/usr/bin/ld: cannot find -ldb1" [09:01] <ajmitch> omega_: the almighty grep tells me we do not link in this file on my computer [09:01] <omega_> um, hmm [09:02] <omega_> indicative of a gtk-config issue somehow [09:02] <ajmitch> johnix: where did this error occur? (paste a few lines above that) [09:02] <johnix> the line above that is a 12-line gcc command, with a bunch of "-I" and "-l" [09:02] <omega_> you sure you have db-devel? [09:02] <johnix> dunno [09:02] <johnix> I'm installing it, we'll see [09:05] <ajmitch> ok, you might see gstreamer announced on fm after the editors go over the announcement ;) [09:06] <ajmitch> where else? [09:07] <johnix> What? The Suse db-3.17 RPM I just downloaded depends on "cygwin1.dll" and kernel32.dll"! [09:07] <omega_> whoops [09:07] <johnix> Where does that come from? [09:07] <omega_> windoze [09:07] mwc (mal...@ab...) left irc: Ping timeout for mwc[abu.ntu.edu.au] [09:07] <omega_> somehow I think you grabbed the wrong rpm <g> [09:07] <johnix> Is that a trojan? ;) [09:07] <omega_> windoze? yeah. [09:08] <johnix> RPM for Windows? [09:08] <johnix> Is that a MS conspiracy??? [09:08] <omega_> it's a sacrilege, is what it is [09:08] <ajmitch> johnix: you need some sort of prepackaged exorcism kit there [09:09] Action: omega_ prefers a shotgun [09:09] <big_T> nay, crusifix and some holy water is what you need [09:10] <big_T> perhaps some explosives too [09:10] <johnix> apt-get install prayer-dev [09:10] <johnix> oops [09:10] <johnix> whong shell [09:10] <omega_> heheh [09:11] <johnix> Ok, I found what I needed. [09:11] <johnix> But I feel that this dependancy should be added to the configure script... [09:11] <omega_> it's not our dependency [09:11] <omega_> it's gtk's [09:11] <omega_> which is gtk-config's job to deal with, and for some reason it's not [09:11] <johnix> omega_: Ok... [09:11] <johnix> Maybe... [09:12] <omega_> it should fail the gtktest, though.... [09:12] <johnix> effectively, there was this package called "gnome-db" [09:12] <johnix> Does it have anything to do with it? [09:12] <ajmitch> nope [09:13] <ajmitch> gnome-db is totally different, iirc ;) [09:13] <omega_> oh, that's evil: http://ars.userfriendly.org/cartoons/?id=20010608 [09:14] <ajmitch> hehe [09:14] <big_T> what does iirc mean? [09:14] <omega_> if I remember correctly [09:15] <big_T> oh, okey... I've been wondering about that abrev for ages ;) [09:15] Topic changed on #gstreamer by ChanServ!s@ChanServ: GStreamer "Critical Mass" 0.2.0 released!: http://gstreamer.net/releases/0.2.0/ [09:16] <johnix> How close to production is it, btw? [09:17] <omega_> 0.2.0? about 0.8 away <G> [09:17] <ajmitch> lol [09:17] <ajmitch> how long has it taken to get this far? :) [09:17] <omega_> um, ~2yrs [09:17] <omega_> but that's very non-linear.... [09:17] <omega_> 1.0 should happen maybe even within this calendar year (no, not the Chinese one...) [09:18] <ajmitch> wow [09:18] <big_T> omega_, you have more developers and testers now than before so... [09:18] <ajmitch> i suppose, with people like me helping out.... ;) [09:18] <omega_> yup. just need to organize them all <g> [09:18] <ajmitch> what's planned for 0.3.0 so far? [09:18] <omega_> stuff [09:19] <big_T> has seeking been implemented yet? The ability to position a stream to an arbitrary position and play from that? [09:19] <omega_> that's one of the major ones in 0.2.x, yeah [09:19] <omega_> x>=1 [09:20] <johnix> omega_: If gstreamer is what my boss wants, then maybe I'll get to spend the next 3 months full-time on gstreamer... [09:20] <omega_> cool [09:20] <ajmitch> nice... ;) [09:20] <ajmitch> omega_: put 0.2.0 on news on the gstreamer SF page? [09:20] <omega_> oh, need to do that [09:21] <johnix> it's there [09:21] <johnix> already [09:21] <omega_> ajmitch: remind me of anything I need to do, and/or do it yourself [09:21] <ajmitch> heh ;) [09:21] <omega_> johnix: a news item? [09:21] <johnix> So I need to write some kind of report in the next 2 hours... [09:21] <ajmitch> just wasn't sure if i had rights to do that ;) [09:21] <johnix> omega_: Yup. [09:21] <omega_> yeah, on gstreamer.net [09:22] <omega_> he's talking about the sf project page [09:22] <johnix> omega_: oh, sorry. [09:22] <johnix> So, what's the structure of g-streamer? [09:23] <omega_> see the slides on the documentation page [09:23] <johnix> except for the fact that it accepts plugins [09:23] <johnix> omega_: [09:23] <johnix> ok [09:24] <johnix> hum... [09:24] <johnix> ...so you use windows, heh... [09:24] <omega_> I had to for those slides [09:25] <omega_> the guadec slides are more recent/correct [09:25] <omega_> and the PLUG slides were presented tonight [09:25] <omega_> er, last night, technically, now.... [09:26] <ajmitch> bother: [09:26] <ajmitch> Permission Denied [09:26] <ajmitch> You cannot submit news for a project unless you are an admin on that project [09:26] <ajmitch> omega_: it's up to you, i guess ;) [09:26] <omega_> ajmitch: where? [09:26] <omega_> I already did the sf news [09:26] <ajmitch> oh, ok [09:26] <ajmitch> i was just slow logging in ;) [09:26] <omega_> quite <g> [09:29] <johnix> A DVD player built with gstreamer would be somewhat slowed down by the need to copy data from filter to filter, wouldn't it? [09:29] <omega_> nope [09:29] <omega_> there's no copies unless you force it to do so by writing plugins wrong [09:29] <johnix> ok [09:30] <big_T> omega_, well that depends... some plugins might produce more/less data as output than it got as an input [09:30] <ajmitch> big_T: like a videoscaler? [09:30] <omega_> big_T: then it's gonna copy anyway [09:30] <omega_> by definition it's gonna traverse all the data and write it back out in a different form [09:31] <omega_> so then the copy (if it needs to be) is "free" [09:31] <ajmitch> ok, gotta go [09:31] <omega_> ajmitch: l8r [09:31] Nick change: ajmitch -> ajbusy [09:31] <johnix> oops, "make check" failed... [09:32] <big_T> omega_, yes you are right. [09:32] <omega_> hmm, doesn't surprise me [09:32] <richardb> Morning! [09:32] <omega_> richardb: yo [09:32] <big_T> morning richardb [09:32] <richardb> We finally have release, then. ;) [09:32] <ajbusy> morning richardb, cya richardb ;) [09:32] <omega_> richardb: it's not legal to be happy and bright and such in the morning, fyi [09:32] <richardb> omega_: you can't _hear_ me. [09:32] <omega_> hehehe [09:33] <richardb> So, do we have a plan for what to do next, in what order? [09:33] <omega_> I'm working on a couple of ridgerun-specific gstreamer projects [09:33] <omega_> first is gobject [09:33] <richardb> I have to get my new computer set up, and compiling [09:33] <richardb> Ah, right. [09:33] <omega_> then a massive attempt at slimming it down (core to 50KB) [09:34] <omega_> then some very work-related stuff (plugins and wrappers specific to the hardware we target) [09:35] <richardb> I should be more active doing actual development now, since I can actually compile in a sane amount of time... [09:36] <richardb> Speaking of which, having some trouble configuring X... :( [09:37] <omega_> whoops [09:37] Action: omega_ builds two copies of gstreamer simultaneously [09:39] <johnix> gstreamer-complete gives me a core dump on startup [09:39] <omega_> you're not supposed to run it alone [09:39] <omega_> and you have to have a completion registry available [09:39] <johnix> ok [09:39] <omega_> there's a man page in there somewhere [09:39] <omega_> man ./gstreamer-complete.1 [09:40] <johnix> ahaaa... [09:41] <richardb> Wow: answering queries by reference to the appropriate docs... Novel. [09:41] <omega_> neat, huh? [09:41] <richardb> It shouldn't core dump anyway, ofc. But hey. [09:41] <omega_> yeah [09:42] <omega_> eesh. I'm doing an analysis of the size of the core.... [09:42] <omega_> gst_pad_set_getregion_function with all debug and such turned off is 144 bytes [09:42] <omega_> that's insane [09:42] <omega_> gst_pad_set_getregion_function (GstPad *pad, [09:42] <omega_> GstPadGetRegionFunction getregion) [09:42] <omega_> { [09:42] <omega_> g_return_if_fail (pad != NULL); [09:42] <omega_> g_return_if_fail (GST_IS_REAL_PAD (pad)); [09:42] <omega_> GST_RPAD_GETREGIONFUNC(pad) = getregion; [09:42] <omega_> GST_DEBUG (GST_CAT_PADS,"getregionfunc for %s:%s set to %s\n", [09:42] <omega_> GST_DEBUG_PAD_NAME(pad),GST_DEBUG_FUNCPTR_NAME(getregion)); [09:42] <omega_> } [09:43] <richardb> Where does it go? [09:43] <omega_> go? [09:43] <richardb> What takes the space up? [09:43] <omega_> finding out [09:44] <big_T> the g_return_if_fail might not be the smallest macros [09:44] <omega_> that's turned off [09:44] <omega_> I thought [09:44] <omega_> um. somehow the debug stuff isn't turned off ;-( [09:45] <omega_> or it's not cooperating, because it's all here in the asm [09:45] <omega_> hrm. [09:45] <richardb> -DG_DISABLE_CHECKS ? [09:45] <omega_> yeah [09:45] <omega_> somehow #define GST_DEBUG_ENABLED 1 [09:45] <omega_> is still in the config.h [09:46] <omega_> turned it off and rebuilding, now tracing why it didn't get undefined [09:47] <omega_> hrm. need to rethink how that's all dealt with in configure.in [09:47] <omega_> ok, that's a lot better. 16 bytes [09:48] <omega_> down from 200KB to 113KB [09:48] <big_T> omega_, should I send the factory-less-alsa-stuff-patch to you or wtay-zZz? [09:48] <omega_> big_T: me, I guess [09:48] <big_T> okey [10:07] <walken> ok cya [10:07] walken (fo...@c1...) left irc: zzZzZzZZzzzzz [10:09] <big_T> omega_, sent the spam^H^H^H^Hpatch a second ago [10:09] <omega_> heh, got it [10:11] <big_T> ok [10:11] <omega_> I'll have to deal with it tomorrow, I'm crashing [10:12] <big_T> okey, no real rush :) [10:18] <omega_> wow. by trimming things like capsnego, autoplugger base object, and xml support, I can tim 30KB off of 110KB [10:18] <omega_> getting closer to 50KB.... [10:27] ShrimpX (marius@131.252.244.168) joined #gstreamer. [10:27] <omega_> yo [10:27] <ShrimpX> hey [10:27] <ShrimpX> is this the "official channel" for gstreamer? [10:27] <omega_> yup [10:28] <ShrimpX> you erik? [10:28] <omega_> yup [10:28] <ShrimpX> hey, good presentation tonight! [10:28] <omega_> not for long though. I change into a punpkin in a few minutes <g> [10:28] <ShrimpX> :) [10:28] <omega_> thanx <g> [10:28] <ShrimpX> a few of the people there were scratching their heads, but that's ok. :) [10:29] <omega_> a few people were nodding off <g> [10:29] <ShrimpX> hehe [10:29] <omega_> like me.... [10:29] <ShrimpX> hehe [10:30] <ShrimpX> gstreamer seems to be it, though. It's a good thing. Slashdotters all agree... [10:30] <omega_> heh [10:30] <omega_> er, is that possible? [10:30] <omega_> /. commenters never agree on anything... [10:30] <ShrimpX> haha [10:31] <ShrimpX> true. let's settle for "most of them" for now. [10:31] <omega_> heh [10:32] <ShrimpX> I use mplayer for a lot of stuff. It seems to have morphed into something useful... [10:32] <omega_> afaik it's still a very monolithic program, though [10:33] <ShrimpX> yes it is. [10:34] <ShrimpX> I'm excited about an mp3 player based off gstreamer. I'm sick of xmms and 550Mhz not being enough for it to run... [10:34] <omega_> uh? [10:35] <ShrimpX> you can use the gstreamer core to write an mp3 player, right? [10:35] <omega_> you can use the plugins to decode mp3's, yes [10:36] <ShrimpX> cool [10:36] <omega_> simple pipeline is: [10:36] <omega_> ./gstreamer-launch disksrc location=music.mp3 ! mad ! osssink [10:36] <ShrimpX> aah... cool [10:37] <omega_> of course, as we've found out, there are various problems with most of the libraries we link plugins against [10:37] <ShrimpX> heh. go figure. [10:37] <omega_> basic problem is that the majority of these people just don't know how to write a decent API [10:38] <ShrimpX> that's the "downside" of Linux. Anyone can just hack stuff together and people accept it as standard... [10:38] <omega_> yup [10:39] <ShrimpX> I just ran 'make' on gstreamer. [10:40] <omega_> how long did it take, or is it still running? <g> [10:40] <ShrimpX> still running [10:40] <omega_> ah, then I have time to sleep before it's ready for you to use <g> [10:40] <ShrimpX> hehe [10:41] <ShrimpX> it's not that bad, it is? [10:41] <omega_> should take about 20min if you have a decent group of the dependant libraries installed [10:41] <ShrimpX> ouch [10:41] <omega_> but there's a really long list of those, so chances are it'll take about 15min [10:41] <omega_> takes about 23.5min on my p3-500, so.. [10:42] <ShrimpX> I got an AMD k6-2/550. It tends to suck ass compared to any p3. [10:42] <omega_> oh. then at least 30min ;-( [10:42] <ShrimpX> damn [10:43] <omega_> there's a lot of code there... [10:43] <ShrimpX> yes there is. I browsed through it a bit... [10:43] <big_T> wc `find -name "*.[ch]"` -l [10:43] <big_T> the plugin-dir is about 130k lines of code [10:43] <ShrimpX> hehe [10:44] <omega_> I get about 210K lines total [10:44] <omega_> for everything [10:44] <big_T> yes, me too :) [10:45] <ShrimpX> hehe [10:45] <big_T> nothing compared to the kernel though [10:45] <ShrimpX> I get 540588 for absolutely everything in the directory [10:45] <omega_> heh [10:46] <big_T> yes, but ten you have a lot of makefiles and such which doesn't count [10:46] <ShrimpX> yup [10:46] <ShrimpX> and docs, etc. [10:47] <omega_> we've got a ways to go to catch the kernel, at about 2.8M lines, but we can make it <g> [10:48] <big_T> more plugins is all we need... about 100more should do it :) [10:48] <ShrimpX> how much of that is drivers though? [10:48] <omega_> plugins? 130K lines of that out of 210 [10:48] <ShrimpX> no, kernel drivers. they make up a lot of it [10:49] <omega_> oh, the kernel? yeah, most of it [10:49] <big_T> yes, but it is more likely that more plugins will be added than more core-stuff [10:49] <ShrimpX> yup [10:49] <ShrimpX> that just means that it was built right :) [10:49] <omega_> I get about 234K lines of init, kernel, lib, mm, and net [10:50] <ShrimpX> heh, that's not a lot [10:50] <omega_> 44K lines of x86 stuff [10:50] <omega_> and lots of that net stuff is not usually compiled, either [10:50] <omega_> 31.6K without net/ [10:51] <big_T> yeah, I created a small kernel as a pet-project in school - and it got to around 16k lines with memoryprotection, multitasking, filesystem, and some drivers... [10:51] <omega_> about 73K lines out of net in your average kernel [10:52] <ShrimpX> cool, I'll get to play with the Linux kernel in school... [10:52] <omega_> so about 150K lines in the basic kernel [10:52] <omega_> less than gstreamer ;-) [10:52] <ShrimpX> hehe [10:52] <big_T> haha [10:53] <omega_> except my current project for work is to get the gstreamer core down to about 50KB [10:53] <omega_> I'm theoretically down to about 80KB so far, by mentally trimming stuff [10:53] <ShrimpX> oh, that's right... for embedded stuff [10:53] <omega_> yeah [10:53] <omega_> gotta get really small [10:53] <omega_> and the trick is that glib/gobject isn't small either... [10:53] <omega_> which is one reason I'd like to see gtk used for embedded guis, if it can be stripped of atk and pango and most of its bulk [10:54] <ShrimpX> I noticed that gstreamer is already a feature on their web site [10:54] <omega_> can use glib/gobject for more than just one package that way [10:54] thomas (th...@21...) joined #gstreamer. [10:54] <omega_> ShrimpX: yup, on the office page [10:54] <ShrimpX> yep [10:54] <omega_> or at least it was, searching the page again [10:55] <omega_> hrm, not there anymore ;-( [10:55] <omega_> it's gonna have to be though, unless it's no longer part of the 'office' stuff and more in the core [10:55] <thomas> hi [10:55] <omega_> yo [10:55] <thomas> omega_: you still up ? [10:56] <ShrimpX> it's on the products page, in the open source components for DSPLinux [10:56] <omega_> barely. need to sleep ASAP [10:56] <omega_> huh? gnome or ridgerun ? [10:56] <ShrimpX> gstreamer [10:56] <omega_> hmm? [10:56] <omega_> you lost me <g> [10:56] <omega_> what url are you looking at? [10:57] <ShrimpX> http://www.ridgerun.com/products/ [10:57] <omega_> ok [10:57] <thomas> omega_: I have a simple question re: that cothread limit issue and cothreads not being cleaned up/rescheduled/whatever ... [10:57] <ShrimpX> oops we were talking about different things. sorry... [10:57] <thomas> ... since I need that for the mixer as well, when is that scheduled to be handled, and can I help ? [10:58] <omega_> thomas: um, soon probably [10:58] <omega_> if you want to look into how to deal with that, sure, it's all in cothreads.[ch] [10:58] <omega_> the key is to do lazy cothread destruction [10:58] <thomas> omega_: I'm new to cothreads, so I'll have to learn a lot about it first ;) [10:58] Nick change: ajbusy -> ajmitch [10:58] <thomas> but since it doesn't work, I cannot break it, right ? [10:58] <omega_> on cothread destruction, set a bit that that 'slot' is free. when trying to allocate a new cothread, if it's going to reclaim that slot, it must acquire the lock first [10:59] <omega_> that keeps the thread from being reused out from underneath the original cothread before it even finishes running [10:59] <omega_> thomas: um, <g> [11:01] <ajmitch> hi again [11:02] <ajmitch> had many visitors lately? :) [11:02] <omega_> whoah [11:02] <omega_> just with compile options I got the *stock* libgst.la down to 71KB [11:02] <ajmitch> wow [11:02] <omega_> turn everything off use -Os [11:02] <ajmitch> what did you start with? [11:03] <omega_> um, checking [11:03] <omega_> unstripped 851KB, stripped 210KB [11:04] <omega_> down to stripped 70KB [11:04] <omega_> with the modifications I've already started listing, I can trivially get it to 50KB [11:04] <omega_> very cool [11:04] <ajmitch> wow, that's a massive different [11:05] <omega_> autoplug 2K, xml 2K, typefind 1.5K, most of gst.o 2.5K [11:05] <omega_> there's down to ~61KB already [11:05] Action: ajmitch wonders what has happened with SF [11:05] <ajmitch> you can still visit the page? [11:05] <omega_> they're sucking again? [11:05] <omega_> yup, no prob [11:05] <ajmitch> yeah, can't get to it with webbrowser [11:05] <omega_> at least you're using the right tool ;-) [11:06] <ShrimpX> I can't get to it either... [11:07] <ajmitch> hmm traceroute works for me tho ;) [11:07] <ajmitch> was going to close a bug that i reported (crashing due to no Xv) [11:12] _gst_newt_ joined #gstreamer. [11:12] <ShrimpX> yup [11:12] <ajmitch> not sure if the download.sourceforge.net server is still up tho :( [11:13] <ShrimpX> looks like it is... [11:14] <ajmitch> gstreamer.net is also hosted on sourceforge, so could be only the main server that's down [11:14] <ShrimpX> yea, probably [11:15] <ajmitch> so where did you find out about gstreamer? at the talk? ;) [11:15] <ShrimpX> at the PLUG meeting tonight. [11:16] <ajmitch> heh, cool [11:16] <ajmitch> it was a good talk? [11:16] <ShrimpX> yea. It seemed that most people were pretty confused when it got a little more tachnical. :) [11:16] <ajmitch> hehe [11:16] <ShrimpX> but I enjoyed it. [11:16] <ajmitch> i am still confused about somethings [11:17] <ShrimpX> are developing for it? [11:17] <ajmitch> a little ;) [11:17] <ajmitch> only minor things at the moment [11:17] <ShrimpX> I wish I could. :) [11:17] <ShrimpX> I should be able to, soon [11:17] <ajmitch> cool [11:18] <ajmitch> what are you planning to do? [11:18] <ShrimpX> notthing for now. Learning how to program in general :) (I'm a freshman in college - cs major) [11:18] <ajmitch> ah, sounds like me ;) [11:18] <ShrimpX> hehe cool. [11:18] <ajmitch> first year at uni here in NZ, cs major [11:19] <ShrimpX> cool [11:19] <ShrimpX> what language are you starting with? [11:19] <big_T> ajmitch, same here, although I've done alot of programming before [11:19] <ajmitch> we learn java (vomit) at uni ;) [11:19] <ShrimpX> haha [11:19] <ajmitch> but i've used C before [11:19] <ShrimpX> cool [11:19] <ajmitch> and am using python for a job in a week or so ;) [11:20] <ShrimpX> we're doing c++, and always reference back to c. It's cool [11:20] <big_T> ajmitch, personally ML and Prolog makes me sick... although nothing really beats Prolog and suckiness [11:20] <ShrimpX> I program Perl for a living. (Web/database stuff) [11:20] <ajmitch> at uni here they used to use pascal [11:20] <ajmitch> ShrimpX: cool ;) [11:20] <ShrimpX> it's ok. It's really slow now, with the economy biting the dust... [11:21] <ajmitch> yeah... [11:21] <ajmitch> i was lucky to be contacted a couple of days ago with an offer of a small job [11:21] <ajmitch> cool, get connection refused from SF now, maybe it's restarting ;) [11:21] <ShrimpX> I got started as an HTML temp. And I just learned from there, and was promoted... [11:21] <ShrimpX> hehe [11:22] <ajmitch> nice... [11:23] <ShrimpX> Pyton is cool. The people at my work hate it though, and refuse to let me use it. I like it. [11:23] <big_T> haven't seen much Python, but heard that it was pretty cool and nifty [11:23] <ShrimpX> s/Pyton/Python/ [11:23] <ShrimpX> it's got the whitespace thing that people hate, but it's sweet otherwise [11:24] <ajmitch> yes, the whitespace thing takes a little getting used to ;) [11:24] <big_T> in sweden we have an expression which goes like "It tastes like pyton" meaning it tastes really really bad :) [11:24] <ShrimpX> hahaha!! [11:24] <ShrimpX> I tastes like Pascal. [11:24] <big_T> got nothing with the language to do though, but with the python-snake(it's called that in english?) [11:24] <big_T> i think anyway ;) [11:25] <ShrimpX> yep [11:25] <ShrimpX> it's funny though, geek-wise. [11:25] Action: ShrimpX has been compiling gstreamer for 40+ minutes... [11:25] <big_T> yes [11:25] <ajmitch> ShrimpX: what speed computer? ;) [11:26] <ShrimpX> AMD k6-2/550 [11:26] <ShrimpX> k6-2's suck ass equally at any speed [11:26] <ajmitch> not a bad machine, my k6-2/400 takes awhile to compile stuff [11:26] <ShrimpX> hehe [11:26] <ajmitch> yeah, would have to agree ;) [11:26] <ShrimpX> I got an Athlon too, but it's in pieces on my floor [11:26] <big_T> K7 is a pretty good architecture - not bad at all [11:27] <ajmitch> how much ram in that k6-2? [11:27] <ShrimpX> k7/athlon rocks [11:27] <ShrimpX> 192MB [11:27] <ajmitch> hmm, should be ok for compiling i guess ;) [11:27] <ShrimpX> ah, just finished :) [11:27] <ajmitch> damned omega compiles gstreamin in ~15min [11:28] <ShrimpX> that's what I hear. on a p3/500. How?? [11:28] <ajmitch> k6-2 is crap ;) [11:28] <ajmitch> maybe doesn't compile all the plugins, i dunno [11:28] <ShrimpX> he modified the compiler to optimize for gstreamer :) [11:28] <big_T> haha [11:29] <ajmitch> i got damned impatient & added stuff to configure so that examples & tests weren't built ;) [11:29] <ShrimpX> hehe [11:29] <big_T> hes got it all tared up and just does tar xf gstreamer-fake-compile.tar ;) [11:29] <ShrimpX> haha [11:29] <ShrimpX> that would seem more reasonable [11:29] <big_T> yeppers [11:30] <ajmitch> big_T: i added an option on here to let people not build plugins, too ;) [11:33] <big_T> ajmitch, that's nice... perhaps should have a 'make menuconfig'... more and more options are added :) [11:33] <ajmitch> lol [11:35] Action: ajmitch watches his cpu overheat as he compiles gstreamer - with compilation process reniced to -20 [11:35] Action: ShrimpX is cooking breakfast on his CPU [11:35] Action: johnix is leaving for the week-end [11:35] johnix (thomas@61.152.140.57) left irc: Client Exiting [11:36] <ShrimpX> I have to work in 4 hours and I haven't slept yet.... [11:36] <ShrimpX> I have to stop doing this... [11:36] <ShrimpX> ls [11:36] <ShrimpX> oops [11:36] <ajmitch> hehe [11:37] Action: ajmitch hands ShrimpX some coffee [11:38] Action: ajmitch decides it is not best to stay up hacking all night before an exam ;) [11:38] <ShrimpX> hehe what class? [11:38] <ajmitch> physics [11:39] <ShrimpX> sorry [11:39] <ajmitch> ;) [11:39] <ShrimpX> I'm taking physics in the fall [11:39] <ajmitch> ah, cool [11:39] <ajmitch> where you studying? [11:40] <ShrimpX> PSU (portland state university) [11:40] <ajmitch> k [11:40] <ShrimpX> I like the CS program here. [11:40] Action: ajmitch is at University of Otago [11:40] <ShrimpX> cool [11:41] <ShrimpX> we use unix/vi/gdb for everything here. :) no useless GUI stuff [11:41] <ajmitch> hehe [11:41] <ajmitch> lucky sods, we'll probably have to use some disgusting windows IDE for java [11:41] <ShrimpX> that sucks. we are taught to hate everything that bears the microsoft label. It's nice. :) [11:42] <ajmitch> ah, a uni with intelligent life within ;) [11:43] <ShrimpX> yep. That's why I chose it... [11:56] johnix (thomas@61.152.140.57) joined #gstreamer. [11:56] johnix (thomas@61.152.140.57) left irc: Client Exiting [11:59] <ajmitch> ok, just about finished build... [12:02] <ajmitch> not bad, 33min on my machine from autogen.sh to finishing the compile... [12:09] <ShrimpX> yea, mine was probably around there, now that I think about it. [12:13] <ajmitch> i used 'time' to check it ;) [12:13] <ajmitch> using gcc 3.0 snapshot, disable tests & examples [12:14] <ajmitch> so compiling the whole thing would no doubt take a bit longer ;) [12:14] <ajmitch> does it run ok for you? [12:22] Action: ajmitch will sleep now... [12:22] Nick change: ajmitch -> ajzzzz [12:33] Nick change: ShrimpX -> ShrimpZzZ [14:04] big_T (th...@fy...) got netsplit. [14:04] big_T (th...@fy...) returned to #gstreamer. [14:18] dobey (do...@dr...) joined #gstreamer. [14:51] steveb-away (st...@no...) left irc: Ping timeout for steveb-away[node1ee4c.a2000.nl] [15:21] taazzzz (dlehn@66.37.66.32) left irc: Ping timeout for taazzzz[66.37.66.32] [15:23] taazzzz (dlehn@66.37.66.32) joined #gstreamer. [15:37] greg_ (gr...@ho...) joined #gstreamer. [15:37] <greg_> hi all [16:04] ajzzzz (aj...@p2...) left irc: Ping timeout for ajzzzz[p2-max5.dun.ihug.co.nz] [16:05] greg_ (gr...@ho...) left irc: Ping timeout for greg_[home.sente.pl] [16:18] <richardb> [16:49] <richardb> Woohoo. Gst compiles in 7 minutes on my new machine. :) [16:49] <richardb> I feel like I can be a developer again. [16:49] Action: richardb smiles [17:16] vektor (ve...@cr...) joined #gstreamer. [17:53] dobey (do...@dr...) left irc: Ping timeout for dobey[dreadnought.ximian.com] [17:56] thomas (th...@21...) left irc: [x]chat [18:13] Uraeus (csc...@c2...) joined #gstreamer. [18:13] <Uraeus> hello [18:13] Nick change: wtay-zZz -> wtay [18:13] <wtay> hi guys [18:14] <Uraeus> hi wtay [18:16] <Uraeus> wtay: where have we sent a release announcement? [18:16] <wtay> fm only AFAIK [18:16] <wtay> gnotices has one too [18:18] <Uraeus> no, not to gnotices I think [18:18] Action: Uraeus prepares for some marketing action [18:19] <wtay> well, the recent software section on gnome.org [18:19] <wtay> hehe [18:19] <wtay> Uraeus: you compile galeon from cvs? [18:19] <Uraeus> wtay: no recently no [18:20] <wtay> never? [18:20] <Uraeus> wtay: the software section of gnome.org was my doing, it did it at work today :) [18:20] <wtay> oh, ok :) [18:20] <Uraeus> wtay: I have compile dit earlier yes [18:20] <wtay> with mozilla cvs too? [18:21] <Uraeus> no never mozilla, always used Blizzards RPMS [18:21] <wtay> ok [18:21] <Uraeus> or the tar binary from mozilla.org [18:21] <wtay> I'm going to attempt a galeoncvs and moz-cvs combination then... [18:21] <Uraeus> wtay: but I have a machine running Solaris right next to me now, so I can from here on test gstreamer compilation issues for us [18:21] <wtay> cool [18:22] <Uraeus> kinda slow (pentium 200) so I guess it will take a while but still :) [18:22] <Uraeus> wtay: just let me know if you need access and I can set up an account for you [18:23] <wtay> o, I probably will when I get this galeon stuff sorted out [18:26] matrix (pro...@ho...) joined #gstreamer. [18:26] <matrix> hi wtay [18:26] <wtay> hi [18:27] <matrix> i am Zeenix [18:28] <matrix> glad to know about the new release [18:28] <wtay> yeah:) [18:28] <matrix> wtay: that gsm path was for the tar ball, i have the rpm [18:29] <matrix> sorry patch not path [18:29] <wtay> Mozilla/5.0 (X11; U; Linux 2.4.5 i686; en-US; Galeon; Debian/GNU) Gecko/20010606 [18:29] <wtay> whoohoo! [18:30] <matrix> wtay: what ? [18:30] <wtay> matrix: yup, you need the recompile from source [18:30] <wtay> galeon-cvs with mozilla-cvs :) [18:30] <matrix> wtay: you mean i should get the tarball [18:30] <wtay> yes [18:31] <matrix> wtay: should i search for it or you already know where to get it, i found the rpm on rpmfind.net [18:34] <wtay> http://members.core.com/~stabro/gsm_mmx.html [18:34] <wtay> http://kbs.cs.tu-berlin.de/~jutta/toast.html [18:36] <matrix> wtay: did you tried the new gstreamer ? [18:36] <Uraeus> wtay: can I add information about Pioneer sponsoring some development in the Gnotice announcement? [18:36] <wtay> matrix: try? I wrote parts of it! <g> [18:37] <wtay> Uraeus: no, don't bother... [18:37] <matrix> wtay: sorry for underestimatinge you [18:38] <wtay> matrix: just kidding, I did about a weeks testing on the tarballs [18:39] <matrix> wtay: is the gstreamer-manual updated or not ? [18:39] <Uraeus> hmm, strange that omega is allowed to use his temple baptis mail address [18:39] <wtay> matrix: not really 100% up to date, I hope to do some more work on that soon [18:39] <wtay> matrix: the API docs are ok [18:40] <matrix> wtay: are you the author of that manual ? [18:40] <wtay> matrix: yes [18:40] <Uraeus> wtay: when we in the announcement say that the media player is relativly stable, are we refering to the command line one or the GUI one? [18:41] <wtay> Uraeus: gui one [18:41] <matrix> wtay: will it be updated completelly after two weeks [18:41] <matrix> ? [18:41] <wtay> matrix: most likely not :( [18:42] <wtay> Uraeus: another way of saying this is "the media player is relatively unstable again" [18:44] <richardb> Hi guys. [18:45] <richardb> Anyone know if taaz was satisfied with the state of the debian packaging stuff. [18:45] Action: richardb tries to make debs on new machine [18:45] <matrix> wtay: where are the rpms for the gstreamer ? [18:45] <wtay> hi richardb [18:45] <wtay> matrix: not made yet [18:49] <matrix> wtay: i want to learn about the basic object model in C without learning about gtk or gstreamer first ? [18:51] <wtay> matrix: there is no object model in C [18:52] <matrix> wtay: i mean how to do object oriented programming in C [18:52] <wtay> matrix: that's what the gtk object model is for.. [18:53] <matrix> wtay: i know, but how would i define my own classes & inherit those... things like that [18:53] <wtay> matrix: in C? [18:54] <matrix> wtay: there is no classes in C nor is any type of inheritence possible [18:55] <wtay> matrix: not in the language, no. You'll need some standards to do inheritance (with structs and stuff) [18:55] <Uraeus> wtay: does glib allow some sort of inheritance? [18:55] <wtay> nope [18:56] <wtay> glib2.0 has gobject which provides an object framework [18:56] <Uraeus> wtay is my intro ok?: http://news.gnome.org/992019219/index_html [18:56] <matrix> wtay: can you give me some overall idea about it [18:57] <wtay> reading... [18:58] <wtay> Uraeus: very good [18:58] <wtay> matrix: about gtk objects? [18:59] <matrix> wtay: yes, classes & inheritence [18:59] ChiefHighwater (pa...@te...) joined #gstreamer. [18:59] <wtay> yo [18:59] <wtay> matrix: gtk objects are implemented using plain C structs [18:59] <ChiefHighwater> ello [19:00] <wtay> matrix: there is a struct for the class instance and the object instance [19:00] <wtay> matrix: inheritance is done by creating a new struct with the first field, the class/object you want to inherit [19:01] <wtay> matrix: gtk objects also provide a standard way of setting properties of the object and attaching signals to it [19:01] <wtay> matrix: gtk objects also provide a nice way to do introspection on them (a bit like in java) [19:02] steveb (st...@no...) joined #gstreamer. [19:02] <matrix> wtay: seperate structs for class & its objects ? whats the big idea, how do they connect [19:03] <wtay> matrix: there is a type system and the object instance has a pointer to the class struct [19:03] <matrix> wtay: type system ? [19:04] <wtay> matrix: you can register a new object type. At runtime you can check for the type of an object [19:04] <Uraeus> wtay: noticed that we are now in the Sourceforge top 100 list [19:05] <wtay> Uraeus: yeah, pretty cool :) [19:06] <Uraeus> wtay: I think we must have a little campaing to get people to rank you and omega so that your names appear on the Sourceforge frontpage [19:06] <wtay> Uraeus: uhm... [19:07] <Uraeus> wtay: 10 people ranking you are enough to get you onto the frontpage :) [19:07] <wtay> Uraeus: is that so? [19:09] <Uraeus> wtay: I did some research the other day and saw that some of the people there today doesn't have more people ranking them [19:09] <Uraeus> wtay: I am updating the roadmap, avidecoder -> green, what is the encoder situation still red? [19:10] <wtay> Uraeus: yes [19:10] <Uraeus> wtay: jpeg was fixed? [19:10] <wtay> Uraeus: more or less... it works on more formats now [19:11] <Uraeus> dec or enc or both? [19:11] <wtay> oly the decoder was tested enough [19:12] <Uraeus> anything else? [19:12] <wtay> that's it for me :) [19:13] <Uraeus> :) [19:13] <wtay> for now <g> [19:14] <Uraeus> wtay: I am moving the roadmap to gstreamer.net, I also renaming it a status-report and remove the future colums from the table, I came to the conclusion that they don't really mean anything [19:17] <Uraeus> I mean it is not like it is a big suprise that we plan on making everything green :) [19:18] <wtay> heh, yeah [19:22] <Uraeus> bah, the Solaris X system doesn't handle the mouse being discontected for a while, arrgh [19:22] <wtay> duh [19:24] <matrix> wtay: does gstplay now play AVI movies [19:24] <wtay> yes [19:24] <Uraeus> matrix: but you need windows dll files [19:25] <matrix> which dlls [19:26] <Uraeus> wtay: do you have a link to the dll's we support? [19:26] <wtay> Uraeus: I think they are still on gst.net [19:27] <Uraeus> wtay: matrix wants em :) [19:27] <Uraeus> wtay: can you give me admin rights on the GStreamer bugzilla? [19:27] <wtay> Uraeus: ? [19:27] <Uraeus> wtay: I don't remember the URL for the dll's [19:28] <wtay> gstreamer.net/binaries.tar.gz [19:29] <Uraeus> wtay: not that it is a urgent need with the number of bugs we have, but being able to triange and close bugs etc. would be nice, also I could assign them to people [19:30] matrix (pro...@ho...) left irc: Ping timeout for matrix[host-67.netzone.net.pk] [19:30] <wtay> Uraeus: can I do that? sec.. [19:30] <wtay> hehe, I can now use galeon to log into SF :) [19:32] Zeenix (pro...@ho...) joined #gstreamer. [19:33] <Uraeus> Zeenix: did you get the URL for the dll files? [19:33] <wtay> Uraeus: updated a lot of stuff... [19:34] <Zeenix> no [19:34] <Uraeus> wtay: seems to work, I can now modify the bugs, thanks [19:34] <Uraeus> Zeenix: gstreamer.net/binaries.tar.gz [19:35] <wtay> Uraeus: you can also admin the feature reqs, support reqs and patches :) [19:36] <Zeenix> Uraeus: do they work on the prev. release of gstreamer [19:39] <Zeenix> ? [19:39] <Uraeus> Zeenix: think so [19:39] <Uraeus> Zeenix: you need to put them in /usr/lib/win32 [19:41] <Zeenix> Uraeus: you mean i should uncompress it to this dir [19:42] <Uraeus> yes [19:54] Nick change: Uraeus -> Ura_away [19:58] <richardb> Was AM_DISABLE_STATIC meant to be commented out in the configure scripts in the release? [19:59] <wtay> I think so [19:59] <richardb> It wasn't. :( [20:00] <richardb> deb builds mutter about it, if I'm understanding the muttering correctly. [20:00] <wtay> oops [20:00] <richardb> Ho hum. [20:01] <richardb> We should make a "release target" in the top level Makefile.am to build a release, so we get all this stuff right nexttime. [20:01] <richardb> Or something like that. [20:02] <wtay> yup [20:03] thomas (th...@ad...) joined #gstreamer. [20:03] <richardb> Hmm, perhaps a release parameter in autogen.sh would be the way. [20:03] <thomas> hi people [20:03] Action: richardb ponders [20:03] <richardb> hi [20:04] vektor (ve...@cr...) left irc: Ping timeout for vektor[cr823975-a.ktchnr1.on.wave.home.com] [20:05] <wtay> hi [20:05] <thomas> ok, anyone have a spare machine lying around on which they want to test my gstreamer toolbox installer ? [20:05] <thomas> <evil grin> [20:07] vektor (ve...@cr...) joined #gstreamer. [20:09] <thomas> no takers then ? ;( ah, I shouldn't preach to the converted ... [20:10] <richardb> Sorry, I've got everything in the world installed anyway... [20:10] <wtay> I might want to give it a try soon... [20:10] <thomas> richardb: oh, so you're porting gstreamer to cobol then ? ;) [20:11] <thomas> well, it still needs some work anyway [20:11] <thomas> wtay: what library is needed for gsm ? I can add that one as well then [20:11] <wtay> libgsm [20:11] <thomas> duh ;) [20:11] <thomas> ok, I'll put that in [20:13] vektor (ve...@cr...) left irc: Ping timeout for vektor[cr823975-a.ktchnr1.on.wave.home.com] [20:13] vektor (ve...@cr...) joined #gstreamer. [20:16] Nick change: Ura_away -> Uraeus [20:25] Zeenix (pro...@ho...) left irc: Ping timeout for Zeenix[host-67.netzone.net.pk] [20:26] Nick change: wtay -> wtay-afk [20:33] <Uraeus> richardb: are you around? [20:34] <richardb> Yup. [20:34] <richardb> I see you're assigning me bugs. [20:35] <richardb> ;( [20:35] <Uraeus> hehe :) [20:36] <Uraeus> richardb: you are the right address for build system fixes right? [20:38] <richardb> Yeah, probably. [20:38] <Uraeus> richardb: well you got away cheaply I think, only one bug assigned to you :) [20:39] <Uraeus> richardb: you get a mail when I assign you a bug? [20:48] Nick change: wtay-afk -> wtay [20:49] <Uraeus> hi wtay [20:50] <wtay> bleagh, a bug assigned to me :( [20:50] <Uraeus> what? how could that be :) [20:51] <wtay> yes, Uraeus, I suspect *you* :) [20:51] <wtay> oh, only a minor bug... [20:51] Action: wtay is trying to use avifile for the windows stuff.. [21:02] <Uraeus> wtay: there is a request for Sorenson reverese-engineering on Gnotices, should I assign it to you ass a wishlist bug? <g> [21:02] <wtay> reverse engeneering?, sounds fun :-) [21:02] <wtay> ooh, you put gst on gnotices :) [21:03] <Uraeus> yup, and jrb from redhat saw it and said he want to port the Nautilus music view to gstreamer :) [21:03] <wtay> cool [21:04] <wtay> we really should use bonobo-media for that [21:05] <Uraeus> wtay: which reminds me, I mailed with Michael Meeks the other day and he said that he will make sure ORBit2 has the multithreading support we need [21:05] <wtay> thanks :) [21:05] Nick change: hds-afk -> hadess [21:05] <hadess> dudes ;) [21:05] <wtay> dude [21:05] <hadess> congrats for the release [21:05] <hadess> yeah [21:06] <wtay> :-) [21:06] <Uraeus> hi hadess [21:06] <hadess> heya Uraeus [21:06] <wtay> hadess: hehe, yeah :) [21:06] <Uraeus> hadess: guess which windowmanager I have started using :) [21:07] <hadess> Uraeus: wmaker =) [21:09] <Uraeus> hadess: so when will you make output plugin which sends video to a WMaker Dockapp? [21:09] <hadess> haha [21:09] <Uraeus> hadess: I am using wmusic btw [21:09] <hadess> it should actually be possible with the xvideosink [21:09] <hadess> wmusic will use rhythmbox at some point [21:10] <Uraeus> hadess: guess if you modify the xscreensaver embeding dockapp you could put GStremer in there instead [21:11] <hadess> nah, there's no need of embedding another process [21:11] <hadess> just make a dockapp mediaplayer, lol [21:11] <Uraeus> heh [21:12] <Uraeus> hadess, ok so you are going to put video and a full set of controls into 64x64 square :) [21:12] <hadess> no, i'm not, you crazy lad [21:13] Action: Uraeus is adding a wishlist bug and assigning it to hadess [21:14] <Uraeus> hadess: actually if you use the wmusic dockapp as a start, the area above the controls looks like it is a perfect fit for widescreen movies :) [21:15] <hadess> *smack* [21:16] <Uraeus> I take that as a yes, cool [21:16] omega_ (om...@om...) joined #gstreamer. [21:17] <hadess> hey omega_ [21:17] <omega_> yo [21:17] <Uraeus> hi omega_ [21:17] <wtay> yo [21:17] hadess (ha...@pc...) left irc: Read error to hadess[pc2-guil2-0-cust150.gui.cable.ntl.com]: EOF from client [21:18] hadess (ha...@pc...) joined #gstreamer. [21:19] <hadess> damn x crashed [21:19] <Uraeus> hehe, the Sun GNOME installer has a GPL license screen with a Accept button at the bottom, thats a first :) [21:20] <omega_> no, I've had to Accept the GPL before for other stuff [21:27] Action: Uraeus is testing his new SPEC before upload [21:28] Kuroyi (rick@65.35.82.223) left irc: Read error to Kuroyi[65.35.82.223]: EOF from client [21:29] <hadess> http://apachetoday.com/news_story.php3?ltsn=2001-06-07-001-01-OS-SW <- me works with stuff like that... [21:30] <Uraeus> looks nice [21:30] <hadess> java xml databases, all buzzwords in one company ;P [21:31] <Uraeus> hmm, sounds like somewhere I work [21:31] <hadess> heh [21:31] ajzzzz (aj...@p3...) joined #gstreamer. [21:32] Nick change: ajzzzz -> ajmitch [21:33] <Uraeus> hi ajmitch [21:33] <ajmitch> hello [21:33] <Uraeus> ajmitch: you need to get added to the gstreamer developers list on sourceforge so I can assign your own bugs to you :) [21:33] <ajmitch> Uraeus: i am added [21:34] <ajmitch> ;) [21:35] <Uraeus> hmm, yes I see, but you are not on the list of people to whom I can assign bugs [21:35] <Uraeus> weird [21:35] <ajmitch> heh [21:35] Action: omega_ will fix that [21:35] Nick change: omega_ -> omega_breakfast [21:35] <ajmitch> oh well, i guess that means i don't have a responsibility to fix bugs ;) [21:35] <Uraeus> just you wait :) [21:36] <omega_breakfast> this user permissions interface on sourceforge is pathetic [21:37] <ajmitch> hehe [21:38] <ajmitch> i was going to close a bug last night but got interrupted [21:39] Action: hadess installing dri on his machine [21:39] makki (yv...@ca...) joined #gstreamer. [21:39] <Uraeus> hi makki [21:39] <makki> hello [21:39] <ajmitch> hi makki [21:39] <wtay> hi [21:42] <ajmitch> Uraeus: you been hard at work with PR for this release? ;) [21:43] <Uraeus> not as hard as I should maybe, but I posted us to icewalkers and gnotices [21:43] <Uraeus> I will make a Slashdot attempt to [21:44] <ajmitch> ok, i posted it to freshmeat yesterday [21:44] <Uraeus> yup, I saw that great [21:45] <ajmitch> heh, no flames on gnotices yet ;) [21:45] <ajmitch> gonna post to any of the gnome lists about it? [21:46] <Uraeus> ajmitch: hmm, didn't omega_ mail it to GNOME announce? [21:46] <omega_breakfast> I haven't mailed it anywhere yet [21:46] <ajmitch> hmm, haven't seen it on any of the gnome lists i'm on [21:47] <ajmitch> (g-hackers-readonly, g-2.0-list, g-list) [21:47] <Uraeus> omega_breakfast: ok, add gnome-announce and gnome-sound to the list of places to mail [21:47] <ajmitch> i should probably subscribe to them... [21:47] <omega_breakfast> yeah, and l-a-d [21:49] hadess (ha...@pc...) left irc: mooooh! [21:49] <Uraeus> and maybe kde-multimedia :) [21:50] <ajmitch> hehe ;) [21:50] <omega_breakfast> um [21:50] <ajmitch> Uraeus: i'll let you do that ;) [21:51] <Uraeus> omega_breakfast: what about ardour-devel? <g> [21:52] <omega_breakfast> sure [21:52] <omega_breakfast> make a list [21:52] Nick change: omega_breakfast -> omega_ [21:52] Nick change: ChiefHighwater -> CHW_lunch [21:52] <CHW_lunch> 8-] [21:53] <ajmitch> hmm, one finishes breakfast, the other starts lunch.... ;) [21:54] <Uraeus> omega_: just out of curiousity, why doesn't RidgeRun demand that you use their emailaddress? would guess they wanted that free marketing [21:54] <omega_> there are problems with using their servers still [21:55] <omega_> a T1 is on the way, at which point I can start doing so [21:55] <omega_> but I need to pick a new mail client that can handle multiple mailboxen and sending addrs [21:55] <Uraeus> evolution? [21:55] <omega_> um, maybe [21:55] <Uraeus> balsa? [21:56] <ajmitch> evolution is ok, i use it, but is broken at times ;) [21:56] <thomas> pine ! [21:57] <CHW_lunch> M$ Outlook Express 5.0 8-] [21:57] Action: CHW_lunch ducks [21:57] <Uraeus> ajmitch: well, like GStreamer it is still a work in progress :) [21:57] <Uraeus> CHW_lunch: no need to duck, Miguel makes no secret of admiring lots of MS stuff :) [21:57] <thomas> gstreamer-launch mailsrc=omega@ridgerun ! festival ! esdsink [21:58] <wtay> thomas: oohh, commit that! <g> [21:58] <CHW_lunch> Uraeus:more worried about omega_'s 2x4 [21:59] <Uraeus> hehe [21:59] <omega_> actually, festival and viavoice or sphinx2 would be really cool [21:59] <makki> thomas: pine sux [21:59] <thomas> you know, you should put some gst-launch pipes on the site for fun, and see if you can get away with them ;) [21:59] <makki> thomas: use mutt [21:59] <thomas> makki: why ? [22:00] <makki> thomas: the license sux :-) [22:00] <thomas> makki: I've always wanted to try that... I really should I suppose [22:00] <thomas> makki: well, I use Outlook as well, so I'm used to licenses that suck ;) [22:00] <makki> hehe [22:01] <makki> and pico isn't that good an editor... [22:02] <Uraeus> omega_: whe should have used my codename for this release, cause it seems to do exactly that :) [22:02] Action: thomas still uses pico for programming stuff; should learn emacs though [22:02] <thomas> whlie on the concept: what IDE's are people using here ? [22:03] <omega_> um, bash [22:03] <ajmitch> emacs [22:04] <ajmitch> bbl [22:04] Nick change: ajmitch -> ajbusy [22:05] <makki> i'm using vim [22:06] <makki> in combination with Makefiles it's all i need for a comfortable setup [22:06] <wtay> vim/gcc/bash [22:16] tris (tris@195.44.6.23) joined #gstreamer. [22:16] <tris> Hi there, Ive just tried gstreamer for the first time, love the idea, seems like a really powerful concept. [22:17] <CHW_lunch> wow, thanks. [22:17] <tris> I was just wondering what an example commandline for gstreamer-launch would be for playing vcds [22:17] <wtay> tris: something like: [22:17] <omega_> -launch vcdsrc ! mpeg1parse video_00! mpeg2dec ! xvideosink [22:17] <omega_> for video only [22:18] <omega_> -launch can't do split pipelines just yet [22:18] <tris> Also wondering if anyone has looked an integrating the codecs needed for H.232 video conferencing , a good video conferencing tool for gnome would be great and this stuff look like just the thing to get it started [22:19] <omega_> I've heard that the openh323 guys were looking at gstreamer, we should ping them again, that would solve that problem <g> [22:19] <tris> ahhh, Id missed out the argument to the mpeg1parse bit [22:19] <omega_> spaces are critical in the -launch syntax [22:24] <tris> I dont seem to have mpeg2dec, -inspect showed mpeg2play, but Im just getting lots of ascertion failure at the moment. [22:24] Action: wtay remember why C++ sucked that much... [22:24] <wtay> tris: substitute mpeg2dec by mp1videoparse ! mpeg_play [22:25] <wtay> or install limpeg2dec [22:25] <wtay> mibmpeg2dec even [22:25] <wtay> er lib* [22:25] <omega_> wtay: it's libmpeg2dec <g> [22:26] <wtay> yeah, yeah, I'm tired... <g> [22:26] <wtay> I'm hacking libavifile into gstreamer now (the decoders, not the demuxer) [22:26] <Uraeus> ok, Slashdot submission made, lets hope they are feeling friendly [22:27] <omega_> the demuxer is the interesting bit, because it does .asf.... [22:27] Action: Uraeus dances around his livingroom for good luck [22:29] <CHW_lunch> omega_:is gobject behaving? [22:29] Nick change: CHW_lunch -> ChiefHighwater [22:29] <omega_> hrm, I didn't know that a micrometer was one 100,000th of an inch... [22:30] <omega_> ChiefHighwater: not causing me troubel yet, but I haven't gotten too far yet [22:30] <ChiefHighwater> huh? [22:30] <omega_> according to http://unisci.com/stories/20012/0608013.htm [22:30] <ChiefHighwater> doh. [22:30] <steveb> is heroinewarrior still doing the gobject port? [22:31] <omega_> not afaict [22:31] <steveb> too hard eh :) [22:31] <omega_> apparently [22:32] <steveb> did he write b2000? [22:32] <omega_> yup [22:34] <steveb> omega_: when dynparms lands you might want some of it moved out to a lib to keep the core small [22:34] <omega_> yeah [22:35] <omega_> wtay: btw, I managed to set compile and configure options (sorta) so that the stripped libgst.so from 0.2.0 is only 70KB [22:35] <omega_> but configure.in is broken, I had to manually disable GST_DEBUG_ENABLED [22:35] <wtay> omega_: yeah, I saw that, pretty amazing [22:35] Nick change: ajbusy -> ajmitch [22:35] tris (tris@195.44.6.23) left irc: [x]chat [22:35] <omega_> so I'm gonna be adding #ifdefs around some of the larger chunks of stuff like autoplugging, xml, caps and capsnego, with configure-time options to turn them off [22:36] <omega_> then users can chose what to keep depending on their app [22:36] <wtay> good idea [22:36] <omega_> if they're doing known static pipelines, they don't need any of the above, including caps and props [22:36] <omega_> so 50KB should be a trivial target to hit on the x86 [22:36] <wtay> caps and props are harder [22:36] Action: ajmitch wonders whether he should bother studying some more before today's exam [22:36] <omega_> yeah [22:37] <omega_> /nick ajstudy [22:37] <ajmitch> study is boring tho ;) [22:37] <wtay> most plugins use them to set their internal state [22:37] <omega_> point [22:37] <steveb> ajmitch: i did that paper :) [22:37] <omega_> wtay: right, so you don't disable when you're using those plugins [22:38] <wtay> I suppose so.. [22:39] <steveb> wtay: could we have a seperate func to set state in the plugins - only called after caps is complete? [22:39] <wtay> steveb: the newcaps function is made for that [22:39] <ajmitch> steveb: which one? [22:39] <steveb> wtay: oh, didn't know about that [22:39] <steveb> ajmitch: physics 101 [22:40] <ajmitch> steveb: heh, didn't find it too hard? ;) [22:40] <wtay> steveb: after the caps are set to the pad, the function (callback) is called [22:40] <steveb> wtay: I should use that in my plugins [22:40] <steveb> ajmitch: aced it ;) [22:41] <wtay> steveb: most plugins do [22:41] <ajmitch> steveb: hehe, i barely studied for the mid-term exam and got 75%, i'm hoping to get a bit higher this time ;) [22:46] <wtay> woa! a 15 minutes avifile hack and it works! [22:46] <omega_> ooooh [22:46] <ajmitch> cool [22:46] <ajmitch> now show us a demo ;) [22:47] <wtay> I'll commit soon ;) [22:47] <makki> wtay: cool [22:47] <wtay> our second C++ plugin :) (first one working) [22:49] Action: ajmitch wonders why the SF download count is still 0 [22:49] <Uraeus> ajmitch: it is 0 for all projects, think it is borked [22:50] <ajmitch> not fair :P [22:50] <wtay> apparently it uses a DirectShow plugin... [22:53] <Uraeus> wtay: so now we support .asf too :) [22:53] <wtay> Uraeus: nope, I'm still not using the avifile demuxers [22:54] Action: Uraeus has no idea what a demuxer is [22:54] <wtay> they don't allow reading from memory.. :( [22:54] Action: Uraeus still has no idea what a demuxer is [22:55] <wtay> a demuxer splits the video into elementary components (audio/video) [22:55] <Uraeus> ah [22:55] <wtay> you then need a decod... [truncated message content] |
From: <wim...@ch...> - 2001-06-10 04:28:03
|
[07:11] vektor (ve...@cr...) left #gstreamer. [07:22] Nick change: taazzzz -> taaz [07:44] Nick change: ajbusy -> ajmitch [07:47] Nick change: ajmitch -> ajbusy [08:15] Nick change: taaz -> taazzzz [08:15] chillywilly (bau...@d1...) joined #gstreamer. [08:17] big_T (th...@fy...) got netsplit. [08:17] big_T (th...@fy...) returned to #gstreamer. [08:25] steveb (st...@no...) joined #gstreamer. [08:56] rowenc (rowenc@152.117.209.53) left irc: Read error to rowenc[152.117.209.53]: EOF from client [09:09] dobey (do...@dr...) left irc: later [09:17] chillywilly (bau...@d1...) left irc: Philosophers and plow men, each must know his part, to sow a new mentality closer to the heart... [10:15] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee4c.a2000.nl] [10:29] Nick change: ajbusy -> ajmitch [11:26] steveb (st...@no...) joined #gstreamer. [11:26] <ajmitch> hey steveb [11:26] <steveb> hi [11:27] <ajmitch> what's up? [11:47] mmicire (mm...@li...) joined #gstreamer. [11:47] <ajmitch> hi [11:48] mmicire (mm...@li...) left irc: [BX] You can breathe without BitchX, but I wouldn't recommend it [11:55] mmicire (mm...@li...) joined #gstreamer. [12:08] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee4c.a2000.nl] [12:09] <ajmitch> hmm [12:09] greg_ (gr...@ho...) joined #gstreamer. [12:09] <ajmitch> hey greg_ [12:09] <greg_> hello ! just woken up ;-) [12:10] <ajmitch> hehe [12:10] <ajmitch> what time is it there? ;) [12:10] <greg_> ATM is 00:12PM, I woken up at 11:30AM ;-) [12:10] <mmicire> .nz Where is that? [12:11] <ajmitch> new zealand [12:11] <greg_> mmicire: New Zeland ? near Australia ??? [12:11] <mmicire> ah. just curoius. :) [12:11] <ajmitch> greg_: sorta close to australia ;) [12:12] <ajmitch> people have tried out gstreamer 0.2.0 release yet? ;) [12:12] <mmicire> compiling as we speak. [12:12] <greg_> is 0.2.0 released ? I've seen some pre1 I think [12:12] <ajmitch> look at topic ;) [12:12] steveb (st...@no...) joined #gstreamer. [12:13] <greg_> Is it the same as sourceforge's CVS version ??? [12:13] <ajmitch> [22:12:53] --- Topic for #gstreamer is GStreamer "Critical Mass" 0.2.0 released!: http://gstreamer.net/releases/0.2.0/ [12:13] <ajmitch> cvs version probably has a fix or too (for avi plugin i think) [12:15] <ajmitch> but it's basically the same as 0.2.0 [12:15] <ajmitch> is it working ok at the moment, or still having problems? [12:16] <mmicire> me? [12:16] <ajmitch> either of you ;) [12:16] <ajmitch> greg_: weren't you having problems earlier? [12:16] <mmicire> still compiling. damn slow machine. :) [12:17] <ajmitch> mmicire: probably faster than mine ;) [12:17] <mmicire> if I can show my ignorance of this project; is it possible to do live streaming content with this package? [12:17] <ajmitch> hmm, ok, steveb is up & committing changes ;) [12:18] <greg_> wtay said yestarday not to make install, cause it DOESN'T work. Did you released such a version ??? [12:18] <ajmitch> greg_: don't use make install if you're installing from cvs, because autogen.sh turns on different configure options than what is used in the tarball [12:19] <ajmitch> if you used make install, then 2 sets of plugins would be found, and the behaviour of gstreamer in that situation is 'undefined' ;) [12:20] <greg_> ajmitch: what i did yesterday: ./autogen.sh ; ./configure; make - don't you think it should have worked ? [12:20] <ajmitch> yes, it should work [12:21] <ajmitch> of course running ./configure is not needed, since autogen.sh runs that anyway ;) [12:22] <greg_> ajmitch: it is said, that autogen.sh runs configure with parameter making impossible to do make install (a error message is generated, when trying to make install) [12:22] <ajmitch> yes, that's basically what i said (because it will look for plugins in the build dir) [12:22] <greg_> to be precise: since I fetched the sources from CVS I did: ./autogen.sh make make install(error) ./configure make make install and didn't worked ;-( [12:23] <ajmitch> what was the error? [12:23] <greg_> ajmitch: it was some message "configured with --inside-build-dir-only.... so cannot do make install" basically [12:24] <ajmitch> yeah, it's part of the makefile apparantly [12:24] <ajmitch> what happened when you then ran ./configure ; make; make install ? [12:25] <greg_> ajmitch: it didn't worked - gstmediaplay did not played any of my mp3, mpeg, avi... [12:25] <ajmitch> it crashed or just did not work? [12:26] <ajmitch> does 'gstreamer-inspect' give a long list of plugins? [12:27] <greg_> AFAIR crashed (cannot test ATM as I did make uninstall and comipiling) [12:27] <ajmitch> ok [12:27] <greg_> ajmitch: yes, [12:27] <ajmitch> for best results when running from CVS, you should only use ./autogen.sh; make [12:28] <ajmitch> and then run from the directory it is compiled in [12:28] <greg_> ajmitch: I will do it this way this time. But it used to work 2 months before... [12:28] <ajmitch> the source tarball has been generated with make dist, so i think that is is designed to be installed [12:29] <ajmitch> it was working for me a few weeks ago when i had it both installed & running from the sourcedir ;) [12:29] <ajmitch> until omega told me off for doing that ;) [12:30] <greg_> ajmitch: as from the user view it is nice to have binaries in separate subdir (where they are installed) and gstmediaplay working fine ;-) [12:31] Nick change: greg_ -> greg-eat [12:31] <ajmitch> greg_: yeah, well, CVS designed for those following development ;) [12:31] <ajmitch> (or so i was told... ;) ) [12:31] <greg-eat> eating: BRB [12:32] <ajmitch> k [12:50] Nick change: greg-eat -> greg_ [12:52] <greg_> where will I find gstmediaplay in source tree after make ? [12:53] <ajmitch> in gstplay dir [12:53] <ajmitch> after the build, you must run tools/gstreamer-register [12:54] <greg_> Cannot save new registry to `/etc/gstreamer/.reg.xml.tmp' [12:54] <greg_> (Permission denied) [12:54] <greg_> I see it requiries root - don't you think it shouldn't be ? [12:56] <greg_> there is NO executable in gstplay dir [12:57] <greg_> I found it in history: configure: WARNING: Couldn't find libglade-config - Can't build gstplay [12:58] <ajmitch> you don't have the glade devel libs installed/ [12:58] <ajmitch> ? [12:59] <greg_> probably no [12:59] <greg_> greg:/home/greg# dpkg -l |grep glade [12:59] <greg_> ii libglade-gnome 0.16-3 Library to load .glade files at runtime (Gno [12:59] <greg_> ii libglade0 0.16-3 Library to load .glade files at runtime. [13:00] <ajmitch> ajmitch:/usr/share/doc/pkg-config# dpkg -l |grep glade [13:00] <ajmitch> rc glade 0.5.11-5 GTK+ User Interface Builder [13:00] <ajmitch> ii glade-gnome 0.6.2-4 GTK+ User Interface Builder (with GNOME supp [13:00] <ajmitch> ii libglade-gnome 0.16-3 Library to load .glade files at runtime (Gno [13:00] <ajmitch> ii libglade-gnome 0.16-3 Development files for libglade (Gnome widget [13:00] <ajmitch> ii libglade0 0.16-3 Library to load .glade files at runtime. [13:00] <ajmitch> ii libglade0-dev 0.16-3 Development files for libglade. [13:00] <ajmitch> ii python-glade 0.6.8-2 Put a bit of python code behind interfaces b [13:00] <ajmitch> you don't need all of those, of course ;) [13:00] <greg_> which one: libglade-gnome-dev or libglade0-dev ??? [13:00] <ajmitch> but the libglade0-dev and possibly libglade-gnome-dev [13:00] <ajmitch> not sure, i have both installed, will check to see which has the needed files [13:01] <ajmitch> install libglade-gnome0-dev and libglade0-dev [13:01] <greg_> ok [13:02] <greg_> libglade-gnome0-dev depends libglade0-dev anyway [13:03] <ajmitch> ok [13:04] <greg_> make disclean && ./autogen.sh && make should do the thing [13:04] <ajmitch> hopefully ;) [13:49] BBB (bb...@bu...) joined #gstreamer. [13:49] <BBB> hey guys, anyone awake yet? [13:50] <ajmitch> about to go to sleep, actually ;) [13:50] <BBB> hm.... [13:50] <BBB> sleepyhead :P [13:51] <BBB> I haven't tried 0.2.0 actually, bad bad me [13:51] <ajmitch> hey, i'm on the other side of the world! :P [13:51] Action: BBB doesn't have access to a linux computer currently [13:51] <ajmitch> bad... [13:51] <BBB> how big is the tarball? [13:51] <ajmitch> only 1.8MB tar.bz2, iirc [13:51] <BBB> let's download it :) [13:51] <ajmitch> you can, i compile from cvs ;) [13:52] <BBB> I don't have CVS access here [13:52] <BBB> winmodem [13:52] <BBB> and windoze won't access CVS [13:52] <BBB> I never found any CVS-utility for windows [13:52] <ajmitch> hehe [13:52] <ajmitch> rockwell winmodem? [13:52] <BBB> so linux has no inet-access here [13:52] Action: BBB wants to go back home soon [13:52] <BBB> no.... [13:52] <BBB> teles ISDN [13:53] <BBB> internal [13:53] <ajmitch> oh [13:53] <BBB> 16.3c ISN PnP [13:53] <ajmitch> fun ;) [13:53] <BBB> nope :) [13:53] <BBB> PS, MJPEG-1.4.0 has been released [13:53] <ajmitch> anyway, must sleep, in the middle of exams here ;) [13:53] <BBB> so I think we'll start to work on librarification and gstreamerification soon [13:53] <BBB> that'll make omega feel good :) [13:54] <ajmitch> cool [13:54] <ajmitch> night [13:54] <BBB> well, goodnight for you [13:54] <ajmitch> ;) [13:54] Nick change: ajmitch -> ajzzzz [13:54] <BBB> those stupid australians sleep all day ;) [13:54] <BBB> :P [13:54] <ajzzzz> australians???? [13:54] <BBB> new-sealandians [13:54] <BBB> :P [13:54] <ajzzzz> don't *ever* make that mistake :) [13:54] <BBB> :) [13:54] <BBB> it's the same continent [13:55] <ajzzzz> hardly... [13:55] <ajzzzz> NZ is 'only' about 3000km to the east [13:55] <ajzzzz> check your map sometime [13:55] <steveb> ajzzzz: the dutch should know - they 'discovered' it [13:56] <ajzzzz> steveb: hehe [13:56] <BBB> ajzzzz: sorry for the mistake - it's early morning here [13:56] <BBB> oh wait 2 pm [13:56] <BBB> and still asleep [13:56] Action: BBB needs coffee [13:56] <steveb> someone say coffee? [13:56] Nick change: steveb -> steveb-coffee [13:56] <ajzzzz> night, anyway [13:57] <BBB> nite nite [14:05] <BBB> I'll tell omega when I get back :) [14:05] Action: BBB will be back [14:05] BBB (bb...@bu...) left irc: [14:53] steveb-coffee (st...@no...) left irc: Ping timeout for steveb-coffee[node1ee4c.a2000.nl] [14:56] steveb-coffee (st...@no...) joined #gstreamer. [14:59] Nick change: richardb-asleep -> richardb [15:19] Uraeus (csc...@c2...) joined #gstreamer. [15:20] <Uraeus> morning [15:20] steveb-coffee (st...@no...) left irc: Ping timeout for steveb-coffee[node1ee4c.a2000.nl] [15:40] steveb (st...@no...) joined #gstreamer. [15:41] rugrat71 (ne...@di...) joined #gstreamer. [15:46] rugrat71 (ne...@di...) left #gstreamer. [15:50] hadess (ha...@pc...) joined #gstreamer. [15:58] <Uraeus> hi hadess [17:15] Action: Uraeus tries to shake wtay-zZz awake [17:33] thomas (th...@ad...) joined #gstreamer. [17:44] <Uraeus> hi thomas [17:48] <thomas> hi uraeus... slow day huh ? [17:48] <thomas> ;) [17:56] <Uraeus> yup [17:56] <Uraeus> I am working on the SPEC file but I get some strange errors [17:57] <Uraeus> thomas: after installation gstreamer-register complains about not being able to initialize: gstputbits, gstvideoscale, gstriff [17:58] <Uraeus> any ideas what could cause such problems? [18:01] <steveb> Uraeus: sounds like libs/... aren't getting installed [18:02] <Uraeus> yeah, I can't find anything not getting installed, but I am doing some serious searching now so maybe something turns up [18:10] <thomas> Uraeus: I just happen to be learning how to build rpm's at this moment ;) [18:11] <thomas> I'm trying to get abcde into a nice package now [18:36] <thomas> hmm... when a package wants to provide and install a modified version of a file that has been provided by a previous package, what do you do ? [18:44] <Uraeus> thomas: you have a case :( unless you can mark the other package in your SPEC file as obsolete [18:44] <Uraeus> alternativly get your package to place the file somewhere else [18:56] <thomas> Uraeus: hmm, well this is the issue [18:56] <thomas> abcde provides /etc/abcde.conf [18:57] <thomas> I want to replace it with my own [18:57] <thomas> right now I'm winging it with pre/post inst/uninst stuff [18:58] <Uraeus> thomas: if you include all of abcde, then just put a Obsolete: abcde in your SPEC file and that RPM gets removed [19:04] <Uraeus> thomas: have you ranked wtay and omega on Sourceforge? if not do so, and we might get them onto the top ranking list :) [19:10] <Uraeus> thomas: still around? [19:11] Nick change: wtay-zZz -> wtay [19:11] <wtay> yo [19:11] <wtay> bleagh, I bought 256MB ram for my PIII and it doesn't work :( [19:12] <greg_> I just wanted to say, that running ./tools/gstreamer-register needs root, which should not be a must, but an option IMHO - don't You think ? [19:12] <Uraeus> hi wtay, the bz2 of 0.2.0 is corrupt [19:12] <Uraeus> greg_: I think we all agree :) [19:12] <wtay> uh? [19:12] <wtay> currupt? [19:12] <Uraeus> wtay: you can't unpack it [19:13] <greg_> Uraeus: OK - just my $0.03 ;) [19:13] <wtay> it works fine here. [19:13] <Uraeus> greg_: it is already a registered bug in the bug system, actually [19:13] <greg_> Uraeus: I am sorry - I didn't know about it. [19:13] <wtay> I think there is a configure option for it too.. [19:14] <wtay> --with-configdir [19:14] <wtay> going to put the memory in this PC now... [19:14] <Uraeus> wtay: huh [19:14] <wtay> brb [19:15] wtay (wi...@ca...) left #gstreamer. [19:15] <greg_> wtay: but if I compiled ./configure --enable-plugin-srcdir (I beliwe I do so, when I don't wanna to make install, but just run from source dir) - it shouldn't try to access any absolute path automaticly ??? [19:33] _gst_newt_ joined #gstreamer. [19:33] wtay (wi...@ca...) joined #gstreamer. [19:33] <wtay> hmm, it works here... [19:33] <wtay> oh well, with 512MB of ram I shouldn't complain [19:34] <Uraeus> wtay: what works? [19:34] <wtay> my 256MB extra ram [19:35] <wtay> I needed it for my other PC, but it didn't work [19:35] <Uraeus> wtay I need your input [19:36] <Uraeus> wtay: ** ERROR **: file gstplay.c: line 322 (gst_play_have_type): assertion failed: (autoplug != NULL) [19:36] <Uraeus> aborting... [19:36] <Uraeus> Aborted (core dumped) [19:36] <Uraeus> I get that when playing something using GStreamer and my RPMS, suggestions to what could be causing it? [19:36] <Uraeus> gstmediaplay that is [19:37] <wtay> it doesn't find the plugins [19:38] <Uraeus> hmm, ok [19:38] <Uraeus> do I need the pkgconfig stuff and which package does it belong in core or devel? [19:39] <wtay> devel [19:40] <greg_> should mpg video be played fine ? I have some video scewed under gst but looking fine under M$win98 player. [19:41] <greg_> does it make any sense to tell you about such things ? [19:41] <Uraeus> greg_: yes it should, if you can file a bug on it and add a link to the video that is causing problems that would be great [19:42] <greg_> Uraeus: I can put it on site right now. Do you want it ? [19:43] <Uraeus> yes [19:45] <greg_> try this: http://home.sente.pl/gst/lisa16.mpg you must be at least 18 to watch this ;-) [19:46] <Uraeus> ok, we are not adding that to out testsuite then :) [19:48] <greg_> "out testsuite" ? Is it a set of files that should be played propoerly ? [19:48] <Uraeus> hmm, is there a command to count the number of files in a directory? [19:48] <Uraeus> greg_: yes [19:48] <Uraeus> greg_: http://gstreamer.net/media [20:12] <wtay> I gotta go now, cya later [20:12] Nick change: wtay -> wtay-snooker [20:12] <Uraeus> see ya [20:12] <wtay-snooker> sorry for being distracted :) [20:14] <Uraeus> hehe [20:34] ChiefHighwater (pa...@su...) joined #gstreamer. [20:35] <Uraeus> hi ChiefHighwater [20:35] <ChiefHighwater> Ello 8-] [20:40] <greg_> hi Ch*water ! [20:40] <ChiefHighwater> ello greg_ 8-] [20:46] <Uraeus> ChiefHighwater, greg_: you should download the following file to see why developing multimedia software maybe should be banned: http://gstreamer.net/media/honolulu.mp3 [21:00] Action: ChiefHighwater goes and gets Uraeus a grass skirt [21:16] <Uraeus> thnx :) [21:16] <Uraeus> ChiefHighwater: scores pretty hight on the horrible song ranking doesn't it :) [21:25] Nick change: ajzzzz -> ajmitch [21:25] <ajmitch> morning all [21:28] <Uraeus> hi ajmitch [21:28] <ajmitch> hmm, i see you're cheating to get your commit count up ;) [21:28] <ajmitch> committing every tiny change to the spec file ;) [21:29] <Uraeus> ajmitch: well, I am done for a couple of days now :) [21:30] <ajmitch> heh, excellent ;) [21:30] <ajmitch> i'll have to do that sort of thing once exams are done to beat you ;) [21:30] <Uraeus> so now I am just creating bugs reports and assigning them to people :) [21:30] <ajmitch> oh no... [21:31] <ajmitch> luckily i can't take responsibility for any of the code... ;) [21:31] <Uraeus> I am assigning: reverse engineer Sorenson codec to you [21:32] <Uraeus> ajmitch: is 3 weeks a reasonalbe timeframe? [21:33] <ajmitch> umm, maybe 3 decades? ;) [21:36] <Uraeus> hmm, they must have serious problems on themes.org, been down for over a week at least [21:36] <ajmitch> was cracked, remember [21:38] rowenc (rowenc@Roaming-209-53.PLU.edu) joined #gstreamer. [21:38] <ajmitch> hi [21:39] <rowenc> hey [21:42] <ajmitch> bbl [21:42] Nick change: ajmitch -> ajbusy [21:44] <Uraeus> rowenc: what distro do you use? [21:44] <rowenc> redhat 7.1 [21:48] <Uraeus> rowenc: would you be willing to test my RPMS? [21:48] <rowenc> sure [21:48] Apoc (Ap...@pp...) joined #gstreamer. [21:49] <Apoc> Yo [21:50] steveb (st...@no...) left irc: [x]chat [21:50] <Uraeus> rowenc: http://www.linuxrising.com/files [21:50] <Uraeus> hi apoc [21:51] <rowenc> Uraeus: wow, there's a bunch of them [21:51] <Uraeus> rowenc: yes, it is GStreamer and all its major depenencies [21:52] <rowenc> your screen shot only shows blue [21:53] <Uraeus> rowenc: yes I know :), you can't make screenshots of GStreamer currently [21:54] <rowenc> heh... too bad :) [22:02] <Uraeus> rowenc: any success with the RPMS? [22:02] <rowenc> I'm working on it... they install fine, I'm trying to get everything to work [22:03] <Uraeus> rowenc: you must ro a gstreamer-register as root before playing [22:03] <Uraeus> rowenc: we also have testfiles at http://www.gstreamer.net/media [22:03] <rowenc> I saw it registering them when the rpm installed [22:03] <Uraeus> rowenc: it doesn't get all, I need to add the registration part to each one [22:04] <rowenc> ok, trying.. [22:04] <rowenc> ok, working now :) [22:06] <Uraeus> rowenc: great [22:08] <rowenc> of course I had most of the supporting libraries installed from before [22:08] <rowenc> but I don't think there'll be too many conflicts [22:08] <Uraeus> rowenc: well I have tested them now on both my own and your system so I guess things are working well <g> [22:09] <rowenc> yea, they seem to work good [22:09] <Uraeus> ChiefHighwater: got Erik hanging around somewhere today? [22:17] BBB (bb...@bu...) joined #gstreamer. [22:19] <BBB> hmm.... quietness here [22:19] <Uraeus> hi BBB [22:20] <BBB> hi Uraeus [22:20] <BBB> I have a stupid question (but a very good one) [22:20] <BBB> which of the decoder pluins could I take as a good example of how to write my own? [22:21] Action: Uraeus points BBB to wtay-snooker -> who as the name implies it out playing snooker [22:21] <BBB> hmm..... [22:21] Action: BBB tips wtay to see whether he's at home [22:21] <Uraeus> BBB: the AVI plugin has just gotten a major overhaul by wtay [22:22] <BBB> hmm... [22:22] <BBB> I don't have CVS access (yet) [22:22] <BBB> that is [22:22] <BBB> I'm not at home [22:22] <BBB> I'm not on a linux computer [22:23] <BBB> so I cannot access CVS or anything - won't work with windoze [22:23] Action: BBB will take a look when he gets back homr [22:25] Action: Uraeus prepares to try to get GStreamer running on his Intel Solaris box [22:25] <BBB> and the quicktime support.... [22:25] <BBB> is that the libquicktime from Adam Williams? [22:25] <BBB> "q4l" [22:26] Action: BBB has a better idea - I will ask my questions on the gst-devel mailing list [22:27] <Uraeus> BBB: kinda, 3ivx has created a quicktime library which is somewhat based upon the one from Adam, they claim it is better coded [22:27] <hadess> yo guys [22:28] Action: hadess is back [22:28] <Apoc> yo hadess [22:28] <hadess> hey Apoc [22:29] <Uraeus> hi hadess [22:29] <hadess> hey Uraeus [22:29] <hadess> my evolution won't start anymore... in fact everything that uses plugins doesn't work any more [22:30] <BBB> hadess: how about "reboot"? [22:30] <BBB> :) [22:30] <hadess> damn libc upgrade [22:31] <Uraeus> shite, anynome know how I can fix it on Solaris if my root user have an invalid shell, and there are no other users on the box? [22:32] <hadess> reboot with the install media, and edit your /etc/passwd file ? [22:33] <Uraeus> hadess: that is how it works with Solaris? [22:34] <hadess> you have another idea ? [22:34] <Uraeus> no [22:34] <Uraeus> can't think how I do something like single logon under Solaris [22:35] <hadess> i don't have a clue about that, otherwise i'd have told you that... [22:37] dobey (dobey@141.154.95.71) joined #gstreamer. [22:37] <hadess> hey dobey [22:37] <dobey> hey [22:38] <dobey> i installed ydl2 [22:40] <hadess> is it good ? [22:40] <dobey> yeah [22:40] <dobey> it's a lot better than linuxppc [22:40] <hadess> that's easy... [22:40] Action: BBB will leave again [22:40] <dobey> heh [22:41] <dobey> brb [22:41] <BBB> let's start doing gstreamer lav-support :) [22:41] <BBB> bye [22:41] BBB (bb...@bu...) left irc: [22:41] <dobey> the installer sucks though [22:43] <Uraeus> dobey: and it is KDE centric [22:44] <dobey> hrmm [22:44] <dobey> ? [22:46] <hadess> dobey: dude, do you know where the evolution snapshots are ? i need to get the debs of the latest one [22:46] <dobey> hadess: on ftp i think [22:46] <dobey> hadess: or in the RC channel [22:46] <hadess> dobey: rc... *smack* [22:47] <dobey> heh [22:47] <hadess> i don't even know if i'll have enough space to compile it [22:49] <dobey> heh, i'll be building me a new evo soon [22:52] <Apoc> Can we make a CDDB plugin for gstreamer ?? Maybe in the cdparanoia plugin ?? [22:53] <hadess> that's the kind of tricky stuff... [22:53] <hadess> "configure: error: You must have libdb3 version 3.1.17 installed to compile evolution" [22:53] <dobey> heh, yep [22:53] <Apoc> hadess : ?? [22:54] <hadess> dpkg -s libdb3 | grep Version [22:54] <hadess> Version: 3.2.9-10 [22:54] <hadess> so it's not "3.1.17 or better" [22:54] <dobey> right [22:55] <hadess> Apoc: talk to omega about the metadata for plugins, there's nothing now [22:57] <Apoc> hadess : Ok ... I'will ... [22:57] <hadess> let's try to install 0.10.0 instead... [23:00] <hadess> cool, seems like it will work this time [23:04] <hadess> *smack* [23:05] <dobey> heh [23:05] <hadess> now the version of pi-sock is wrong [23:06] Action: hadess rebuilds with pilot support disabled [23:11] Nick change: ajbusy -> ajmitch [23:11] <hadess> hey ajmitch [23:12] <ajmitch> hey [23:12] <dobey> hola [23:13] <ajmitch> dammit, snowing here... [23:18] <greg_> ajmitch: seriously ? oh - yeah - you're on the other side of the earth ;-)) [23:19] <ajmitch> greg_: hehe [23:19] <ajmitch> greg_: yes, where we all have to hang on so we don't fall off :) [23:20] <greg_> ajmitch: same here from your point of view ;-0 [23:21] <ajmitch> everyone knows that NZ is actually in the top half of the world - it only stands to reason that all those heavy continents slid down your way ;) [23:22] thomas-away (th...@ad...) left irc: Client Exiting [23:23] <greg_> ajmitch: you've got very international... no: global sense of humor ;-) I am very impressed. [23:24] <ajmitch> hehe ;) [23:24] <Apoc> bye all ... ;-) [23:25] <ajmitch> bye [23:25] <ajmitch> greg_: you get gstreamer working? [23:25] Apoc (Ap...@pp...) left irc: [23:27] <greg_> ajmitch: your jokes look vell-trained. Is it because people usually make jokes about "the other side" or so ? [23:27] <ajmitch> yeah, all too often ;) [23:27] <greg_> /s/vell/well/ [23:31] <greg_> ajmitch: I thought so. Answering your question: in some sense - yes. But gstmediaplay hangs or crashes after pressing stop, seeking desn't work in most cases... [23:31] <ajmitch> anyway, cya, i'm off to study for exams.. [23:31] <ajmitch> yeah, known problems with gstmediaplay sorry ;) [23:31] <ajmitch> l8r... [23:31] Nick change: ajmitch -> ajstudy [23:39] omega_ (om...@om...) joined #gstreamer. [23:41] <hadess> hey omega [23:41] <omega_> yo [23:45] mmicire (mm...@li...) left irc: Ping timeout for mmicire[lightman.csee.usf.edu] [23:48] <Uraeus> hi omega_ [23:48] <omega_> yo [23:49] <Uraeus> omega_: the SPEC file now works, but the different mpeg releated plugins need sorting out, but I need more info to be able to do so cause now I am unsure of which belongs to what [23:49] <Uraeus> omega_: should I upload my RPMS anywhere? [23:50] <omega_> yeah, into gstreamer.net/release/0.2.0/rpms/ [23:52] <Uraeus> omega_: do you want to look over the SPEC and see of there is anything you want changed first? [23:52] <hadess> gahhhhhhh [23:52] <Uraeus> what? [23:52] <hadess> evolution didn't compile [23:54] <omega_> Uraeus: I'd put all the %package and %files sections togehter, if that's legal [00:00] --- Sun Jun 10 2001 [00:01] <Uraeus> omega_: it might be, but it is not how 'everyone' else does it [00:02] <omega_> so? [00:02] <omega_> 'everyone' isn't building 30+ packages [00:02] <omega_> it should be obvious which files are in which package [00:03] <Uraeus> ok, but could you make me small letter with where to but the different libgstmp*** libraries, and I will fix both issues in one big swoop [00:04] <Uraeus> sorry, I mean libmp*** [00:04] <omega_> um [00:05] <omega_> odd. my gstreamer build is fragged [00:09] <Uraeus> hmm, seems RPM is okey with moving the %files sections [00:17] <hadess> omega_: did the changes from the egocam got into the main tree ? [00:25] mmicire (mm...@li...) joined #gstreamer. [00:25] <hadess> Uraeus: heh, i'm browsing the photos from guadec ;) http://www.sslug.dk/moede/hygge/2001-04-07/pix/index.php?viewone=18 [00:26] Apoc (Ap...@pp...) joined #gstreamer. [00:27] <omega_> hadess: dunno [00:29] <Uraeus> hadess: :) [00:30] <hadess> http://tigert.gimp.org/photos/index.php3?galerie=Guadec2&snimek=68, <g> [00:30] Action: hadess watches photos of anna until his gf comes back [00:30] <dobey> you got another gf? [00:31] <hadess> yeah [00:31] <dobey> you suck. ;-) [00:34] <Apoc> omga : Can we make a CDDB plugin for gstreamer ? Maybe in cdparanoia plugin ?? [00:34] mattsm (ma...@ad...) joined #gstreamer. [00:35] <omega_> not sure it's really cdparanoia's job [00:35] <Apoc> Is it gstreamer job ?? [00:35] <omega_> it can provide the discid, but it shouldn't go to the net and get the data [00:36] <omega_> that's the apps' responsibility [00:36] <mattsm> is it possible that your gstreamer-0.2.0 gzipped tarball is notacually gzipped? I have to run tar xvf - instead of tar zxvf - [00:37] <omega_> um, checking [00:37] <omega_> oh, netscape sometimes has a habit of silently gunzipping things during download [00:37] <omega_> without renaming them [00:38] <Apoc> omega : What about a new plugin just near the cdparanoia one ?? [00:38] <mattsm> omega_: oh, this could be it, it is only 2.3 mb on the server so I thought it was weird too be that small yet still not be gzipped [00:39] <omega_> Apoc: gstreamer's API doesn't have anything to do with cddb [00:39] <omega_> mattsm: yeah, N$ braindamage [00:39] <omega_> the idea is that each plugin does exactly one thing well [00:39] <omega_> in this case, cdparanoia reads audio from the disc. it can construct the discid from this information as well, but it shouldn't go off making network connections [00:42] <Apoc> if cdparanoia procide discid ... it can send it to next plugin in the pipeline which make network connection ?? Is it possible [00:42] <Apoc> ?? [00:42] <omega_> possible, but not what gstreamer is about [00:43] <omega_> if you want to write shared code, write a separate library or component that handles that, the app grabs the discid from the cdparanoia element and hands it to that library [00:43] <hadess> like libcdaudio for example... [00:44] Action: omega_ goes to find more food [00:44] <hadess> it would just be a read-only argument in the cdparanoia plugin... [00:45] <Apoc> yes ... [00:48] ChiefHighwater (pa...@su...) left irc: Ping timeout for ChiefHighwater[sub18-226.member.dsl-only.net] [00:54] <dobey> hadess: you know what gnome-vfs does with cdparanoia? [00:56] <hadess> dobey: yeahm crap [00:56] <dobey> heh [00:58] <mattsm> has someone made libmad and other plugin dependecy packages somewhere? (i remeber seeing them is why I ask) [00:58] <rowenc> there are links to them at the main download page [00:59] <mattsm> i dont see them at: http://gstreamer.net/download.shtml is this the right place too look? [00:59] <mattsm> nm, i've found a link on another page [01:00] <rowenc> I got mine off of sourceforge's download link I believe [01:08] <Uraeus> mattsm: http://www.gstreamer.net/releases/0.2.0/support/ [01:09] <Uraeus> mattsm: those packages are what you need which isn't provided by either Red Hat or Ximian [01:09] Nick change: hadess -> hds-afk [01:10] <mattsm> Uraeus: thanks [01:10] <Uraeus> mattsm: I am uploadin RPMS of GStreamer and all plugins very soon now [01:10] <mattsm> Uraeus: cool [01:11] <omega_> btw Uraeus: gstreamer-register --gst-mask=0 is what you want [01:11] <taazzzz> i'll use that too ;) [01:12] <omega_> even better would be the ability to list certain plugin files, to do an incremental update of the registry [01:12] <omega_> load the registry, load those files, save the registry [01:12] <omega_> easy hack for someone to do [01:12] <mattsm> is there any plans to have a per user registry? [01:13] Nick change: taazzzz -> taaz-away [01:14] <omega_> eventually, yes [01:14] <Uraeus> omega_: seen the Sun mail I forwarded? [01:14] <mattsm> can someone throw up a few gstreamer-launch examples so I can get the gist of it? (i dont see any docs on the web site) [01:15] <Uraeus> mattsm: think the man page covers it [01:15] <mattsm> Uraeus: ok great [01:15] <omega_> no, it doesn't ;-( [01:15] <omega_> or didn't last I looked [01:15] greg_ (gr...@ho...) left irc: Read error to greg_[home.sente.pl]: EOF from client [01:16] <mattsm> the readme in the tools dir has some examples [01:16] <omega_> except that's old, it's a '!' istead of a '|' [01:16] <omega_> and a few other things different ;-( [01:17] <mattsm> can someone paste an example then? [01:18] <omega_> to do what, for instance? [01:18] <mattsm> omega_: well, i was going to see if I could decode mpg and encode them into ogg fiels in one fell swoop [01:19] <mattsm> s/mpg/mp3 [01:19] <omega_> ./gstreamer-launch disksrc location=music.mp3 ! mad ! vorbisenc ! disksink location=music.ogg [01:19] <omega_> that was the example I used at the PLUG meeting, as per recent -devel mail [01:19] <mattsm> omega_: excellent thank you alot! [01:20] Action: Uraeus starts hitting head in the table as he realise he sent out the gstreamer mail without any URL [01:20] <mattsm> omega_: works like a charm [01:21] <omega_> oops [01:21] <dobey> bbl [01:21] dobey (dobey@141.154.95.71) left #gstreamer (eh). [01:21] <mattsm> ./gstreamer-launch disksrc location=music.ogg ! vorbisdec ! osssink play an ogg? [01:21] <omega_> should [01:22] <mattsm> omega_: and it does =) thanks alot, really nice work also [01:24] <Uraeus> mattsm: would you update that readme file with correct syntax? [01:24] <omega_> Uraeus: shouldn't it be in the man page, really? [01:25] <Uraeus> probably, but I for one have no real idea what writing a man page actually demands [01:25] <omega_> nor do I, which is why I was gonna write it up raw and have taaz or someone mark it up <g> [01:25] <mattsm> Uraeus: sure, i can try [01:26] <Uraeus> think there is/was a program called gnome-manedit but I have no idea if it is usefull or not [01:26] <Uraeus> mattsm: great :) [01:26] <mattsm> is gst still capable of playing dvds? [01:26] derek (de...@cp...) joined #gstreamer. [01:27] <derek> anyone here to help jog my memory? [01:27] <derek> on how to make gstreamer work from command line? [01:27] <omega_> maybe ;-) [01:27] <Uraeus> mattsm: should be, it plays VOB files perfectly, but I don't know if the deccs stuff still work [01:27] <omega_> ask mattsm <g> [01:27] <mattsm> derek: what are you trying too do? [01:27] <derek> omega_: i cant remember the syntax [01:27] <derek> i know it was [01:27] <Uraeus> derek: mattsm is writing a README on the subject as we speak :) [01:28] <derek> osssink | vorbisenc | dissink [01:28] <derek> or something [01:28] <derek> but --help didnt help :) [01:28] <derek> and it only worked [01:28] <derek> when i passed a bunch of other params [01:28] <mattsm> derek: are you trying to play a ogg? or encode it? [01:28] <Uraeus> ./gstreamer-launch disksrc location=music.ogg ! vorbisdec ! osssink play an ogg? [01:28] <derek> which i dont recall [01:28] <derek> mattsm: bot [01:28] <derek> both [01:28] <derek> encode at first [01:28] <derek> omega_: i bought the converters [01:28] <mattsm> ./gstreamer-launch disksrc location=music.mp3 ! mad ! vorbisenc ! disksink location=music.ogg [01:28] <mattsm> that takes an mpg and makes it an ogg [01:29] <derek> so tonight im gonna pipe from the mixing board to ogg :) [01:29] <mattsm> s/mpg/mp3 [01:29] <derek> libmad wont compile for me [01:29] <derek> so right now only concerned about ogg [01:29] <Uraeus> derek: which distro? [01:29] <mattsm> derek: i dont think you can grab the sound from osssink, its a sink not a src [01:29] <derek> rh6.2 but i dont wish to trouble shoot that right now [01:30] <mattsm> derek: so you need a source file [01:30] <derek> no i want to stream [01:30] <derek> omega_: you here? [01:30] <derek> you know what im trying to do :) [01:30] <mattsm> derek: stream from what? [01:30] <mattsm> derek: http? [01:31] <derek> ./gstreamer-launch osssink ! vorbisenc ! disksink location=some.ogg [01:31] Action: omega_ is here, but not feeling very good at all (stomach flu?) [01:31] <derek> i think is what i need [01:31] <omega_> osssrc ! vorbisenc .... [01:31] <derek> ah osssrc :) [01:31] <derek> ok [01:31] <derek> now taz or someone helped me get to work first time [01:32] <derek> as that wont work for me out of box [01:32] <derek> have to pass like [01:32] <derek> channel [01:32] <mattsm> derek: can you explain to be waht that does [01:32] <derek> and some other params [01:32] <derek> mattsm: it will take input from sound card and convert to .ogg [01:32] <derek> in my case [01:32] <derek> i am going from mixing board from live band [01:32] <derek> into sound card [01:32] <derek> into .ogg [01:32] <mattsm> derek: neat =) [01:33] <derek> but osssrc doesnt like me [01:33] <derek> it wants me to set defaults [01:33] <derek> on channel etc [01:33] <derek> i dont recall what they were [01:33] <derek> hoping someone knows though [01:33] <omega_> freq=44100 channels=2 etc. [01:33] <derek> as have to leave in 30 minutes to go setup [01:33] <derek> omega_: thats it :) [01:33] <omega_> do a -inspect on osssrc to find out what they are [01:33] <derek> i think there was one more [01:34] <omega_> depth=16 [01:34] <derek> do i pass as [01:34] <omega_> but verify with -inspect what the arg names are [01:36] <derek> yep all those are valid [01:36] <derek> how do pass them again? [01:36] <derek> ./gstreamer-launch --freq=44100 osssrc ! vorbisenc ! disksink location=some.ogg [01:36] <derek> ? [01:36] <derek> or [01:37] <omega_> no, osssrc freq=... channels=... ! vorbisenc [01:37] <derek> ./gstreamer-launch osssrc freq=44100 ! vorbisenc ! disksink location=some.ogg [01:37] <mattsm> Uraeus: let me know once you have those rpms up [01:37] <Uraeus> ok [01:37] <omega_> they're specific to the osssrc element [01:37] <derek> right [01:37] <derek> but how do i pass them in the string? [01:38] <omega_> just like that [01:38] <omega_> ./gstreamer-launch osssrc freq=44100 channels=2 depth=16 ! vorbisenc ! disksink location=file.ogg [01:39] mmicire (mm...@li...) left irc: changing servers [01:39] <mattsm> 'gstutil: could not find argument "freq" in the `GstOssSrc' class ancestry [01:39] <mattsm> gstutil: could not find argument "depth" in the `GstOssSrc' class ancestry [01:39] <derek> yikes [01:39] <derek> dneighbo@latitude:~/downloads/sound/gstreamer/tools$ ./gstreamer-launch osssrc freq=44100 channels=2 depth=16 ! vorbisenc ! disksink location=file.ogg [01:39] <omega_> check with -inspect, those must be the wrong names [01:39] <derek> INFO ( 1549:-1) Initializing GStreamer Core Library [01:39] <derek> INFO ( 1549:-1) CPU features: (0c040882) MMX SSE [01:39] <derek> gstutil: could not find argument "freq" in the `GstOssSrc' class ancestry [01:39] <derek> gstutil: could not find argument "depth" in the `GstOssSrc' class ancestry [01:39] <derek> RUNNING pipeline [01:39] <derek> going into gtk_main() [01:39] <derek> setting sound card to 4000KHz 16 bit stereo (1078122452 bytes buffer) [01:39] <derek> Segmentation fault (core dumped) [01:39] <derek> dneighbo@latitude:~/downloads/sound/gstreamer/tools$ [01:40] <derek> i didnt have this problem before [01:40] <derek> its not freq [01:40] <derek> ./gstreamer-launch osssrc frequency=44100 channels=2 depth=16 ! vorbisenc ! disksink location=file.ogg [01:40] <derek> will work [01:40] <derek> whew [01:40] <derek> scared me for a minute [01:42] <derek> sweet it worked! [01:42] <mattsm> derek: all those options are the defaults, the ones listed anyways [01:42] <mattsm> I have one question, this action records what is being played right? [01:42] <omega_> hmm? [01:43] <mattsm> omega_: does it record the mic input? this could be my problem [01:43] <omega_> it records whatever your mixer feeds the A/D [01:43] <mattsm> omega_: ok, my bad then [01:43] <omega_> you have to tweak the mic and line in levels to get what you want [01:43] <Uraeus> hmm, RPM complains that /usr/bin/gstreamer-register --gst-mask=0 is two files on one line, anyone know how I can mark them as one [01:44] <mattsm> '/usr/bin/gstreamer-register --gst-mask=0' ? [01:44] <Uraeus> tried [01:44] <Uraeus> doesn't work [01:44] <omega_> Uraeus: huh? [01:44] <derek> and ./gstreamer-launch disksrc location=file.ogg ! vorbisdec ! osssink [01:45] <derek> will play back [01:45] <omega_> yup [01:45] <derek> mattsm: they might be the default but on my machien it wont record if i dont pass them for some reason [01:45] <omega_> derek: could be that setting the params forces a capsnego, which means there's a bug in the osssrc [01:45] <derek> could be [01:46] <Uraeus> argh, just me being stupid [01:46] <Uraeus> search and replace also replaced the gstreamer-register entry in the %files list of course ;) [01:46] <derek> omega_: thank you SOOOO much [01:47] <omega_> derek: lemme know how it goes [01:47] <omega_> Uraeus: doh [01:47] <mattsm> ./gstreamer-launch gnomevfssink location=http://mcclintock.net/cmop.mp3 ! mad ! osssink am I missing something? why doesnt this work. gnomevfssink says: error in element 'gnomevfssink0': opening file "http://mcclintock.net/cmop.mp3" [01:48] <omega_> gnomevfssrc [01:48] <mattsm> omega_: ok thanks [01:51] <Uraeus> omega_: is it possible to have two CVS systems sync with eachother? [01:51] <mattsm> Uraeus: galeon does this with gnome-cvs and sf cvs [01:51] <mattsm> Uraeus: you might ask yaneti on irc.gnome.org#galeon how he did it [01:52] <omega_> Uraeus: not afaik [01:52] <Uraeus> mattsm: thnx, cause that the exact same thing we would want to do [01:52] <omega_> hmm, ok, maybe ;-) [01:52] <omega_> why?? [01:52] <mattsm> Uraeus: this also bypasses the anongnomecvs lag [01:53] <mattsm> he is afk atm [01:54] <Uraeus> omega_: well in regards to getting support from Sun QA etc. we need to be in GNOME CVS [01:55] <omega_> hmm, ok [01:55] <omega_> how is the history going to work, since the names in sf cvs won't match gnome cvs? [01:56] <Uraeus> omega_: let me talk with yaneti and see how they have done it [01:56] <omega_> ok [01:57] <Uraeus> actually I saw the other day that KDE has a very strict policy that everything that is to be bundled with KDE has to be in KDE CVS [01:57] <mattsm> where are the example video files on the http server? [01:58] <Uraeus> mattsm: http://www.gstreamer.net/media [02:00] <Uraeus> mattsm:do the bot in #galeon store email addresses? [02:00] derek (de...@cp...) left irc: [x]chat [02:00] Apoc (Ap...@pp...) left irc: Ping timeout for Apoc[ppp-24.dialup-12.worldonline.fr] [02:02] <mattsm> can you not stream ogg files over http with gnome-vfs? [02:03] <omega_> stream over http or icecast ? [02:03] <mattsm> http [02:03] <omega_> no reason you shouldn't be able to [02:03] <mattsm> vorbisdec: end [02:03] <mattsm> Segmentation fault (core dumped) [02:03] <mattsm> omega_: I get that when trying this: [02:04] <mattsm> omega_: grr I waws doing something wrong [02:04] <mattsm> ./gstreamer-launch gnomevfssrc location=http://mcclintock.net/cmop.ogg ! vorbisdec | osssink [02:04] <mattsm> I still get errors [02:04] <mattsm> tons of these errors: [02:04] <mattsm> ** CRITICAL **: file gstpad.c: line 1464 (gst_pad_push): assertion `peer != NULL' failed. [02:05] <omega_> can't have a |, must be ! [02:05] <mattsm> omega_: oh wow, im on a roll =( [02:06] <mattsm> omega_: this works, im sorry [02:08] <mattsm> what do I need installed too get dvdsrc and css-descramble and other dvd things to work? [02:09] <omega_> need to bug taaz to get it rebuilt <g> [02:09] <Uraeus> mattsm: you need to get hold of the deccs code, copy it into the deccs directory in the GStreamer tree and recompile [02:10] <Uraeus> mattsm: I will make a deccs enabled RPM soon [02:10] <omega_> Uraeus: that doesn't really work anymore, because that code is really really hard to find [02:10] <omega_> dvdsrc needs to be updated to use a decss library, which taaz has a lead on [02:10] <Uraeus> omega_: it is publicly posted with a link on linuxrising.org :) [02:11] <omega_> oh [02:11] <Uraeus> omega_: where no-one else dare post it, I do :) [02:11] <omega_> n/m [02:11] <mattsm> what decoders do I need to use to play vob and mpg files? [02:11] <omega_> mpeg2parse, mpeg2dec, ac3dec [02:12] <mattsm> Uraeus: I dont see decss at linuxrising.org [02:12] <omega_> see the website for a screenshot of the editor showing the pipeline you want [02:12] <omega_> be warned that -launch doesn't have a sufficient syntax yet [02:12] <omega_> you're best off modifying mpeg2parse4 in test/ to use a dvdsrc instead of a disksrc [02:12] <Uraeus> mattsm: DVD for every OS banner, right under it [02:13] Action: Uraeus needs to put in a few hours to update that website [02:19] <mattsm> what is the sdl video sink called? [02:19] <Uraeus> sdlsink [02:19] <omega_> sdlvideosink [02:19] <Uraeus> ops [02:19] <mattsm> ./gstreamer-launch disksrc location=~/pond-mpeg2enc.mpg ! mpeg2parse ! mpeg2dec ! sdlvideosink [02:19] <mattsm> gives [02:20] <mattsm> Couldn't create a 'sdlvideosink', no such element or need to run gstreamer-register? [02:20] <omega_> then you must not have sdl [02:23] <Uraeus> omega_: do you know of MovieMaker? [02:24] <omega_> no [02:25] <Uraeus> omega_: just that there was some guy posting on Gnotices saying he was making a GStreamer based app called MovieMaker and he was glad we made a release so he felt sure we hadn't died [02:25] <omega_> hmm, so why hasn't he talked to us?? [02:26] <mattsm> how do you play an mpg file? [02:26] <omega_> what kind? [02:27] <mattsm> pond-mpeg2enc.mpg from the media folder [02:27] <mattsm> ./gstreamer-launch disksrc location=~/pond-mpeg2enc.mpg ! mp1videoparse ! sdlvideosink [02:27] <omega_> no, mp1videoparse doesn't decode [02:28] <omega_> if you have mpeg2dec, that will play mpeg1 videos [02:28] <mattsm> omega_: yea, i tried a few other things, but I just cant get it quite right [02:28] <omega_> otherwise put mpeg_play between videoparse and sdlvideosink [02:29] <omega_> bleagh, /me needs to take a nap [02:29] Nick change: omega_ -> omega_nap [02:29] <mattsm> omega_nap: how does mpeg2dec work on mpg1 files? [02:29] <mattsm> does it need a parser? [02:29] <omega_nap> it decodes them [02:29] <omega_nap> no [02:30] <mattsm> ./gstreamer-launch disksrc location=~/pond-mpeg2enc.mpg ! mpeg2dec ! sdlvideosink [02:30] <mattsm> this does nothing [02:31] <omega_nap> I don't think you can use ~ in the launch cmdline [02:31] <omega_nap> the shell won't expand it [02:31] <mattsm> ./gstreamer-launch disksrc location=~/pond-mpeg2enc.mpg ! mp1videoparse ! mpeg_play ! sdlvideosink [02:31] <omega_nap> but you can tab-complete that into an absolute path [02:31] <omega_nap> just mpeg2dec [02:31] <mattsm> omega_nap: well, the ~ was working for everything else, but i'll leave you alone now [02:32] <omega_nap> hmm dunno [02:32] <omega_nap> be back later [02:32] <mattsm> Uraeus: timecop is an asshole [02:39] <mattsm> Uraeus: its pointless arguing with him [02:49] <mattsm> how do you play vob files? [02:49] <mattsm> something like: ./gstreamer-launch disksrc location=~/CLIP04.VOB ! mpeg2dec ! sdlvideosink [02:50] <mattsm> but I does not like that ( i am mising something ) [02:52] <hds-afk> mattsm: you forgot mpeg2parse to seprarate audio and video [02:52] Action: hds-afk goes to bed [02:56] <Uraeus> mattsm: the RPMS are ready, I am upload em now [03:05] <mattsm> hds-afk: if I add mpeg2parse before or after mpeg2dec it does not work [03:07] greg_ (gr...@ho...) joined #gstreamer. [03:10] Action: mattsm pulls his hair out [03:13] <Uraeus> mattsm: http://gstreamer.net/releases/0.2.0/rpms/ [03:13] <omega_nap> -launch disksrc ... ! mpeg2parse video_00! mpeg2dec ! sdlvideosink [03:13] <Uraeus> omega_nap: the bz2 file of GStreamer 0.2.0 is corrupt [03:14] <omega_nap> fix it ;-) [03:15] <Uraeus> omega_nap: but I want to get to bed it is 0314 here :( [03:15] <mattsm> omega_nap: am I supposed to rename video_00 to something else? [03:16] <Uraeus> night [03:16] Uraeus (csc...@c2...) left #gstreamer. [03:16] <omega_nap> mattsm: no, that's the name of the pad [03:17] <omega_nap> it might be video_0, -inspect mpeg2parse to find it, I forget [03:17] <mattsm> Couldn't create a 'video_00', no such element or need to run gstreamer-register? [03:17] <mattsm> omega_nap: ok [03:19] <omega_nap> you must keep the spacing I give you, it's fundamental [03:19] Nick change: omega_nap -> omega_dinner [03:21] <mattsm> omega_dinner: that gets me farther but it still crashes... [03:36] mattsm (ma...@ad...) left irc: Client Exiting [03:47] dobey (dobey@141.154.95.71) joined #gstreamer. [04:18] FuRR (Jo...@cp...) joined #gstreamer. [04:19] <FuRR> anyone have compile issues with 0.2.0 ? [04:23] <greg_> FuRR: what error messages do you get ? [04:41] greg_ (gr...@ho...) left irc: Read error to greg_[home.sente.pl]: EOF from client [05:09] walken (fo...@c1...) joined #gstreamer. [05:09] <walken> booh [05:15] dobey (dobey@141.154.95.71) left irc: sleep [05:35] ChiefHighwater (pa...@su...) joined #gstreamer. [05:41] <FuRR> greg_: i keep getting errors (on both the release and cvs) looking for gst/meta/audioraw.h [05:48] FuRR (Jo...@cp...) left irc: Client Exiting [06:08] Nick change: wtay-snooker -> wtay [06:08] <wtay> hit me! [06:09] Action: walken hits wtay [06:09] <wtay> lots of activity here... need to read that when I wake up... [06:12] ChiefHighwater (pa...@su...) left irc: [06:14] derek (de...@cp...) joined #gstreamer. [06:22] Nick change: wtay -> wtay-zZz |
From: <wim...@ch...> - 2001-06-11 04:28:38
|
[06:32] Nick change: ShrimpZzZ -> ShrimpX [06:32] <ShrimpX> yo [07:07] mattsm (ma...@ad...) joined #gstreamer. [07:53] mattsm (ma...@ad...) left irc: Client Exiting [08:40] walken (fo...@c1...) left irc: l8r [08:58] steveb (st...@no...) joined #gstreamer. [09:22] <steveb> nice to see mp3 painting itself into a corner http://www.techreview.com/web/kiang/kiang060701.asp [09:54] omega_dinner (om...@om...) left irc: Read error to omega_dinner[omegacs.net]: EOF from client [10:19] Nick change: ajstudy -> ajmitch [10:45] ShrimpX (marius@131.252.244.168) left irc: ircII EPIC4-1.0 -- Are we there yet? [11:05] Nick change: ajmitch -> ajzzzz [11:12] rowenc (rowenc@Roaming-209-53.PLU.edu) left irc: Client Exiting [11:48] ajzzzz (aj...@p3...) left irc: Ping timeout for ajzzzz[p38-max7.dun.ihug.co.nz] [12:14] ajzzzz (aj...@p9...) joined #gstreamer. [12:42] greg_ (gr...@ho...) joined #gstreamer. [13:56] greg_ (gr...@ho...) left irc: Ping timeout for greg_[home.sente.pl] [13:58] greg_ (gr...@ho...) joined #gstreamer. [13:59] Nick change: wtay-zZz -> wtay [13:59] <wtay> hi [14:05] <steveb> hi [14:05] <steveb> i hope i haven't broken gstparse [14:07] Action: steveb flies to london [14:07] steveb (st...@no...) left irc: [x]chat [15:01] hds-afk (ha...@pc...) left irc: mooooh! [15:27] Uraeus (csc...@c2...) joined #gstreamer. [15:27] <Uraeus> hi [15:29] <Uraeus> hello? [15:46] <greg_> hello Uraeus. Everyoone is sleeping I belive. [16:02] <Uraeus> seems so [16:29] wtay (wi...@ca...) left irc: Ping timeout for wtay[cable-195-162-214-58.upc.chello.be] [16:30] wtay (wi...@ca...) joined #gstreamer. [16:30] <Uraeus> hi wtay [16:52] thomas (th...@ad...) joined #gstreamer. [16:52] <thomas> yo [16:52] <Uraeus> yo [16:54] <thomas> how's the RPM's going ? [16:55] <thomas> and btw are you also doing rpm's for the libraries that plugins depend on ? [16:55] <Uraeus> they are ready, if you go to the download page you see both a link to the GStreamer RPMS and RPMS for some of the prereq [16:56] <Uraeus> s/download page/release page/ [16:56] <thomas> Uraeus: ok, I'll try them out on a new machine then ;) [16:56] <Uraeus> cool [16:57] <Uraeus> thomas: the prereq packages are what you don't get from RH+Ximian [16:57] <Uraeus> still a few things missing, still need to get a lame RPM made for instance [16:58] <thomas> Uraeus: I hmade a lame 3.70 rpm yesterday [16:58] <thomas> and omega gave me an 3.88 one once which I still have I think [16:59] <thomas> you need those ? [16:59] <Uraeus> do you have the SRPMS for both ? [16:59] <thomas> no, only the 3.70 [16:59] <thomas> but I suppose that doesn't help, I'm not sure if 3.70 works for gstreamer [17:00] <Uraeus> don think so, guess I try asking omega for the 3.88 SRPM [17:01] <Uraeus> damn, the Sourceforge file release system sucks when you are releasing lots of files, you need to approve and reload the page for each file [17:07] <Uraeus> thomas: if I manage to get it to compile I will make Solaris pkg also :) [17:07] <thomas> Uraeus: heh [17:07] <thomas> Uraeus: for the trendy people among us :-) [17:08] <thomas> ok, I'll try again on this rpm issue I'm having... [17:08] <thomas> You know, the hardest part is having two machines to do it with ;) [17:08] <thomas> one to write and one to test [17:09] <Uraeus> yeah, luckily we are getting so many here that it is possible to find testers [17:18] taaz-away (dlehn@66.37.66.32) left irc: Ping timeout for taaz-away[66.37.66.32] [17:19] taaz-away (dlehn@66.37.66.32) joined #gstreamer. [17:23] <wtay> hi all [17:23] Action: wtay is distracted again [17:24] <thomas> hi wtay [17:27] <wtay> hello [17:30] Nick change: wtay -> wtay-eating [17:46] <Uraeus> wtay-eating: I have now uploaded and added the Status table to the GStreamer website it is in the right menu [17:52] <Uraeus> thomas: any success with my RPMS? [18:04] Zeenix (pro...@ho...) joined #gstreamer. [18:04] <Uraeus> hi Zeenix [18:04] <Zeenix> hi [18:04] <Zeenix> hi everyone [18:05] <Zeenix> I am happy & sad today [18:06] <Zeenix> I am going to karachi for 2 weeks tour, & i know i wont be seeing Linux or any type of unix there [18:07] <Zeenix> I'll be missing my Linux [18:07] <Zeenix> thomas: are you here ? [18:10] <Uraeus> Zeenix: he is here, but I think he is not siting at his desk just at the moment [18:11] <Zeenix> Uraeus: i m sure you & many others on this channel must be quite unhappy with my stupid questions [18:12] <Uraeus> Zeenix: not at all, we are a very inclusive gang here :) [18:14] Zeenix (pro...@ho...) left irc: Ping timeout for Zeenix[host-141.netzone.net.pk] [18:16] Zeenix (pro...@ho...) joined #gstreamer. [18:16] <Zeenix> Uraeus: sorry got disconnected for a while [18:17] <richardb> Hi guys [18:17] <richardb> Hmm: manedit looks a useful thing. [18:17] Action: richardb prepares to redo the man pages for gstreamer properly. [18:17] <Zeenix> richardb: whats manedit [18:18] <Zeenix> ? [18:18] <Uraeus> Zeenix: tool for editing Unix man pages [18:18] <richardb> A program for editing / writing man pages. [18:19] <Zeenix> richardb: i thought its something very hard [18:19] <Zeenix> to do [18:19] <richardb> That's why manedit seems to be useful. [18:20] <Zeenix> is that included in redhat7 box [18:20] <richardb> No idea. I use debian [18:22] <Zeenix> richardb: who is the mentainer of standard gstreamer & www.gstreamer.net website ? [18:23] <richardb> Several people: omega, wtay in particular, but any project developer can modify it. [18:24] <Zeenix> can I ? [18:25] <Uraeus> Zeenix: not unless Omega gives you an account, but the Wiki pages can anyone edit [18:25] <Zeenix> what are wikki pages [18:26] <Uraeus> Zeenix: there is a link on the gtreamer.net page called Wiki, it leads to a set of dynamic pages where content can be added by anyone [18:26] <Zeenix> i'll see them for myself [18:29] <Zeenix> does omega,wtay & others only meet eachother online ? [18:31] <Uraeus> Zeenix: many of us including omega, richardb, wtay and myself met eachother at GUADEC 2 in Copenhagen not so long ago [18:36] <Zeenix> Uraeus: So when did the idea of gstreamer came to you minds ? [18:37] <Uraeus> Zeenix: GStreamer is came from a project Erik did at a school long time ago and which he evolved into GStreamer [18:38] <Zeenix> you is Erik ? [18:38] <Uraeus> Zeenix: Omega is Erik [18:40] <Zeenix> what is wtay eating, a dinosaur [18:42] <Uraeus> <g> [18:42] Nick change: wtay-eating -> wtay [18:42] <wtay> hi [18:43] <Zeenix> wtay: hi [18:43] <wtay> hi Zeenix [18:43] <Zeenix> wtay: go to http://home.sprynet.com/~cbagwell/AudioFormats-4.html [18:43] <Zeenix> it says about gsm that its too costly in terms of performance [18:44] <wtay> yeah, so it seems... [18:46] <wtay> we need CELP too... [18:46] <Uraeus> Zeenix: it says it is costly, not TOO costly :) [18:46] <wtay> not as costly as mp3 :) [18:47] <Zeenix> wtay: do still say its the best for my app. [18:50] <wtay> yes, sure [18:51] <Zeenix> wtay: ? [18:52] <wtay> Zeenix: one of the best codecs you can get for IP phone [18:53] <Zeenix> wtay: i am going out for 2 weeks from tommorow, i could be on this channel when i am away, but there wont be a linux or unix there [18:53] <wtay> :( [18:53] <Zeenix> here also, i see linux only on my home PC [18:55] <Zeenix> wtay: thats another big reason for my being totally missfit in this country beside others i told you yesterday [18:56] <wtay> I only see linux on my home PC too... [18:58] <Zeenix> wtay: too strange [18:59] <Zeenix> wtay: cant you make an rpm for libgsm, which includes the MMX optimization stuff ? [19:00] <wtay> Zeenix: isn't there one allready? [19:00] <Zeenix> wtay: i dowloaded the libgsm rpm from rpmfind.net, i dont know whether it has got MMX or not [19:01] <wtay> Zeenix: I use debian, not redhat... [19:01] <Uraeus> is there a patch or parameter? [19:01] <Uraeus> I can put it in my RPM [19:01] <wtay> Uraeus: yes, probably not hard [19:02] <Zeenix> wtay: dont you use rpms on debian, what else then ? [19:02] <Uraeus> wtay: what do I need to do? [19:03] <wtay> Zeenix: .deb [19:03] <wtay> Uraeus: search for libgsm MMX on google, you'll find a patch for stock libgsm [19:03] <wtay> Uraeus: maybe you can get libgsm-src.rpm to start from [19:04] <Uraeus> wtay: I already have a src.rpm, that is why we offer gsm.rpm from GStreamer.org :) [19:04] <Zeenix> wtay: "stock" ? [19:04] <wtay> Uraeus: oh, cool.. then just patching it would do I think [19:04] <wtay> Zeenix: standard [19:05] <wtay> Uraeus: beware, it only works on MMX capable CPUs, no fallback [19:05] <Zeenix> wtay: nowadays i dont see any non MMX CPUs [19:07] <wtay> yup [19:08] <Zeenix> Uraeus: send that rpm to me when its ready [19:09] <Uraeus> wtay: have you tested the patch? [19:09] <wtay> Uraeus: not at all... [19:09] <Uraeus> wtay: I got a message saying Reversed (or previously applied) patch detected! Assume -R? [n] [19:10] <wtay> maybe it is already patched... [19:10] <wtay> Uraeus: that's the rpm on gst.net? [19:10] <Uraeus> yes [19:10] <Uraeus> wtay: well that is the one I made a couple of days ago, so mmx might be in it [19:11] <Uraeus> Zennix: http://gstreamer.net/releases/0.2.0/support/ [19:11] <Zeenix> Uraeus: i'm there, then ? [19:11] <Uraeus> Zeenix: think so [19:12] <Uraeus> wtay: anyway to check the binary to find out if it uses mmx or not? [19:12] <wtay> lemme check the src.rpm [19:12] <wtay> Uraeus: yup [19:12] <wtay> I'll check the src.rpm [19:13] <wtay> bleagh I can look into it.. [19:13] <wtay> s/can/can't [19:13] <Uraeus> wtay: I can dcc you the sourcefile [19:14] <wtay> I can't accept DCC (firewall) [19:14] <wtay> going to extract it on my old P120... [19:15] <wtay> whoa doesn't work either... [19:15] <Zeenix> Ureaus: do you want to tell me that the libgsm rpm on support page contains MMX [19:15] <Zeenix> ? [19:15] <wtay> Uraeus: can you mail me the src? [19:15] <Uraeus> wtay: ok [19:15] <wtay> Zeenix: we're trying to find out... [19:18] <Uraeus> wtay: mail sent [19:19] bln (bj...@h1...) joined #gstreamer. [19:19] <wtay> yo [19:19] <Uraeus> wtay: I was thinking about mailing the developers of the libraries we use that are GPL and ask if they would be willing to relicense it to LGPL for us [19:20] <wtay> why? [19:21] <Uraeus> wtay: well having lots of GPL plugins undermines having a LGPL framework since the framework is not useable as LGPL as long as you 'need' a GPL plugin [19:22] <wtay> you don't *need* the plugins at all [19:22] <wtay> glibc is also LGPL but that doesn't mean the apps have to be.. [19:23] <Zeenix> wtay: did you find it out [19:23] <wtay> no MMX in there [19:24] <Zeenix> Uraeus: will you be able to make one ? [19:25] <wtay> Uraeus: I have another patch... [19:27] <wtay> Uraeus: ok, that patch works, sending it to you... [19:27] <Uraeus> Zeenix: yes, at least with wtay's new patch [19:28] <Uraeus> wtay: we are a librart like glibc, that is why I think it would be nice to be as LGPL as possible [19:28] <Uraeus> s/librart/library/ [19:28] <Zeenix> Uraeus: plz send it to me when ready on pro...@li... [19:28] <Uraeus> Zeenix: ok [19:29] <wtay> Uraeus: the plugins are going to be separated from the core soon, does it still matter then? [19:30] <wtay> I mean they are just components built against a LGPL lib... [19:31] <wtay> and the plugins are not libraries so it doesnt make sense to license them as LGPL IMO [19:31] <Zeenix> wtay: whats that LGPL & GPL prob., i thought LGPL is just a seperate GPL for libs [19:32] <wtay> indeed [19:32] <bln> wtay, hard to make commercial applications based on gstreamer then the foundation (the plugins) are GPL'ed.. [19:33] <wtay> hmm, that's true.. [19:33] <Uraeus> wtay: also I have this deccs idea which demands our DVD code being all LGPL [19:34] <wtay> hmm [19:34] <Uraeus> wtay: cause the we could release a closed-source application for DVD playing-only which contained the deccs code and the RIAA would have a much harder time giving us pain [19:36] <Uraeus> of course the idea hinges on it being possible to have the deccs code in the application instead of linking it directly into our DVD playback plugins [19:36] <wtay> doesn't LGPL require you to distribute the code too? [19:36] <Uraeus> wtay: no on the app, only the library [19:36] <wtay> but interesting parts of the app are the plugins.. [19:37] <Uraeus> wtay: so they need to be LGPL [19:37] <wtay> so you also need to distribute the code, no? [19:37] <Uraeus> only for the plugins, not the GUI+deccs app [19:37] <wtay> oh and decss would go into the app then? [19:38] <Uraeus> <Uraeus> of course the idea hinges on it being possible to have the deccs code in the application instead of linking it directly into our DVD playback plugins [19:38] <Uraeus> <g> [19:38] <wtay> my fault :) [19:38] <Uraeus> is it possible? [19:38] <wtay> yes [19:39] <wtay> so.. the close app cannot use GPL plugins then? [19:39] <Uraeus> no [19:40] <Uraeus> that is why I want as many of our plugins to be LGPL, not only for closed source apps, but for any apps not using the GLP [19:40] <Uraeus> err GPL [19:40] <wtay> I understand [19:40] <Uraeus> like BSD, MIT, MPL whatever [19:41] <wtay> we basically need to make sure that all the libs we link against (libmpeg2, ac3, etc..) are LGPL then [19:41] <wtay> the plugin itself is of course no prob [19:41] <Uraeus> wtay: yes, and they aren't today so I want to mail the authors asking if they are willing to relicense [19:42] <wtay> it makes a lot of sense for libs to be LGPL so it seems a very good idea to me [19:42] <Uraeus> another advantage is making those hackers aware of us and maybe they even decide to start helping out :) [19:44] <wtay> hehe [19:45] <Zeenix> wtay: can you send me a picture of yours ? [19:45] <wtay> Zeenix: sec.. [19:47] <wtay> http://gstreamer.net/Gstreamer_team_at_Guadec.jpg [19:48] <Zeenix> wtay: what is GUADEC ? [19:49] <Zeenix> wtay: you are a very smart looking guy [19:49] <Zeenix> wtay: who is that zaheer, is he an indian ? [19:50] RageMax (ra...@c1...) joined #gstreamer. [19:50] <Uraeus> Zeenix: guadec.gnome.org -> GNOME user and developer conference [19:51] <RageMax> is this app going to turn into a video editing solution? [19:51] <RageMax> something like premiere for instance [19:51] <Uraeus> RageMax: kinda, there are video editing solutions planned built on top of GStreamer [19:52] <RageMax> ok, good [19:52] <Uraeus> RageMax: gstreamer itself will just be the foundation [19:53] <RageMax> neat [19:55] <wtay> the idea is to build a video editing library on top of gstreamer that introduces the concepts of timelines etc.. [19:55] <wtay> also transitions etc.. [19:56] <wtay> and then we only need a gui on top of that [19:57] <Zeenix> wtay: is this whole team working part-time on gstreamer ? [19:57] <wtay> Zeenix: yup, only omega works on it full time [19:58] <Zeenix> wtay: from where he gets his living [19:58] <Zeenix> then ? [19:58] <wtay> yes [19:59] <wtay> omega earns money doing gstreamer stuff :) [19:59] <Zeenix> how ? [20:00] <wtay> he works for ridgerun (they want to use gstreamer in embeded systems) [20:01] <Zeenix> arent redhat people interested in gstreamer yet or they dont know about it yet ? [20:01] <bln> do there exist a OpenDivX (MPEG-4) plugin? [20:01] <wtay> bln: not yet [20:01] <wtay> Zeenix: dunno [20:02] <Uraeus> Zeenix: actually Sopwith which has been here works for Red Hat and Jonathan Blandford another Red Hat hacker has said he make use of GStreamer in Nautilus [20:02] <bln> wtay, ok.. i have looked at the encore librarys api and its very simple so i shouldnt be hard to make one. [20:03] <Uraeus> bln: 3ivx is already making one for us [20:03] <wtay> Uraeus: but that's closed [20:03] <Uraeus> wtay: is it? [20:03] <Uraeus> my mistake then [20:03] <wtay> Uraeus: yes, very [20:03] <bln> uraeus, using the OpenDivX library or the win32 dll's? [20:04] <Uraeus> bln: they are writing their own [20:04] <wtay> bln: their own code [20:04] <bln> ok.. i see.. [20:04] <Zeenix> wtay: i am a poor student, i would like to get some money out of my app., can i ? [20:04] <wtay> Zeenix: yeah, sure :) [20:05] <Zeenix> wtay: any advice, how could i ? [20:05] <wtay> Zeenix: make a f*cking good product people want to pay money for [20:06] <wtay> Zeenix: I don't believe in consumer apps that cost mobey though.. [20:06] <wtay> s/mobey/money [20:06] <Zeenix> wtay: then what would i do for my living ? [20:07] <wtay> Zeenix: get a job :) [20:08] <Zeenix> wtay: if every software gets free of cost, there wont be any software houses & no jobs then [20:08] <wtay> Zeenix: untrue, I believe in asking money for support and app customisation [20:09] <wtay> my vision is like this: [20:09] <wtay> I make a video editing app someday, that will be free for all [20:09] <Zeenix> wtay: i understand you point now , you said consumer apps., not apps on demand [20:10] <wtay> hopefully some big company wants to use it, they need support and specific apps to integrate it into their existing setup [20:10] <wtay> that'll cost money [20:12] <Zeenix> wtay: as my app will be using other people's code with no WARANTY of any kind, how come i can be sure that i could give good costumer support for my app ? [20:12] <wtay> Zeenix: 'cause you have the source. [20:13] <Zeenix> wtay: i am getting too late, i got to reach home or my mom will come here with a dager in her hand [20:13] <Zeenix> s/dager/pistol [20:14] <Zeenix> wtay: bi [20:14] <wtay> cya [20:14] <Zeenix> wtay: you are a fine looking guy, can attract any gal before she does [20:15] <wtay> mwha [20:15] <Zeenix> wtay: i wish i could become that healthy, bi [20:15] Zeenix (pro...@ho...) left irc: [20:16] Action: Uraeus puts back his eyeballs after they popped out [20:16] Action: wtay is lucky Zeenix lives far away... [20:17] Action: wtay considers kicking Zeenix next time... [20:17] Action: wtay *is* gonna kick him next time... [20:18] <thomas> aah, "an open source project never chooses it's developers, developers choose it" is what my grandma used to say when she washed me in the kitchen sink ;) [20:18] <wtay> alegorically speaking, I may hope :) [20:19] <thomas> actually, she did wash me in the kitchen sink when I was very little ;) but she didn't say that [20:19] <bln> poor boy.. :-) [20:19] Action: thomas learned that pico is only provided by pine [20:19] <thomas> bln: not really, I seem to remember I liked that. They had a huge sink. [20:19] <thomas> never mind [20:20] <thomas> Uraeus: I'm reinstalling my box with kickstart disks to get lynx installed ;) [20:20] <thomas> that's why I couldn't get the RPMS [20:21] <Uraeus> aha [20:22] <thomas> why do it the easy way right ? [20:22] <thomas> Uraeus: your rpm's do kind of render my toolbox project obsolete though ;) [20:22] <Uraeus> ok, licensing mails sent and already we probably have two more LGPl libs (due to my mistake in the first palce that is :) [20:22] <Uraeus> thomas: toolbox project? [20:22] <thomas> anyway, as soon as they're installed I'll do the RPMS [20:22] <thomas> Uraeus: a copy of the apache toolbox project, a nice script to automatically download and install gstreamer and libraries on which it depends [20:23] <thomas> it just looks flashy for a text-based script [20:23] <Uraeus> aha [20:23] Nick change: thomas -> thomas-install [20:26] <Uraeus> wtay: remember to tell maYam how healthy and attractive you look ;) [20:27] <wtay> hehe [20:29] <Uraeus> wtay: just forwarded you the mail from the libmp3lame author, he has some questions and suggestions [20:30] <wtay> ok [20:31] <wtay> aha a better api! [20:32] <Uraeus> heh [20:34] <wtay> we still use the old one, so we should upgrade soon... [20:35] mattsm (ma...@ad...) joined #gstreamer. [20:35] <Uraeus> hi [20:35] <wtay> yo [20:35] <wtay> we still use the .a file instead of the shared lib AFAIK [20:35] <wtay> yup, we do. [20:36] <Uraeus> wtay: but we will switch? [20:36] <wtay> we will, yes. [20:36] <wtay> 3.86 currently doesn't have a .so file on my system [20:40] <Uraeus> wtay: ok, I mail him saying we haven't been able to compile anything higher than 3.86 then and that we use the .a currently [20:41] <wtay> Uraeus: that sounds like we failed to compile it.. [20:41] <wtay> Uraeus: better say that we currently only use 3.86 [20:41] <Uraeus> but we (I) did fail to compile it [20:41] <wtay> and thus the .a file and the old api to set the params [20:41] <wtay> Uraeus: ah <g> [20:41] <wtay> I didn't try it [20:41] <bln> is it normal to get alot of warnings from gcc then building gs 0.2.0 ? [20:42] <wtay> bln: not really, where? [20:42] <bln> wtay, well, i can send you a mail with the errors. [20:43] <wtay> bln: ok [20:43] <wtay> errors? [20:44] <bln> no errors, just warnings.. [20:45] <wtay> ok [20:47] <bln> i know the gcc 2.96 shipped with redhat 7.1 are pedantic with C syntax than stock 2.95.x [20:48] <bln> cat: gs-020-errors.txt: No such file or directory [20:48] <wtay> I compiled with 3.0 too [20:48] <bln> gstelement.c: In function `gst_element_class_init': [20:48] <bln> gstelement.c:142: warning: assignment from incompatible pointer type [20:48] <wtay> ok, that's normal [20:49] <bln> shouldnt it be fixed? ;) [20:49] <wtay> yeah [20:50] <wtay> just a missing typecast [20:50] <wtay> doesn't harm [20:50] <wtay> I think I already fixed it in CVS.. [20:52] <bln> ok.. myself i'm very pedantic on Quality Assurance and never leave code with warnings in it, always use -Werror :-) [20:52] <wtay> yeah, I tend to be like that too :) [20:53] RageMax (ra...@c1...) left irc: later [20:59] <Uraeus> mattsm: did you finnish that README update? [20:59] <mattsm> Uraeus: I have some stuff in it, but I failed too get enough info to finish it up nicely [20:59] <Uraeus> :( [21:02] Nick change: thomas-install -> thomas [21:02] <thomas> Uraeus: ok, I tried some of the rpms... MAN there are a lot ;) [21:03] <thomas> first off : gstreamer depends on Hermes right ? is that necessary ? [21:03] <thomas> I had to search all over for a Hermes RPM which was served through http:// [21:03] <thomas> anyway, nice work ! [21:04] <wtay> Hermes is used for colorconversion [21:05] <Uraeus> thomas: thanks, it was not a few hours I spent hacking that SPEC file :) [21:05] <wtay> if you're lucky you don't need it [21:06] <thomas> wtay: what do you mean, if you're lucky ? [21:06] <thomas> Uraeus: are you planning on making the supporting rpm's available as well ? [21:06] <Uraeus> thomas: isn't the Hermes RPM on our site using http? [21:06] <Uraeus> thomas: already have [21:06] <wtay> thomas: if you have the Xv extension, most things will work [21:06] <thomas> Uraeus: hmmm, I should check [21:06] <Uraeus> thomas: http://gstreamer.net/releases/0.2.0/support/ [21:06] <thomas> wtay: well, it's on my gbox/davedina box... I'm still looking for a G400 [21:07] <thomas> Xv doesn't work for the Vesa device. [21:07] <Uraeus> thomas: the 0.2.0 release page links to that directory [21:07] <wtay> then you'll need hermes [21:07] <thomas> wtay: btw, you don't happen to want to switch your G400 for a G450 ? [21:07] <wtay> thomas: I have a G450 [21:07] <thomas> wtay: bummer ;) [21:07] <wtay> I could use a second one though :) [21:08] <thomas> heh [21:08] <thomas> I have a trident isa card you can have [21:08] <wtay> no thanks... <g> [21:09] <thomas> OK, I need to find a good way of doing post-install RPM stuff on my box... [21:09] <thomas> btw should we move all non-gstreamer chatter to #gstreamer-dating ? [21:09] Action: thomas ducks [21:09] Action: wtay hits thomas with a big basball bat [21:10] Action: thomas passes out [21:10] <thomas> wtay: you're getting too personal today ;) [21:10] <Uraeus> rofl [21:10] <wtay> the guy is a moron [21:11] <thomas> wtay: you should cut that out of the logs [21:11] <wtay> he can read that [21:12] <Uraeus> hmm :) [21:13] <wtay> well, he lives in a weird country.. [21:13] <thomas> wtay: did he also ask questions about your religion ? [21:14] <wtay> not really.. just about the general life in Europe and stuff.. [21:14] <wtay> private chat [21:14] <thomas> wtay: yeah, me too a month ago [21:15] <wtay> thomas: oh, he takes all of us into private chat or what? [21:15] <wtay> looks like he needs help to get out of that country... [21:15] <thomas> wtay: Don't know. it was my first private chat ever, being an irc newbie ;) [21:16] <wtay> heh [21:16] <wtay> help = me marying him or something [21:16] <thomas> wtay: I'd watch out though, some religions allow for him to have multiple men ;) [21:17] <wtay> rotfl [21:18] <wtay> weird shit [21:26] Action: Uraeus notices that he is kicking wtay's ass when it comes to CVS commits :) [21:29] <wtay> Uraeus: you just wait and see :) [21:31] <thomas> is it possible to document plugins in the same way that the gstreamer api is now documented ? [21:31] <thomas> and is that hard to do if it is possible ? [21:31] <wtay> thomas: I did try that at some point, using perl to extract args, templates etc... [21:31] <thomas> wtay: how is it done for the gstreamer api ? [21:31] <wtay> but I kinda failed bacause my PERL sucks [21:32] <wtay> thomas: using docbook and gtkdocs [21:32] <wtay> thomas: check out docs/plugins/gstdoc* [21:32] <thomas> ok... [21:33] <wtay> using scanobj, it would extract all plugin info into some file that can be used to create a doc then using gtkdoc/bocbook [21:33] <wtay> s/boc/doc [21:34] <thomas> ok, running that only gives errors [21:34] <wtay> right [21:34] <mattsm> would buffering be in the realm of gstreamer? [21:34] <wtay> mattsm: using a queue [21:34] <thomas> hmmm... isn't it a better idea to have everyone use docbook for this ? [21:34] <wtay> mattsm: and a thread [21:34] <wtay> thomas: for what? [21:35] <thomas> Uraeus,wtay: ok, I'm catching up on your GPL/commercial thread up there [21:35] <thomas> wtay: for documenting plugins... I was reading Uraeus's mail [21:35] <mattsm> wtay: umm, its this beyond gstreamer-launch? only available to apps? [21:35] <thomas> one question: if I write a mixer plugin and others now, and I write a mixer app to be used in a commercial closed product, is that wrong ? [21:35] <wtay> mattsm: not at all, use queue ! { ... } [21:36] <mattsm> wtay: will it automaticall take care of buffering? [21:36] <wtay> mattsm: yes, a queue does that, it also has some parms to control the amount of buffering [21:36] <mattsm> wtay: awesome [21:37] <wtay> mattsm: we'd laike to extend the queue to allow it to buffer let's say X secs of data, or X bytes of data too [21:37] <wtay> s/laike/like [21:37] <mattsm> wtay: nice.. =) [21:38] <wtay> mattsm: the args are already in gstqueue.c but not the code :-) [21:39] <thomas> wtay: don't you need mimetype-based functions to calculate bitrate then ? [21:39] <wtay> thomas: ? [21:39] <thomas> wtay: well, how do you decide what's one second's worth of data ? [21:39] <wtay> thomas: the timestamps [21:40] <thomas> wtay: oh, so plugins will be forced to generate timestamps ? [21:40] <mattsm> whats an mp3encode enc? [21:40] <wtay> thomas: yes, timestamps are an integral part of steaming media [21:41] <thomas> hmmm, ok, I suppose that's planned real soon then ;) [21:41] <wtay> mattsm: you mean an mp3 encoder? lame and mpegaudio are the ones we currently have [21:41] <mattsm> wtay: when I throw queue into the fray I get a: [21:41] <wtay> thomas: most of the plugins already have timestamps [21:41] <mattsm> ** CRITICAL **: file gstpad.c: line 616 (gst_pad_connect): assertion `sinkpad != NULL' failed. [21:41] <mattsm> wtay: but it still seems too work [21:42] <wtay> mattsm: what's the syntax you use? [21:42] <mattsm> gstreamer-launch queue level=100 ! gnomevfssrc location=http://mcclintock.net/cmop.mp3 ! mad ! osssink [21:42] <wtay> mattsm: you need gnomevfssrc location=... ! queue level=100 ! { mad ! osssink } [21:43] Nick change: Uraeus -> Ura_MIB [21:43] <mattsm> wtay: what do the brackets do? [21:43] <wtay> mattsm: the queue buffers the data comming from gnomevfssrc [21:43] <wtay> mattsm: that's a thread [21:43] <mattsm> wtay: yea that worked, and it pulled alot more data from the stream at first [21:43] <wtay> mattsm: queues be on thread boundaries (DECOUPLED element) [21:44] <wtay> mattsm: queues must be on ... [21:44] <bln> does mpeg-1/mpeg-2 encoding work? [21:45] <wtay> bln not very well tested, it used to work though [21:45] <wtay> bln: I'm currently revising all of the encoding plugins so that they work again [21:46] <bln> wtay, i playing with the idea of having a m1v stream and a wav or mp3 steam merged into a m1s to m2s.. [21:46] <wtay> bln: the muxer is broken.. [21:46] <bln> wtay, ok.. [21:47] <bln> wtay, what need to be done on the muxer? [21:47] <wtay> bln: I hope to get all this working again soon... [21:47] <wtay> bln: no clear idea of what's wrong really.. [21:48] <wtay> bln: I think it just needs a rewrite [21:48] <bln> wtay, ok.. [21:48] <mattsm> Ura_MIB: are you there? [21:49] <wtay> bln: test/video2mp1 should give a general idea of how to encode things.. [21:49] <wtay> bln: but again: it's outdated... [21:49] <mattsm> wtay: can you explain the video_0! stuff? does the grab separate stuff out of a stream? [21:50] <bln> wtay, ok.. [21:50] <wtay> mattsm: the mpeg1parse plugin separates audio and video, yes [21:50] <wtay> mattsm: mpeg1parse has so called "sometimes" pads, they are created dynamically [21:50] <mattsm> wtay: so I would have to grab the video and the audio and handle them separatly? [21:51] <wtay> mattsm: yes [21:51] <mattsm> wtay: could you give an example on how to do this? [21:51] <wtay> mattsm: video using a video decoder (mpeg2dec) and audio using mpg123 or mad [21:51] <wtay> mattsm: mpeg2parse3 implements a little vob player as an example [21:52] <mattsm> wtay: ok, do I just do audio after video by doing audio_0 (or whatever its called) and do I need the thread (aka brackets) around anything? [21:52] <wtay> mattsm: you can't do that with -launch, it doesn't handle multiple paths [21:53] <mattsm> wtay: ahh ok.. [21:53] <mattsm> wtay: so I cant get sound and video from a mpg1? [21:53] <wtay> mattsm: nope [21:53] <wtay> mattsm: not with -launch, that is [21:53] <mattsm> wtay: ok [21:54] <mattsm> can anyone tell me what AlienSong.mpg is encoded with? [21:54] <wtay> mattsm: you mean what encoder? [21:54] Nick change: taaz-away -> taaz [21:54] <wtay> yo [21:54] <mattsm> wtay: yea what would I need to decode it with? [21:54] Nick change: ajzzzz -> ajmitch [21:55] <wtay> mattsm: oh, just mpeg_play [21:55] <bln> mattsm, look at the mpeg signature in the file. [21:55] <wtay> mattsm: gstreamer-launch gnomevfssrc location=aliensong.mpg ! mpeg1parse video_00! queue ! { mp1videoparse ! mpeg_play ! sdlvideosink } [21:56] <wtay> bah, steveb broke -launch :( [21:57] <mattsm> wtay: for some reason I get lots of: Error!!!! 000001ba skipping.. [21:57] <mattsm> wtay: but still gstmediaplay can play them file however [21:57] <wtay> mattsm: what cmd line do you use? [21:58] <mattsm> gstreamer-launch gnomevfssrc location=http://www.gstreamer.net/media/AlienSong.mpg ! queue ! {mp1videoparse ! mpeg_play ! sdlvideosink} [21:58] <wtay> <wtay> mattsm: gstreamer-launch gnomevfssrc location=aliensong.mpg ! mpeg1parse video_00! queue ! { mp1videoparse ! mpeg_play ! sdlvideosink } [21:58] <wtay> mattsm: you need a demuxer (mpeg1parse) [21:58] <mattsm> wtay: ohh jeez, im a moron =) [22:01] <mattsm> awesome! [22:01] <mattsm> gstreamer-launch gnomevfssrc location=http://www.gstreamer.net/media/AlienSong.mpg ! mpeg1parse audio_00! queue ! {mad ! osssink} [22:01] <mattsm> works for sound [22:01] <wtay> yup [22:02] <wtay> if we can find a syntax for both pipelines you can do both :) [22:02] <mattsm> wtay: ok, i have 1 more question and I will be quiet for a while, how do I decode vob file? [22:02] <wtay> mattsm: gstreamer-launch gnomevfssrc location=aliensong.mpg ! mpeg2parse video_0! queue ! { mpeg2dec ! sdlvideosink } [22:03] <derek> are there tools to see inputs/outputs etc [22:03] <derek> from gstreamer? [22:03] <mattsm> wtay: and for audio? [22:03] <wtay> derek: gstreamer-inspect mpeg2parse? [22:03] <mattsm> ac3dec ! mad ! osssink ? or something? [22:03] <wtay> mattsm: depends on the audio in the vob, assuming its ac3: [22:04] <wtay> mattsm: ac3dec ! osssink [22:04] <derek> hmmm [22:04] <derek> i dont think thats what im looking for [22:04] <derek> if i run [22:04] <wtay> audio_0! queue ! { ac3dec ! osssink } [22:04] <wtay> derek: ? [22:04] <derek> ./gstreamer-launch osssrc ! vorbisenc ! disksink location=foo.ogg [22:05] <derek> i want to be able to control that [22:05] <derek> i.e. see what the levels are coming [22:05] <derek> in [22:05] <derek> and tweak them etc [22:05] <derek> preferably via a gui [22:05] <wtay> derek: you need to write an app then [22:05] <derek> ok [22:05] <derek> im not against doing that [22:05] <derek> i just asked is there one :) [22:05] <ajmitch> heh ;) [22:06] <wtay> nope, we don't have decent apps yet :) [22:06] <ajmitch> you want some sort of equalizer? [22:07] Nick change: taaz -> taaz-away [22:07] <derek> ajmitch: well that would be nice [22:07] <derek> but for now [22:07] <derek> i wouldnt mind just visuals [22:07] <derek> w/o being able to tweak [22:07] <Ura_MIB> mattsm: yes [22:07] <derek> to verify you are getting an audible signal [22:07] <derek> etc [22:08] <wtay> we need an fft plugin [22:08] <mattsm> Ura_MIB: did you ever get the dvd enabled stuff? dvdsrc or whatever it was? [22:08] <Ura_MIB> mattsm: if you are asking about a deccs enabled plugin, no not yet [22:08] <ajmitch> wtay: what does the spectrum visualisation plugin do? [22:08] <wtay> ajmitch: crash <g> [22:09] Action: ajmitch volunteers Ura_MIB for an fft plugin ;) [22:09] <wtay> Ura_MIB: MIB == Man In black? [22:09] <mattsm> gstreamer-launch gnomevfssrc location=http://www.gstreamer.net/media/CLIP04.VOB ! mpeg2parse video_00! queue level=10 ! {mpeg2dec ! sdlvideosink} [22:09] <mattsm> [22:09] <Ura_MIB> ajmitch: how did you know I was a fast fourier transform expert :) [22:09] <Ura_MIB> wtay: yes [22:09] <mattsm> this gives a segfault... [22:09] <wtay> mattsm: yup, you need video_0 [22:10] <mattsm> wtay: works now [22:10] <mattsm> star trek, mmm [22:10] <ajmitch> Ura_MIB: oh, since you are an expert in every other field, i thought you must be in this ;) [22:10] <wtay> streaming even... [22:10] <mattsm> wtay: its a little laggy, [22:10] <wtay> mattsm: I'm not surprised [22:10] <mattsm> wtay: for some reason the queue is not helping [22:11] <wtay> mattsm: nope, you don't have enough bandwidth, that's all [22:11] <Ura_MIB> well back to MIB for me [22:11] Action: mattsm brags about his bandwidth [22:11] <wtay> mattsm: the queue could be extended to really buffer first [22:11] <mattsm> my bandwidth is huge [22:11] <mattsm> wtay: this would be nice =) [22:11] <wtay> like? [22:12] <wtay> how huge? [22:12] <mattsm> 150 kb/s is all [22:12] <wtay> mine too [22:12] <wtay> way too slow for mpeg2 [22:12] <mattsm> wtay: its not half bad = [22:12] <mattsm> =) [22:12] <mattsm> wtay: if it had a little bit of real buffering [22:13] <wtay> mattsm: it needs to download at least half of the video before starting to play it I think [22:13] hadess (ha...@pc...) joined #gstreamer. [22:13] <wtay> yo [22:14] <mattsm> wtay: well, not how I viewed it, but I am not sure if I downloaded sound [22:14] <ajmitch> hey hadess [22:14] <hadess> hey dudes [22:15] <bln> gstreamer-launch disksrc location=testfile.mpg ! mpeg1parse video_0! queue ! {mpeg2dec ! sdlvideosink} shouldnt this work for playing video from a mpeg-1 system stream? [22:15] <hadess> how you doing gang ? [22:15] <mattsm> bln: try video_00 [22:15] <ajmitch> got an exam in ~1hr :( [22:15] <wtay> bln: video_00, yes [22:15] <hadess> good luck then [22:15] <wtay> hadess: pretty good :) [22:15] <mattsm> gstreamer-launch gnomevfssrc location=http://www.gstreamer.net/media/CLIP04.VOB ! mpeg2parse audio_1! queue ! {ac3dec ! osssink} [22:15] <mattsm> is not grabbing the sound, I know I am doing something wrong [22:15] <wtay> ajmitch: yeah, good luck [22:16] <ajmitch> thx [22:16] <wtay> mattsm: checking... [22:16] <bln> its says RUNNING pipeline but no window shows.. [22:16] <bln> Gtk-WARNING **: invalid cast from (NULL) pointer to `GstPad' [22:16] <bln> Segmentation fault [22:16] <bln> *ugh* [22:17] <wtay> bln: that final segfault is normal [22:17] <mattsm> bln try: gstreamer-launch gnomevfssrc location=http://www.gstreamer.net/media/AlienSong.mpg ! mpeg2parse video_00! queue level=10 ! {mpeg2dec ! sdlvideosink} [22:18] <bln> mpeg2parse::parse_packet: unknown packet id 0xBE !! [22:18] <bln> it prints alot of these. [22:18] <mattsm> bln: but it should still show the video [22:19] <wtay> bln: AlienSong is an mpeg1 system stream, you need mpeg1parse instead of mpeg2parse [22:19] <bln> yes it does.. [22:19] <hadess> wtay: dude, do you know what "bus error" means ? [22:20] <bln> wtay, mpeg2parse seams to work nice, video show great here.. [22:20] <mattsm> what filters would I need for this? http://www.gstreamer.net/media/Matrix_f900.avi [22:20] <wtay> mattsm: CVS version? [22:21] <mattsm> wtay: im using 0.2.0 [22:21] <wtay> mattsm: ok [22:24] <wtay> mattsm: doesn't work [22:25] <mattsm> wtay: ok.. [22:25] <wtay> gstreamer-launch disksrc location=/opt/data/Matrix_f900.avi ! avidecoder video_00! queue ! { colorspace ! xvideosink } [22:25] <wtay> mattsm: it does work actually like that [22:25] <wtay> lots of criticals though... [22:26] <wtay> hadess: bus error? [22:27] <hadess> wtay: yep, playing some mp3s with madplay, it segfaults when i use the mad plugin from gst [22:28] <wtay> hadess: hmm [22:28] <hadess> i could put the mp3 up on my site [22:29] <wtay> In general, in NetBSD the 'normal' way of [22:29] <wtay> doing things is "SEGV if it's an OK access that you don't have perms [22:29] <wtay> for or have not mapped, BUS if it'd never be a valid access." [22:29] <wtay> hadess: ok [22:29] <mattsm> does gstreamer-complete work? [22:29] <wtay> yup [22:29] <mattsm> how then? ive run gstreamer-compprep and added those lines too my bashrc [22:30] <mattsm> gstreamer-complete segfaults when I run it alone [22:31] <mattsm> ok, you have to run gstreamer-compprep as root [22:31] <wtay> yeah [22:31] <mattsm> wow that rocks =) [22:32] <wtay> too bad file completion doesn't work anymore.. [22:36] <hadess> http://213.105.184.150/guadec , wtay, the mp3 file and the MD5SUM for it [22:36] <mattsm> why cant you use { } anywhere? [22:38] <ajmitch> hadess: hmm, doesn't work thru squid... tries to go to kara.hadess.net/guadec [22:38] <hadess> add a / at the end [22:38] <richardb> Boo! [22:39] <hadess> hey richardb [22:39] <richardb> I mean, "Hi guys" [22:39] <richardb> Anyone writing an fft plugin? [22:39] <ajmitch> not yet... [22:39] <hadess> thank god you just volunteered ;) [22:39] <wtay> hehe [22:39] <ajmitch> you want one? [22:39] <wtay> hi richardb [22:40] <richardb> I'm in the mood for doing something with the visualisation plugins... [22:40] shitowax (yann@AAnnecy-101-1-1-62.abo.wanadoo.fr) joined #gstreamer. [22:40] <ajmitch> ah, ok [22:40] <hadess> cool [22:40] <ajmitch> gonna use fftw or something like that? [22:40] <hadess> salut froggie ;P [22:40] <richardb> Just checking fftw's license... [22:40] <ajmitch> GPL [22:40] <shitowax> hello everybody [22:41] <richardb> I've got a fft library I wrote myself lying around. [22:41] <hadess> shitowax: beau boulot sur nautilus [22:41] <wtay> hi shitowax [22:41] <wtay> hadess: what file? [22:41] <hadess> wtay: beastie boys song [22:41] <richardb> Could make two fft plugins: an LGPL one base on my lib, and a GPL one based on fftw (which will probably be faster. By far.) [22:42] <wtay> hadess: ok [22:42] <mattsm> gstreamer-launch gnomevfssrc location="http://213.105.184.150/guadec/Beastie%20Boys%20-%20Super%20Disco%20Breakin'.mp3" ! mad ! osssink [22:42] <mattsm> hadess: =) [22:43] <richardb> What mime type should spectral domain data be? [22:43] <wtay> richardb: good question... [22:43] <wtay> application/x-fft? [22:43] Action: richardb doesn't know anything about mimetypes. [22:43] <richardb> That'll do, I imagine. [22:44] <wtay> along with a bunch of props for the number of bands [22:44] <richardb> Suggestion: we should organise the plugins according to license, when we move into the new structure. [22:45] <richardb> So: top level dirs would be GPL, LGPL, closed-source, etc. [22:45] <wtay> perhaps, yeah [22:46] <hadess> is gst GPL or LGPL ? [22:46] <richardb> Would help to avoid license breaking by mistake, or thinking that we supported features entirely in free software that we actually rely on nonfree for. [22:46] <richardb> LGPL. [22:46] <richardb> I think. [22:46] <wtay> hadess: LGPL for the core libs [22:46] <hadess> ok, so what's in the core distro should be LPGL only i guess [22:47] <richardb> That probably makes sense, yes. [22:47] <wtay> hadess: the core is gonna have 0 plugins [22:47] <Ura_MIB> hadess: problem is that all the DVD stuff is GPL [22:47] <hadess> Ura_MIB: then we put it in a GPL package [22:47] Nick change: Ura_MIB -> Uraeus [22:48] <Uraeus> hadess: yeah, but better is getting it re-licensed [22:48] Action: hadess needs a new hd badly [22:48] <hadess> Uraeus: fair enough [22:50] <hadess> wtay: did you get the mp3 ? [22:50] <wtay> yup, got it... [22:51] <wtay> plays fine [22:51] <hadess> at the end i get the "bus error" [22:51] <wtay> oh, that could be the EOS thing not working [22:52] <hadess> btw, you can also grab the fellowship.mpeg at the same place, it plays fine with the windows media player (afaik) but acts weird with gstmediaplay and plaympeg [22:52] <hadess> wtay: no, try using madplay on it [22:52] <wtay> hmm ok [22:52] <wtay> oh [22:53] <wtay> hadess: I have to wait till the end, ight? [22:53] <hadess> yep, about 1:30 minutes... [22:54] <hadess> frustrating if you don't like the beastie boys <g> [22:54] <wtay> yeah [22:54] <wtay> I used to like their old stuff though [22:55] <hadess> more "punk" maybe, rather than rap [22:55] <wtay> yeah [22:55] <wtay> Title/Songname/Content description: Super Disco Breakin' [22:55] <wtay> Bus error [22:55] <wtay> heh [22:56] <hadess> voila [22:56] <wtay> hadess: only for this mp3? [22:56] <hadess> yep [22:56] <wtay> you need to file a bug report to the mad author I feel [22:56] <hadess> right... [22:57] <hadess> remind me to do that when i have a working email client [22:57] <wtay> it generally is a bad memory access somewhere [22:57] <hadess> ok [22:58] <wtay> uhm [22:58] <mattsm> gnome vfs via smb works with gnomevfssrc =) [22:58] <hadess> of course it does, that's the whole point ;) [22:58] <mattsm> hadess: its nice =) [22:59] <hadess> mattsm: you can also play icecast streams you know [22:59] <mattsm> hadess: ive never used icecast streams =) [23:03] Nick change: ajmitch -> ajbusy [23:03] <mattsm> do all parses split the media into separate streams? [23:03] <mattsm> parsers rather [23:05] <wtay> mp3parse doesn't split, it just splits the data in mp3 frames [23:06] <mattsm> wtay: why cant you use the brackets anywhere? [23:06] <wtay> mattsm: brackets are threads [23:06] <wtay> mattsm: you need queues to separate threads [23:06] <mattsm> wtay: so only queue can separate threads? [23:07] <wtay> yes [23:07] <wtay> any DECOUPLED element in fact [23:10] <bln> whats the pipeline for playing a wav file? [23:10] <wtay> disksrc location=... ! parsewav ! osssink [23:11] <bln> tnx.. [23:11] Action: hadess is too slow [23:11] <mattsm> wtay: do you know how too convert an mp3 into a wav file off the top of your head? [23:11] <hadess> mattsm: there's no wav encoder [23:11] <mattsm> hadess: ok [23:11] <wtay> mattsm: we don't have a wav encoder yet [23:11] <hadess> mattsm: write one, thanks =) [23:11] big_T (th...@fy...) got netsplit. [23:12] <mattsm> hadess: i'll start with updating the tools/README =) [23:12] big_T (th...@fy...) returned to #gstreamer. [23:12] Action: wtay is happy mattsm volunteered for the wav encoder [23:13] <mattsm> is there a program that can make an mp3 into a wav file? besides xmms because its fubared here [23:13] Action: mattsm would have to start looking into the source for this [23:13] <hadess> mattsm: mpg123 [23:14] <mattsm> hadess: how do you get mpg123 to do this? [23:14] <hadess> mattsm: don't remember [23:15] <wtay> mpg123 --help [23:15] <mattsm> ive tried mpg123 -s music.mp3 > music.wav [23:16] <hadess> mattsm: mpg123 -w foo.wav foo.mp3 [23:16] <mattsm> hadess: just found out about --longhelp [23:17] bln (bj...@h1...) left irc: sleep [23:17] <hadess> it is in --help dude... [23:19] <shitowax> bye [23:19] <hadess> cya [23:19] <mattsm> hadess: dam its right there and like the biggest thing, grr =) [23:19] <mattsm> ok what are hte mp3 encoders again? [23:20] <wtay> cya [23:20] <wtay> mpegaudio and lame [23:22] <thomas> mattsm: try if afsink works for you to write a wav file [23:22] <thomas> I'm not sure if it still works [23:23] shitowax (yann@AAnnecy-101-1-1-62.abo.wanadoo.fr) left #gstreamer. [23:23] <mattsm> how do you set paramters that are enums? type=Wave or type=4 or what? [23:23] <wtay> hadess: I should better remove that illegal mp3 now :-) [23:24] <wtay> mattsm: type=4 [23:24] <mattsm> wow mpegaudio is slow =) [23:24] <wtay> it is? [23:24] <hadess> wtay: haha ;) [23:24] <mattsm> wtay: well i did the default bitrate of 384 [23:24] <wtay> hmm [23:24] <wtay> mpegaudio is a crappy encoder [23:25] <wtay> doesn't even handle 128Kbps [23:25] <mattsm> i dont have lame installed.. [23:25] <wtay> lame is so much better [23:25] <wtay> but it doesn't do layer1/2, needed for mpeg1 [23:26] <mattsm> Audio File Library: could not open file '(null)' [error 3] [23:26] <mattsm> open: Bad address [23:27] <thomas> mattsm: also use location="..." [23:27] <thomas> mattsm: do you have rpm's ? I have a lame 3.88 rpm if you want [23:27] <mattsm> oh its a sink also [23:27] <mattsm> thomas: yea sure [23:28] <mattsm> Uraeus: there is no gstreamer-lame =) [23:28] <thomas> mattsm: http://urgent.rug.ac.be/thomas/download [23:30] <mattsm> yikes i messed up that stream somewhere [23:30] <mattsm> gstreamer-launch disksrc location=music.mp3 ! mad ! afsink type=4 location=new.wav [23:30] <mattsm> is something wrong here? [23:31] <thomas> no, looks good... [23:31] <mattsm> thomas: them afsink does not work [23:31] <thomas> mattsm: hmmm, afsink should work since it's fairly recent... if it doesn't then I'll check it tomorrow ;( [23:31] <mattsm> thomas: i got all fuzz in the new.wav [23:31] <thomas> mattsm: hmmm, that sounds like endianness trouble [23:31] seer (se...@ga...) joined #gstreamer. [23:31] <thomas> fuzz in which you hear your audio, or just static fuzz [23:31] <thomas> ? [23:31] <seer> hey there.. anyone home? [23:31] <wtay> yo seer [23:32] <mattsm> thomas: static fuzz, it did not resemble the orignal mp3 in any way I could tell [23:32] <seer> my compiled 0.2.0 seems to have an error : ** WARNING **: osssink: unable to open the sound device (in use ?) [23:32] <seer> but i use ESD [23:32] <thomas> mattsm: what arch are you on ? [23:32] <seer> how can I tell gtsmediaplay to use esd? [23:32] <hadess> Uraeus: dude, btw, i'll try to do some testing next week with gst on macos x [23:32] <mattsm> seer try esdsink [23:32] <mattsm> thomas: x86 [23:32] <seer> how [23:32] <mattsm> seer: are you using gstreamer-launch? [23:32] <wtay> seer: you need to modify the gstplay.c file [23:33] <seer> i'm using gstmediaplay ... i haven't gotten -laungh to do much yet :-) [23:33] <thomas> mattsm: hmmm, ok, I'll check afsink tomorrow... sorry for that ;) [23:33] <wtay> seer: gstplay.c line 126, modify osssink to esdsink [23:34] <seer> wtay: thanks! done! I'm there! [23:34] <thomas> ok, I'm off to bed, night all [23:34] <wtay> cool [23:34] <Uraeus> mattsm: problem making a lame RPM has also hindered me in making a GStreamer-lame RPM [23:34] <thomas> wtay: don't dream weird stuff tonight now ok ? ;) [23:34] <mattsm> thomas just listed a link with a lame rpm [23:34] <wtay> thomas: shut up :-) [23:34] <Uraeus> hadess: cool about macosX, I am busy trying to get a usefull Solaris x86 env for testing [23:34] <mattsm> http://urgent.rug.ac.be/thomas/download [23:35] <hadess> Uraeus: _still_ trying ? [23:35] <thomas> wtay: heh [23:35] thomas (th...@ad...) left irc: Client Exiting [23:35] <hadess> Uraeus: you got your user pb fixed ? [23:36] <Uraeus> hadess: user pb? [23:36] <seer> hmnm.. while i'm at it, I have a Matrox G450 and have been using V4L thru it... how would I enable V4L as a video source? [23:36] <hadess> Uraeus: root passwd [23:36] <Uraeus> hadess: yes, true the magic of a reinstall :) [23:36] <hadess> Uraeus: hehe [23:36] <Uraeus> s/true/through/ [23:36] <hadess> seer: gstmediaplay /dev/video might work [23:36] <hadess> seer: or break... [23:37] <seer> i have been using Xawtv, but I'd prefer something unified. [23:37] <Uraeus> seems I need to reinstall once again, since I am missing all the devel stuff [23:37] <wtay> seer: v4l is a bit broken currently :( [23:37] <hadess> Uraeus: pkgadd packages, iirc [23:38] <Uraeus> hadess: if I was missing 3-4 packages yes, since I am missing everything I think a reinstall is easier [23:38] <seer> wtay: thanks... you rock! [23:39] <mattsm> gstreamer-launch cdparanoia start_track=1 end_track=1 ! mpegaudio ! disksink location=track1.mp3 [23:39] <mattsm> gives failed opening cd [23:39] <mattsm> , it defaults too /dev/cdrom and this is my cdrom with a audio cd in it, any ideas? [23:39] <seer> i'm very happy at gstreamer's ablitiy to play "broken" files. [23:39] <hadess> Uraeus: heh, right [23:39] <wtay> seer: it does? [23:40] <wtay> mattsm: no idea... [23:40] <wtay> mattsm: and without the start/end_track? [23:41] <mattsm> wtay: same thing [23:41] <richardb> mattsm: that works for me... [23:41] <Uraeus> I need to get some sleep,workday tommorow see ya [23:41] Uraeus (csc...@c2...) left irc: syntax error - user imploded [23:41] <seer> yeah.. at least the CVS shortly after "lots of duct tape" played all the broken mpg's much better then xanim or Xtheater [23:42] <seer> but I notice now that "VideoLan Client" also plays them fairly well... [23:42] <wtay> seer: neat.. that's a coincidence really.. [23:42] <wtay> seer: yeah, vlc is very nice IMO [23:42] <seer> i agree. [23:42] <hadess> i'll do a dvd players round-up next week for the ppc [23:42] <mattsm> richardb: you run that exact command? [23:43] ChiefHighwater (pa...@su...) joined #gstreamer. [23:43] <wtay> hi [23:43] <hadess> see what comes up as working [23:43] <ChiefHighwater> ello [23:43] <richardb> mattsm: yes [23:43] <hadess> hey CHW [23:43] <wtay> hadess: good luck :-) [23:43] <hadess> wtay: gst plays dvds right ? [23:43] <ChiefHighwater> ello hadess 8-] [23:43] <wtay> hadess: if plays vobs, yes [23:43] <wtay> s/if/it [23:43] <hadess> wtay: i know that, but dvds ? [23:44] <wtay> that's not a realy DVD [23:44] <wtay> hadess: sorta, only the vob of the DVD is decoded [23:44] <wtay> hadess: DVD has lots of other neat sruff, like navigation and such [23:44] <ChiefHighwater> hadess: it kinda does dvds, but there is no mechanism to sync the audio and video sinks yet, afaik [23:44] <mattsm> richardb: you dont have to mount the cd right? you cant mount audio cd's right? [23:45] <richardb> mattsm: correct. [23:45] <hadess> ChiefHighwater: you tried it ? [23:45] <ChiefHighwater> Omega plays a Matrix clip on his [23:45] <richardb> mattsm: well, you can mount them with a special filesystem (I forget the name), but I didn't do that... [23:45] <wtay> ChiefHighwater: it's as good as the mpeg1 sync [23:46] <wtay> ChiefHighwater: well, no, not really, cause the mpeg2parser doesn't handle the timestamps right [23:46] <ChiefHighwater> I'm just repeating what Omega told the person who asked that question at his talk on Thursday [23:46] Action: ChiefHighwater knows nothing [23:47] Action: ChiefHighwater is a windoze user [23:47] <ChiefHighwater> Omega even played the DVD clip out the ascii art sink [23:47] <ChiefHighwater> that was fun to watch [23:48] <ChiefHighwater> the bullet scene from Matrix in ascii [23:48] <wtay> ChiefHighwater: you were on his talk? [23:48] <ChiefHighwater> I was there, yes. [23:48] <wtay> ChiefHighwater: cool [23:48] <ChiefHighwater> I was his ride, Omega doesn't drive [23:48] <wtay> yeah :) [23:50] <wtay> ChiefHighwater: can we voluteer you to do some development now? <g> [23:51] <ChiefHighwater> lol...sure on my win98se laptop or win2kpro desktop? [23:51] <wtay> whoohoo! ChiefHighwater volunteers to port gstreamer to win98se *and* win2kpro !!! [23:52] <ChiefHighwater> hehe [23:52] <mattsm> gstreamer-launch gnomevfssrc location=http://www.gstreamer.net/media/AlienSong.mpg ! mpeg1parse video_00! mpeg_play ! sdlvideosink [23:52] <mattsm> grr I can seem to remeber how to do things, whats wrong here? [23:52] <ChiefHighwater> I hear about you guys breaking code all the time...I can do that 8-] [23:52] <wtay> mattsm: you need a queue between the parser and the decoder [23:53] <mattsm> wtay: oh a queue is required? hrmm [23:53] <wtay> ChiefHighwater: yup...but we fix it eventually too <g> [23:53] <ChiefHighwater> wtay:details, details [23:53] <wtay> mattsm: yup, video_00 is dynamically created [23:53] <mattsm> gstreamer-launch gnomevfssrc location=http://www.gstreamer.net/media/AlienSong.mpg ! queue ! { mpeg1parse video_00! mpeg_play ! sdlvideosink } [23:53] <mattsm> something is still wrong here [23:53] <wtay> yes, very [23:54] <wtay> mattsm: it's in the logs [23:54] <wtay> <wtay> <wtay> mattsm: gstreamer-launch gnomevfssrc location=aliensong.mpg ! mpeg1parse video_00! queue ! { mp1videoparse ! mpeg_play ! sdlvideosink } [23:55] <wtay> write it down this time :-) [23:55] <mattsm> wtay: can you explain why you need mp1videoparse after mpeg1parse? [23:56] <wtay> mpeg_play expects the video frames to be on sync boundaries, mp1videoparse does that [00:00] --- Mon Jun 11 2001 [00:08] mattsm (ma...@ad...) left irc: Ping timeout for mattsm[adsl-216-62-180-169.dsl.hstntx.swbell.net] [00:20] <wtay> hadess: fellowship1 is quite interesting... [00:21] <hadess> yep, is really weird shit, i don't know what the guy used to convert it... [00:21] <wtay> looks sqaushed and all, plays fine with plaympeg though [00:21] <hadess> loops after 30 seconds in plaympeg [00:21] <wtay> squashed even [00:22] <wtay> yup, that's in the stream, no error [00:22] <hadess> oh, so the stream is broken then [00:22] <wtay> yeah, pretty broken I would say [00:22] <hadess> hehe [00:23] <wtay> but we should at least get the size right... [00:23] <wtay> hmm, I got to sleep now :( [00:23] <wtay> cya all [00:23] Nick change: wtay -> wtay-zZz [00:23] <hadess> cya dude [00:32] ChiefHighwater (pa...@su...) left irc: Ping timeout for ChiefHighwater[sub18-226.member.dsl-only.net] [00:32] UnNamed (un...@27...) joined #gstreamer. [00:33] <UnNamed> hi [00:33] <hadess> hey dude [00:33] <UnNamed> I have a doubt, does it support ASF audio now? [00:34] <UnNamed> the version I have has problems with it... aka lots of messages, and no audio [00:34] <hadess> i don't think it does [00:34] <UnNamed> :[ [00:36] BBB (BB...@uc...) joined #gstreamer. [00:36] <hadess> hey BBB [00:36] <BBB> hi hadesss [00:36] <BBB> finally back at my own sweet comp [00:36] <BBB> I hate having to live w/o linux for two weeks [00:37] <UnNamed> linux syndrome ;] [00:38] <BBB> computer-addicted? [00:38] <BBB> or linux addicted [00:39] <UnNamed> unixy, more or less [00:40] seer (se...@ga...) left irc: [x]chat [00:42] <UnNamed> time to go, bye [00:42] UnNamed (un...@27...) left irc: sleep [00:45] hadess (ha...@pc...) left irc: sleep, and nobody to snore [01:16] Action: BBB is away: zzz [01:16] Nick change: BBB -> BBB-zZz [01:24] greg_ (gr...@ho...) got netsplit. [01:24] iGN__ (ig...@lo...) got netsplit. [01:24] BBB-zZz (BB...@uc...) got netsplit. [01:24] derek (de...@cp...) got netsplit. [01:24] big_T (th...@fy...) got netsplit. [01:24] taaz-away (dlehn@66.37.66.32) got netsplit. [01:24] richardb (ri...@ix...) got netsplit. [01:24] iGN__ (ig...@lo...) returned to #gstreamer. [01:24] derek (de...@cp...) returned to #gstreamer. [01:24] greg_ (gr...@ho...) returned to #gstreamer. [01:24] BBB-zZz (BB...@uc...) returned to #gstreamer. [01:25] richardb (ri...@ix...) returned to #gstreamer. [01:25] taaz-away (dlehn@66.37.66.32) returned to #gstreamer. [01:25] big_T (th...@fy...) returned to #gstreamer. [01:26] big_T (th...@fy...) got netsplit. [01:27] big_T (th...@fy...) returned to #gstreamer. [01:27] Kuroyi (ri...@ub...) joined #gstreamer. [01:34] ChiefHighwater (pa...@su...) joined #gstreamer. [02:03] ChiefHighwater (pa...@su...) left irc: [02:24] greg_ (gr...@ho...) left irc: Read error to greg_[home.sente.pl]: EOF from client [04:06] johnix (thomas@61.152.140.57) joined #gstreamer. [04:06] <johnix> Hi [04:46] Nick change: ajbusy -> ajmitch [05:11] arik (ar...@po...) joined #gstreamer. [05:11] <arik> hi [05:11] <ajmitch> hey arik [05:11] <arik> i got fullscreen sort of working [05:12] <arik> need to commit it [05:12] <ajmitch> cool [05:12] <arik> heh [05:12] <ajmitch> bet it doesn't work right here ;) [05:12] <arik> hah [05:13] <ajmitch> requires decent Xv support? [05:13] <arik> heh [05:13] <arik> it's actually pretty gheto how i did it [05:13] <arik> galeon style gdk trips [05:13] <ajmitch> hehe [05:13] <arik> tricks [05:29] Kuroyi (ri...@ub...) left irc: Read error to Kuroyi[ubr-35.82.223.wmelbourne.cfl.rr.com]: EOF from client [05:33] ChiefHighwater (pa...@su...) joined #gstreamer. [05:34] <ajmitch> ello ;) [05:34] <ChiefHighwater> Ello 8-] [05:34] <ajmitch> what's up? [05:35] <arik> bbiab [05:35] arik (ar...@po...) left irc: leaving [05:36] <ChiefHighwater> just got home from Church, lookin for Omega [05:36] <ajmitch> ah, ok ;) [05:37] <ajmitch> was fun walking to church here yesterday, was snowing & icy ;) [05:37] <ChiefHighwater> brrr [05:38] <ajmitch> wasn't too bad inside the church tho, the heaters were on ;) [05:38] <ChiefHighwater> good thing 8-] [05:41] Kuroyi (ri...@ub...) joined #gstreamer. [05:42] <ajmitch> ChiefHighwater: been hacking on gstreamer lately? ;) [05:52] <ChiefHighwater> umm, I don't hack [05:53] <ajmitch> aww, why not? it's fun ;) [05:53] <ChiefHighwater> they haven't found a title for what I do yet [05:53] <ajmitch> ChiefMotivator [05:53] <ChiefHighwater> lol [05:53] <ChiefHighwater> sometimes |
From: <wim...@ch...> - 2001-06-13 04:28:50
|
[06:30] big_T (th...@fy...) left irc: Ping timeout for big_T[fysgr763.sn.umu.se] [06:30] <ajmitch> like are we assured of salvation once we accept Jesus? ;) [06:31] <ChiefHighwater> yes. [06:31] <ChiefHighwater> got another 8-] [06:32] <ChiefHighwater> or maybe you want some backup on my answer? [06:32] Action: taaz has some salivation going on waiting for dinner to cook... [06:32] <ajmitch> give me reasons why... [06:32] <taaz> oh wait... that's not what you meant <g> [06:32] <ChiefHighwater> sure [06:32] <ajmitch> also, another tricky question, why is my computer so slow when running nautilus? ;) [06:33] big_T (th...@fy...) joined #gstreamer. [06:34] <ChiefHighwater> ajmitch: [06:34] <ChiefHighwater> John 10:27-29 [06:34] <ChiefHighwater> 27 My sheep listen to my voice; I know them, and they follow me. [06:34] <ChiefHighwater> 28 I give them eternal life, and they shall never perish; no one can snatch them out of my hand. [06:34] <ChiefHighwater> 29 My Father, who has given them to me, is greater than all; no one can snatch them out of my Father's hand. [06:34] <ChiefHighwater> (NIV) [06:36] <ajmitch> and how does that prove it? [06:37] <ChiefHighwater> If God says he's given you salvation and no one can take you from God's hand, how do plan to lose it? [06:37] <ajmitch> so no matter what someone does once they have accepted, they are still saved? [06:38] <ChiefHighwater> here's another scripture along those lines [06:38] <ChiefHighwater> Rom 8:38-39 [06:38] <ChiefHighwater> 38 For I am convinced that neither death nor life, neither angels nor demons, neither the present nor the future, nor any powers, [06:38] <ChiefHighwater> 39 neither height nor depth, nor anything else in all creation, will be able to separate us from the love of God that is in Christ Jesus our Lord. [06:38] <ChiefHighwater> (NIV) [06:39] <ChiefHighwater> I would not use these scriptures as a license to do whatever I want [06:40] <ChiefHighwater> that was actually an issue addressed by the apostle Paul. Let me look it up [06:40] mattsm (ma...@ad...) left irc: Client Exiting [06:40] <ajmitch> oops, we're scaring away the natives ;) [06:40] <ChiefHighwater> hehe [06:40] <ChiefHighwater> here it is [06:40] <ChiefHighwater> Rom 6:1-2 [06:40] <ChiefHighwater> 1 What shall we say, then? Shall we go on sinning so that grace may increase? [06:40] <ChiefHighwater> 2 By no means! We died to sin; how can we live in it any longer? [06:40] <ChiefHighwater> (NIV) [06:40] <ajmitch> wait, we are the natives in here, we live here ;) [06:41] <ChiefHighwater> if we're buggin folks we can msg each other [06:41] <ChiefHighwater> but i don't hear anyone complainin...yet [06:41] <ChiefHighwater> make good reading for the logs [06:42] <ajmitch> heh [06:42] <ajmitch> pity i only got KJV on here, it's harder to read ;) [06:42] <ChiefHighwater> yup, sure is [06:43] <ajmitch> older english... [06:43] big_T (th...@fy...) left irc: Ping timeout for big_T[fysgr763.sn.umu.se] [06:43] <ChiefHighwater> I think it is safest to say: Christianity isn't about a single moment of your life where you made a commitment to [06:43] <ChiefHighwater> God, but rather [06:44] <ChiefHighwater> a lifetime of choices each day as we seek to walk with Him [06:44] <ajmitch> yes, we can make the choice not to follow at any time [06:45] big_T (th...@fy...) joined #gstreamer. [06:45] <ChiefHighwater> that is called sin, yes [06:46] <ChiefHighwater> here's a relevant passage [06:46] <ChiefHighwater> 1 Cor 3:12-15 [06:46] <ChiefHighwater> 12 If any man builds on this foundation using gold, silver, costly stones, wood, hay or straw, [06:46] <ChiefHighwater> 13 his work will be shown for what it is, because the Day will bring it to light. It will be revealed with fire, and the fire will test the quality of each man's work. [06:46] <ChiefHighwater> 14 If what he has built survives, he will receive his reward. [06:46] <ChiefHighwater> 15 If it is burned up, he will suffer loss; he himself will be saved, but only as one escaping through the flames. [06:46] <ChiefHighwater> (NIV) [06:46] <ajmitch> 1 Corinthians 10 [06:46] <ajmitch> 11 Now all these things happened unto them for ensamples: and they are [06:46] <ajmitch> written for our admonition, upon whom the ends of the world are come. [06:46] <ajmitch> 12 Wherefore let him that thinketh he standeth take heed lest he fall. [06:47] <ajmitch> mmm, KJV english... ;) [06:47] <ChiefHighwater> hehe, np [06:48] <ChiefHighwater> keep going in that text [06:48] <ChiefHighwater> here: [06:48] <ChiefHighwater> 1 Cor 10:11-13 [06:48] <ChiefHighwater> 11 These things happened to them as examples and were written down as warnings for us, on whom the fulfillment of the ages has come. [06:48] <ChiefHighwater> 12 So, if you think you are standing firm, be careful that you don't fall! [06:48] <ChiefHighwater> 13 No temptation has seized you except what is common to man. And God is faithful; he will not let you be tempted beyond what you can bear. But when you are tempted, he will also provide a way out so that you can stand up under it. [06:48] <ChiefHighwater> (NIV) [06:48] <taaz> so uh... what does this have to do with gstreamer? [06:48] <ChiefHighwater> taaz:nothin [06:48] <ajmitch> taaz: preventing the developers from going wacko ;) [06:48] <derek> taaz: it means you better be able to pray [06:49] <ajmitch> hi derek ;) [06:49] <derek> if you have a chance in hell of getting gstreamer to work :) [06:49] <ChiefHighwater> derek:hehe [06:49] <ajmitch> hmm, where is chillywilly? he enjoys these discussions too [07:11] mwc (ma...@ly...) left irc: [x]chat [07:32] Nick change: ajmitch -> ajbusy [07:32] <ChiefHighwater> Happy Birthday ajbusy! [07:54] steveb (st...@no...) joined #gstreamer. [08:02] Nick change: ajbusy -> ajmitch [08:05] Nick change: taaz -> taazzzz [08:09] <ChiefHighwater> omega_dinner:ping [08:10] Nick change: omega_dinner -> omega_ [08:11] <ajmitch> hey, that was a long dinner ;) [08:12] <derek> rofl [08:12] <omega_> yeah, well, watched a movie too [08:12] <derek> omega_: well bummed streaming from board wasnt a huge success [08:12] <derek> but i dont think its gstreamers fault [08:12] <omega_> sound quality problems? [08:13] <derek> um worse [08:13] <omega_> hmm [08:13] <derek> the signal was so faint [08:13] <derek> could barely hear it [08:13] <derek> i tee'd off the tape in [08:13] <derek> and tape was fine [08:14] <derek> but the pc was barely audible [08:14] <derek> so dont know exactly the issue [08:14] <omega_> ah, level problems are the bane of a sound-guy's existence [08:14] <derek> so my big gripe [08:14] walken (fo...@c1...) joined #gstreamer. [08:14] <omega_> btw ajmitch: happy birthday <g> [08:14] <derek> no way in gstreamer to see [08:14] <derek> what the input level is [08:14] <derek> :) [08:14] <derek> until you are done and play it back [08:14] <derek> also i got a depth error [08:14] <omega_> right. need a simple recording app that does this, and includes a levelmeter plugin [08:14] <omega_> yo walken [08:15] <derek> so next week i hope to play some more [08:15] <derek> and pipe off own channel instead of piggy backing [08:15] <ajmitch> thx omega_ ;) [08:16] <omega_> that'll get wtay thinking about getting the tee behaving somehow with capsnego <g> [08:20] <ajmitch> argh, damned net 'radio' stations that use WMP formats!! [08:23] <steveb> would g_mem_chunk make a good basis for a bufferpool? [08:23] <omega_> it's already used for that afaik [08:23] <omega_> but the real point of bufferpools is for hardware buffers [08:23] Nick change: ajmitch -> ajstudy [08:24] <steveb> so should I write my own if I want a float bufferpool? [08:25] <omega_> bufferpool's only function is to provide a specific memory space, has nothing to do with the format at all [08:26] <steveb> I thought the space had a type associated with it [08:29] <omega_> nope [08:29] <omega_> the bufferpool is associated with a pad, which implicitely associates it with a type [08:29] <omega_> but the bufferpool itself is unaware [08:29] <omega_> we need to add basic features to support things like requesting a buffer for a given timestamp, etc. [08:29] <omega_> but that doesn't change the basic idea [08:31] <steveb> using a bufferpool would avoid constant calls to g_new to create new buffers - thats a good thing right? hardware buffers aside [08:31] <omega_> yes [08:32] <omega_> that's why the default uses a memchunk afaik [08:32] <steveb> where is the default one? [08:32] <omega_> it's implicit [08:33] <steveb> but where is the code? [08:33] <omega_> in buffer_new [08:34] <omega_> there may not be a general case memchunk, but.... [08:34] <omega_> consider that memchunk is designed for uniform allocs [08:34] <omega_> and the default is not [08:34] <omega_> so a better solution would be something that creates a generic bufferpool with a memchunk of a given size default [08:35] <steveb> so i'm already using bufferpools without knowing it? [08:35] <omega_> not really [08:35] <omega_> I'm not sure what wtay did to the default case [08:37] <walken> pouet [08:37] <steveb> it seems the buffer comes from a mem_chunk, but not the data [08:37] <omega_> right [08:38] <omega_> buffer itself is always the same size, data isn't [08:39] <steveb> yep, but there could be a real performance boost if the data mem came from a pool [08:40] <omega_> yes, but can only be of use if the pool is tuned to the same size [08:40] <omega_> else it falls back to a malloc [08:41] <steveb> yep, presumably the common case would be the same size [08:41] <omega_> but you can't assume that [08:41] <omega_> and memchunk requires a common size [08:42] <steveb> no, but a really smart pool could adapt to different sizes - when it gets requests for a different size it could start a new pool [08:42] <omega_> that's what malloc() is [08:42] <steveb> and start harvesting old pools when they stop being used [08:43] <steveb> if that is not too muck like reinventing gc [08:43] <omega_> it's better for a plugin like mad to construct a custom memchunk bufferpool to get 4608 bytes at a time, than to add complexity to the default case [08:43] ChiefHighwater (pa...@su...) left irc: [08:44] <steveb> ok, so should I write a bufferpool which only int2float uses, or should I write one which all elements needing new float buffers can use? [08:45] <omega_> it has nothing to do with floats at all [08:45] <omega_> it is a memory allocation mechanism, nothing more [08:45] ChiefHighwater (pa...@su...) joined #gstreamer. [08:45] <steveb> ok, so should I write a bufferpool which only int2float uses, or should I write one which all elements needing new audio buffers can use? [08:45] <omega_> ChiefHighwater: um, don't aim that sniper rifle at your power strip..... [08:46] <omega_> steveb: write a bufferpool constructor that creates one pre-configured with a memchunk of a given size and block-count [08:46] <omega_> bufpool = gst_bufferpool_new_with_size(int size,int count) [08:46] <steveb> where should I put it? [08:46] <omega_> it creates a 'custom' bufferpool that uses the memchunk [08:46] <omega_> in gstbufferpool.c [08:46] Action: walken has an idea for float2int [08:46] <walken> I even think its a good one [08:46] <omega_> mmx/sse ? [08:46] <steveb> omega_: ok [08:46] <walken> yup :) [08:47] <steveb> walken: ? [08:47] <walken> simple mmx actually [08:47] <omega_> double-width with sse2 [08:47] <omega_> walken: huh? [08:47] <omega_> mmx doesn't do floats at all [08:47] <omega_> or are you thinking of fabricating the ieee format in the int regs? [08:47] <omega_> eeew [08:47] <steveb> walken: do you mean for the actual conversion of float 2 int ? [08:48] <walken> well bias the floats so that they are between, say, 3 and 5 instead of being between -1 and 1 [08:48] <walken> yes, float->int conversion [08:48] <walken> no wait [08:48] <walken> make that between 5 and 7 [08:48] <omega_> walken: why not just grab the sign bit and separate it out? [08:48] <omega_> huh? [08:48] Action: omega_ doesn't have any idea what float32 looks like [08:49] <walken> now you know the sign and exponent and everything of your expected floats [08:49] <steveb> I will eventually integrate this snippet to do the conversion: [08:49] <steveb> inline int fi_round(float f) { [08:49] <steveb> f += (3<<22); [08:49] <steveb> return *((int*)&f) - 0x4b400000; [08:49] <walken> and all the information is in the mantissa bits - which are the lower bits if you look at it as an integer [08:49] <steveb> it came from someone on l-a-d [08:49] <walken> hmmm. OK. well that was basically the same idea :) [08:49] <omega_> walken: right, why move the zero point at all? just move the sign bit into the right place and shift the int bits up to meet the decimal? [08:50] <walken> omega. thats harder :) [08:50] <omega_> it's harder than moving the zero point? [08:50] <walken> you have a shift that is controlled by the exponent for starters [08:50] <omega_> how do you plan to re-zero everything afterward? [08:50] <walken> so you cant do lots of them at the same time [08:51] <steveb> is this something that could live in libgstaudio so everyone can use it? [08:51] <omega_> imo it's something that would go in the equiv of libcodec [08:51] <walken> I guess. except you want to do it on whole buffers and it would be faster if the plugin that writes the floats does the bias too. [08:51] <omega_> I'm starting to think that there really should be several specialized libs with a common framework... [08:52] <steveb> yep [08:52] derek (de...@cp...) left irc: Read error to derek[cpe-24-221-112-50.az.sprintbbd.net]: EOF from client [08:52] <omega_> walken: hmmm [08:52] <walken> you can even get rid of the shift at all if you premultiply by the right constant [08:53] <omega_> but float2int and int2float are standalone elements [08:54] <steveb> om :could you elaborate? [08:54] <walken> so all you have left is an mmx substract (gets rid of the sign and exponent bits, and an mmx pack double to word [08:55] <omega_> walken suggested that the code preceeding the conversion do the bias, but that's not possible with int2float or float2int being separate elements, in a general sense [08:55] <walken> no [08:55] <big_T> morning [08:55] <omega_> besides, I'd much much rather be able to specify that pro-audio is float32 with a range of -1 to +1, for simplicities sake... [08:55] <omega_> big_T: yo [08:56] <walken> I think its good in general to have an output volume setting for most plugins [08:56] <walken> so thats the premultiply [08:56] <walken> the bias is harder to justify in general though [08:56] <omega_> walken: right [08:56] <omega_> walken: you familiar with intercept/slope as a terminology for this? [08:56] <walken> no [08:57] <omega_> intercept is the zero point (bias), slope is the range [08:57] <omega_> so normal is 0,1 [08:57] <omega_> you're proposing 6,1, for instance [08:57] <walken> ok [08:58] <omega_> what I don't think would be very efficient in the general case would be for all elements to handle arbitrary intercept/slope on both input and output [08:59] <walken> we dont need it on input at least [08:59] <omega_> but then again having just a intercept/slope converter element wouldn't be very efficient either [08:59] <omega_> walken: true, I guess as long as each output do any s/i, you just match the default input s/i [08:59] <walken> yes [09:00] <steveb> is all this just to optimise the float2int operation? [09:00] <omega_> steveb: do you think that's reasonable in the general case? [09:00] <omega_> steveb: I hope not <g> [09:00] <omega_> I'd much rather standardize on 0,1 [09:01] <steveb> i got lost - what is being proposed? [09:01] <omega_> anyone know where I can read on on the ieee spec float32 follows? [09:01] <omega_> steveb: in general, that one of either the input or the output (as a global std) be capable of using an arbitrary s/i [09:01] <omega_> for any given float32->float32 filter [09:02] <omega_> because the output of any filter can then conform to any input [09:02] <omega_> or vice versa [09:02] <steveb> well the caps are in place to support that [09:02] <walken> omega. 0,1 is natural for input IMHO [09:02] <walken> and 0,level is natural for output [09:03] <walken> and bias,level is only useful when you want to convert to int just after that [09:03] <omega_> right, or you've got a funky element [09:03] <steveb> when you say input&output do you mean for every element in a float pipline? [09:03] <omega_> right [09:03] <omega_> either all the input pads are capable of it, or all the output pads are capable of it [09:04] <omega_> not both [09:04] <walken> omega. you can do it the lazy way too [09:04] <omega_> one can always deal with the variance in the other [09:04] <walken> like 0,1 is standard and everyone supports it [09:04] <omega_> right [09:04] <steveb> so is there an implicit conversion of level->1 between each element? [09:04] <omega_> steveb: no [09:04] <walken> and you have caps for people who want to support other formats [09:04] <omega_> walken: rigt [09:04] <omega_> steveb: assume output must handle any s/i [09:04] <omega_> input can be fixed [09:05] <walken> damn [09:05] <walken> I was so surprised at work [09:05] <omega_> if the fundamental s/i of the two filters don't match, the output will be required to compensate as part of its processing for the downstream filter [09:05] <walken> I had some algorithm that was roughly a brensenham derivative [09:05] <omega_> if all filters do that on the output side, then everyone is happy [09:05] <walken> I was surprised it needed an explanation [09:05] <walken> lol [09:06] <omega_> walken: hmm [09:07] <steveb> I thought it would be the other way around - so that filters that scale as a side-effect can compensate by just setting the output slope. then the input would have to adapt. That was Apples original intention for using s/i [09:07] <omega_> either direction works equally well [09:07] <omega_> but all filters have to work that same direction for the whole concept to work [09:08] <omega_> steveb: but there's no functional difference except the chance that it's fractionally faster on average over many cases to do the scaling pre- or post- filter [09:08] <omega_> if everyone optimizes their code, there should be zero speed difference [09:08] <steveb> in that case I think I prefer elements setting their own output s/i and adapting to any input [09:08] <omega_> then again is this ever going to be a problem?? [09:09] <omega_> imo elements that can't conform to 0,1 are gonna be very very rare [09:09] <omega_> if they're source, just fix the algo [09:09] <omega_> if they're binary, well, imo stick a patch element around them [09:10] <omega_> the cost to do this for the general case is non-trivial (extra code, etc.) [09:10] <steveb> well yeah - good point. also many elements that might do this are ladspa and they standardise on 0,1 [09:10] <omega_> yup [09:11] <omega_> to have zero live cost, you have to have two instances of every filter's code, one that assumes 0,1, the other can handle anything [09:11] <omega_> and use caps to switch between them [09:11] <omega_> but that's way too much work imo to deal with the pathological case [09:11] <steveb> yeah - lets stick with 0,1 for now and wait for a real reason to change [09:14] <steveb> i really should go to work [09:14] <omega_> we can use unpack and CVTDQ2PS for int2float [09:15] <omega_> movq, unpack word to doubleword into xmm, cvtdq2ps, movdq [09:15] <steveb> i would agree if i knew what you were talking about <g> [09:15] <omega_> handles 4 samples at a time, in 4 insns [09:15] <omega_> walken has an idea what I'm talking about, I hope <g> [09:15] <steveb> sounds good [09:16] <steveb> l8er [09:16] Nick change: steveb -> steveb-away [09:20] <omega_> great. waited a couple days and now the price for the ticket to OLS has gone up $200 [09:25] <omega_> I'm gonna go to a local travel agent tomorrow morning and see what they can dig up [09:36] <walken> I still have to get my ticket too [09:36] <omega_> better *hurry* [09:36] <walken> I guess [09:36] <omega_> it was $286, it's now $422+ [09:36] <walken> eeeeeeeeeeeeeeeeeeeeeeek [09:36] <omega_> right, hurry [09:36] <omega_> check with a local agent first if you can [09:37] <omega_> I've got one I trust [09:37] <walken> omega. well no I dont know the ss2 instructs so I have no idea what you were talking about :) [09:37] <omega_> oh..., look it up ;-) [09:38] <walken> I most usualy go with cheaptickets [09:38] <omega_> travelocity gave me this price [09:38] <omega_> checking cheaptickets [09:39] <omega_> signing away my firstborn again... [09:40] <omega_> ah, neat [09:41] <omega_> delta and continental can both do it for $249.40 [09:41] <omega_> no surcharges and pfc's included though [09:41] <omega_> but lots better than $450 [09:41] <walken> heh [09:42] <omega_> except delta is sold out [09:42] <omega_> continental has one [09:43] <omega_> odd. they say "do NOT use a debit card" [09:43] <omega_> that's all I've got, maybe I'll use my parents or somesuch [09:43] <omega_> as soon as I figure out why they say that [09:44] <walken> heh. I got that problem at a car rental once [09:44] <walken> for the warranty [09:44] <omega_> wonder if I should just get this, or wait till morning and see what the local agent can dig up? [09:44] <walken> I have a card that does ATM and credit - and I couldnt use it for warranty as their dumb machine would then debit on it [09:45] <omega_> neat [09:45] <walken> so I went somewhere else [09:45] <walken> heh [09:45] <omega_> wow. would be on 4 different plane types for 4 different hops [09:45] <walken> haha [09:45] <walken> looks like your flight is not very simple [09:45] <omega_> 737-800, ERJ Jet, Saab 340 turboprop, and 757 [09:45] <walken> I dont think I'll have to do more than 2 hops. hopefully. [09:45] <omega_> yikes [09:46] <omega_> pdx -> ewr(newark) -> yow [09:46] <omega_> yow -> dtw(detroit) -> pdx [09:46] <walken> oh, with the return flights you mean [09:46] <omega_> no, those are all four flights, two in two out [09:47] <omega_> yeah, I wouldn't get a price like this on 4 hops each way.... [09:47] <omega_> more specifically, I wouldn't do it [09:47] <walken> I guess :) [09:47] <omega_> in ottawa 24th at 6:30pm, leaving 7:35pm the 29th [09:48] <omega_> home 12:39am the 30th [09:48] <walken> oh right you fly from .de ? [09:48] <omega_> huh? [09:48] <walken> no :) [09:48] <omega_> pdx == portland, OR [09:48] <walken> I keep being confused :) [09:48] <walken> its guenter thats in .de [09:48] ChiefHighwater (pa...@su...) left irc: Ping timeout for ChiefHighwater[sub18-226.member.dsl-only.net] [09:48] <walken> lol [09:48] <omega_> taazzzz has it easy, he's a lot closer to ottawa [09:49] <omega_> no tz switches, etc. [09:49] <walken> last year I met a cute girl in the plane [09:49] <walken> too bad I was not single back then [09:49] <omega_> lucky yo [09:49] <walken> or something [09:49] <omega_> er, you [09:50] <walken> we even had to go to the hotel because of a flight delay [09:50] <walken> I cant say I wasnt tempted :) [09:50] <omega_> mu [09:53] <omega_> you think I should just buy this ticket, or wait to talk to an agent tomorrow morning? [09:53] <walken> I'm not really an expert there :) [09:54] <walken> what I always do is that I call cheaptickets and buy whatever they tell me [09:54] <walken> (no, this is not paid advertisement :) [09:58] <omega_> bleagh, I think I'll just get it [09:58] <omega_> there's basically zero chance they can beat this price [09:58] <omega_> except they want a credit card [10:00] <omega_> I have no choice though, since I can't use someone else's card [10:00] <omega_> neat [10:04] <omega_> great. my binutils doesn't understand sse [10:14] arik (ar...@po...) joined #gstreamer. [10:14] <arik> hi all [10:14] <omega_> yo [10:14] <omega_> you going to ols? [10:14] <arik> don't know [10:15] <arik> i doubt it ;-( [10:15] <omega_> I'm arranging stuff now [10:15] <omega_> before the ticket prices get even more obscene [10:15] <arik> oh [10:15] <arik> yeah [10:15] <arik> i got fullscreen working right [10:15] <arik> still done wrongly though [10:16] <omega_> ooh, I can save $0.13 by using expedia instead <g> [10:16] <arik> heh [10:19] <omega_> oh well, attempting via cheaptickets.com, if that fails, I'll use expedia [10:19] <arik> heh [10:20] <omega_> something about debit cards not working in some cases [10:20] <arik> ek [10:23] Action: omega_ goes to ridgerun.com to look up his work phone [10:23] <arik> hahaha [10:23] steveb (ste...@21...) joined #gstreamer. [10:24] <arik> they don't have a blurb about you on the site do they? [10:24] <omega_> also xfering money to make sure that my debit card has enough balance to handle $260 for the conference, $290 for the ticket, etc. [10:24] <omega_> no, I'm not speaking [10:24] <arik> heh [10:24] <omega_> oh, on rr.com? [10:24] <arik> no i mean on the ridgerun site [10:24] <omega_> they did [10:25] <arik> oh [10:25] <omega_> before the new site went up [10:25] <arik> heh [10:25] <omega_> it pointed to the /. article [10:25] <omega_> complete with a thumbnail pic of me and everything ;-) [10:25] <arik> heh [10:25] <omega_> oops, yeah, xfering money would be a good idea.... $370 in there [10:25] <arik> heh [10:26] <omega_> and the credit union tech supt still hasn't figured out what I suggested: put the *names* of the accounts next to the numbers in the drop-downs for quick-transfer [10:26] <arik> hehe [10:26] <omega_> I have to visually xfer the numbers to the names [10:27] <omega_> there, registered for ols [10:27] thomas (th...@21...) joined #gstreamer. [10:27] <arik> neet [10:28] Nick change: thomas -> thomas-morning [10:28] <thomas-morning> hi [10:28] <omega_> yo [10:29] <omega_> doh. website claims the ticket is 41.50 and the govt taxes are $250 [10:29] <omega_> pfff, that's a new one [10:29] <arik> hahahahaahaha [10:29] <omega_> it mangled the first digist, swapped them [10:29] <omega_> er, digits [10:40] <walken> hehe [10:41] Action: omega_ watches a test by Carmack and friends of various materials not to wear while messing with hydrogen peroxide.... [10:41] <omega_> moral of the story: DON'T wear leather [10:42] <walken> hmmm [10:42] <walken> where's that ? [10:42] <thomas-morning> that's a good average life moral as well imo ;) [10:42] <omega_> hehe [10:42] <walken> wow. I found a wierd bug in ac3dec. [10:42] <walken> doesnt trigger actually, but doesnt work as designed :) [10:42] <walken> lol [10:42] <omega_> the shoe they put it on spontaneously burst into flame after a little over a minute [10:42] <ajstudy> hmm, is there midi playback support yet? i remember it being discussed... [10:43] Action: walken doesnt deal with peroxyde often, but I'll remember :) [10:43] <omega_> only via the xmms wrapper [10:43] <omega_> but that works better than xmms itself does, on my machine [10:46] <arik> night all [10:46] <thomas-morning> arik: bye [10:47] <omega_> l8r [10:47] arik (ar...@po...) left #gstreamer. [10:47] <ajstudy> heh, ok, i should try that then, if i have the xmms wrapper working ;) [10:47] <ajstudy> hmm, last time it got compiled was in march, i guess i need to install the xmms stuff... [10:48] <thomas-morning> omega_: I'm looking at how to tie caps to a sort of GST_FRAME_BYTE_SIZE macro ... [10:48] <walken> eeeek ! my / is filling up, again [10:48] <thomas-morning> ... should this be a macro or function in mime type factories ? [10:48] <thomas-morning> ... or should I just only do it for audio right now ? [10:48] <omega_> I'd do it only for audio atm [10:49] <omega_> walken: stop typing then <g> [10:49] <omega_> wow. /me is glad he moved some money between accounts [10:49] <omega_> I got twice as much interest in one account as another on roughly the same amount of $ [10:51] <walken> hmmmm, its because of my local cvs repository of test streams and reference results [10:51] <walken> binary cvs repositories get big [10:52] <omega_> yup [10:53] <omega_> yuck [10:53] <omega_> my stocks are worth 1/27th of their peak at the height of linux stock mania [10:53] Action: omega_ gets depressed [10:54] <walken> heh [10:55] <omega_> well, /me will have rr stock options as soon as /me finds and signs the forms [10:55] Action: omega_ hopes there's a friends&family thing when the rr IPO happens <g> [11:01] <walken> heh [11:02] Action: walken never got any of the linux stock bonuses [11:02] <omega_> I got some cash out of RHAT before it tanked [11:03] <omega_> but even with that I'm still overall down from my initial $10k total ;-( [11:03] <walken> huh [11:04] <omega_> but I have most of almost 8mo of nice salary in the bank, now, which I didn't then [11:04] <omega_> oh wait, not most of <g> [11:05] <walken> 8 month ? thats really alot [11:05] Action: walken doesnt have 8 month in advance like that :) [11:05] <omega_> well, 3.5mo at digital mercury, then however long I've been at rr [11:05] <omega_> at least half of that is in the bank, I think [11:05] <walken> wow [11:06] <walken> I cant save that much [11:06] <walken> 25% maybe (best case) [11:06] <omega_> and with my current livig and hw expenses, I could live for many years off that <g> [11:06] <omega_> helps to live with parents with low rent <g> [11:06] <walken> not with what I bought this weekend :) [11:06] <omega_> dare I ask? [11:06] <walken> oh nothing special really [11:06] <omega_> just $$$ [11:06] <omega_> <g> [11:07] <walken> just lots of stuff I should have bought before [11:07] <omega_> um, that was informative <g> [11:07] <walken> small furniture that adds up [11:07] <omega_> ah [11:07] <omega_> no Ikea? [11:07] <walken> no ikea [11:07] <omega_> doh [11:07] <walken> actually thats about the only shop I didnt go yet :) [11:07] <omega_> heheh [11:07] <walken> for this round of furniture [11:11] <walken> ok, gotta sleep I guess :) [11:12] <omega_> good idea [11:12] <walken> no, you're not allowed [11:12] <omega_> um [11:13] <walken> go fix more bugs and optimize mpeg parsing :) [11:13] <omega_> watch me <g> [11:13] <omega_> heh [11:13] <omega_> trust me, the conversations at OLS will be very very deep into that [11:13] <walken> (did you got any results with that ?) [11:13] <omega_> I have some bizarre ideas [11:13] <omega_> 've been stalled for a few days [11:13] <walken> hehe [11:13] <walken> speed optimization sucks [11:13] <walken> you can do it for weeks and get 0.1% [11:13] <omega_> sick over the weekend, and have work stuff to do [11:14] <walken> and then someday you have a good idea and you get 20% :) [11:14] <omega_> yup [11:14] <omega_> it's almost a game of chance ;-( [11:14] <walken> hmmmm [11:14] <walken> so basically I'm still king of the hill ? :) [11:14] <omega_> haven't complete/timed my hierarchical method yet [11:14] <omega_> but so far, yes [11:15] <walken> yay [11:15] <walken> now I can sleep in peace :) [11:15] <omega_> pff [11:15] <walken> :p [11:15] <omega_> l8r [11:15] <walken> cya [11:15] walken (fo...@c1...) left irc: l8r [11:15] omega_ (om...@om...) left irc: killall -SLEEP omega_ [11:50] steveb (ste...@21...) left irc: quit has steveb, yes [12:02] johnix (thomas@61.152.140.57) left irc: Client Exiting [12:21] ajstudy (aj...@p3...) left irc: Ping timeout for ajstudy[p31-max7.dun.ihug.co.nz] [12:21] ajstudy (aj...@p3...) joined #gstreamer. [12:32] hadess (ha...@de...) joined #gstreamer. [12:32] <hadess> hi guys [12:43] <ajstudy> hi [12:47] <hadess> hey aj [12:58] hadess (ha...@de...) left irc: Ping timeout for hadess[dekker-ext.frontwire.com] [13:01] hadess (ha...@de...) joined #gstreamer. [13:02] <ajstudy> wb [14:06] Nick change: ajstudy -> ajzzzz [14:47] ajzzzz (aj...@p3...) left irc: Ping timeout for ajzzzz[p31-max7.dun.ihug.co.nz] [14:48] ajzzzz (aj...@p3...) joined #gstreamer. [15:14] ajzzzz (aj...@p3...) left irc: Ping timeout for ajzzzz[p31-max7.dun.ihug.co.nz] [15:38] Uraeus (csc...@c2...) joined #gstreamer. [15:38] iGN___ (ign@216.167.121.81) joined #gstreamer. [15:39] thomas_ (th...@21...) joined #gstreamer. [15:39] <Uraeus> thomas-morning: awake? [15:39] <Uraeus> thomas_? [15:39] thomas-morning (th...@21...) got netsplit. [15:39] iGN__ (ig...@lo...) got netsplit. [15:40] <hadess> hey Uraeus [15:40] <Uraeus> hi hadess! [15:40] ajzzzz (aj...@p1...) joined #gstreamer. [15:40] <thomas_> Uraeus: yeah I'm here [15:41] <Uraeus> thomas_: think you latest checkins broke the build [15:41] <thomas_> hi all [15:41] <thomas_> Uraeus: you sure ? where does it break ? [15:41] <Uraeus> the mad plugin [15:41] <Uraeus> nome-xml -g -O6 -Wall -I../../.. -I../../../include -Wp,-MD,.deps/gstmad.pp -c g [15:41] <Uraeus> stmad.c -fPIC -DPIC -o gstmad.lo [15:41] <Uraeus> gstmad.c: In function `gst_mad_init': [15:41] <Uraeus> gstmad.c:149: structure has no member named `framestamp' [15:41] <Uraeus> gstmad.c: In function `gst_mad_input': [15:41] <thomas_> Uraeus: what's the error ? [15:41] <Uraeus> gstmad.c:193: structure has no member named `framestamp' [15:41] <Uraeus> gstmad.c: In function `gst_mad_output': [15:41] <Uraeus> gstmad.c:270: structure has no member named `framestamp' [15:41] <Uraeus> gstmad.c:273: structure has no member named `framestamp' [15:41] <Uraeus> make[4]: *** [gstmad.lo] Error 1 [15:42] <thomas_> Uraeus: you're right... hang on, let me change that [15:42] Nick change: Uraeus -> tinderbox [15:42] <thomas_> uraeus: ok, check out gstmad.h again [15:43] Nick change: tinderbox -> Uraeus [15:43] <Uraeus> ok [15:50] iGN__ (ig...@lo...) got lost in the net-split. [15:50] thomas-morning (th...@21...) got lost in the net-split. [15:58] steveb (ste...@21...) joined #gstreamer. [16:00] ajzzzz (aj...@p1...) left irc: Ping timeout for ajzzzz[p1-max6.dun.ihug.co.nz] [16:01] Uraeus (csc...@c2...) left irc: Ping timeout for Uraeus[c224s9h5.upc.chello.no] [16:26] hadess (ha...@de...) left irc: Client Exiting [16:29] steveb (ste...@21...) left irc: quit has steveb, yes [17:57] Nick change: taazzzz -> taaz [17:58] dobey (do...@dr...) joined #gstreamer. [18:15] ChiefHighwater (pa...@te...) joined #gstreamer. [18:23] Nick change: steveb-away -> steveb [18:33] Nick change: ChiefHighwater -> CHW-meeting [18:56] <steveb> th:ping [18:56] <steveb> thomas_: ping? [19:12] Nick change: dobey -> dobeylunch [19:33] Nick change: dobeylunch -> dobey [19:53] greg_ (gr...@ho...) joined #gstreamer. [20:09] Nick change: CHW-meeting -> ChiefHighwater [20:44] <greg_> Hi ! Have you looked at this ? http://www.linuxpower.org/display.php?id=211 ??? [20:44] <greg_> The guy is talking about X, then at the end he goes into some audio-video/remote stuff... [20:45] <ChiefHighwater> yup. christian was in here and gave a prerelease [20:46] <greg_> He says he doesn't like any existing sound servers... "I lump all existing audio servers I've looked at (there are one or two I've not looked at) into this category, for one reason or another." [20:47] <ChiefHighwater> both Wim and Omega had various comments..should be in the logs, somewhere [20:47] <ChiefHighwater> lemme see if i can find it [20:49] <ChiefHighwater> was awhile back [20:52] <ChiefHighwater> into last month..still lookin [20:59] <ChiefHighwater> ahha [20:59] <ChiefHighwater> 6/6 [21:00] <greg_> ChiefHighwater: what are you talking about ? [21:01] <ChiefHighwater> Uraeus came in here on 6/6 and gave a url with that interveiw on it [21:01] <ChiefHighwater> Omega and Wim both had comments [21:01] <ChiefHighwater> they're in the logs for that date [21:02] <greg_> ChiefHighwater: what is "6/6" ? how to get to those comments ? [21:02] <ChiefHighwater> June 6th [21:03] <ChiefHighwater> gst-daily mail list archives [21:03] <ChiefHighwater> time index 23:44 [21:04] <ChiefHighwater> you see it? [21:04] <ChiefHighwater> here's a direct link [21:04] <ChiefHighwater> http://www.geocrawler.com/lists/3/SourceForge/12388/0/5909684/ [21:05] <greg_> thanks - going to [21:06] <ChiefHighwater> best comment, imo, is at 23:57 [21:08] iGN___ (ign@216.167.121.81) got netsplit. [21:08] richardb-away (ri...@ix...) got netsplit. [21:08] richardb-away (ri...@ix...) returned to #gstreamer. [21:08] iGN___ (ign@216.167.121.81) returned to #gstreamer. [21:09] <ChiefHighwater> Sopwith gets into it also at 00:30ish [21:10] iGN___ (ign@216.167.121.81) got netsplit. [21:10] richardb-away (ri...@ix...) got netsplit. [21:10] richardb-away (ri...@ix...) returned to #gstreamer. [21:10] iGN___ (ign@216.167.121.81) returned to #gstreamer. [21:10] Nick change: wtay-zZz -> wtay [21:10] <wtay> yo [21:11] <greg_> yo [21:14] omega_ (om...@om...) joined #gstreamer. [21:14] <wtay> hi [21:14] <omega_> yo [21:14] <greg_> hi [21:15] <wtay> ../tools/gstreamer-launch disksrc location=/opt/data/armageddon1.mpg ! mpeg1parse sync=false video_00! queue ! { mp1videoparse ! mpeg_play ! winenc compression=DIV3 bitrate=100 quality=1000 keyframe=50 \!video_%02d avimux ! disksink location=test.avi } [21:15] <omega_> that works? [21:15] <wtay> yeah, pretty cool [21:16] <omega_> wow [21:16] <wtay> looks like the winenc crashes on big sizes though [21:16] <wtay> somewhere in pe_loader: This_Is_A_Hack :-) [21:17] <omega_> what's the \!video_%02d thing? [21:17] <wtay> a request pad [21:17] <omega_> why the \ ? [21:18] <wtay> 'cause bash thinks !video is an event or something [21:18] <omega_> oh [21:18] <omega_> that's why the code mashes the cmdline together and reparses it: so you can wrap it in ' ' [21:18] <wtay> I encoded a little mpeg movie to 200bps <g> [21:18] <omega_> whoops [21:19] <wtay> it's still viewable [21:19] <wtay> even 100bps is [21:19] <omega_> div3 ? [21:19] <wtay> yeah [21:19] <wtay> I've got about 2 minutes of video in 2MB :) [21:20] <omega_> whoah [21:20] <omega_> what framesize? [21:20] <wtay> 352x288 [21:21] <wtay> I should upload it [21:21] <wtay> lemme do Smirnoff [21:22] <omega_> wtay: doh [21:22] <omega_> I just realized something. does the colorspace filter build at all without Hermes? [21:23] <wtay> nope [21:23] <omega_> whoops [21:23] <omega_> that must be fixed [21:23] <omega_> colorspace is needed for yuv->rgb even if hermes isn't around for rgb->rgb [21:23] <wtay> yup [21:24] <wtay> Smirnoff is too much action... it drops frames to meet the bitrate then.. [21:25] <omega_> ok, flight to OLS purchased [21:25] <wtay> it did compress to 10:1 though :-) [21:25] <omega_> with zero loss of visual quality, right? <g> [21:25] <wtay> not at all [21:28] Action: taaz needs to buy an OLS flight... or decide to drive [21:29] <omega_> eesh [21:30] <wtay> 3:1 is sorta acceptable.. [21:46] ajmitch (aj...@p3...) joined #gstreamer. [21:49] <wtay> yo [21:49] <ajmitch> hi [21:50] <ajmitch> fscking computer had locked up - memory issues with 2.4.5 [21:50] <omega_> leak->oom, or just insanity? [21:50] <steveb> it would be nice if audio elements had access to some global settings like default sample rate, buffer size (in frames) etc. where could this settings api live? I was thinking libgstaudio [21:50] <wtay> yup, 2.4.5 sucks [21:50] <omega_> that's the apps' job [21:51] <ajmitch> omega_: dunno, but X was basically non-reponsive, so i used the alt-sysrq keys to kill it [21:52] <steveb> but if these settings aren't set explicitly in the elements, it would be nice if the elements chose default values that matched. This would need a settings api for the elements to use [21:52] <omega_> steveb: I think the defaults should be CDish, and the real values will percolate through the pipeline anyway [21:53] <omega_> wtay: that's a good test case: create a pipeline of audio elements and set the caps on one end [21:53] <wtay> omega_: ? [21:53] <steveb> how about buffer size? [21:53] <omega_> watch the capsnego go all the way through the pipeline [21:53] <omega_> steveb: part of caps [21:54] <wtay> omega_: you tried that? [21:54] <omega_> no [21:54] <steveb> so you think we should add buffersize to all audio caps? [21:55] <omega_> dunno, maybe as an optional property? [21:56] Nick change: ajmitch -> ajbusy [21:56] <steveb> how about we use GConf and let elements get some default values from that? [21:57] <omega_> no, I don't think that's the element's job at all [21:57] <omega_> the application's responsibility is to set up the pipeline as it needs to be [21:57] <omega_> capsnego will handle most of those cases automatically, the rest are the apps problem [21:58] <wtay> you can use gst_element_load_thyself to set the properties from an XML file in the app [21:59] <wtay> you can create a little control panel to adjust the defaults [21:59] <omega_> but that's still part of the app [21:59] <wtay> of course [21:59] <wtay> but it's not hard [22:00] <steveb> i guess so [22:00] <wtay> you can't expect the core to load "foobar.XML" and set the apps settings.. [22:01] <wtay> s/apps/plugins [22:03] <steveb> i was just thinking of a simple defaults api for elements to use. It would be up to the app to populate the defaults [22:03] greg_ (gr...@ho...) left irc: [x]chat [22:03] <omega_> we don't need one more API ;-) [22:03] <omega_> just pick some defaults as part of the pro-audio type document, and leave it at that [22:04] <omega_> let the app handle the complexity of picking formats and such [22:04] <steveb> but new APIs are the solution to every problem :) [22:04] <omega_> um [22:04] <omega_> remember, I'm trying to fit this thing into 50KB <g> [22:04] <wtay> removing API is a more interesting problem [22:05] <steveb> it wouldn't need to be anything more than a global GObject with some get/set args [22:05] <wtay> uhm... [22:06] <steveb> singleton, not global [22:06] <wtay> If I would add it, it would be like when an element is instantiated, a /etc/gstreamer/<elementname>.xml is loaded with the settings [22:07] <omega_> that doesn't seem too interesting, really... [22:08] <wtay> why not? [22:08] <omega_> what's that gonna do that can't be coded into the element as a default up front? [22:08] <omega_> is anyone really gonna change the default params on a global elementname basis? [22:08] <wtay> you can create a gui to edit the xml [22:08] <omega_> right...? [22:08] <omega_> why would I want to set the default bitrate of an mpeg encoder? [22:09] <wtay> you can also have in the users home dir [22:09] <omega_> the app is *always* gonna override it anyway [22:09] <omega_> these things are app-specific settings, always [22:09] <wtay> I tend to agree... [22:10] <wtay> but since steveb brough up the idea I thought he had a good reason to make something global.. [22:10] <wtay> that's what defaults are all about, right? [22:13] <steveb> brb - doing dishes (hard to type with wet rubber gloves) [22:14] <omega_> not recommended, either [22:16] <wtay> btw, CVS gstplay gracefully fallback when colorspace is not found.. [22:22] <steveb> bask [22:22] <steveb> back [22:24] <steveb> ok, I can see instances where an app doesn't know about all of its elements and their quirks and it might neglect to set a sample rate or preferred buffersize through the element's args... [22:24] <omega_> then that's either the app's or the element's problem, imo [22:24] <steveb> but each element will know that there is a default value for certain things so can just get that at each null_to_ready transition [22:25] <omega_> if it fails to set the samplerate somewhere, that's a bug in one of the elements, or the app [22:25] <omega_> right, the default should be hardwired into the pro-audio conventions document [22:26] <steveb> bytestoread and buffersize are the same thing - we would have to document some standards for arg introspection to work [22:26] <omega_> yeah [22:26] <omega_> bytesperread is a really stupid name I came up with, I want to replace it <g> [22:27] <steveb> i suppose i could live with it if args were standardised [22:28] <omega_> args have to be standardized, esp in the pro-audio/video case [22:29] <steveb> anyway, slightly different subject. in int2float I am currently taking the same buffersize that comes from the int pad - but am thinking of always creating the new float buffer using a size set as an arg (this will make bufferpool easier). good idea? [22:31] <omega_> depends [22:31] <omega_> it can introduce extra latency [22:32] <steveb> if the int pad data is interleaved then there is a bit of buffer munging going on anyway - biggest impact might be on mono [22:33] <omega_> no, not processing time [22:33] <omega_> if you have buffers of 6KB coming in [22:33] <omega_> and 4KB going out [22:33] <omega_> you'll get b.bb b.bb b.bb [22:34] <omega_> where each . is latency where the last 2KB of the prev incoming buffer is held up waiting for the first 2KB of the next buffer = latency [22:35] <steveb> so latency can be introduced if each _chain call doesn't result in a _pad_push? [22:35] <omega_> yes [22:35] <omega_> because the second incoming buffer will have some time delay [22:36] <steveb> well, that answers my question [22:36] <omega_> and that delay will push the second of three outgoing buffers back in time, meaning you have extra latency *and* jitter [22:36] <omega_> but you could still reframe it, as an option [22:37] <steveb> i'll try using the bufferpool in a way which copes with varying buffer sizes first [22:37] Zeenix (bundanache@61.11.234.86) joined #gstreamer. [22:38] <Zeenix> hello [22:38] <omega_> yo [22:38] <steveb> hi [22:38] <Zeenix> wtay: hello [22:39] <Zeenix> where is Uraeus today [22:41] <steveb> time for sleep [22:41] steveb (st...@no...) left irc: [x]chat [22:41] <omega_> haven't seen Uraeus today [22:42] <Zeenix> omega_ : are you omega( Erik ) ? [22:44] <omega_> I think so ;-) [22:44] Action: omega_ looks in the mirror [22:44] <omega_> yup <g> [22:45] <Zeenix> omega_ : what if i want to be a member of gstreamer devels team [22:45] <omega_> just start hacking on something <g> [22:46] <wtay> yo Zeenix [22:46] <Zeenix> wtay: what you all mean by 'yo' ? [22:47] <omega_> shortening of 'hello' [22:47] <Zeenix> wtay: yo [22:49] <Zeenix> wtay: here i am in karachi, many time away from my Linux PC sitting on Win98 [22:50] <wtay> haha [22:51] <Zeenix> i thought i'll install linux here but this PC has got a 1.2GB harddisk with not enough free space [22:52] <Zeenix> so sad, i wont be able to do anything for ... [22:53] <Zeenix> wtay: bi for today. have nothing to discuss today [22:53] <wtay> right, cya [22:53] Zeenix (bundanache@61.11.234.86) left irc: [22:55] <ajbusy> bah, yet another exam to sit today... [22:55] <ajbusy> what can i do on gstreamer tomorrow afternoon? [22:56] <omega_> stuff <g> [22:58] <ajbusy> stuff all, you said? ;) [22:58] <omega_> heh [23:05] <ajbusy> bbl, time to fry brain ;) [23:05] Action: omega_ hands ajbusy some cooking oil [23:16] thomas (th...@ur...) joined #gstreamer. [23:17] <wtay> hi [23:17] <thomas> evening ! [23:18] <thomas> I have another question about timestamps... two actually. [23:18] <thomas> 1) it seems to me like mad sent out the same timestamp (from the input buffer) for every output buffer it generated from that input buffer ... [23:18] <thomas> is that the case or am I reading the code wrong ? [23:18] <omega_> it would be wrong code [23:19] <wtay> almost right.. [23:19] <thomas> ok, I'll have to double-check. Anyway, it didn't seem that that part of the code was working. [23:19] <thomas> second question : shouldn't there be a way to send along a timestamp when EOS hits ? [23:20] <wtay> why? [23:21] <omega_> could be... [23:21] <thomas> well ... it's not absolutely necessary but it would be nice for plugins down the line to know what time the last bit of data should be sent out [23:21] <omega_> makes sense, mark the end of the stream, which won't be at the last timestamp, it'll be after the last timestamp [23:22] <thomas> I was just thinking about how to do that since timestamps now are tied to the buffers [23:22] <omega_> the events may have timestamps too [23:22] <thomas> omega_: ok, cool, that would fix it [23:22] <thomas> is that implemented with a macro already ? [23:23] <omega_> which? [23:23] <thomas> event timestamps ? [23:23] <omega_> no, there are no events yet at all [23:23] <thomas> omega_: oh, so you mean event in a bigger sense than signals then ? [23:23] <omega_> yes [23:24] <thomas> ok, so I suppose the best thing would be to give that some thought together with the eos problems ? [23:24] <omega_> yup [23:29] <omega_> whee, gstobject.c compiles with gobject! [23:30] <ChiefHighwater> woohoo [23:30] <omega_> 1 down, 28 to go ;-( [23:30] <wtay> cool [23:30] Action: thomas is away - Automatically set away. - messages will be saved. [23:30] Uraeus (csc...@c2...) joined #gstreamer. [23:30] <Uraeus> hello [23:30] <wtay> hi [23:30] <omega_> yo [23:31] <Uraeus> heh, seems my Jim Gettys interview just got slashdotted :) [23:31] <omega_> yup [23:31] <omega_> I'm getting some HTTP noise in the middle of the interview page [23:32] <omega_> ntent-Transfer-Encoding: binary [23:32] <omega_> just above "Talking with..." [23:32] sweeze (jba...@sw...) joined #gstreamer. [23:33] <omega_> yo [23:33] <Uraeus> omega_: I see it, must be the add banner misfunctioning [23:33] <omega_> nope, it's in the html source [23:33] <omega_> <TD WIDTH=85%> [23:33] <omega_> ntent-Transfer-Encoding: binary [23:33] <omega_> <head> [23:33] <omega_> and a little detail: you have nested <HEAD>'s [23:34] shitowax (yann@AAnnecy-101-1-1-102.abo.wanadoo.fr) joined #gstreamer. [23:34] <Uraeus> omega_: hmm, well not much to do with it, Linuxpower is very probably shutting down soon [23:34] <omega_> uh? [23:34] <shitowax> yo [23:34] <wtay> yo [23:35] <Uraeus> omega_: well the origianl creators aren't intereted anymore and I don't feel I need a site of my own, so I will contant some other sites and check out if they are interested in me writing for them [23:35] <omega_> hmm [23:35] <Uraeus> s/contant/contact/ [23:35] <wtay> shitowax: I promise, tomorrow I'll look at QT :-! [23:36] <wtay> Uraeus: 10 commits today, can you match that? <g> [23:36] <Uraeus> wtay: hmm ;) [23:37] <omega_> Uraeus: if we get a bridge between gst cvs and gnome cvs, do we then end up on the top of the gnome weekly commit count? [23:37] <omega_> that's enough reason right there <g> [23:37] <Uraeus> omega_: yup :) [23:37] <wtay> yeah :) [23:37] <wtay> I got to sleep now :( [23:37] <wtay> cya all [23:37] <omega_> wimp [23:37] Nick change: wtay -> wtay-ZzZ [23:37] <omega_> it isn't even midnight yet [23:38] <Uraeus> omega_: but after seing the galeon solution I am not very positive :( [23:38] <omega_> hmm [23:38] <wtay-ZzZ> omega_: enough stuff screwed up for today :-) [23:39] <Uraeus> omega_: I feel that the Galeon solution doesn't really buy us much [23:39] <sweeze> so i'm sure this has been asked many times, but approximately how long unitl .deb releases? [23:39] <Uraeus> sweeze: as long as it takes our laze deb makers to catch up to their super effective RPM brethren :) [23:39] <Uraeus> s/laze/lazy/ [23:40] <sweeze> Uraeus : :) [23:40] Action: omega_ pings taaz with a nerf gun [23:40] Action: Uraeus shows is nice RPMS to taaz and smiles in triumph [23:41] <sweeze> that's what i figured... i just though perhaps said deb makers were here and could give an estimated target [23:41] <taaz> who what where? [23:41] <Uraeus> sweeze: as you can see who what where is all they can say :) [23:41] <thomas> night all [23:41] thomas (th...@ur...) left irc: I'm outta here! [23:42] <shitowax> bye [23:42] shitowax (yann@AAnnecy-101-1-1-102.abo.wanadoo.fr) left #gstreamer. [23:43] <Uraeus> taaz: any debs manifesting themselves anytime soon? :) [23:43] <taaz> yeah yeah... i've been trying to put out other fires recently [23:43] <taaz> i'll try to get that done tonight [23:43] <omega_> bigger fires? [23:44] Action: omega_ throws gasoline onto the debs [23:44] <Uraeus> wow, according to our statistics my RMPS has increased our download rather drastically [23:45] <Uraeus> before: 36 d/l a day; after 816 d/l a day :) [23:45] <omega_> I think that also had something to do with 0.2.0 [23:45] <omega_> maybe [23:45] <taaz> bunch of lazy users... ;) [23:45] <sweeze> can gstmediaplay read fromDVD? [23:46] <omega_> nope [23:46] <sweeze> is video output plugin configureable? i haven't seen any options for such [23:46] <omega_> we have several, but haven't done anything to gstmediaplay yet to let you pick one [23:46] <sweeze> k, just wondering :) [23:46] <omega_> need more hardware-specific output elements before that has any real use, though [23:47] <Uraeus> sweeze: you have to use the command line interface to use other output plugins [23:47] <sweeze> and what command line options would those be? (--help shows me ...little of use in regard to taht) [23:48] <omega_> someone was writing some docs, they're ready, but I haven't seen a copy yet [23:48] <Uraeus> omega_: isn' there some examples in your slides? [23:49] <omega_> yeah [23:49] <Uraeus> sweeze: if you look at the slides on the webiste you see some examples, you should be able to experiment some from that [23:49] <sweeze> k, thanks... the guadec slides? [23:49] <omega_> yeah [23:50] <omega_> or check the irc logs in the gstreamer-daily list archives from the last couple days [23:50] <omega_> search for 'launch' in them [23:52] dobey (do...@dr...) left irc: home [23:53] <Uraeus> well, tommorow is a new day with boring work so I need to sleep ;) see ya [23:53] Uraeus (csc...@c2...) left irc: syntax error - user imploded [23:56] sweeze (jba...@sw...) left irc: Client Exiting [00:00] --- Wed Jun 13 2001 [00:02] <taaz> so who's going to write a plugin to rip live realaudio streams to disk for me? ;) [00:03] <omega_> that'd be me, at some point [00:03] <omega_> I've written Real plugins before <shame> [00:05] <taaz> well i wish i had this plugin about 20 mins ago ;) oh well [00:06] Kuroyi (ri...@ub...) joined #gstreamer. [00:38] <omega_> gstautoplug.c compiles! [00:40] <omega_> now to figure out arguments [00:53] dobey (do...@ip...) joined #gstreamer. [00:55] <omega_> taaz: have you arranged somewhere to stay for ols? [00:59] <taaz> nope [01:00] <omega_> and if you haven't arranged a flight already, you may be stuck with driving <g> [01:00] <taaz> why do you say that? [01:00] <omega_> all the cheap flights for me were almost gone [01:01] <omega_> only barely got one less than $300 ;-( [01:04] <taaz> driving is about 14 hours from here and 11 from my parents house. that's what i did last year... but guy i went with is flying from west coast this year. kind of a long trip if i'm alone [01:04] <omega_> ah [01:04] <taaz> at least i think he is... [01:05] <taaz> i'm trying to get him to get his company to pay for a hotel room for me to crash at ;) [01:05] <omega_> heh [01:07] <taaz> why do these web flight reservation systems suck so much? [01:07] <omega_> because [01:07] <omega_> I used cheaptickets.com [01:07] <taaz> i could write a sql query to find exactly the flight i want... but these forms are stupid [01:07] <omega_> they found some stuff no one else could find [01:10] <taaz> $251.19 DCA->YOW [01:11] <omega_> ouch [01:11] <omega_> that's as much as I paid from PDX [01:11] <omega_> lemme check cheaptickets for you [01:11] <taaz> yeah, its been ~$250 for last 3 months. its not going to get cheaper ;) [01:11] <omega_> 24 -> 29 ? [01:11] <taaz> yeah [01:12] <taaz> heh... i was too lazy to create a login on cheaptickets ;) they really shouldn't require that [01:12] <omega_> turned out to be worth it [01:12] <omega_> yeah, $280 [01:12] <omega_> the *stupid* thing is that you can't say "any time during this day, I don't care!!" [01:13] <taaz> man i'm going broke on airfare this summer... [01:13] <taaz> yeah, exactly! [01:13] <omega_> that's the only major annoyance I've had [01:13] <omega_> expedia gives you +-6hrs options [01:14] <taaz> so does travelocity [01:14] <omega_> rather, expedia gives you am,noon,pm [01:17] <taaz> blah... >$90 more to fly out of an airport close to me. might as well drive 4 hours visit parents and use a DC area airport [01:17] <omega_> quite [01:18] <omega_> save the parking - get a ride to DCA <g> [01:18] <taaz> yup, this trip too long to pay for parking ;) which is why i have to schedule times so someone can drive me ;) [01:19] <taaz> usually i try to get the trips that leave and arrive at 3-4am ;) [01:19] <omega_> heh [01:23] <taaz> and why is flying from A->B->C less expensive than getting on the same plane for just the B->C leg?? [01:24] <omega_> they're theives [01:24] <omega_> no other answer fits [01:25] <omega_> a real airline like southwest doesn't do those things [01:26] <taaz> $240.79 IAD->YOW no stops [01:26] <omega_> IAD ? [01:27] <taaz> dulles [01:27] <omega_> ok [01:27] <omega_> not bad [01:28] <taaz> what's an appropriate comparison for driving? gas + wear&tear about $0.30/mile? that works out to $175 [01:28] <omega_> sounds about right [01:29] <omega_> oh, yeah, some mail I just got off a list reminded me [01:30] <omega_> without knowing it, I was showing software at NAB a couple months ago [01:30] <taaz> nab? [01:30] <omega_> National Associaton of Broadcasters conference [01:30] <omega_> bcast2k and Kino were there, both sporting shiny new DV decoders <g> [01:31] <omega_> there's a good chance that that's where heroines got his job [01:32] <taaz> so is the gobject conversion as difficult as s/he made it seem? [01:32] <omega_> no [01:33] <omega_> also, I got a hint from the PLUG meeting that he's younger than I am [01:33] <omega_> apparently he got $1000 to get gear to work on bcast2k from linuxfund [01:34] <taaz> punk [01:34] <omega_> but fyi linuxfund was literally begging to give away money at PLUG [01:35] <taaz> last year at ols alan cox said he thought it was an interesting application but had way too many bugs or something [01:35] <omega_> they're launching (or now have launched) a new campaign to get all the LUGs to spread the linuxfund credit card [01:35] <omega_> taaz: sounds about right <g> [01:35] <taaz> money for what? dvd players? <g> [01:35] <omega_> so if you can write up ~500 words explaining what you'd be doing, there's a good chance you gan grab some [01:35] <omega_> linuxfund is in pdx, btw [01:36] <taaz> heh... does that reduce my liability to the evil media corporations? [01:36] <omega_> no [01:36] <omega_> just increases theirs ;-( [01:37] <taaz> dumb question: what timezone is ottawa in? [01:37] <omega_> yours [01:39] <taaz> really? flight is only 90mins then... hmm... 90 min flight vs 11 hours driving to save ~$50 i think i'll fly [01:40] ChiefHighwater (pa...@te...) left irc: Ping timeout for ChiefHighwater[temple-baptist.com] [01:41] rowenc (rowenc@Roaming-209-53.PLU.edu) joined #gstreamer. [01:41] <omega_> yeah [01:41] ChiefHighwater (pa...@te...) joined #gstreamer. [01:49] <rowenc> where do you want mp3's / video's that don't work reported? [01:49] <omega_> here [01:49] <omega_> and on the list, with a url if you can post them [01:49] <rowenc> ok [01:53] <rowenc> here's one: http://www.helpingfoot.com/AVI/clip0009.avi [01:54] <rowenc> I have some more on my hdd [01:59] <rowenc> has anything been done to use the bundled wine instead of an already installed version? [02:00] <omega_> dunno [02:00] <omega_> I haven't touched that stuff <g> [02:00] <rowenc> I had to remove my wine *.h files before gstreamer would compile... it was suggested to somehow put gst's path first [02:00] <omega_> hmm [02:01] <rowenc> I was getting a ton of incompatible errors with a cvs build of wine I had [02:01] ChiefHighwater (pa...@te...) left irc: Read error to ChiefHighwater[temple-baptist.com]: Broken pipe [02:02] <omega_> I would expect that we're entirely self-contained [02:02] <o... [truncated message content] |
From: <wim...@ch...> - 2001-06-14 04:29:09
|
[06:39] mwc (ma...@ly...) got netsplit. [06:40] iGN___ (ign@216.167.121.81) left irc: Ping timeout for iGN___[216.167.121.81] [06:41] mwc (ma...@ly...) returned to #gstreamer. [06:41] mwc (ma...@ly...) got netsplit. [06:41] mwc (ma...@ly...) returned to #gstreamer. [07:07] <johnix> 'later [07:07] johnix (thomas@61.152.140.57) left irc: Client Exiting [07:35] Nick change: taaz -> taazzzz [07:45] walken (fo...@c1...) joined #gstreamer. [07:51] steveb (st...@no...) joined #gstreamer. [07:58] <krappie> i see a bunch of typos in the application devleopment manual [08:06] iGN___ (ig...@lo...) joined #gstreamer. [08:19] ajmitch (aj...@p3...) left irc: Ping timeout for ajmitch[p3-max1.dun.ihug.co.nz] [08:19] ajmitch (aj...@p3...) joined #gstreamer. [08:40] Nick change: ajmitch -> ajstudy [08:44] rowenc (rowenc@Roaming-209-53.PLU.edu) left irc: Client Exiting [09:07] Uraeus (csc...@c2...) joined #gstreamer. [09:07] <Uraeus> hello [09:07] Uraeus (csc...@c2...) left irc: Read error to Uraeus[c224s9h5.upc.chello.no]: EOF from client [09:08] Uraeus (csc...@c2...) joined #gstreamer. [09:08] Uraeus (csc...@c2...) left irc: Read error to Uraeus[c224s9h5.upc.chello.no]: EOF from client [09:09] Uraeus (csc...@c2...) joined #gstreamer. [09:11] Uraeus (csc...@c2...) left irc: syntax error - user imploded [09:19] krappie (j@24.167.84.194) left irc: [09:59] sam (sa...@zo...) joined #gstreamer. [09:59] <walken> hi sam :) [10:00] <sam> 'morning/'evening/'whatever dudes [10:00] <sam> walken :) [10:00] <walken> 'nite :) [10:10] <walken> ok, gotta sleep now [10:10] walken (fo...@c1...) left irc: l8r [10:21] omega_ (om...@om...) joined #gstreamer. [10:26] mwc (ma...@ly...) left irc: [x]chat [10:53] <thomas_> omega_: still there ? [11:13] <omega_> no, going to sleep now.... [11:14] <omega_> cya tomorrow... (oh, that's today ;-( ) [11:14] <thomas_> ok, sleep tight [11:14] omega_ (om...@om...) left irc: [x]chat [11:33] Nick change: ajstudy -> ajmitch [11:35] johnix (thomas@61.152.140.57) joined #gstreamer. [11:47] <johnix> Hi [11:48] <thomas_> hi [11:50] <ajmitch> evening... [11:54] ajmitch (aj...@p3...) left irc: Ping timeout for ajmitch[p3-max1.dun.ihug.co.nz] [12:04] johnix (thomas@61.152.140.57) left irc: Read error to johnix[61.152.140.57]: EOF from client [12:08] ajmitch (aj...@p1...) joined #gstreamer. [12:24] sweeze (jba...@sw...) joined #gstreamer. [12:24] sweeze (jba...@sw...) left #gstreamer (Client Exiting). [12:34] ajmitch (aj...@p1...) left irc: Ping timeout for ajmitch[p11-max2.dun.ihug.co.nz] [12:35] ajmitch (aj...@p1...) joined #gstreamer. [13:07] Zeenix (bundanache@61.11.233.106) joined #gstreamer. [13:08] <Zeenix> where is Ureaus these days [13:08] <ajmitch> dunno... [13:14] Zeenix (bundanache@61.11.233.106) left irc: Ping timeout for Zeenix[61.11.233.106] [13:39] big_T (th...@fy...) left irc: Ping timeout for big_T[fysgr763.sn.umu.se] [15:22] Nick change: ajmitch -> ajzzzz [15:43] Nick change: taazzzz -> taaz [15:45] dobey (do...@dr...) joined #gstreamer. [15:48] ajzzzz (aj...@p1...) left irc: Ping timeout for ajzzzz[p11-max2.dun.ihug.co.nz] [15:50] ajzzzz (aj...@p1...) joined #gstreamer. [17:56] Nick change: dobey -> dobey-eat [18:29] <thomas_> [18:30] <steveb> [18:46] magnesium (mag...@ni...) joined #gstreamer. [18:46] <magnesium> is there going to be a .2.1 release of some kind to address the AVI issues? [18:47] Nick change: wtay-ZzZ -> wtay [18:47] <wtay> yo [18:47] <wtay> magnesium: AVI issues? [18:47] <magnesium> hi [18:48] <magnesium> wtay: yeah, I thought I saw people having problems compiling avi support (I couldn't get it to compile myself) [18:48] <wtay> magnesium: hmm, yeah avifile issues I guess... [18:52] Nick change: thomas_ -> thomas-feelslikegoinghome [18:52] thomas-feelslikegoinghome (th...@21...) left irc: [x]chat [18:52] Nick change: richardb-away -> richardb [18:52] <richardb> Afternoon all. [18:53] <steveb> hi [18:53] <wtay> hi [18:53] <richardb> I think another release in short order, just before doing the gobject stuff, would be a good plan. [18:54] <steveb> 0.2.5? [18:54] <richardb> There've been a few build fixes as well as the avifile stuff, as well. [18:54] <richardb> fullscreen support in gstplay is probably worth a release on its own. ;-) [18:55] <steveb> how about doing the source tree reorg before that [18:55] <richardb> Are there any documents written describing how we're going to reorganise the tree? [18:56] <steveb> thomas did some stuff on the plugin hierarchy - thats about it [18:57] Nick change: dobey-eat -> dobey [18:58] <wtay> wow, I forgot to patch automake after an upgrade... 1GB of virtual memory consumed... [18:58] Action: richardb is trying to understand the code in CVS automake for conditionals. [18:59] <richardb> It's very confusing. [18:59] Action: wtay is away now for half an hour [18:59] Nick change: wtay -> wtay-away [19:05] big_T (th...@fy...) joined #gstreamer. [19:30] Nick change: wtay-away -> wtay [19:36] <steveb> man, I can't download avifile - their site is screwed [19:36] <wtay> we have the tarball on gst.net... [19:37] <wtay> http://gstreamer.net/avifile_0.6.0.20010518-1.tar.gz [19:38] <wtay> that's the one I currently build against [19:38] <steveb> ok, ta [19:43] PiKolO (luisca@213.97.75.132) joined #gstreamer. [19:43] <wtay> yo [19:43] <PiKolO> hi [19:44] <PiKolO> where can i find debs of gstreamer?? [19:44] <wtay> PiKolO: on taazs HD :-) [19:44] <PiKolO> i've tried to build from cvs... but it failed several times [19:45] Action: taaz hides [19:45] <taaz> ok, i have a question... is the lame plugin useful to anyone? ;) [19:46] <PiKolO> xD [19:46] <taaz> seriously... [19:46] <wtay> taaz: yes, very [19:46] <taaz> blah... cause they kinda suck ass [19:46] <taaz> they have shared libs disabled by default [19:46] <wtay> lame? [19:47] <wtay> yeah [19:47] <taaz> so if i try to depend on lame packages or even self-installs there is no chance the shared libs will be there [19:47] <PiKolO> are any plan to make 0.2.0 debs for sid?? (an put them in the official distro) [19:47] <taaz> so the only option is to do a ugly static link of a 400k library... that is very lame [19:48] <taaz> that's the reason i havent put the debs up yet... didnt know how to deal with such crap. is there some reason they dont use shared libs? [19:48] <PiKolO> ops [19:48] <wtay> then leave lame out for now [19:49] <PiKolO> i see [19:50] <taaz> well... i can just leave the huge plugin there too... [19:50] <taaz> that's also an option for mpeg2dec plugin. staticly link it. right now it'll probably just segfault or something if you haven't manually installed mpeg2dec libs [19:51] <taaz> which i should probably pop up a big ol warning about during the install processes [19:51] magnesium (mag...@ni...) left irc: [x]chat [19:52] <steveb> taaz: mpeg2dec doesn't build for me so, yeah [19:52] shitowax (yann@AAnnecy-101-1-2-176.abo.wanadoo.fr) joined #gstreamer. [19:53] <shitowax> yo [19:53] <taaz> steveb: it doesn't? i haven't seen your bug report on livid-mpeg2dec list. (hint hint) [19:54] <wtay> yo [19:54] <steveb> bah! its my birthday - i don't need this :) [19:55] <wtay> steveb: happy birthday!! [19:55] <shitowax> has somebody looked at the quicktime plugin ? [19:55] <wtay> I'm at it now [19:55] <steveb> cheers - off to dinner [19:56] Nick change: steveb -> steveb-away [19:56] <shitowax> have you seen the streams in htdocs/media ? [19:57] <wtay> yup, I have all of the QT ones [19:57] <wtay> shitowax: are you sure the demuxer works? [19:57] <shitowax> ... what do you mean ? [19:58] <wtay> you seem to use a lot of pullregion stuff, does that still work? [20:00] <shitowax> it seems so, I got the jpeg plugin decompress the first picture of the MPJG-A stream [20:00] <wtay> oh, ok [20:01] <shitowax> the picture was not displayed but it was decompressed [20:02] <shitowax> I had to add a pullregion(...,offset,0) to fixe an offset bug ... [20:04] <wtay> right [20:05] <shitowax> can you see the horrible hack in gstquicktimedemux.c line 388 ? [20:06] <wtay> ick [20:06] <shitowax> I've hardcoded the size of the sample here =) [20:07] <wtay> it doesn't look like pullregion works ... [20:08] <shitowax> what source are you using, I use disksrc, gnomevfssrc may not work ... [20:08] <wtay> right, that's true [20:08] PiKolO (luisca@213.97.75.132) left irc: ta luego [20:09] <shitowax> I was notable to simply retrieve the size of the file from the sinkpad of the muxer ... any idea ? [20:10] <wtay> I think the maxsize property of the buffer is for that [20:11] <wtay> but it's not set in disksrc AFAIK [20:12] <shitowax> hummm, have you succeed in linking the demuxer to the jpeg plugin ? [20:13] <wtay> not yet, going trhough debug line to see what happens [20:15] <shitowax> by the way, the switch on the fourcc in gstquicktimetypes.c line 186 doesn't seem to work neither that's why I've hardcoded the linking to the jpeg plugin ... :( [20:15] <wtay> hmm [20:16] <wtay> first of all, we need to remove *all* of the hacks and fix things that don't work [20:16] <shitowax> it was just to see a picture ;) [20:18] <wtay> help me understand how the openwt API works..: [20:18] <wtay> openqt even [20:19] <wtay> 1. in open_pad you create a structure with read/seek callbacks.. [20:19] <shitowax> yes, what's the problem ? [20:19] <shitowax> yes [20:19] <shitowax> and codec handling also [20:19] <wtay> also a callback for init_video and init_audio [20:20] <wtay> how does one start the qt loop then? [20:21] Uraeus (csc...@c2...) joined #gstreamer. [20:21] <shitowax> ... there is a quicktime_demux_loop function ? [20:22] <shitowax> which calls a quicktime_demux_chain [20:22] <wtay> ok [20:23] <wtay> in a loop? [20:23] <wtay> but you don't pull a buffer? [20:23] <Uraeus> hmm, 48 on the Sourceforge most active projects list :) [20:23] <wtay> aha, a quicktime_read_frame call... [20:23] <wtay> Uraeus: hehehe [20:24] <shitowax> that's it an this one calls the callback [20:24] <wtay> ok, I see [20:25] <wtay> do I have to use current CVS for the openquicktime lib? [20:25] <shitowax> yes, please [20:25] <shitowax> I'm sorry but I have to leave ... I will be back in about 3 hours ... will this be to late to continue ? [20:25] <wtay> I'll spend the whole evening on this.. [20:26] <wtay> I might end up rewriting and rearanging a lot of stuff.. [20:26] <wtay> if you don't mind. [20:26] <shitowax> you have my agreement ;) [20:27] <wtay> cool [20:27] <shitowax> ok, then, I will do as fast as I can ... [20:27] <shitowax> CU everybody ... [20:27] shitowax (yann@AAnnecy-101-1-2-176.abo.wanadoo.fr) left #gstreamer. [20:27] sully (dev...@hs...) joined #gstreamer. [20:35] <wtay> yo [20:41] greg_ (gr...@ho...) joined #gstreamer. [20:44] ChiefHighwater (pa...@te...) joined #gstreamer. [20:48] <wtay> hi [20:49] <ChiefHighwater> Ello [20:50] <greg_> hi [20:51] <ChiefHighwater> Ello [20:52] <greg_> Is there any place I can find #IRC channel about ReiserFS ??? [20:52] <wtay> dunno [20:54] <greg_> I have 80MB HDD and mkreiserfs on its only partition eats half of it - dont know how to change it. [20:54] <Uraeus> greg_: use XFS :) [20:54] <greg_> oh - very funny. It is less tested than XFS. [20:55] Action: Uraeus uses XFS :) [20:56] <greg_> Uraeus: good If I had more than one processor. For one reiserfs is good enough. Anyway - where to find *live* people to talk about reiser ??? [20:56] <ChiefHighwater> do a google search for it [20:57] <ChiefHighwater> search on: reiserfs irc channel produced several links [21:00] <greg_> yeah, but nothing about #reiser or simillar [21:00] thomas (th...@ad...) joined #gstreamer. [21:00] <thomas> hi [21:01] <greg_> hello [21:02] <wtay> yo [21:04] <Uraeus> yo [21:08] RagingMind (RM@209.69.192.87) joined #gstreamer. [21:08] <wtay> hi [21:08] <Uraeus> hello [21:08] <RagingMind> hi [21:08] <dobey> Uraeus: hey [21:08] <wtay> dobey! [21:09] <dobey> hola wtay [21:09] RagingMind (RM@209.69.192.87) left #gstreamer. [21:10] <Uraeus> hi dobey [21:16] Nick change: ajzzzz -> ajmitch [21:16] <Uraeus> hmm, I just figured out how to put us on top of the Sourceforge activity list, I just file 500 wishlist bugs :) [21:16] <ajmitch> heh [21:17] <ajmitch> morning all [21:18] <thomas> hi ajmitch [21:18] <thomas> cool, I didn't break the CVS tree for once, I must be learning ;) [21:25] <Uraeus> <g> [21:26] <wtay> actually you did leave a bit of mess behind in mad :-) [21:31] <greg_> Uraeus: at #kernelnewbies I was told 32MB is minimum size of journal for ReiserFS. Do you know if XFS has SIMILAR behaviour ??? [21:33] <Uraeus> greg_: no idea, but you probably can easily find out on the XFS site [21:37] <greg_> Uraeus: any URL for it ? [21:38] <Uraeus> greg_: www.sgi.com [21:38] <greg_> going to [21:38] <greg_> thx [21:47] omega_ (om...@om...) joined #gstreamer. [21:47] <wtay> yo [21:47] <ajmitch> hey omega_ [21:47] <wtay> omega_: I found a problem (and have a fix...) [21:47] Action: ajmitch wonders when 0.3.0 is coming out ;) [21:47] <omega_> which problem? [21:48] <wtay> pullregion stuff [21:48] <wtay> when it pulls a buffer, it takes the one from the bufpen, which is not allways the right region [21:49] <omega_> ah [21:49] <omega_> well, that's all going to be depracated soon anyway [21:49] <wtay> when the other element is scheduled first for example [21:49] <wtay> I know.. [21:49] <wtay> anyway, I did a do while on the pull, until the ranges match.. [21:50] <wtay> the second attempt ot pulling a buffer always works [21:50] <omega_> yup [21:50] <wtay> I'll leave it like this for now? [21:50] <omega_> sure [21:50] <wtay> ok [21:50] <omega_> if you want to help out on gobject, we can get it done before you sleep I think [21:50] Action: wtay is working on the quicktime decoder [21:51] <omega_> for libgst.la at least [21:51] <wtay> omega_: I promised to get the QT decoder working.. [21:51] <omega_> ok [21:51] <wtay> for jpeg at least, I'm almost there I think... [21:51] <Uraeus> omega_: have you managed a GTK+1.2 & GObject (Glib 2.0) solution? [21:52] <omega_> I think that's possible, yes [21:52] <omega_> only doing gobject right now though [21:52] <Uraeus> heh [21:53] <omega_> as soon as I get the gobject-based system working, I'll start on a shim [21:53] <omega_> wtay: I need you to look over the object destruction stuff I'm not doing in gobject [21:53] <Uraeus> omega_: are you also aware that we now have climed to position 48 on the Sourceforge ranking :) [21:53] <omega_> cool [21:53] <Uraeus> what is: gstxa ? [21:54] <wtay> omega_: you can leave the destruction for now until I have time to look at it [21:54] <omega_> the 'xaudio' mp3 decoder shim [21:54] <omega_> ok [21:54] <ajmitch> Uraeus: #48, eh? [21:54] <Uraeus> yup [21:54] <ajmitch> ah, i see, they got the download tracker working ;) [21:54] <omega_> Uraeus: where do you see that? [21:55] <Uraeus> omega_: if you go to the www.soureforge.net frontpage and click on the activity ranking [21:55] <omega_> cool [21:56] <Uraeus> 0.3.0 will surely put us in the top ten :) [21:56] <omega_> heh [21:56] <ajmitch> bochs looks popular... [21:56] <ajmitch> why are we not in the top 5 already? ;) [21:57] <omega_> 1.2.1 yesterday (bochs) [21:57] <Uraeus> not enough downloads I think [21:57] <ajmitch> bochs gets several thousand page hits a day [21:57] <Uraeus> that too, I think packages of interest to windows users have an advantage [21:58] <ajmitch> and daily download count somewhere between 1000 & 2000... [21:58] <omega_> Uraeus: when the debs go up, we'll jump again [21:58] <omega_> Uraeus: I wonder if we should have an 0.2.0-rpm and 0.2.0-deb section in sourceforge? [21:58] <omega_> or not, that'll only allow us one 'current' release [21:59] <ajmitch> how about submitting news for the main sf page? ;) [21:59] <Uraeus> omega_: debs == vaporware ;) [21:59] <ajmitch> hehe [21:59] <ajmitch> rpms == vapourware ;) [21:59] <omega_> ajmitch: go see the download page [21:59] <Uraeus> ajmitch: rpms already out :) [21:59] <ajmitch> :P [21:59] Action: omega_ goes to the closet to get something to shoot at taaz [21:59] <Uraeus> ajmitch: including prereqs [22:00] Action: ajmitch runs dpkg-buildpackage [22:00] <Uraeus> ajmitch: I will have solaris .pkg's ready before any debs are seen :) [22:01] <ajmitch> hehe [22:02] <ajmitch> attempting to build debs from taaz's rules now... [22:02] <wtay> omega_: a question about the structure of a loop function... [22:03] <omega_> yeah? [22:03] <wtay> can it have { init; do { pull; push } while (!stopping) [22:03] <wtay> or do the init somewhere else? [22:03] <omega_> do the init in state change, ideally [22:04] <wtay> ok, the cothread stopping event can occur at all times I suppose [22:04] <omega_> and remember that the loop function can cease to exist at any time [22:04] <omega_> yes [22:04] <omega_> you can also be switched out and never switched back to at any push or pull [22:04] <wtay> like after evey push/pull [22:04] <omega_> during, not after [22:04] <wtay> ok [22:05] <wtay> we cannot assume we can do a pull in the READY/PAUSED state, right? [22:06] <omega_> righ [22:06] <wtay> so my only option to do init is at the start of the loop function (if init==FALSE) and then set init=TRUE [22:06] <wtay> state change sets the init to FALSEE [22:06] <omega_> what init? [22:07] <wtay> init in QT is pulling a few buffers, getting the number of streams [22:07] <omega_> that's just part of the stream, just like mpeg2 [22:08] <wtay> but it only has to happen once (the first time) [22:08] <omega_> don't you do it when you get the first buffer? it has the header, so you figure out what's going on? [22:09] Nick change: steveb-away -> steveb [22:09] <wtay> omega_: heh, the header is *at the end* :( [22:09] <omega_> neat. so much for streaming [22:09] <wtay> yup [22:09] <omega_> this is qt, right? [22:09] <omega_> that can't be the case [22:10] <wtay> maybe it has other options but in the test case I have, it is at the end [22:10] <omega_> if you grab a file designed for streaming, it should be at the front [22:10] <omega_> it kinda has to be, unless the streaming server dishes it up out of order [22:11] <wtay> I hope it is.. [22:11] <omega_> but qt also has been suggested and accepted as the meta-file format for mpeg-4 stuff [22:11] <omega_> and they're going to *require* that the headers be first, else it can't be streamed broadcast [22:12] <wtay> indeed, It probably can do that too [22:13] <wtay> I can see from the pullregion offsets that it pulls the index/info from the tail of the stream.. [22:13] <omega_> neat [22:13] <wtay> and it looks like openquicktime was not designed for streaming either.. [22:14] <omega_> nope [22:14] <wtay> instead of getting some sort of callback when a frame has been decoded, you have to do a get_next_frame etc.. [22:14] <omega_> right [22:14] <omega_> it's designed inside-out [22:15] <wtay> I wonder how I'm gonna do audio and video like this... [22:15] <omega_> you actually should have an API that a) aids in parsing atoms, and b) has a stack of functions that specifically handle those atoms [22:15] <omega_> not the other way around (hand it data and it does something) [22:15] Action: omega_ should write a book on good multimedia API design [22:15] <wtay> yes, please [22:16] <omega_> or just design APIs for people [22:16] <wtay> a template [22:16] <omega_> no such thing [22:17] <wtay> I had best results with a struct of callbacks and a method _parse_buffer... [22:17] <wtay> hand it buffers, when it has something, it does the callback [22:18] <omega_> yeah [22:18] <omega_> that's what my bitstream code does, walken didn't like the idea that much [22:19] <omega_> but I think I can persuade him [22:19] <wtay> mpeg2dec - the vo stuff is very nice IMHO [22:19] <wtay> - == minus [22:19] <omega_> uh? [22:20] <wtay> libmpeg2dec? [22:20] <omega_> yeah, the libvo stuff? [22:20] <omega_> oh, ok, your syntax was odd [22:20] <omega_> mpeg2dec has some serious problems [22:20] <wtay> yup, that's the only ugly thing in it [22:20] <wtay> like? [22:21] <omega_> specfically, it's *not* callback based, which has a noticable performance impact [22:21] <wtay> howso not? [22:21] <omega_> his bitstream system has to be fed, in exactly the way you described openqt [22:21] <omega_> my proposed API for libac3dec for him was using my bitstream code [22:21] <wtay> ah, that's an inside issue [22:21] <omega_> which you give a callback to, which handles the gathering of new data [22:22] <omega_> libmpeg2dec has to do a byte-by-byte copy with inspection on the entire mpeg data, without doing anything else, before it can decode [22:22] <omega_> I made that 3x faster a while ago, but he hasn't merged it in yet [22:22] <omega_> it's still 1-2% of the whole decode, and much worse on smaller caches [22:22] <wtay> yeah, it first seeks for the sync by copying everything over [22:22] <omega_> right [22:23] <omega_> and with a callback-based bitstream, that's entirely unnecessary [22:23] <thomas> make[4]: Entering directory `/home/thomas/gst/gstreamer/plugins/filters/deinterlace' [22:23] <thomas> make[4]: *** No rule to make target `all'. Stop. [22:23] <thomas> m [22:23] <thomas> hmmm... that's new right ? [22:23] <omega_> bad wtay... [22:23] <wtay> thomas: autogen.sh ? [22:23] <thomas> just clean the source out ? [22:23] <thomas> wtay: did that [22:23] <wtay> hmm [22:24] <thomas> wtay: did you commmit Makefile.am in filters ? [22:24] <wtay> thomas: I did, but I forgot the configure.base file [22:24] <thomas> yeah, that's it [22:24] <thomas> ok, I'll do that [22:24] <wtay> thomas: commit it too then.. [22:25] <thomas> ok, done. [22:25] <thomas> Hmm, I have a question about the event thing. What other kinds of events where you planning besides eos ? [22:26] <wtay> seek for one [22:26] <thomas> wtay: how would seek work in regard to plugins ? could you give an example ? [22:27] <wtay> thomas: a plugin would 'listen' to an event, find out it is a seek event and do something [22:28] <thomas> wtay: but would plugins themselves ask for a seek ? or only the app ? [22:28] <wtay> a plugin would [22:28] <wtay> if it has no eventhandler, the event is passed on to all the sink/src pads [22:28] <wtay> I think [22:28] <thomas> hmmm... ok, so while an eos event would travel left to right, a seek would probably travel right to left then ? [22:29] <wtay> yes [22:29] <thomas> hm, ok, so what makes the eos problem "hard" currently ? the scenarios in docs/random/eos seem like they should work, no ? [22:29] <wtay> some sort of inter-plugin communication mechanism [22:30] <wtay> thomas: with incsched they should be a bit easier I think.. [22:30] <wtay> easier to remove/disable elements etc.. [22:31] <wtay> omega_ also wants state changes to happen with eos IIRC [22:31] <thomas> wtay: I think left to right signals should be maneagable... [22:31] <thomas> If you'd force plugins to either send out a buffer or send out an event ... [22:32] <thomas> ... and have that travel down the line ... [22:32] <thomas> then it should work. [22:32] <wtay> that's how eos signals travel currently [22:32] <thomas> right to left seems harder... is that set_eos function implemented already ? [22:32] <wtay> thomas: yes [22:32] <wtay> thomas: so is qos (right to left) [22:33] <thomas> but not the eos management of the bin in which elements are ? [22:33] <wtay> thomas: that's why we need a simple more general api [22:33] <wtay> thomas: nope, it was removed with incsched [22:33] <thomas> wtay: hm, ok, shouldn't we start from a list of reasonable events then ? [22:34] <wtay> I suppose [22:36] <thomas> what ideas were there for an event-api ? anything that has been looked at before ? [22:38] <wtay> ask omega_ :) [22:38] <thomas> ok... omega ? ;) [22:38] <omega_> um, haven't put much thought into [22:38] <omega_> it [22:38] <omega_> the moment I do, I'm gonna start coding it, though <g> [22:38] <omega_> and that's why I've been putting it off [22:38] <omega_> that's next after gobjectification, imo [22:39] <thomas> uh ok... then I won't get in your way on that one ;) [22:39] <omega_> heh [22:39] <wtay> a good start would be to define what data we want for eos/qos/seek etc.. [22:39] <omega_> gobject is coming together [22:39] <thomas> qos is quality of service in this case ? [22:39] <omega_> but I'm gonna need a perl script to rewrite the GTypeInfo stuff, it's getting tedious, and will get really painful for ~90 plugins [22:39] <ajmitch> fsck it! [22:40] <ajmitch> silly me upgraded automake, it's not patched... [22:40] <omega_> yup [22:40] <ajmitch> (when building debs, autogen.sh isn't run & doesn't check for old automake) [22:40] Action: Uraeus just sent out a proposal for a gnome.no soccer team ;) [22:40] <thomas> so, that confuses me. Why would a plugin down the stream want to ask for a different qos upstream ? [22:41] <ajmitch> i got suspicious when xmms started skipping & mouse movement became jerky... [22:41] <omega_> it's not asking for anything [22:41] <omega_> it's screaming at the upstream element saying "LESS DATA PLEASE!" [22:41] <thomas> ok, that's a good thing to allow ;) [22:41] <wtay> or more.. [22:41] <omega_> or more, yes [22:41] <wtay> thomas: you need it! [22:42] <thomas> so in this case, you actually want to send more data than just the event id itself [22:42] <omega_> 21 of 29 .c files in libgst.la compiled [22:42] <wtay> yep [22:43] <wtay> qos is typically more/less fps, bytes/sec, percentage [22:43] <omega_> 22.. [22:44] <thomas> one other thing : if you write an audio plugin, can you always assume that the buffer size is a whole multiple of the frame size ? [22:44] <thomas> or should a plugin be forced to deal with those issues itself ? [22:44] <omega_> frame size or sample size? [22:45] <thomas> frame size... for example if you have stereo data but your buffer is, for example in this case, 2 bytes [22:45] <thomas> (16 bit) [22:45] <omega_> that's sample size [22:45] <thomas> I meant : 16 bit data [22:45] <thomas> omega_: yeah, so it's half a frame in this case [22:45] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee02.a2000.nl] [22:45] <omega_> oh, you think of a frame as 1sample * Nchannels [22:46] <thomas> omega_: yeah, is that the wrong way ? [22:46] <omega_> no, I think it's 'correct', but then we need a different name for the default number of frames processed at a time [22:46] steveb (st...@no...) joined #gstreamer. [22:47] <omega_> for instance, 48 frames per buffer (192 bytes) would be about 1.1ms of data [22:47] <omega_> or 1.0, for DAT [22:47] <thomas> omega_: yeah, that's how I count. [22:47] <thomas> it won't be long before people will want 5+1 audio with gstreamer ;) [22:47] <thomas> so imo that makes more sense [22:48] <omega_> yup [22:48] <thomas> are there instances in gstreamer where it's interpreted differently ? [22:49] <omega_> what is? [22:49] <thomas> well, the definition of frame, since we got mixed up just now ? [22:49] <omega_> that terminology isn't in the code [22:50] <thomas> ok. so given that definition, can a buffer not be a whole multiple of frames (which would make it harder on plugin writers) ? [22:50] <omega_> we can state up front "no" and leave it at that, I think [22:51] <omega_> one more check for the audio-specific element validation program [22:51] <thomas> omega_: whereas for video, it's probably not a good idea to say no, I suppose ? [22:51] <omega_> how would you represent a partial frame? [22:51] <omega_> it's got spatial as well as temporal coordinates... [22:51] <omega_> I would like to be able to do that, but I can't think of any sane way to do it offhand [22:52] <thomas> omega_: so for video, you generally use bigger data buffers than 4096 bytes ? [22:52] <omega_> for raw video a buffer *is* a frame [22:53] <thomas> ok, makes perfect sense then ;) [22:54] <omega_> for coded data, there may or may not be a sensible framing structure, and if so, there may or may not be a boolean property stating whether the data is on those boundaries or not [22:54] <omega_> part of the data-format specs in the appendix of the GStreamer book <g> [22:58] rowenc (rowenc@Roaming-209-53.PLU.edu) joined #gstreamer. [22:59] <Uraeus> hi [22:59] <wtay> hi [22:59] <rowenc> hi [23:00] <omega_> hi [23:00] <thomas> hi [23:00] <wtay> hi [23:00] <omega_> hi [23:00] <wtay> yo [23:00] Action: omega_ runs in terror [23:00] <thomas> hihi [23:00] <ajmitch> hi [23:00] <thomas> we sound married ;) [23:01] <taaz> ajmitch: having fun? debs build just fine, and its not supposed to run autogen... i got yelled at for that [23:01] <omega_> taaz: if the debs build just fine, where are they on sf? [23:01] <ajmitch> taaz: yeah, i think it might be building now ;) [23:02] <taaz> blah! [23:03] <taaz> omega_: did you see my bitching earlier about the lame lame people? [23:03] <omega_> yes, and I agree entirely, I think [23:04] <omega_> they do quite a few things wrong [23:04] Nick change: ajmitch -> ajbusy [23:05] <wtay> whooao! mjpeg quicktime works! [23:05] <thomas> congrats [23:05] <taaz> suggestions on how to handle this? just leave it as a 450k plugin? insist people install lame shared libs? drop lame altogether? [23:05] <omega_> um. what does Uraeus do? [23:06] <taaz> huh? [23:06] <omega_> he has rpms that depend on liblame [23:07] <omega_> I think he built his own lame rpm based on mine to solve the problem [23:08] <taaz> i cant really do that [23:08] sully (dev...@hs...) left irc: Ping timeout for sully[hs2-236.magma.ca] [23:08] <omega_> then do a static link for now ;-( [23:08] <omega_> make a note of it, and we need to work on getting the lame people to fix these things [23:08] <Uraeus> nope [23:09] <Uraeus> I am still working on lame, we are having some build problems but I am working with the lame hackers on it [23:09] <taaz> i see no gst-lameass plugin at all [23:09] <taaz> Uraeus: beg them to enable shared libs by default [23:09] <omega_> ok, then taaz and Uraeus should work with the lame people to fix the brokenness [23:09] <Uraeus> taaz: shared libs doesn't work properly yet [23:09] <omega_> Uraeus: do you have a list of things that they need to fix? I can add to it [23:10] <taaz> Uraeus: why not? they too lame to fix it? [23:10] <omega_> taaz: by definition, yes <g> [23:10] Action: taaz not going to stop with the lame puns till they are no longer lame [23:10] <Uraeus> the next version will solve most of our problems I think [23:10] <ChiefHighwater> go taaz, go [23:11] <Uraeus> actually the lame people have been very forthcomming in trying to help me out [23:11] <omega_> ok, then... taaz: just need to get them to fix what you need fixing [23:12] <Uraeus> taaz: do you want me to forward you the correspondece up to now? [23:12] <omega_> Uraeus: near the top of my list (the first thing I hit whenever I grab lame) is that the tarball is lame3.88 and the package is lame-3.88. they have to intentionally rename the tarball from what automake generates, and it's just annoying [23:12] <omega_> Uraeus: me too [23:16] <Uraeus> omega_, taaz: ok mail forwarded, only the replies I got, think my mails are all part of theirs [23:16] <omega_> ok [23:16] <Uraeus> haven't replied yet to the latest as I have still not determined that all problems are solved :) [23:16] <Uraeus> actually I am pretty sure they are not [23:18] <omega_> 25 down, 4 to go [23:19] <thomas> wtay, omega: I can see the gst_pad_set_eos/gst_pad_set_eos_function combination ... [23:19] <thomas> ... so I suppose standard plugins should just copy the eos through ? [23:19] <thomas> ... but I suppose not all of them do that ? [23:19] <omega_> for now, I think so [23:20] <wtay> thomas: if a plugin does not set the oes function, the default handler is used [23:20] <thomas> ok, so for lame it would be handy to have a sort of sync event ? [23:20] <wtay> thomas: this one just passes the event to all of the oposite pads of the eleme [23:20] <wtay> nt [23:20] <thomas> a sync event would cause lame to flush out it's curent audio buffer and start again with the next buffer in this case [23:21] <wtay> thomas: that's a buffer property [23:22] <thomas> wtay: hmmm, so lame doesn't have it then ? [23:22] <wtay> GST_BUFFER_FLUSH [23:22] <wtay> thomas: I don't think so [23:22] <omega_> will be an event soon [23:22] <wtay> yup [23:23] <wtay> mpg123 handles the buffer flush for example [23:30] steveb (st...@no...) left irc: [x]chat [23:33] <thomas> yay, afsink works again with a little eos magic [23:34] <thomas> what's a good binary comparison utility ? [23:35] <omega_> cmp <g> [23:38] dobey (do...@dr...) left irc: home [23:46] greg_ (gr...@ho...) left irc: [x]chat [23:48] <omega_> [omega@omicron .libs]$ ldd libgst.so.0.0.0 [23:48] <omega_> libglib-1.3.so.5 => /usr/lib/libglib-1.3.so.5 (0x40044000) [23:48] <omega_> libz.so.1 => /usr/lib/libz.so.1 (0x400a2000) [23:48] <omega_> libxml.so.1 => /usr/lib/libxml.so.1 (0x400b0000) [23:48] <omega_> libc.so.6 => /lib/libc.so.6 (0x40120000) [23:48] <omega_> /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x2aaaa000) [23:48] <wtay> oooh [23:49] <wtay> only 11 libs less than me :-) [23:49] <omega_> only [23:49] <wtay> is libz needed? [23:50] <omega_> for libxml [23:50] <wtay> oh, yes [23:50] <omega_> I'm going to add defines to remove all the pipeline save/restore stuff as needed [23:50] <wtay> libpthread? [23:50] <omega_> partway there in HEAD [23:50] <omega_> oh, um.. hrm [23:51] <omega_> I'm not being careful about setup, so gthread isn't listed in configure.base yet [23:51] <wtay> libm? [23:51] <omega_> math [23:51] <omega_> needed for float stuff. part of libc [23:52] <omega_> committed [23:53] <wtay> for people wanting to see naked men: gstmediaplay douche.PHOTO_JPEG.mov (on gstreamer.net/media) [23:54] <wtay> time for bed [23:54] shitowax (yann@AAnnecy-101-1-2-176.abo.wanadoo.fr) joined #gstreamer. [23:54] <wtay> cya all [23:54] Nick change: wtay -> wtay-zZz [23:54] <omega_> l8r, and check out object destruction when you next have time [23:54] <wtay-zZz> shitowax: QT works! [23:54] <wtay-zZz> omega_: ok will do [23:54] <omega_> on gobject branch, `grep -B 1 destroy *.c` [23:54] <shitowax> so wtay has just gone or what ? [23:55] <wtay-zZz> shitowax: yeah, I need sleep [23:55] <omega_> shitowax: poor timing <g> [23:55] <shitowax> success with quicktime ? [23:55] <wtay-zZz> shitowax: but I fixed quicktime first [23:55] <wtay-zZz> shitowax: mjpeg plays fine [23:56] <shitowax> coll I will check this just now [23:56] <wtay-zZz> shitowax: had to fix the pullregion bug in the core too [23:56] <shitowax> ok, then a big checkout [23:56] <wtay-zZz> shitowax: hope it works for you too.. if not I'll be here tomorrow (as usual) [23:57] <shitowax> ok, then good night ;) [23:57] <wtay-zZz> I tried with douche.PHOTO_JPEG.mov [23:57] greg_ (gr...@ho...) joined #gstreamer. [23:57] <wtay-zZz> MJPEG_A has issues (jpegdec cannot decompress it) [23:58] <shitowax> hum... stange the QT4L codec are the same ... [00:00] --- Thu Jun 14 2001 [00:13] shitowax (yann@AAnnecy-101-1-2-176.abo.wanadoo.fr) left #gstreamer. [00:26] <thomas> night all [00:26] <omega_> l8r [00:27] thomas (th...@ad...) left irc: sleep [00:46] dobey (do...@ip...) joined #gstreamer. [00:46] <Uraeus> night [00:47] Uraeus (csc...@c2...) left irc: syntax error - user imploded [01:27] sienap (sy...@ip...) joined #gstreamer. [01:27] <dobey> hey sienap [01:27] <sienap> hej guys [01:27] <sienap> it is this far again [01:27] <sienap> i am drunk :) [01:28] <dobey> heh [01:28] <sienap> it must be fun too drink with ameriacns [01:28] <sienap> that just got 21 [01:28] <sienap> :) [01:29] <dobey> ? [01:29] <dobey> i don't drink to get drunk [01:29] <dobey> i drink to enjoy the beverage which i intend to consume [01:30] <sienap> getting drunk is a nice side effect [01:30] <sienap> of wine [01:30] <sienap> :) [01:30] <omega_> wow. nearly fully automated conversion from GtkObject to GObject for elements <g> [01:31] <ChiefHighwater> kewl [01:31] <sienap> omega damn [01:31] <sienap> automated ? [01:31] <sienap> explain dude [01:31] <sienap> :) [01:31] <omega_> sed/perl [01:31] <sienap> you are writing code ai's ? :) [01:31] <sienap> lol [01:31] <sienap> mothafucking ape :) [01:32] <sienap> sick sick sick :) [01:32] <sienap> have to see if keneth is online [01:33] <sienap> always had a great talk with him when we where drunk :) [01:33] <sienap> damn [01:33] <sienap> not online [01:33] <sienap> BAH [01:33] Nick change: sienap -> drunknap [01:34] <drunknap> so so so [01:34] <drunknap> parapraxi has been on lately ? [01:39] <drunknap> btw i had some good fucks this weekend [01:39] <drunknap> hell yeah [01:39] <drunknap> :) [01:40] <omega_> drunknap: we do not need to know this, please do not share it with us [01:41] <drunknap> oh [01:41] <drunknap> okie [01:41] <drunknap> :) [01:41] <drunknap> sorry for the disinfotmaton [01:41] <drunknap> infromation [01:41] <drunknap> aah forget it [01:41] <drunknap> but has paripraxi been on lately ? [01:42] greg_ (gr...@ho...) left irc: [x]chat [01:42] <dobey> haven't seen him [01:42] <drunknap> damn [01:42] <drunknap> i need the dude [01:42] <drunknap> bad [01:42] <drunknap> :) [01:44] <drunknap> ehm [01:44] <drunknap> why is everyone ignoring me ? :) [01:44] <dobey> you aren't saying anything [01:44] <drunknap> ooh [01:44] <drunknap> ok [01:44] <drunknap> :) [01:57] drunknap (sy...@ip...) left irc: drunknap has no reason [02:12] omega_ (om...@om...) left irc: [x]chat [02:18] ajbusy (aj...@p1...) left irc: Ping timeout for ajbusy[p11-max2.dun.ihug.co.nz] [02:20] Nick change: ChiefHighwater -> CHW_away [02:34] ajbusy (aj...@p1...) joined #gstreamer. [03:03] mwc (ma...@ly...) joined #gstreamer. [03:04] johnix (thomas@61.152.140.57) joined #gstreamer. [03:06] Nick change: ajbusy -> ajmitch [03:08] <johnix> hi [03:09] <dobey> hey aj [03:10] <ajmitch> hey [03:22] Kuroyi (ri...@ub...) left irc: Ping timeout for Kuroyi[ubr-35.82.223.wmelbourne.cfl.rr.com] [03:22] ajmitch (aj...@p1...) left irc: Ping timeout for ajmitch[p11-max2.dun.ihug.co.nz] [03:46] ajmitch (aj...@p1...) joined #gstreamer. [03:51] ajmitch (aj...@p1...) left irc: Ping timeout for ajmitch[p11-max2.dun.ihug.co.nz] [04:03] Kuroyi (ri...@ub...) joined #gstreamer. [04:06] dobey (do...@ip...) left irc: later [04:27] Kuroyi (ri...@ub...) left irc: Read error to Kuroyi[ubr-35.82.223.wmelbourne.cfl.rr.com]: EOF from client [04:46] walken (fo...@c1...) joined #gstreamer. [04:48] Kuroyi (ri...@ub...) joined #gstreamer. [05:08] ajmitch (aj...@p6...) joined #gstreamer. |
From: <wim...@ch...> - 2001-06-15 04:29:25
|
[06:31] Nick change: ajmitch -> ajbusy [06:39] ajbusy (aj...@p6...) left irc: Ping timeout for ajbusy[p6-max10.dun.ihug.co.nz] [06:39] ajbusy (aj...@p6...) joined #gstreamer. [06:44] CHW_away (pa...@te...) left irc: Read error to CHW_away[temple-baptist.com]: Connection reset by peer [07:01] rowenc (rowenc@Roaming-209-53.PLU.edu) left irc: Ping timeout for rowenc[Roaming-209-53.PLU.edu] [07:01] rowenc (rowenc@Roaming-209-53.PLU.edu) joined #gstreamer. [08:25] steveb (st...@no...) joined #gstreamer. [08:41] CHW_away (pa...@te...) joined #gstreamer. [08:53] CHW_away (pa...@te...) left irc: Ping timeout for CHW_away[temple-baptist.com] [09:45] rowenc (rowenc@Roaming-209-53.PLU.edu) left irc: Client Exiting [10:41] mwc (ma...@ly...) got netsplit. [10:41] big_T (th...@fy...) got netsplit. [10:41] sam (sa...@zo...) got netsplit. [10:41] richardb (ri...@ix...) got netsplit. [10:41] walken (fo...@c1...) left irc: zzzZZZzzz [10:41] richardb (ri...@ix...) returned to #gstreamer. [10:41] sam (sa...@zo...) returned to #gstreamer. [10:41] big_T (th...@fy...) returned to #gstreamer. [10:41] mwc (ma...@ly...) returned to #gstreamer. [10:53] mwc (ma...@ly...) got netsplit. [10:53] sam (sa...@zo...) got netsplit. [10:53] richardb (ri...@ix...) got netsplit. [10:53] big_T (th...@fy...) got netsplit. [10:54] richardb (ri...@ix...) returned to #gstreamer. [10:54] sam (sa...@zo...) returned to #gstreamer. [10:54] big_T (th...@fy...) returned to #gstreamer. [10:54] mwc (ma...@ly...) returned to #gstreamer. [11:13] mwc (ma...@ly...) left irc: [x]chat [12:13] johnix (thomas@61.152.140.57) left irc: Client Exiting [13:22] Nick change: ajbusy -> ajzzzz [15:02] Kuroyi (ri...@ub...) left irc: Read error to Kuroyi[ubr-35.82.223.wmelbourne.cfl.rr.com]: Connection reset by peer [15:02] Kuroyi (ri...@ub...) joined #gstreamer. [15:14] thomas (th...@21...) joined #gstreamer. [15:14] <thomas> hi [15:17] Kuroyi_ (ri...@ub...) joined #gstreamer. [15:17] Kuroyi (ri...@ub...) left irc: Read error to Kuroyi[ubr-35.82.223.wmelbourne.cfl.rr.com]: Connection reset by peer [15:42] gomez (dr...@wm...) joined #gstreamer. [15:44] gomez (dr...@wm...) left irc: BitchX-1.0c17 -- just do it. [15:44] sienap (sy...@ip...) joined #gstreamer. [15:45] sienap (sy...@ip...) left irc: sienap has no reason [15:46] sienap (sy...@ip...) joined #gstreamer. [15:46] <sienap> heydieh0w [16:27] dobey (do...@dr...) joined #gstreamer. [16:28] <sienap> dopey!@ [16:29] <dobey> hi [16:50] Kuroyi_ (ri...@ub...) left irc: Read error to Kuroyi_[ubr-35.82.223.wmelbourne.cfl.rr.com]: EOF from client [17:03] ShrimpX (marius@131.252.244.168) joined #gstreamer. [17:17] <thomas> yay ! I just completed an NFS-based Redhat 7.1 install with gstreamer packages all in ! [17:28] pimlottc (pim...@mo...) joined #gstreamer. [17:29] pimlottc (pim...@mo...) left irc: BitchX: the quilted quicker picker upper [17:31] sam (sa...@zo...) left irc: Ping timeout for sam[zoy.org] [17:39] sam (sa...@zo...) joined #gstreamer. [17:58] Uraeus (csc...@c3...) joined #gstreamer. [17:58] <Uraeus> hiho [17:59] <Uraeus> more good press: http://www.lwn.net/2001/0614/devel.php3 [17:59] <thomas> hi uraeus [18:00] <thomas> Uraeus: I just did a kickstart install with redhat 7.1 with modified rpm's, including gstreamer, all without a single keypress ;) [18:00] <thomas> so thanks for the rpms [18:01] sam (sa...@zo...) left irc: Ping timeout for sam[zoy.org] [18:03] <Uraeus> my pleasure :) [18:03] <dobey> hola Uraeus [18:04] <Uraeus> hi dobey [18:04] <dobey> man, i need to fix my keys, buy a hard drive and keyboard, and some different keys [18:04] <thomas> Uraeus: when is gstreamer scheduled to run under windows ? [18:05] <dobey> ewww [18:05] <Uraeus> thomas: there has been some people interested, but no code yet [18:05] <Uraeus> thomas: I think they based that assumption upon the roadmap/status table [18:07] <Uraeus> need to go again, see ya later [18:07] Uraeus (csc...@c3...) left irc: syntax error - user imploded [18:19] ajzzzz (aj...@p6...) left irc: Ping timeout for ajzzzz[p6-max10.dun.ihug.co.nz] [18:37] ajzzzz (aj...@p6...) joined #gstreamer. [18:46] Nick change: wtay-zZz -> wtay [18:47] <wtay> yo [18:47] <dobey> yo [18:56] <thomas> evening [18:56] <thomas> wtay: I have a thread/queue question ;) [18:56] <wtay> yes? [18:57] <thomas> hang on, gathering evidence ... [18:57] <wtay> uh oh... [18:58] <thomas> ok, I'll start at the beginning... [18:59] <thomas> I added a thread to my bin... I suppose it's not a good idea to set the thread explicitly to playing ? [18:59] <wtay> not really, no [19:00] <wtay> are you using a pipeline as the top level element? [19:00] <thomas> I don't think so, just a bin. [19:00] <wtay> uhm, that's a mistake [19:01] <wtay> a bin doesn't have a scheduler [19:01] <thomas> ah ok... so why does the two-threaded decoder example in the manual use it ? [19:01] <thomas> just outdated, or ... ? [19:01] <thomas> hang on, I'm using a pipeline [19:02] <thomas> main_bin = gst_pipeline_new ("bin"); [19:02] <wtay> outdated for sure [19:02] <thomas> damn... before I had the problem that when I tried to pause the whole pipeline, the thread wouldn't pause, it gave an assert error ... [19:02] <thomas> ... in a block of code which mentions that at the moment that was illegal [19:03] <thomas> now my code hangs when it tries to pause the main_bin... [19:03] <wtay> ok, looks like the player then [19:03] <wtay> sorta [19:04] <thomas> hmmm, this is totally weird... [19:05] <thomas> now my main bin keeps iterating, but the other thread seems to be locked setting the main_bin to paused [19:05] <thomas> hmmm... should I explain what it does before asking questions ? ;) [19:13] <wtay> sorry, I got distracted.. [19:13] <wtay> hm, go on :) [19:13] <thomas> well, it's the cutter example [19:13] <thomas> it works the way it is in cvs now [19:14] <thomas> what it does is listen to osssrc, and when the level goes up, it starts recording to disk [19:14] <wtay> ok [19:14] <thomas> when it goes down, it stops recording [19:14] <thomas> when it goes up again, it starts again with a new file [19:14] <thomas> so that works [19:14] <thomas> of course, I want to add a queue to provide buffering for the soundcard [19:14] sienap (sy...@ip...) left irc: Ping timeout for sienap[ipc379c096.dial.wxs.nl] [19:14] <thomas> so I do osssrc -> queue -> thread (cutter -> afsink) [19:14] <wtay> right [19:15] <thomas> ok, suppose we start silent. it's just sending recorded data to dev/null [19:15] <thomas> audio goes up, so cutter plugin fires signal, main app catches it, and starts signal handler [19:15] <thomas> signal handler pauses pipeline [19:15] <thomas> and should then change location and restart, which worked in the non-threaded version [19:15] <thomas> in the threaded version, it cannot pause the pipeline [19:15] <thomas> I'm now going through debug output to see why [19:16] <wtay> ok [19:16] <wtay> I have no idea what's wrong actually.. [19:16] <thomas> DEBUG(25695: 1)gst_element_set_state:783: [bin] setting state from PLAYING to PAUSED [19:16] <thomas> DEBUG(25695: 1)gst_element_set_state:783: [audio_src] setting state from PLAYING to P [19:17] <thomas> ok that's where it starts to pause right ? [19:17] <thomas> lets see ... [19:17] <thomas> DEBUG(25695: 1)gst_element_change_state:869: PLAYING->PAUSED: element "audio_src" has parent "bin" and sched 0x808f838 [19:17] <wtay> yup [19:17] <thomas> DEBUG(25695: 1)gst_thread_change_state:400: peer element isn't DECOUPLED [19:17] Last message repeated 1 time(s). [19:17] <thomas> DEBUG(25695: 1)gst_thread_change_state:373: [thread] sync(25695): element "cutter" [19:17] <thomas> DEBUG(25695: 1)gst_thread_change_state:412: [thread] sync(25695): element "cutter" has pad cross sched boundary [19:17] <thomas> DEBUG(25695: 1)gst_thread_change_state:400: peer element isn't DECOUPLED [19:17] <thomas> DEBUG(25695: 1)gst_thread_change_state:421: [thread] sync(25695): waiting for thread to stop spinning [19:17] <thomas> what's that last message for ? [19:17] <wtay> paused should stop the thread from spinning [19:18] <wtay> can you try to add the queue *inside* the thread [19:18] <thomas> hmmm... I should probably rebuild first because I have been trying things in gstthread.c [19:18] <thomas> wtay: ok, I'll try that [19:19] <wtay> 'cause when the queue is blocked, the thread cannot stop.. well not when the queue unlock code is malfunctioning [19:19] <thomas> hmmm, seems to do the same thing. [19:19] <thomas> I'll rebuild first to make sure I'm chasing the right problem here [19:34] <thomas> wtay: ok, here I am again, and here's my error [19:34] <thomas> when I have audio, this is what happens : [19:34] <thomas> ** ERROR **: file gstthread.c: line 351 (gst_thread_change_state): assertion failed: (!pthread_equal(self, thread->thread_id)) [19:34] <thomas> aborting... [19:34] <thomas> Aborted (core dumped) [19:35] <thomas> looking in the code, it says that should be allowed, but isn't [19:36] <wtay> hmm [19:38] <wtay> it says here that it is not allowed currently (but should be in the future) [19:39] <thomas> yeah ... has anyone ever paused a bin with a thread before ? [19:39] <wtay> not in the callback of an element of the thread itself [19:40] <thomas> wtay: hmmm... you mean it might be a problem because when the signal fires, it's fired from a plugin inside the thread ? [19:41] <thomas> in that case, I could set a global variable in the signal handler, and check that variable between iterations [19:42] <wtay> the problem is because you change the threads state inside the threads context [19:42] <thomas> wtay: ok, so if I set a variable and wait for the program to go outside the thread's context, I'm all right ? [19:42] <wtay> yup [19:42] <thomas> ok, I'll try that. [19:43] <thomas> I was wondering about something else... I have the feeling that gstreamer tries to start up as quickly as possible at first ... [19:43] <thomas> ... but when you're doing recording, you'd want the app to set up everything first and get all this stuff out of the way before recording. [19:43] <thomas> is there some way to force gstreamer to first load everything it's gonna need and get into a stable state before actually starting any work ? [19:44] <wtay> like making sure the plugins are loaded? [19:44] <thomas> wtay: yeah, stuff like that. I don't know what gstreamer does behind the scenes other than that [19:46] <wtay> gst_plugin_load was meant to do that but I don't know it'll work [19:46] <wtay> anyway, lazy binding happens at runtime [19:49] <thomas> wtay: and would it help to set a bin to playing and paused first, then sleep a bit ? [19:49] <wtay> not really.. [19:49] <wtay> I think.. [20:01] <thomas> wtay: ok, cool, now it works ! [20:01] <thomas> wtay: hmm, I'll check if it solves my latency issues tomorrow ... [20:01] <wtay> cool [20:02] <thomas> hmm... so that thread problem I have should be allowed in the future then ? [20:03] <wtay> thomas: I hope so [20:05] <thomas> ok, I'm going home. [20:05] <thomas> thanks for the help ! [20:05] <wtay> cya [20:05] thomas (th...@21...) left irc: [x]chat [20:56] greg_ (gr...@ho...) joined #gstreamer. [21:12] <wtay> yo [21:13] <greg_> ello [21:14] ChiefHighwater (pa...@te...) joined #gstreamer. [21:14] <wtay> hi [21:14] <ChiefHighwater> Ello [21:14] <ChiefHighwater> I wouldn't expect to mega in here any time soon [21:14] <wtay> ? [21:15] <ChiefHighwater> we were up till 4am playing with a new dual Xeon machine last night [21:15] <wtay> ah :) [21:17] <greg_> ChiefHighwater: UT must run well on such a speedy horse ;0 Arent you tired playing ? ;-) [21:18] <ChiefHighwater> hehe [21:19] <ChiefHighwater> they're only PII 450s with 512k cache, so not the speediest horse at the track [21:19] <ChiefHighwater> but fun to play with, regardless [21:20] <wtay> oh yeah.. obsolete... donate it to a poor belgian hacker instead <g> [21:20] <greg_> what kind of fun are talking about ? Had no occassion to play with SMP machine yet... [21:21] <ChiefHighwater> wtay:was donated to the church by an Intel employee that upgraded [21:21] <wtay> ChiefHighwater: cool [21:21] <ChiefHighwater> greg_:it's fun just to watch the thing chew through numbers compiling [21:21] <wtay> steveb: ? [21:21] <ChiefHighwater> throw on some graphs for cpu/ram and such 8-] [21:22] <wtay> steveb: you broke dynamic pads in -launch... [21:22] <greg_> CHW: ;-)) That's a pity the numbers 450 + 450 dont just add... [21:23] <ChiefHighwater> hehe [21:27] ChiefHighwater (pa...@te...) left irc: Read error to ChiefHighwater[temple-baptist.com]: Connection reset by peer [21:27] ChiefHighwater (pa...@te...) joined #gstreamer. [21:32] Action: ChiefHighwater will brb [21:32] ChiefHighwater (pa...@te...) left irc: [21:36] sully (dev...@hs...) joined #gstreamer. [21:36] omega_ (om...@om...) joined #gstreamer. [21:37] <wtay> hi [21:37] <omega_> yo [21:39] Action: wtay is wrting a reply to the event mail on -devel.. [21:40] ChiefHighwater (pa...@te...) joined #gstreamer. [21:41] <richardb> Evening all. [21:42] <richardb> I'm writing a plugin to draw a histogram of values output from an FFT. [21:42] <richardb> What format would you recommend as easiest to output video in? [21:42] <richardb> ie, what FOURCC to use? [21:43] <omega_> I'd stick with rgb [21:43] <omega_> use rgb32 maybe [21:44] <wtay> hi richardb [21:44] dobey (do...@dr...) left #gstreamer (eh). [21:46] hadess (ha...@pc...) joined #gstreamer. [21:46] <hadess> hey guys [21:47] <richardb> So, that would be GST_MAKE_FOURCC('R','G','B',' '). [21:47] <richardb> Where's the depth defined? [21:47] <richardb> That's not in the FOURCC? [21:48] <omega_> nope [21:48] <wtay> richardb: nope, in the properties [21:48] <richardb> Ah. [21:48] <omega_> fourcc is of relatively limited use [21:48] <wtay> richardb: xvideosink has an example of RGB properties [21:48] ChiefHighwater (pa...@te...) left irc: Ping timeout for ChiefHighwater[temple-baptist.com] [21:48] <richardb> Anyone got a list of what properties ... [21:48] <richardb> ah. [21:48] ChiefHighwater (pa...@te...) joined #gstreamer. [21:52] <ChiefHighwater> ello? am I back? [21:54] <wtay> ChiefHighwater: yup [21:56] <ChiefHighwater> wtay:txs [21:56] <ChiefHighwater> our dsl is having connectivity issues [21:57] <ChiefHighwater> proly due to the fact the people running it are morons [21:57] <wtay> great... [21:57] <ChiefHighwater> right omega_? [21:57] <omega_> yup [21:57] <wtay> omega_: dynamic pads in -launch are broken currently... [21:58] <omega_> yup [21:59] Nick change: ajzzzz -> ajmitch [21:59] <ajmitch> morning all [21:59] <wtay> ajmitch: 'morning [22:00] <ChiefHighwater> ello [22:00] <ajmitch> ello ChiefHighwater [22:00] <wtay> omega_: how did you install glib1.3 to compile the GOBJECT1 branch? [22:00] <omega_> from cvs [22:00] <wtay> from the build dir? [22:00] <omega_> no, installed [22:00] <ajmitch> ooh, GOBJECT1 branch? /me runs off to chek out & install... [22:01] <wtay> no conflicts with other glib install? [22:01] <omega_> [omega@omicron omega]$ pkg-config glib-2.0 --cflags --libs [22:01] <omega_> -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -L/usr/lib -lglib-1.3 [22:01] <omega_> it puts them in those subdirs now by default [22:04] <wtay> ok, trying [22:04] <wtay> right, I need pkgconfig [22:18] <wtay> omega_: commit gstmarshall.h plz.. [22:18] <omega_> oh [22:18] <omega_> ok [22:19] Action: ajmitch tries to build it [22:21] <ajmitch> argh, autogen.sh has slowed down my game of freeciv :( [22:21] <wtay> some undefined symbols when linking an app: g_module_* stuff... [22:22] <ajmitch> you got glib compiled ok? [22:22] <wtay> yup, no prob there [22:23] <omega_> wtay: new configure.base [22:23] <wtay> okie [22:23] Nick change: hadess -> hds-busy [22:25] thomas (th...@ad...) joined #gstreamer. [22:25] <thomas> hi [22:25] <wtay> hi again [22:26] <wtay> we should get -launch and friends working ASAP [22:26] ChiefHighwater (pa...@te...) left irc: Ping timeout for ChiefHighwater[temple-baptist.com] [22:26] <omega_> working on it [22:26] <wtay> ok [22:26] <omega_> there are some mis-conversions of params that I'm tracking down [22:26] <wtay> ok [22:27] <wtay> I spotted some warnings in autoplugger stuff.. will try to fix them [22:30] ChiefHighwater (pa...@te...) joined #gstreamer. [22:35] <ajmitch> hmm, need realaudio replacement, who's going to reverse-engineer the format for me? ;) [22:39] sully (dev...@hs...) left irc: [x]chat [22:42] <steveb> wtay: i'm back - what did I break? [22:42] <wtay> steveb: dynamic pads in -maunch [22:42] <wtay> -launch even [22:43] <steveb> can you give an example? [22:44] <wtay> disksrc loaction=/some.mpg ! mpeg1parse video_00! queue ! { mp1videoparse ! mpeg_play ! sdlvideosink } [22:45] <wtay> the previous version of -launch used to set up a new_pad signal on the mpeg1parse element when the pad was not found, doing the connection to the queue when the signal was emitted [22:45] <wtay> I tried to fix it but I cannot follow the code anymore :) [22:45] <steveb> hrm [22:46] <steveb> ok, so where does video_00 come from? [22:46] <wtay> it's a pad from mpeg1parse, dynamically created t runtime [22:47] <wtay> heh, great.. gstobject is abstract :) [22:47] <omega_> yup [22:48] <wtay> so far for my refcount test.. [22:48] <steveb> try video_[1-16] instead [22:48] <wtay> steveb: uh? [22:49] <wtay> steveb: that pad will never be created.. [22:49] <steveb> if a pad with that name doesn't exist, it tries calling _element_request_pad_by_name [22:49] <wtay> steveb: the pad video_[1_16] doesn't exist either [22:49] <wtay> it's a template [22:49] <wtay> mpeg1parse has just one pad: sink [22:50] <steveb> try it [22:50] <steveb> video_[1-16] [22:50] <wtay> ok [22:51] <wtay> INFO ( 622:-1) Initializing GStreamer Core Library [22:51] <wtay> INFO ( 622:-1) CPU features: (c1c7f9ff) MMX 3DNOW MMXEXT [22:51] <wtay> Couldn't create a 'video_', no such element or need to run gstreamer-register? [22:52] <steveb> ok [22:52] <wtay> steveb: it doesn't work like that, you need to set up a new pad signal when the pad is not found [22:53] <steveb> that might have worked if you had a request_new_pad func [22:53] <wtay> steveb: nope [22:54] <wtay> steveb: well, it might have, but that's not how mpeg1parse works.. it only creates the pad it there is something to push on that pad [22:54] <steveb> so what is supposed to send a signal? [22:54] <wtay> steveb: mpeg1parse [22:55] <wtay> steveb: the code is still in there, but it is never reached [22:55] <steveb> i see it now [22:56] <omega_> ok, /me has sinesrc ! osssink working [22:56] <wtay> woo [22:56] <wtay> bah, element is abstract too... [22:56] <omega_> yup [22:56] <wtay> aha, bin works [22:57] <wtay> heh, good thing I have lots of ram now. :-) [22:58] <ChiefHighwater> lots of ram is good 8-] [22:58] <ajmitch> yep ;) [22:58] Action: ajmitch has 24mb swap free... [22:58] <wtay> Mem: 513388 500508 12880 0 14692 415052 [22:58] <wtay> -/+ buffers/cache: 70764 442624 [22:58] <wtay> Swap: 498004 100936 397068 [22:59] <ajmitch> my... [22:59] <ajmitch> total used free shared buffers cached [22:59] <ajmitch> Mem: 384496 374844 9652 0 45820 255712 [22:59] <ajmitch> -/+ buffers/cache: 73312 311184 [22:59] <ajmitch> Swap: 275296 250532 24764 [22:59] <ajmitch> can you tell i'm running 2.4.5? ;) [23:00] <wtay> yup, same prob here :-) [23:00] <taaz> hmm... latest cvs lame does build shared libs [23:01] <wtay> taaz: yup [23:02] <taaz> deb build rules dont include them though [23:02] <omega_> wtay: update gst/ and plugins/oss/ [23:02] <taaz> screw it... if you want to use lame deb plugin i guess you'll have to build it from cvs until they fix things ;) [23:03] Action: taaz is too lazy to deal with this... [23:03] Action: taaz is getting a number of emails asking where the debs are ;) [23:03] <steveb> wtay: do you mind if I fix this tomorrow? i'm a bit tired [23:04] <wtay> steveb: not at all :-) [23:04] <wtay> steveb: enjoy your sleep [23:04] <steveb> night [23:04] <taaz> poll: should i build debs from latest cvs or stock 0.2.0? [23:04] steveb (st...@no...) left irc: [x]chat [23:04] <wtay> cya [23:04] <wtay> taaz: hmmm [23:04] <taaz> i'm thinking just 0.2.0 for now. [23:04] <omega_> to get -launch working I ust brute-force commented out all gtk stuff [23:05] <taaz> if anything important has been done we should do a 0.2.1 [23:05] <wtay> omega_: only option IMO [23:05] <ajmitch> is latest cvs as good or better than 0.2.0? [23:05] <wtay> taaz: yup [23:05] <wtay> ajmitch: better :-) [23:05] <taaz> i'll do 0.2.0 [23:06] <wtay> we don't want deal with different flavors of 0.2.0 [23:06] <ajmitch> wtay: what has been fixed? [23:06] <wtay> ajmitch: avi stuff, and quicktime [23:06] <ajmitch> ah, ok [23:06] CHW (pa...@te...) joined #gstreamer. [23:06] <ajmitch> what has been broken? ;) [23:07] <omega_> CHW: office link is up,down,up,down [23:07] <omega_> now it's down again [23:07] ChiefHighwater (pa...@te...) left irc: Ping timeout for ChiefHighwater[temple-baptist.com] [23:08] thomas (th...@ad...) left irc: naptime [23:08] <taaz> dont forget some gstmediaplay commits... [23:08] <wtay> ajmitch: nothing yet [23:09] <ajmitch> wow, people can't have been busy if nothing's been broken ;) [23:09] Action: ajmitch can't wait to wake up & start work on it.... ;) [23:09] CHW (pa...@te...) left irc: Ping timeout for CHW[temple-baptist.com] [23:10] Uraeus (csc...@c3...) joined #gstreamer. [23:10] <Uraeus> hey [23:10] <ajmitch> hey Uraeus [23:10] <ajmitch> just in time to volunteer for a new job... [23:11] ChiefHighwater (pa...@te...) joined #gstreamer. [23:11] <Uraeus> stage the GStreamer karaoke contest? [23:12] <wtay> hi Uraeus [23:12] <Uraeus> hi wtay [23:12] <taaz> Uraeus: in that status page, it's rather confusing to have "CVS" and "0.2.0" columns where the 0.2.0 is an improvement over CVS... [23:13] <Uraeus> taaz: I know, I plan on switching them around at my next update [23:13] <taaz> and there is a dangling "<" tag after Frontends/applications heading [23:13] <taaz> its not just switching it around... the data is wrong [23:13] <Uraeus> taaz: I know the status table is just a quick hack of my roadmap [23:13] <Uraeus> taaz: outdated you mean? [23:14] <taaz> um.. like incsched... yellow for CVS and green for 0.2.0? what does that mean? [23:15] <Uraeus> taaz: hmm, yeah thought I had fixed those, think I did some errouneous table clearings [23:15] <wtay> omega_: gtk_main becomes? [23:17] <Uraeus> btw, why have I lost my rights on the gstreamer bugzilla? [23:17] ChiefHighwater (pa...@te...) left irc: Ping timeout for ChiefHighwater[temple-baptist.com] [23:18] <wtay> Uraeus: ? [23:18] <Uraeus> wtay: i can no longer assign or close bugs etc. [23:19] <wtay> Uraeus: checking... [23:21] <wtay> Uraeus: recheck... [23:22] <wtay> omega_: ok sinesrc ! osssink sorta works [23:23] <wtay> piiiep [23:24] <ajmitch> In file included from gstcolorspace.c:23: [23:24] <ajmitch> gstcolorspace.h:24: gdk/gdk.h: No such file or directory [23:24] Action: ajmitch wonders what happened there... ;) [23:24] <wtay> ajmitch: only gst/ works [23:24] <wtay> ajmitch: and plugins/oss [23:24] <ajmitch> ah, ok [23:24] <omega_> wtay: sorta? [23:25] <wtay> works fine, I had to run register first to hear something [23:25] <wtay> footprint is half the size of HEAD one :) [23:26] <omega_> footprint? [23:26] <wtay> probably not fair because head has way more plugins... [23:26] <wtay> memory footprint [23:26] <omega_> ah [23:27] <wtay> 1.8MB vs 3.6MB [23:29] <Uraeus> ok, I am updating the status table, first of all is capsnego still bad in 'adder'? [23:29] <wtay> Uraeus: it's imroved [23:30] <wtay> Uraeus: it's ok now [23:30] <Uraeus> status of audioscale? [23:31] <wtay> same [23:31] <Uraeus> aviencoder? [23:31] <wtay> winenc is ok [23:31] <wtay> aviencoder is almost ok (yellow) [23:32] <wtay> omega_: hmm, need to install gtk2.0 too to compile gstplay [23:32] Nick change: hds-busy -> hadess [23:33] <wtay> hadess: dude [23:33] <hadess> hey wtay [23:34] <Uraeus> wtay: colospace plugin, does it still have bad capsnego? [23:36] <wtay> Uraeus: yup [23:36] Action: wtay checks out the gtk+-enterprise-edition :) [23:37] <Uraeus> disksrc -> still mmapped disk reads problems? [23:37] <Uraeus> and still issues with gsmnfilter? [23:38] <Uraeus> s/gsmnfilter/gsmfilter/ [23:38] <hadess> mmap'ed reads problems ? [23:38] <wtay> Uraeus: disksrc: no change [23:38] <wtay> Uraeus: what's wrong with gsm? [23:39] <Uraeus> wtay: you comment from when I made the entry was: 'Somewhat works' you know better than me what you meant by that :) [23:39] <wtay> it works [23:40] <Uraeus> sinesrc, does it still tie audio caps to core elements? [23:40] <wtay> yup [23:41] <omega_> wtay: so why does sinesrc not work if -register hasn't been run? [23:41] <wtay> omega_: no idea... could be my fault.. it suddenly worked after doing -register [23:41] <omega_> no, I see the same [23:42] <wtay> it is iterating though... [23:42] <omega_> yeah. I haven't read the plugin code recently, but I think I want to rework it a bit [23:43] <wtay> good idea [23:43] <wtay> it looks like shit now [23:43] <omega_> not sure exactly how, but I want to have the shells of the plugins (from the registries) look just like loaded ones, except for a bit [23:44] <wtay> that's how it is now [23:44] <omega_> and a routine that is called before any given operation that requires the plugin to be loaded [23:44] <Uraeus> v4lsrc, still orange? [23:44] <omega_> yup [23:44] <wtay> Uraeus: yep [23:44] <omega_> and that checks the bit, if it's not set, it loads hte plugin, and sets the bit, continues on its way [23:44] <Uraeus> stereo, volume, median, smooth any change? [23:45] <ajmitch> brb [23:45] <wtay> omega_: it has some ugly stuff to remove the old shell and replace it with the newly registered real one [23:45] ChiefHighwater (pa...@te...) joined #gstreamer. [23:45] Nick change: ajmitch -> ajbusy [23:45] <wtay> Uraeus: no change [23:45] <taaz> blah... lame build fails with latest cvs [23:45] <Uraeus> volenv? [23:45] <omega_> ghost plugin structs, basically [23:46] <wtay> Uraeus: same [23:46] <wtay> omega_: need a way to cleanly modify the bit... [23:46] <Uraeus> jpegenc still orange? [23:46] <omega_> wtay: how do you mean cleanly? [23:46] <wtay> Uraeus: yellow [23:46] <Uraeus> jpegdec still yellow? [23:46] <omega_> where'd the problem be? [23:47] <wtay> omega_: when the code was written, the plugin just did add_factory [23:47] <omega_> hmm [23:47] <taaz> oh jeez... all the lame flags are set via funcs now [23:47] <omega_> ok, I"ll look at it later then [23:47] <wtay> omega_: that piece of code added the factory and removed the old one [23:47] <omega_> wtay: right [23:48] <omega_> wtay: are you on gobject now? [23:48] Action: taaz sighs... just wanted to build the debs. why is it so hard? ;) [23:48] ajbusy (aj...@p6...) left irc: Ping timeout for ajbusy[p6-max1.dun.ihug.co.nz] [23:48] ChiefHighwater (pa...@te...) left irc: Ping timeout for ChiefHighwater[temple-baptist.com] [23:48] <wtay> omega_: yes, trying to compile gtk2.0 now [23:48] Action: omega_ can't get mail because the mail server's DSL is going up,down,up,down,down,down,down,u..,down,down,down [23:48] <Uraeus> mpg123, still orange? [23:48] <wtay> Uraeus: no change [23:48] sweeze (jba...@sw...) joined #gstreamer. [23:49] <wtay> yo [23:49] <sweeze> any chance of .deb's in the next 4 days? [23:49] <omega_> taaz: ? [23:49] <Uraeus> system_encode any change? [23:49] <wtay> Uraeus: no change [23:49] <taaz> omega_: ? [23:49] <Uraeus> mpeg2enc any change? [23:49] <omega_> taaz: debs? [23:49] <wtay> Uraeus: no change [23:50] <taaz> i'm trying damnit! ;) [23:50] <wtay> Uraeus: only ask for plugins with CVS commits done on them :-) [23:50] <taaz> any volunteers to update lame-ass plugin to latest api? [23:50] <Uraeus> wtay: don't rember all CVS commit, but I will try to filter more [23:50] <Uraeus> rtjpegenc and dec? [23:50] <wtay> no change [23:51] <Uraeus> ladspa? [23:51] <wtay> Uraeus: no change AFAIK [23:51] <taaz> sweeze: i was going to put the debs up right now but now it appears i hit compatability problem between lame and gst lame plugin [23:51] <taaz> sweeze: what happens in 4 days? [23:52] <sweeze> taaz:i go away for the summer ;-) [23:52] <omega_> taaz: upload the rest of the debs for now, we can fix lame later [23:53] <Uraeus> hmm, was sure I saw some ladspa work recently [23:53] <Uraeus> ok, Quicktime? [23:53] <taaz> omega_: that's a hassle ya know.. i have to disable building it and blah blah blah... would be easier if it just freaking built [23:54] <taaz> let me take a 5 min stab at getting it to work [23:54] <omega_> wtay: colorspace depends on gdk.. [23:55] <Uraeus> wtay: quicktime? [23:55] <omega_> wtay: so does videoscale [23:56] <wtay> omega_: yup [23:56] <wtay> Uraeus: works [23:57] <wtay> omega_: we should now check for gtk I guess.. [23:57] <omega_> why do those need gdk? [23:57] <omega_> they're inline filters, not sinks [23:57] <wtay> omega_: ancient history with metadata [23:58] <omega_> no, can that be removed then? [23:58] <wtay> omega_: I guess the include can be removed [23:58] <omega_> no, GdkVisual is referenced in the code of colorspace [23:58] <wtay> ok [23:58] <wtay> we don't even use colorspace anymore... [23:58] <omega_> we don't? [23:58] <wtay> not as a lib, no [00:00] --- Fri Jun 15 2001 [00:03] <Uraeus> ok, status table is now organized better and plugin statuses should be more correct [00:04] <omega_> wow. two scripts and a couple manual fixes, and dvdec is ported to gobject [00:06] <wtay> omega_: make sure to commit the scripts too... [00:07] sweeze (jba...@sw...) left irc: Client Exiting [00:07] sweeze (jba...@sw...) joined #gstreamer. [00:08] <Uraeus> well, I need to sleep, company party tommorow, which is my last chance/hope/attempt to win over a certain girl [00:08] <Uraeus> see ya [00:08] Uraeus (csc...@c3...) left irc: syntax error - user imploded [00:08] <wtay> gtk2.0 works... [00:10] <wtay> bleagh, I need gnome2.0 too... [00:12] <wtay> pff, I'm not gonna build the complete gnome CVS.. [00:13] <omega_> why do you need it? [00:14] ChiefHighwater (pa...@te...) joined #gstreamer. [00:14] <wtay> omega_: gstplay [00:14] <wtay> a bit premature, I agree, but still... [00:15] shitowax (yann@AAnnecy-101-1-3-16.abo.wanadoo.fr) joined #gstreamer. [00:15] <shitowax> yo [00:15] <wtay> yo [00:16] <shitowax> I've played a little with the quicktime plugin ... [00:16] <shitowax> I've think I've understood why the mjpeg codec is broken [00:17] ChiefHighwater (pa...@te...) left irc: Ping timeout for ChiefHighwater[temple-baptist.com] [00:17] <wtay> shitowax: ah? [00:17] <shitowax> there are specific data which encapsulate the jpegs in quicktime (I've found this in the Quicktime norme) [00:18] <shitowax> I wonder if I should add this in the jpeg codec or in the quicktime plugin ... [00:18] <omega_> qt [00:18] ChiefHighwater (pa...@te...) joined #gstreamer. [00:19] <shitowax> why, this seems to be part of the jpeg norm ... [00:19] <shitowax> sepecific to mjpeg [00:19] <omega_> if it's in the jpg qt atom, then it goes in jpegdec [00:20] <omega_> but if it's in a different atom, the qt demux has to deal with it somehow, either by merging it into the jpeg stream, or something [00:20] <wtay> shitowax: qttypes.c was made for that.. [00:21] <wtay> shitowax: does this extra data have a functionality? [00:21] ChiefHighwater (pa...@te...) left irc: Ping timeout for ChiefHighwater[temple-baptist.com] [00:22] sweeze (jba...@sw...) left #gstreamer (Client Exiting). [00:23] <shitowax> wait, I cannot refind the URL =) [00:24] ChiefHighwater (pa...@te...) joined #gstreamer. [00:25] <omega_> pff. he joins for 5 seconds, then it goes down again [00:25] <shitowax> got it : http://developer.apple.com/techpubs/quicktime/qtdevdocs/QTFF/qtff-64.html [00:27] ChiefHighwater (pa...@te...) left irc: Ping timeout for ChiefHighwater[temple-baptist.com] [00:27] <wtay> shitowax: looks like it does interlaces mjpeg somehow, this explains why the picture is squashed [00:29] <shitowax> so, I had it in the quicktime plugin ? [00:29] <wtay> add, you mean? [00:29] <shitowax> yep, sorry :) [00:30] <wtay> yup [00:30] <hadess> i broke dpkg and cvs by downgrading glibc... [00:30] <wtay> although I dont know you even can do that.. [00:31] <shitowax> ok, got another question for the 3ivx plugin this time ... [00:31] <wtay> hadess: well done, hadess [00:31] <wtay> shitowax: yup? [00:31] <taaz> someone want to tell me what the mode_fixed equivalent is in the new lame api? [00:32] <shitowax> when you ask for a new buffer to a pool, is the first address you receive always valid, or do you have to ask for a new one each time ? [00:34] greg_ (gr...@ho...) left irc: Ping timeout for greg_[home.sente.pl] [00:34] <hadess> ok, off to bed [00:34] <hadess> cya guys [00:34] hadess (ha...@pc...) left irc: sleep *cough* sleep [00:35] <taaz> shitowax: it should be valid until its refcount hits 0 and it goes back to the pool [00:35] <shitowax> and what happens if I keep the refcount always > 0 ? [00:36] <taaz> why would you do that? [00:36] <shitowax> to avoid a new allocation [00:36] <taaz> but that's what the pools do [00:36] <taaz> or.. should do ;) [00:37] <shitowax> sorry, but when I use a new buffer I have to do much more copy than using always the same buffer [00:37] <taaz> for example the mpeg2dec plugin.. it asks for frames from its src pads. it draws into them, pushes them on, then asks for new ones [00:38] Action: taaz pokes wtay [00:38] <wtay> sorry, was away [00:38] <taaz> i'm probably confusing shitowax... stop me soon ;) [00:38] <shitowax> to explain everything, there is an option in 3ivx to only copy the moving blocks, this helps to speed the decoding process [00:39] greg_ (gr...@ho...) joined #gstreamer. [00:39] <omega_> shitowax: where is this copy and when is it happening? [00:39] <wtay> shitowax: you cannot assume to have direct access to the framebuffer [00:40] <shitowax> there is no way to have access to the frame buffer ? [00:40] <wtay> so if you request a new buffer, you cannot assume the data of the previous one is still in it [00:40] <taaz> that's going to be rough... you basically want a static framebuffer? that doesnt work too well if downstream wants to do something with it [00:40] <wtay> indeed [00:41] <wtay> hmm [00:41] <shitowax> But why can't I assume the data is still in the buffer if it hasn't been released ? [00:42] <wtay> the buffer you pushed might be inside a queue to be shown at some point [00:42] <shitowax> so it's not a real "direct hardware access" ? [00:42] <wtay> or some other plugin might have messed with it in place [00:43] ajbusy (aj...@p3...) joined #gstreamer. [00:43] <wtay> not always, no [00:43] <shitowax> ok, then 3ivx will be a bit slower ;) [00:43] <taaz> you're not supposed to know what happens to your buffers. makes hooking pieces together easier [00:43] <wtay> maybe a bufferpool should have a flag indicating exclusive access [00:44] <wtay> omega_: ideas? [00:44] Nick change: ajbusy -> ajmitch [00:44] <taaz> wtay: that makes things really complex ;) what if you ahve an encoder downstream that requires the last N frames to be buffered? you'd have to have caps negotiation insert buffer copier or something [00:45] <wtay> taaz: not really, when you get a buffer from the pool, there could be a way to find out that it is the same you had before... [00:46] <taaz> so all plugins would have to check that? [00:46] <wtay> not really [00:46] <wtay> well, time for bed... [00:46] <ajmitch> night [00:46] <wtay> I'll dream about it :-) [00:46] <shitowax> I've think about you comments about audio in the QT decoder, and you're right, It need to completly rethink ... You cannot have threads in a gstreamer plugins ? [00:47] <wtay> shitowax: you can, but it's not recommended at all [00:47] Nick change: wtay -> wtay-zZz [00:47] <shitowax> ok, then I got an idea how to do it ... ;) [00:48] <taaz> i think omega_ was working on composite elements, which i guess could have component threads [00:48] <shitowax> what is a "composite element" ? [00:49] <taaz> a element that is actually a container for other elements [00:50] <shitowax> like the avi decoder ? or like a seperate pipe ? [00:50] <shitowax> Do you know if there is a cropping filter in gstreamer ? [00:52] <taaz> no idea.. amusing that there is so much stuff you cant even tell ;) [00:53] <shitowax> Tazz, what is you speciality inside Gstreamer ? [00:54] <taaz> who is Tazz? ;) [00:55] <shitowax> sorry, need another keyboard =) [00:55] <taaz> hell if i know... i'm doing the debs, if i had time i'd be writing a dvd player since i'm a part of the LiViD crew working on OMS [00:56] <taaz> i'll write python bindings at some point too ;) [00:56] <shitowax> hum ... is OMS still living ? (no offense ;)) [00:57] <taaz> gpl code never dies [00:57] <taaz> its just waiting for us to work on it [00:58] <shitowax> you work on the interface, or maybe the core ? [00:59] <taaz> whatever needs work [00:59] <taaz> i really dont care how this stuff works... i just want to watch my dvds ;) [00:59] <shitowax> We (3ivx team) have been thinking of creating plugin for OMS ... [00:59] <taaz> and i find it much more fun to integrate libraries into a greater whole than to fiddle with sync and codec issues [01:00] <shitowax> What do you use now for your DVD ? [01:00] <taaz> so i was trying to put gstreamer into oms... turns out that kinda kills 80-90% of the current oms code [01:01] <taaz> this is just my personal opinion here: don't write OMS plugins. no one is currently working on OMS and IMHO alot of the code is crap compared to what you can do with gstreamer [01:02] <shitowax> That's why we have chosen Gstreamer ;) [01:02] <taaz> oms is kind of lacking a leader right now. dent seems to think he is still working on oms but no cvs checkins for like 8 months is not encouraging [01:03] <shitowax> I think libmpeg2 was one cool thing to keep from Livid [01:03] <taaz> i wouldn't mind just taking over and throwing my stuff into the mix but it tramples all over the code that's there and is a completely different approach [01:03] <shitowax> DVD are not playable for the moment with gstreamer ? [01:03] <taaz> yeah mpeg2dec and ac3dec are great. walken is working on ac3dec now but for some reason doesnt use public cvs... i have no idea why not [01:04] <taaz> no dvdsrc is broken [01:04] <shitowax> is CSS integrated ? [01:05] <taaz> one of the other dvd projects will be releasing code in about 2 weeks ago ;) they have a neat little library to do the dvd access and css crap (based on libcss). its a really nice interface that fits right into dvdsrc [01:06] <shitowax> ... what DVD project, Xine, VLC, ... ? [01:06] <taaz> i tried it with a private release of the lib and got gstreamer to play dvds... except for some bug that makes it halt after a few seconds [01:06] <taaz> i dont think it has a name [01:07] <taaz> they have menus and so on working i think [01:08] <omega_> is this open-source or not? [01:08] <taaz> kind of discouraging to try and work on oms when vlc and this other project have active developers that are good at reverse engineering dvd format stuff [01:08] <taaz> its not released yet [01:08] <omega_> right, is it going to be open-source or not? [01:08] <shitowax> cool ... I've tried some VOB with the current CVS snapshot I have on my linux box, and it was not synchronized and has a big memory leak ;) [01:08] <taaz> yeah it should be [01:09] <taaz> omega_: i know you want to help me... [01:09] <taaz> grab latest lame cvs and tell me what func call that mode_fixed parameter setting should use [01:10] <taaz> i updated the lame plugin to use the get/set function api vs the struct member api from before... but this one parameter is confusing me [01:12] <taaz> oh... whoops, i found it [01:12] <taaz> they removed it ;) [01:12] <omega_> uh? [01:12] <taaz> removed mode_fixed variable. If calling program selects a mode [01:12] <taaz> (mono, stereo, jstereo) then that mode is used, otherwise [01:12] <taaz> lame picks best mode to use. [01:13] <omega_> ah [01:13] <shitowax> Is somebody creating a generic encoding tools using Gstreamer (some kind of Non Linear Video Editor) ? [01:14] <omega_> not yet, but that's the plan [01:15] <taaz> should gst_lame_change_state be used? [01:15] <omega_> yes [01:15] <omega_> it has to be [01:15] <taaz> its not called anywhere [01:15] Action: taaz thinks -Werror should be on to make this sort of thing obvious [01:16] <shitowax> ciao [01:16] <taaz> what's the one liner patch to fix this? gst_element_add_change_state_function() or something? [01:16] shitowax (yann@AAnnecy-101-1-3-16.abo.wanadoo.fr) left #gstreamer. [01:16] <omega_> check anohter plugin <g> [01:17] <taaz> well, its not like anyone noticed... this is 0.2.0 code [01:17] <taaz> the test suite didnt report any problems so it must work <g> [01:17] <omega_> if there's nothing to do on state change, then it doesn't matter, but if there is, it needs to be added [01:17] <omega_> taaz: if you're gonna keep making jokes about a test suite, please help write one [01:19] <taaz> yeah, sorry. got to finish a thesis first [01:21] <omega_> cause a test suite is the next thing on my list of things to do for ridgerun [01:39] <omega_> oooooh [01:39] <omega_> have a perl script that converts plugins [01:39] <omega_> it converted mad with no problems at all [01:40] <omega_> and BRANCH-GOBJECT1 is playing mp3s! [01:40] <ajmitch> nice ;) [01:41] <ajmitch> gstreamer-launch compiled for you, or do i need to update? [01:41] <omega_> gstmpg123 compiled fine too [01:41] <omega_> I did a brute-force removal of all the gtk code [01:41] <omega_> I don't want to commit that yet, because I want to make it work [01:41] <ajmitch> ah, ok ;) [01:42] <omega_> mpg123 works too [01:42] <omega_> but things are much more fragile now if -register isn't up to date [01:42] <omega_> don't think that's gobject's fault at all, though [01:42] <ajmitch> oh? [01:42] <omega_> try sinesrc ! osssink on HEAD without a recent registry [01:43] <omega_> see if that makes noise or not [01:43] <ajmitch> how can it be worse than usual? ;) [01:43] <omega_> um [01:43] <omega_> yeah, well, that's on my hitlist now [01:44] <ajmitch> [ajmitch @ ajmitch tools] ./gstreamer-launch sinesrc ! osssink [01:44] <ajmitch> INFO (27889:-1) Initializing GStreamer Core Library [01:44] <ajmitch> INFO (27889:-1) CPU features: (808029bf) MMX 3DNOW [01:44] <ajmitch> ** WARNING **: gstplugin: registry needs rebuild: run gstreamer-register [01:44] <ajmitch> RUNNING pipeline [01:44] <omega_> but no sound? [01:44] <ajmitch> no sound [01:44] <omega_> and if you do it with a valid reg.xml ? [01:45] <ajmitch> i get a headache ;) [01:45] <ajmitch> don't try it with freq=15000 ;) [01:45] <omega_> yeah, don't do that [01:47] <omega_> eos even worked [01:48] <ajmitch> hmm [01:48] greg_ (gr...@ho...) left irc: Read error to greg_[home.sente.pl]: Connection reset by peer [01:49] <ajmitch> so you going to try & convert all the plugins? [01:49] <omega_> yup [01:49] <ajmitch> what changes are required? [01:50] <ajmitch> i see ;) [01:51] <omega_> perl is your friend [01:52] <ajmitch> perl == evil ;) [01:54] <taaz> um... so why are all my plugins linked to libasound? [01:54] <omega_> good question [01:55] <omega_> did someone add asound to the wrong var in configure ? [01:55] <taaz> not sure... haven't figured out why yet [01:56] <taaz> this whole build system is confusing ;) [02:07] Action: taaz 's head explodes [02:07] <taaz> the generated alsa configure code is not very clear [02:08] <omega_> you mean in configure.sh itself? [02:08] <taaz> yea [02:09] <omega_> that isn't supposed to be clear [02:09] <taaz> i think its adding alsa crap to LIBS direct [02:09] <omega_> grrrr [02:10] <taaz> that's fixable... just need to save/restore LIBS and CFLAGS [02:10] <omega_> yea [02:10] <omega_> ick [02:11] <taaz> i'm pretty sure -lasound isnt supposed to be in LIBGST_LIBS ;) [02:11] <omega_> right [02:12] <taaz> but we are doing a save/restore... i'm getting confused [02:16] <taaz> hmm... looks like maybe alsa is also doing the save/restore using the same vars and some contention there... maybe [02:17] <omega_> grr [02:46] Nick change: ajmitch -> ajbusy [03:22] <taaz> 'fixed' alsa issue... it still has some redef of ALSA_{LIBS,CFLAGS} cause of the dual check... but it doesnt spam every lib with -lasound now. [03:22] <taaz> shared lib lame check needs -lm looks like [03:23] <taaz> takes way too long to mess with this stuff ;) I make a change, run script to test, go off do something else for 20 mins ;) [03:32] <omega_> I think -lm should be linked to everything anyway [03:32] sienap (sy...@ip...) joined #gstreamer. [03:32] sienap (sy...@ip...) left #gstreamer. [03:43] <taaz> well it didn't here. failed the check cause no pow() [03:43] <taaz> before it was doing static link for me so i guess the static lame lib was ok [04:39] johnix (thomas@61.152.140.57) joined #gstreamer. [05:17] jerwin (er...@66...) joined #gstreamer. [05:37] jerwin (er...@66...) left irc: [05:37] jerwin (er...@66...) joined #gstreamer. [06:06] jerwin (er...@66...) left #gstreamer. [06:09] mwc (ma...@ly...) joined #gstreamer. |
From: <wim...@ch...> - 2001-06-16 04:29:51
|
[06:31] _gst_newt_ joined #gstreamer. [07:00] pimlottc (pim...@mo...) joined #gstreamer. [07:44] walken (fo...@c1...) joined #gstreame= r. [07:48] steveb (st...@no...) joined #gstreamer. [07:49] <taaz> ack! [07:49] <taaz> building these debs is sucking! [07:53] big_T (th...@fy...) got netsplit. [07:55] big_T (th...@fy...) returned to #gstreamer. [08:22] <big_T> 1/msg nickserv identify <his not so secret password> [08:22] <big_T> opps [08:22] <big_T> damn :) [08:23] <big_T> too bad this channel is logged and sent out on the intern= et too [08:23] <big_T> :) [08:23] <big_T> hoho [08:23] <johnix> big_T: you're good for a new pass! [08:23] <big_T> perhaps I should change the password? ;) [08:23] <omega_> email wim...@ch... and he'll remove it [08:23] <big_T> nay, I'll change it instead - see it as an easteregg in t= he logs ;) [08:23] <omega_> otherwise it'll be line #0 of the next daily log mailing= ;-) [08:24] <johnix> if I'm quick enough, I can use it an blast everyone off = that chan [08:24] <big_T> there, new password [08:25] <walken> honk honk [08:25] <omega_> http://www.lwn.net/2001/0614/devel.php3 [08:26] <johnix> note that this was a pretty weak password... 'hope the n= ew one is better [08:27] <big_T> johnix, well - it don't see this one as an important one [08:28] <big_T> have enough of _real_ passwords to keep track of, without= counting the countless number of useless passwords [08:28] <big_T> and I don't view me loosing my nick as very important eit= her [08:31] <big_T> anyhow, just sent the mail to wtay :) [08:31] <big_T> omega_, have you forgotten my alsa-patch or rejected it? [08:32] <omega_> um [08:32] <omega_> probably forgotten [08:32] <big_T> hoho [08:32] Action: big_T teases omega_ a little [08:33] <steveb> omega_: can you confirm if i have fixed gstparse? [08:33] <omega_> um, what are you trying to fix? [08:37] <steveb> mpeg1parse video_00! was broken [08:37] <omega_> ok, um, well, just try it [08:38] <steveb> i don't think I have an appropriate mpeg file [08:38] <omega_> gstreamer.net/media/AlienSong.mpg [08:39] <steveb> what is a full pipeline? [08:39] <omega_> disksrc location=3D.. ! mpeg1parse video_00! mpeg2dec ! = sdlvideosink [08:39] <omega_> or xvideosink [08:41] Action: omega_ will be back in ~10min [08:42] <steveb> i can't build mpeg2dec atm [08:42] Nick change: steveb -> steveb-away [08:49] big_T (th...@fy...) left irc: Ping timeout for big_= T[fysgr763.sn.umu.se] [08:50] ajbusy (aj...@p3...) left irc: Ping timeout f= or ajbusy[p39-max7.dun.ihug.co.nz] [08:52] ajbusy (aj...@p3...) joined #gstreamer. [08:59] big_T (th...@fy...) joined #gstreamer. [09:03] ajbusy (aj...@p3...) left irc: Ping timeout f= or ajbusy[p39-max7.dun.ihug.co.nz] [09:28] makki (yv...@ma...) joined #gstreamer. [09:29] <taaz> got the debs built [09:31] <makki> there are debs around ? [09:32] <omega_> taaz: you're a release tech, so go ahead and release the= m [09:32] <taaz> there will be soon [09:34] <taaz> a what? [09:34] <omega_> you can do the sf release of the files yourself [09:34] <taaz> um... i didnt plan on using the official sf release stuff.= .. was going to just plop things in a dir [09:34] <omega_> well, should do both [09:35] <taaz> 1) there are probably many little problems i'll have to cl= ean up. no reason to 'release' new debs every day. [09:35] <omega_> ok [09:36] <taaz> 2) 99% of the people using the debs will do so via apt. g= oing to be kind of weird to figure out how to set that up with download d= irs i cant control [09:37] Action: taaz is too tired to remember what his other points were = ;) [09:47] <taaz> blah... i swapped files between arts and artsd plugins [09:49] <makki> taaz: np [09:56] <johnix> aliensong: very good! [09:56] <taaz> omega_: gst is going to have to start using proper lib ver= sioning soon [09:57] <omega_> yup [10:04] arik (ar...@po...) joined #g= streamer. [10:04] <arik> hi [10:06] walken (fo...@c1...) left irc: zzzZZz= Zzz [10:40] Nick change: taaz -> taazzzz [10:41] mwc (ma...@ly...) left irc: [x]chat [10:53] omega_ (om...@om...) left irc: Read error to omega_[omegacs= .net]: EOF from client [11:17] makki (yv...@ma...) left irc: Ping timeout for makk= i[makki.wait-a-bit.com] [11:19] thomas (th...@21...) joined #gstreamer. [11:22] makki (yv...@ma...) joined #gstreamer. [11:41] arik (ar...@po...) left irc:= Ping timeout for arik[pool0357.cvx19-bradley.dialup.earthlink.net] [11:42] <thomas> don't you just *love* how a built gstreamer cvs tree amo= unts to 215 MB of stuff ? ;) [12:26] hadess (ha...@pc...) joined #gst= reamer. [12:43] johnix (thomas@61.152.140.57) left irc: Weekend! [13:30] makki (yv...@ma...) left irc: Back to the real worl= d [14:12] Nick change: hadess -> hds-sick [14:33] richardb (ri...@ix...) left irc: richardb [14:34] ajbusy (aj...@p3...) joined #gstreamer. [14:35] sienap (sy...@ip...) joined #gstreamer. [14:42] sienap (sy...@ip...) left irc: bbl [14:43] <thomas> ls -l [14:53] Nick change: ajbusy -> ajzzzz [16:45] lsetia (lsetia@203.200.123.32) joined #gstreamer. [16:45] <lsetia> hello! [16:46] <lsetia> has anyone successfully uncompressed the gstreamer-0.2 b= z2 file ? [16:46] <lsetia> I downloaded it twice and got errors both times [16:46] <lsetia> gstreamer-0.2.0/plugins/mpeg2/ac3dec/downmix_i386.S [16:46] <lsetia> gstreamer-0.2.0/plugins/mpeg2/ac3dec/imdct512_kni.S [16:46] <lsetia> bzip2: Data integrity error when decompressing. [16:46] <lsetia> Input file =3D gstreamer-0.2.0.tar.bz2, output f= ile =3D (stdout) [17:01] <lsetia> hey is anyone around :) [17:02] <lsetia> ok, see you guys later :) [17:02] lsetia (lsetia@203.200.123.32) left irc: noone's watching me [17:04] sienap (sy...@ip...) joined #gstreamer. [17:04] <sienap> boe [17:25] _gst_newt_ joined #gstreamer. [17:54] sienap (sy...@ip...) left irc: No windows for th= is server [17:58] synk (synk2@D5E0BB7C.kabel.telenet.be) joined #gstreamer. [17:59] synk (synk2@D5E0BB7C.kabel.telenet.be) left #gstreamer. [18:12] ChiefHighwater (pa...@te...) joined #gstreamer. [18:24] ajzzzz (aj...@p3...) left irc: Ping timeout f= or ajzzzz[p39-max7.dun.ihug.co.nz] [18:25] ChiefHighwater (pa...@te...) left #gstreamer. [18:26] ChiefHighwater (pa...@te...) joined #gstreamer. [18:26] #gstreamer: mode change '+o ChiefHighwater' by ChanServ!s@ChanSer= v [18:27] omega_ (om...@te...) joined #gstreamer. [18:30] ajzzzz (aj...@p3...) joined #gstreamer. [18:30] #gstreamer: mode change '-o ChiefHighwater' by ChiefHighwater!pau= l...@te... [18:33] Nick change: wtay-zZz -> wtay [18:33] <wtay> hi [18:33] <omega_> yo [18:33] <wtay> oh, how I hate avifile.. :( [18:40] Nick change: wtay -> wtay-nap [18:44] <steveb-away> omega_: early morning or late night? [18:44] <omega_> what's the diff? [18:44] Nick change: steveb-away -> steveb [18:44] <steveb> sleep(1) [18:51] ChiefHighwater (pa...@te...) got netsplit. [18:51] steveb (st...@no...) got netsplit. [18:51] hds-sick (ha...@pc...) got netsp= lit. [18:51] ShrimpX (marius@131.252.244.168) got netsplit. [18:51] iGN___ (ig...@lo...) got netsplit. [18:51] pimlottc (pim...@mo...) got netsplit. [18:51] wtay-nap (wi...@ca...) got netsplit. [18:51] taazzzz (dlehn@66.37.66.32) got netsplit. [18:51] omega_ (om...@te...) got netsplit. [18:55] wtay-nap (wi...@ca...) returned to #gs= treamer. [18:55] taazzzz (dlehn@66.37.66.32) returned to #gstreamer. [18:55] iGN___ (ig...@lo...) returned to #gstreamer. [18:55] ShrimpX (marius@131.252.244.168) returned to #gstreamer. [18:55] pimlottc (pim...@mo...) returned to #gstream= er. [18:55] steveb (st...@no...) returned to #gstreamer. [18:55] hds-sick (ha...@pc...) returned = to #gstreamer. [18:55] ChiefHighwater (pa...@te...) returned to #gstreamer. [18:55] omega_ (om...@te...) returned to #gstreamer. [18:55] #gstreamer: mode change '+o ChiefHighwater' by ChanServ!s@ChanSer= v [18:56] #gstreamer: mode change '-o ChiefHighwater' by ChiefHighwater!pau= l...@te... [18:57] greg__ (gr...@ho...) joined #gstreamer. [18:58] <greg__> hi all [18:58] <omega_> yo [18:59] <ChiefHighwater> ello [18:59] Topic changed on #gstreamer by ChanServ!s@ChanServ: LWN covers GS= treamer release: http://www.lwn.net/2001/0614/devel.php3 [19:00] <greg__> whatcha doing ? [19:00] <ChiefHighwater> reading [19:00] <omega_> porting the rest of the plugins to gobject [19:02] <greg__> greg is cutting debian potato to fit on a very very smal= l disk (firewall/iptables/socks/http-ftp-proxy/...) [19:02] <omega_> cool [19:02] <omega_> slicing a potato? <g> [19:03] <omega_> or dicing? [19:03] <greg__> yeah - it is friday today - time to get to work ;-) if o= nly exams were done. [19:03] Nick change: hds-sick -> hadess [19:04] <greg__> omega_: I know what slicing is, but dicing ??? what do y= ou mean ? [19:04] <omega_> variant of the same thing, from the culinary world [19:06] <greg__> omega_: ah - you got practise from USA i think. Answerin= g your question: Need to fit it on 80MB disk, with some place left for lo= gs which are planned to take much place. Just installed and removing usel= ess stuff (locales, unneded kernel modules)... [19:06] <omega_> greg__: I'm a native... ;-) [19:07] <greg__> It is funny, but at beggining I cut it from 50MB to 22MB= , where 8M in those 22MB were text files used by apt... [19:07] <omega_> is that fitting on a hard drive of that size, or a CD ? [19:07] <greg__> omega_: HDD [19:07] <omega_> ok [19:07] <hadess> hey guys [19:07] <omega_> thin client? [19:07] <omega_> hadess: yo [19:07] <greg__> omega_: ups...! I thought you're from europe somewhere -= would have my hand cut for that ;-) [19:07] <hadess> hey omega_ [19:08] <hadess> omega_: take that as a compliment ;) [19:08] <greg__> omega_: just a firewall to play with and have some fun/e= xpirience. [19:08] <greg__> I used LRP till yesterday. [19:08] <hadess> greg__: and don't ever say that anglo-saxons know how to= cook in front of me [19:09] <greg__> hadess: I was talking about omega's dictionary not about= cooking ;-) [19:09] <ChiefHighwater> hadess: I may not know how to cook, but I sure k= now how to eat 8-] [19:09] <hadess> CHW: fair enough dude [19:09] <omega_> last I checked, most people knew how to eat, else..... [19:09] <greg__> CHW: same here in Europe ;-)) [19:09] lsetia (lsetia@203.195.140.217) joined #gstreamer. [19:19] <thomas> lsetia: the bz2 has errors, get the .tar.gz [19:19] <thomas> hi [19:20] <lsetia> ah ok! [19:20] <lsetia> I'm getting it now :) [19:20] <thomas> omega_: I'm having a gstreamer program run all weekend l= ong, hope that works ;) [19:20] <thomas> and for a few months on end after that [19:22] <hadess> hope for you you won't get eos ... [19:22] <hadess> `We are not here to give users what they want'=A0 -- RMS= , at GUADEC 2001 [19:22] <hadess> i just love cactus ;) [19:26] <greg__> hadess: do I see right: You call RMS a cactus ??? [19:43] Action: hadess smacks greg__ with a cluestick and some community = knowledge [20:09] lsetia (lsetia@203.195.140.217) left irc: Read error to lsetia[20= 3.195.140.217]: EOF from client [20:19] big_T (th...@fy...) left irc: Read error to big_T[f= ysgr763.sn.umu.se]: Connection reset by peer [20:20] Nick change: wtay-nap -> wtay [20:20] <wtay> yawn [20:23] Thunderbird (rod...@7d...) joined #gstreamer. [20:23] <wtay> yo [20:25] <steveb> hi [20:25] <Thunderbird> hi [20:28] Nick change: hadess -> hds-afk [20:29] <steveb> wtay: can you check i unbroke gstparse? [20:29] <wtay> steveb: ok [20:44] Thunderbird (rod...@7d...) left #gstreamer. [20:45] rowenc (rowenc@Roaming-209-53.PLU.edu) joined #gstreamer. [20:45] <wtay> hi [20:45] <rowenc> hey [20:51] makki (yv...@ca...) joined #gstreame= r. [20:53] <makki> hello [20:53] <wtay> hi [20:53] thomas (th...@21...) left irc: [x]chat [21:00] <wtay> grrrr... gnome build script has set the LD_BIND_NOW env va= r.. [21:03] <wtay> steveb: it works again (with lots of WARNINGS first..) [21:05] <steveb> yeah, gstparse either needs to check that there really i= s a new_pad func or the warning in gstpad needs to be more subtle [21:05] <wtay> yup, fixing some stuff now [21:06] rowenc (rowenc@Roaming-209-53.PLU.edu) left irc: reboot [21:18] greg__ (gr...@ho...) left irc: Read error to greg__[home.s= ente.pl]: Connection reset by peer [21:19] greg__ (gr...@ho...) joined #gstreamer. [21:24] Nick change: taazzzz -> taaz [21:26] <wtay> yo [21:27] Nick change: hds-afk -> hadess [21:33] mattsm (ma...@ad...) joined = #gstreamer. [21:36] <mattsm> does anyone want this updated tools/README I was unable = to give to someone before I had to go away for a week [21:37] <wtay> mattsm: i can do that, just mail it to me [21:37] <wtay> wim...@ch... [21:37] <mattsm> wtay: ok [21:42] mattsm (ma...@ad...) left ir= c: Client Exiting [21:43] ChiefHighwater (pa...@te...) left irc: Ping timeout f= or ChiefHighwater[temple-baptist.com] [21:49] omega_ (om...@te...) left irc: [x]chat [21:50] makki (yv...@ca...) left irc: Back t= o the real world [21:56] ChiefHighwater (pa...@te...) joined #gstreamer. [22:01] <taaz> i think the debs probably work now [22:03] <wtay> taaz: cool [22:16] steveb (st...@no...) left irc: [x]chat [22:18] Nick change: ajzzzz -> ajmitch [22:18] ChiefHighwater (pa...@te...) left irc: Read error to = ChiefHighwater[temple-baptist.com]: Connection reset by peer [22:20] ChiefHighwater (pa...@te...) joined #gstreamer. [22:47] mattsm (ma...@ad...) joined = #gstreamer. [22:48] mattsm (ma...@ad...) left ir= c: Client Exiting [22:50] <ChiefHighwater> kewl...16M micron CF for US$9 8-] [22:50] <ChiefHighwater> used though [22:51] <ajmitch> heh [22:52] <ChiefHighwater> i got 2...omega got 8 [22:53] <ChiefHighwater> he must know some folks that need some cf [22:53] <wtay> CF? [22:54] <ChiefHighwater> Compact Flash [22:54] <wtay> ah [22:54] <ChiefHighwater> fun stuff 8-] [22:59] richardb (ri...@ix...) joined #gstreamer. [22:59] <richardb> yo [22:59] <wtay> hi [23:01] <richardb> The "crash at end of file" bug is due to signals not b= eing properly designed and implemented, isn't it? [23:02] <richardb> I suppose its therefore going to be a while before it = goes away. Or is there a quick fix possible? [23:02] <wtay> richardb: EOS simply is not implemented well [23:03] <wtay> A signal on the disksrc is not enough [23:03] <wtay> all elements/threads need to stop at the right time [23:04] <richardb> I notice that gstplay goes wrong if you let it get nea= r the end and then seek around. [23:04] <richardb> I'm guessing that once eos fires at the disksrc, seek = can't undo it. [23:04] <wtay> yup [23:04] <richardb> :( [23:05] <wtay> seek is also implemented with a hack btw [23:05] <richardb> No surprise there then... [23:05] <wtay> not really, just anoying [23:06] <wtay> a lot of extra stuff needs to be in place before we can fi= nally make something that *should* work [23:06] <wtay> more specifically, events [23:06] <richardb> I'm making a plugin which takes a chunk of audio data = and draws the waveform to a videoframe and outputs that. [23:06] <wtay> neat [23:07] <richardb> It only wants to deal with a small amount of the audio= data, and skip over the rest. [23:07] <wtay> smoothwave is supposed to do that too [23:07] <richardb> This sounds like it could be addressed by QOS. [23:07] <wtay> hmm, that's not really QoS [23:08] <richardb> The skipping of data frames so that only a reasonable = number get drawn? [23:08] <wtay> can't the plugin just drop buffers it doesn't need? [23:08] <richardb> Yes, but how does it know how many it needs. [23:08] <richardb> For the moment, I'll give it an FPS parameter, and let= it go on that. [23:09] <wtay> when it has too much data, it just starts dropping? [23:09] <richardb> Yes. [23:09] <wtay> I suppose a QoS event could be used too.. [23:09] <richardb> I'll try and get it working at all before worrying abo= ut this too much, I think. :) [23:10] <richardb> Some combination of a user setable framerate and QoS s= eems correct to me. [23:10] <richardb> smoothwave draws to a widget. [23:11] <wtay> although you should not rely on upstream elements being ab= le to handle QoS.. [23:11] <wtay> richardb: true [23:27] omega_ (om...@om...) joined #gstreamer. [23:28] <ChiefHighwater> ello 8-] [23:28] <omega_> yo [23:28] <richardb> evenin' [23:29] Nick change: taaz -> taaz-away [23:30] Action: omega_ has a dual to work with now [23:32] <richardb> omega_: read the article on multithreading near top of= slashdot yet? ;-) [23:32] <omega_> working on it [23:34] <omega_> my position for a long time is that all applications of = any size should be multithreaded, to take advantage of mp machines, which= are only becoming *more* common [23:34] <omega_> esp with SMT catching on [23:47] <wtay> omega_: how's gobject going? [23:47] <omega_> getting there, you need to look at anything that has gdk= in it and try to remove those deps [23:48] <omega_> um, libs/colorspace is used by flx [23:48] <wtay> it can be removed [23:49] <omega_> ok, can you handle that? [23:49] <wtay> yup [23:49] <omega_> I'm not up to speed on any of that stuff [23:50] <wtay> while I'm at it, I should remove libs/winloader and libs/v= ideoscale too.. [23:53] pimlottc (pim...@mo...) left irc: =02[=02BX=02= ]=02 *SUBLIMINAL*MESSAGE* USE BITCHX *SUBLIMINAL*MESSAGE* [00:00] --- Sat Jun 16 2001 [00:00] <wtay> btw, vlc has an altivec IDCT [00:01] ChiefHighwater (pa...@te...) left irc: Ping timeout f= or ChiefHighwater[temple-baptist.com] [00:01] <hadess> wtay: i found better than that [00:02] <wtay> something for libcodec.. [00:02] <hadess> wtay: opendivx has g3 and altivec opts for yuv2rgb and i= dct [00:03] <wtay> cool [00:04] <wtay> omega_: any ideas for editor and gstplay? [00:05] <wtay> I've been trying to compile gnome2.0 but it failed.. [00:06] <hadess> and the license under which it is looks just like a GPL = with advertising clause... maybe we can get the guy who wrote the code to= dual-license... [00:11] ChiefHighwater (pa...@te...) joined #gstreamer. [00:20] big_T (th...@fy...) joined #gstreamer. [00:21] <wtay> hi [00:25] <ChiefHighwater> to me? yet again... [00:42] <richardb> wtay: what are the {red,green,blue}_mask properties in= xvideosink.c? [00:43] <richardb> valid values / meaning? [00:44] <wtay> richardb: they specify where in the longword/word the bits= for RGB are [00:44] <richardb> my guess was right. ;-) [00:48] <richardb> According to the docs, for RGB 32 bit, blue is stored = at the lowest address byte, followed by green, followed by red, then an i= gnored byte. [00:48] <richardb> so, if I specify endianness =3D G_LITTLE_ENDIAN, then = blue_mask should be 0xff [00:49] <richardb> if I specify endianness =3D G_BIG_ENDIAN, blue_mask sh= ould be 0xff000000 [00:49] <wtay> uhm [00:50] <richardb> and if I specify endianness =3D G_BYTE_ORDER, I'm a bi= t confused. [00:50] <hadess> flamefest on gnome2 list... [00:50] <wtay> richardb: so am I [00:50] <richardb> Is the mask against the memory addresses or the words. [00:50] <richardb> Presumably the words. [00:51] <wtay> endianness doesn influence the bit masks [00:52] <richardb> Problem is that the specs (at http://www.webartz.com/f= ourcc/fccrgb.htm#BI_RGB) imply that blue is at the lowest address, not th= e low byte of words. [00:52] <richardb> What _does_ endianness influence, then? [00:52] <wtay> the meaning of the props are the same as the GdkVisual one= s.. [00:53] <richardb> Ah. Any pointer to docs on that? ;-) [00:53] Action: richardb goes to look. [00:53] <wtay> .. which are the same as XVisuals [00:53] <ajmitch> hadess: hehe, gconf vs bonobo-config, which will win? [00:54] Action: richardb picks up a big X book. [00:54] <hadess> ajmitch: i hope gconf and hp [00:54] <wtay> man XVisualInfo [01:05] ajmitch (aj...@p3...) left irc: Ping timeout = for ajmitch[p39-max7.dun.ihug.co.nz] [01:07] <omega_> wtay: you're working on removing gdk and friends from th= e relevant plugins? [01:09] <wtay> omega_: nope [01:09] <omega_> that's the only think left to do atm in the plugins [01:09] <wtay> I can't compile them [01:09] <omega_> you can't? [01:09] <omega_> oh, I need to commit the updates [01:09] <omega_> ok, gimme 2min [01:10] <wtay> :-) [01:12] <omega_> beware, your next update of GOBJECT1 will take a while <= g> [01:12] <omega_> just like my precommit checks are taking a long time [01:14] ajmitch (aj...@p3...) joined #gstreamer. [01:14] <ChiefHighwater> ello [01:16] <wtay> yo [01:16] Action: omega_ builds GOBJECT1 branch to check to see if there ar= e any dangling symbols [01:17] Action: wtay adds more stuff into HEAD to tease omega [01:18] <omega_> I want to get this merge happening ASAP so we have less = to worry about merging.... [01:18] <omega_> so I'm gonna focus on a gtk 1.2 bridge tonight [01:18] <wtay> yes please [01:19] <ajmitch> nice... [01:22] <omega_> committing... [01:22] <wtay> ah [01:22] Action: omega_ watches his network light have a cow [01:22] <omega_> done [01:24] <omega_> videosink [01:24] <omega_> visualization/smoothwave [01:24] <omega_> visualization/synaesthesia [01:24] <omega_> xmms [01:24] <omega_> xvideosink [01:24] <omega_> those have gdk references in them [01:24] <omega_> others refer to libs that have gdk references [01:24] <wtay> ok [01:24] <omega_> while you work on that I'll start thinking about a gobje= ct->gtkobject bridge [01:25] <wtay> ok [01:26] Action: ChiefHighwater is going home [01:26] <ChiefHighwater> l8rz y'all [01:26] <omega_> l8r [01:26] <wtay> cya [01:26] ChiefHighwater (pa...@te...) left irc:=20 [01:27] <omega_> bleagh, this is gonna be a pain: mentally having to reve= rse some 30 sed lines [01:30] <richardb> smoothwave and synaesthesia are broken anyway... [01:30] <omega_> YUP [01:30] <omega_> sorry caps lock was on [01:30] <omega_> trying to type lots of macro names ;-) [01:30] <wtay> I'll just make sure they build [01:30] Action: richardb grins [01:31] <wtay> they need a rewrite anyway [01:31] <richardb> I intend to write at least synaesthesia [01:31] <omega_> cool [01:31] <richardb> s/write/rewrite/ [01:31] <wtay> can I just remove the offending code then? [01:31] <omega_> get the original program rewritten as a graphics-library= -independent library if you can [01:31] <omega_> wtay: as long as the remaining code works, yes [01:32] <omega_> for plugins, just yank 'em out of Makefile.am [01:32] <wtay> oh [01:32] <wtay> syn has a getter for a gtk widget, that's not easily fixab= le [01:32] <omega_> I thought I patched that... [01:32] <omega_> oh well, I'll deal with the gobject stuff for that later [01:33] <omega_> just shut it off for now until richardb can get around t= o fixing it [01:37] <wtay> xvideosink has other issues: /usr/X11/lib is not in the LD= _FLAGS [01:40] <hadess> /usr/X11R6/lib is the only path you can count on [01:41] <hadess> /usr/X11/ doesn't even exist on my machine... [01:43] <wtay> ladspa got converted to something that doesnt compile [01:44] greg__ (gr...@ho...) left irc: [x]chat [01:44] <ajmitch> hadess: nice flamewar on there isn't it? [01:46] <omega_> ajmitch: ? [01:46] <ajmitch> omega_: gnome-2.0-list, people arguing about gconf vs a= 'new' system [01:47] <omega_> oh [01:49] <hadess> yeah, they want to use bonobo-config as the default conf= iguration system, instead of sticking with gconf, and a lot of ppl disapp= rove the whole thing [02:02] <hadess> cya guys [02:02] <omega_> l8r [02:02] hadess (ha...@pc...) left irc: s= leep [02:22] ajmitch (aj...@p3...) left irc: Ping timeout f= or ajmitch[p3-max2.dun.ihug.co.nz] [02:26] ajmitch (aj...@p6...) joined #gstreamer. [02:26] <wtay> omega_: for avi, I need the head version [02:26] <omega_> ok, um [02:27] <omega_> do a partial merge, don't bother tagging [02:27] <omega_> but make it very clear in the commitlog on GOBJECT1 [02:27] <omega_> what rev from HEAD you took, etc. [02:28] <wtay> ok, will do that tomorrow when I'm less sleepy [02:28] <omega_> heh [02:32] <wtay> it feels like the plugins compile faster [02:33] <omega_> very well could be [02:33] <omega_> I'll time it compared to HEAD [02:36] <wtay> hehe, mp3 plays :) [02:37] <omega_> you doubted? [02:37] <wtay> naah <g> [02:37] <omega_> you fixed arts? wow. did you tell stw? he'd love to know= , I think. [02:37] <wtay> hehehe [02:37] Action: omega_ ducks [02:38] <wtay> no need to duck, stw is not around :) [02:38] <omega_> point [02:38] <wtay> the way we mess with the classes in xmms, ladspa and other= s needs some carefull inspection [02:38] <omega_> yeah [02:40] <wtay> oh yeah, and xmms plugins link against old glib/gtk, so th= ey're quite useless now [02:40] <omega_> oh [02:40] <omega_> yick [02:41] <wtay> yup, no xmms until they upgrade all of their plugins [02:42] <omega_> I wish the linker was capable of dealing with that case [02:42] <omega_> would RTLD_LAZY do the trick? [02:42] <wtay> we always use LAZY binding [02:43] <wtay> this glib thing goes very deep, even old gtk doesn't work = anymore [02:43] <richardb> You could fork off a separate process ... [02:43] <omega_> are there any linker tricks we can play to allow the xmm= s plugins to be loaded, or is it a lost cause? [02:43] <wtay> fork is an option, yes [02:44] <richardb> gives you error trapping for "free" [02:44] <wtay> with some custom interface using pipes or something [02:44] <wtay> or shm [02:44] <omega_> no, make pipefilter a semi-abstract class and derive fro= m it [02:44] <wtay> we need to do that anyway at some point [02:44] <omega_> no point in doing the work thrice or more [02:45] <wtay> you need something to pass caps too [02:45] <omega_> yeah, not too big a deal with a pipefilter subclass [02:46] <omega_> just redesign the pipefilter a little, and it'll work, w= ith whatever communication mechanism pipefitler ends up using [02:46] <wtay> pipefilter is only half the solution, you need something t= o sit around it too [02:46] <omega_> why? [02:46] <wtay> the ides would be to have the xmms plugin write/read from = pipes [02:46] <wtay> then you also need something to feed the pipes [02:47] <wtay> or would you use fdsink/fdsrc? [02:47] Action: omega_ should read pipefilter and gstxmms [02:48] <wtay> you'd need an element that embeds another element basicall= y, the other element lives in another process, the outher element proxies= all of the gst stuff to the other process [02:49] <omega_> oh, I kinda see what you mean [02:49] <wtay> a GstElementBox if you want [02:49] <omega_> right [02:49] <wtay> but then again, we could use corba then for IPC [02:49] <omega_> yup [02:50] <richardb> Change of subject question: is there a way to change t= he size of a buffer, or should I simply free the old buffer and make a ne= w one. [02:50] <wtay> may use MCOP/DCOP :) [02:50] Action: wtay ducks [02:50] <wtay> richardb: realloc the buffer? [02:50] <richardb> And if I should make a new one, should I copy timestam= ps over? [02:51] <wtay> richardb: yup, better realloc it [02:51] <omega_> richardb: depends [02:51] <omega_> currently there's a alloc size independent of the data s= ize [02:51] <omega_> I dunno if it's in use anywhere or not [02:51] <omega_> buf if that's big enough, you can just grow in place [02:51] <richardb> It's just for an element which takes audio/raw in and = video/raw out. So the sizes are completely independent. [02:51] <omega_> but I'm tempted to remove that field anyway, since no on= e has used it [02:51] <wtay> MAX_SIZE? [02:52] <omega_> richardb: oh, then alloc a new one anyway [02:52] <omega_> for that kind of stuff you really do not want to reuse t= he incoming buffer [02:52] <omega_> esp if the audio buffer is also going off to the sound c= ard [02:52] <omega_> better to leave it where it is and just unref it, let th= e ref carry to the sound card unmolested [02:52] <richardb> Right. Copy the timestamp of the audio buffer into th= e video buffer? [02:52] <omega_> yeah [02:53] <richardb> Is there any other data I shoudl copy over? [02:53] <omega_> and remind me that I told you this, later, when I'm writ= ing docs [02:53] <omega_> um... don't think so [02:53] <richardb> great. ;-) [02:53] <wtay> if you use bufferpools, chances are that you don't need an= alloc at all... [02:54] <omega_> yup, from the video output [02:54] <wtay> but that's an optimisation for later [02:54] <richardb> yup [02:57] <wtay> I'm off to bed again.. cya later [02:57] <omega_> ok, l8r [02:57] Nick change: wtay -> wtay-ZzZ [03:34] ajmitch (aj...@p6...) left irc: Ping timeout = for ajmitch[p61-max6.dun.ihug.co.nz] [03:38] ajmitch (aj...@p1...) joined #gstreamer. [03:54] Parapraxis (Sp...@us...) joined #gs= treamer. [03:54] <omega_> yo [04:20] omega_omicron (om...@om...) joined #gstreamer. [04:20] omega_omicron (om...@om...) left irc: [x]chat [04:51] mattsm (ma...@ad...) joined = #gstreamer. [04:52] <omega_> yo [04:54] mattsm (ma...@ad...) left ir= c: Client Exiting [05:00] ajmitch (aj...@p1...) left irc: Ping timeout = for ajmitch[p14-max1.dun.ihug.co.nz] [05:11] Parapraxis (Sp...@us...) left irc: = Client Exiting [05:35] ajmitch (aj...@p2...) joined #gstreamer. [05:35] richardb (ri...@ix...) left irc: =02[=02BX=02]=02 = Occifer, take me drunk, I'm home |
From: <wim...@ch...> - 2001-06-17 04:27:13
|
[06:44] ajmitch_ (aj...@p8...) joined #gstreamer. [06:45] ajmitch (aj...@p2...) left irc: Ping timeout = for ajmitch[p22-max2.dun.ihug.co.nz] [06:47] Nick change: ajmitch_ -> ajmitch [07:12] ChiefHighwater (pa...@su...) joined #gstre= amer. [07:14] <ajmitch> ello :) [07:15] <ChiefHighwater> ello 8-] [07:16] <ajmitch> what's up? [07:17] <ChiefHighwater> tryin to get a dvd player to work [07:17] <ChiefHighwater> Creative Encore 5x w/ dxr2 decoder card under wi= n2kpro [07:18] <ajmitch> why win2k? [07:18] <ChiefHighwater> thats my OS..cuz it works 8-] [07:18] <ajmitch> why not a proper OS? ;) [07:19] <ChiefHighwater> cuz I like run mainstream apps and such [07:19] <ChiefHighwater> and /me isn't a programmer [07:26] _gst_newt_ joined #gstreamer. [07:59] ChiefHighwater (pa...@su...) left irc:=20 [08:21] ajmitch (aj...@p8...) left irc: http://www.fre= edevelopers.net [08:44] steveb (st...@no...) joined #gstreamer. [08:44] <omega_> yo [08:44] <steveb> yo [08:44] Action: omega_ is almost done with his gobject->gtk bridge [08:44] <steveb> will that make MAIN foreward compatible? [08:45] <steveb> i mean HEAD [08:45] <omega_> it means that when GOBJECT1 is merged, we will be able t= o compile against either glib-2.0 or gtk-1.2 [08:45] <omega_> which sets us very much apart from the rest of the gnome= crowd <g> [08:46] <steveb> very nice - wouldn't gnomes want this sort of tool? [08:46] <omega_> the editor and player are a different matter, those coul= d be very difficult to handle both [08:46] <omega_> yeah, I'm gonna post these tools up on a page on gstream= er.net and a link to gtk-list and gnome-devel-list [08:47] <omega_> one script to convert from gtkobject to gobject, a heade= r and a source file to make it compatible back the other way [08:48] <steveb> cool [08:48] <steveb> so is there anything you like less about gobject? [08:49] <omega_> it's got a few things that I think it could do without [08:49] <omega_> and I've seen benchmarks claiming that things like signa= l emission are 10x slower [08:49] <omega_> but these are things that IMO can be fixed [08:49] <steveb> hmm [08:57] <steveb> is glib1.2 source compatible with glib1.3? [08:57] <omega_> the core glib is, for what we use of it [08:57] <omega_> so far, at least [08:57] <omega_> nothing's bit me [08:59] <steveb> should I risk overwriting my 1.2 install with 1.3.6? [08:59] <omega_> if you have 1.2.10 or better, it lives in its own direct= ories called glib-1.2 [08:59] <omega_> and 1.3 will live in glib-2.0 [08:59] <omega_> so they coexist [09:01] <steveb> it seems I do, but I must have an older version as well [09:01] <omega_> upgrade to glib/gtk 1.2.10 or better first [09:01] <omega_> then you need pkg-config [09:02] <omega_> and then the whole glib,pango,atk,gtk stack [09:03] <steveb> i was hoping to get away with just glib for now [09:03] <omega_> you won't get too far with just that [09:03] <omega_> because we have to port all the other stuff to use gtk-2= .0 and gnome-2.0 [09:03] <omega_> which means I have to go build gnome-libs-2.0 stuff soon [09:04] <steveb> right [09:12] <omega_> cool. libgst.la is compiling with the shim in place [09:15] <omega_> there will always be warnings in a few places, but they'= re benign [09:16] <steveb> nice [09:23] <steveb> what is atk? [09:23] <omega_> accessibility toolkit [09:23] <steveb> oh, tat [09:23] <steveb> that [09:24] <omega_> great, configure breaks just as I need it to work ;-( [09:44] <omega_> oooooh [09:44] <omega_> RUNNING pipeline [09:44] <omega_> fakesrc: ******* (fakesrc0:src)>=20 [09:44] <omega_> fakesink: ******* (fakesink0:sink)< (0 bytes)=20 [09:46] Action: steveb fakes excitement [09:46] <omega_> doh [09:46] <omega_> that word apparently just made it into the dictionary, e= ven [09:47] <steveb> what word? [09:47] <omega_> doh [09:47] <steveb> thats a bit sad [09:47] <omega_> true [10:36] <omega_> [omega@omicron tools]$ ./gstreamer-launch disksrc locati= on=3D~/music/creed.mp3 ! mad ! osssink =20 [10:36] <omega_> RUNNING pipeline [10:36] <omega_> aaaaaah [10:37] <omega_> BRANCH-GOBJECT1 with gobject2gtk bridge [10:48] walken (fo...@c1...) joined #gstreame= r. [10:48] <walken> hey [10:50] <omega_> yo [10:50] Action: omega_ just did the impossible [10:52] <walken> what ? [10:52] <walken> faster parsing ? :) [10:52] <omega_> nope. ported gstreamer to gobject-2.0 [10:52] <omega_> then write a bridge to compile that on back on gtk+-1.2 [10:53] <walken> oh [10:53] <walken> today I discovered the joy of objdump -dS [10:53] <omega_> heroines said it couldn't be done [10:53] <omega_> Sopwtih said it couldn't be done [10:54] <omega_> -dS: hrm, looks useful [10:54] <walken> so heroines hangs out on irc now ? [10:54] <omega_> nope [10:54] <walken> I never saw him [10:54] <walken> ok [10:54] <omega_> he went away [10:54] <walken> ok [10:54] <walken> yes, -dS rulez [10:54] <omega_> convinced his manager that it'd take 1+yrs to port gstre= amer to gobject, and that they should do their own streaming system [10:54] <omega_> that was a week ago, roughly [10:55] <walken> whatever :) [10:55] <omega_> yeah, that's what I said [10:56] Action: walken bought yet more furniture [10:56] <omega_> doh (to quote OED) [10:57] <walken> soon my place will look like something [10:57] <walken> oed ? [10:57] <omega_> http://dictionary.oed.com/public/news/0106b.htm#doh [10:59] <walken> now I want the definition for serial monogamy [11:04] <walken> ok, gotta sleep, cya :) [11:04] <omega_> l8r [11:04] <walken> you wont see me next 4 days either [11:04] <omega_> doh [11:04] <omega_> (to beat a dead horse...) [11:04] <walken> going to tahoe and yosemite [11:04] <omega_> fun [11:04] <walken> should :) [11:04] <walken> ok cya [11:04] <omega_> don't fall in [11:04] walken (fo...@c1...) left irc: zzZzzZ= zzzz [11:22] omega_ (om...@om...) left irc: [x]chat [13:03] greg__ (gr...@ho...) joined #gstreamer. [13:08] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee02.a2000.nl] [13:50] greg__ (gr...@ho...) left irc: [x]chat [13:56] steveb (st...@no...) joined #gstreamer. [14:00] Nick change: wtay-ZzZ -> wtay [14:00] <wtay> yo [14:00] <steveb> yo [14:12] Uraeus (csc...@c3...) joined #gstreamer. [14:12] <Uraeus> howdy [14:12] <wtay> hi [14:13] <Uraeus> wtay: have you gone ballistic? <g> [14:13] <Uraeus> wtay: according to the Sourceforge stats we have 61 and = 64 CVS commits the last few days :) [14:13] <wtay> ? [14:13] <wtay> :) [14:14] <wtay> you can colour smooth/media green :) [14:14] <wtay> median [14:21] <Uraeus> ok, fixed :) [14:24] <Uraeus> wtay: I see you have some LADPA fixes commited, any chan= ge to LADSPA status? [14:25] <Uraeus> wtay: is 'cobin' a plugin? [14:33] <wtay> Uraeus: it were just compile fixes for GOBJECT1 branch [14:33] <wtay> cobin is not a plugin [14:36] <Uraeus> wtay: I am thinking about trying to make a man page out = of the README matthm updated, is that something I should talk with richar= db about? [14:37] <wtay> Uraeus: you can do that if you want [14:37] <wtay> there are some man pages allready, they just need to be fl= eshed out I thinl [14:38] <Uraeus> wtay: I know, I started looking at them the other day wi= th gmanedit [14:38] <wtay> you want to add a few examples to them? [14:39] <Uraeus> wtay: that is the main priority yes, but also some more = fleshy descriptions where in order [14:40] <wtay> go for it ! [14:41] <Uraeus> :) [14:42] <Uraeus> think I will get to it this evening, have a few other ob= ligations to take care of first (like fixing the RPMS as per Erik's sugge= stions and some gnome.org work) [14:42] <wtay> sure [14:43] <Uraeus> and of course my main priority today; clean my appartmen= t; getting rather tired of living ontop of a pile of junk etc. [14:43] <wtay> heh [14:46] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee02.a2000.nl] [14:47] magnesium (mag...@ni...) joined #gs= treamer. [14:47] steveb (st...@no...) joined #gstreamer. [14:48] <magnesium> what's the "configure" parameter I need to pass to no= t compile in the windows avi stuff? [14:48] <wtay> magnesium: you can't disable it for now [14:49] <magnesium> oh [15:05] <Uraeus> wtay: if magnesium gets the CVS version and don't have a= vifile installed, will not that actually mean the AVI stuff gets buildt w= ithout the windows stuff? [15:06] <wtay> I think so [15:06] <Uraeus> magnesium: there is your solution then :) [15:07] <wtay> or just comment the avi dir in the Makefile [15:28] shitowax (yann@AAnnecy-101-1-3-163.abo.wanadoo.fr) joined #gstrea= mer. [15:28] <shitowax> yo [15:28] <wtay> hi [15:29] <shitowax> I'm working on the quicktime encoder and have some que= stions ... [15:29] <wtay> ah [15:29] <shitowax> In quicktime, there is a keyframe table ... [15:30] <wtay> an index? [15:30] <shitowax> i wonder if there is a way to add simple information o= n a buffer to say if it's a keyframe or not ? [15:31] <shitowax> yep, an index on the frame table [15:31] <wtay> nope [15:31] <wtay> that's what metadata will be used for I think [15:31] <shitowax> so, could I manage to do it using caps ? [15:31] <wtay> that's expensive [15:32] <wtay> we plan to add properties to buffers too [15:32] <shitowax> but metadata have been replaced by caps ... ? [15:32] <wtay> it has been replaced for type negotiation, yes [15:33] <wtay> I think the concept of adding extra info to buffers is sti= ll valid [15:34] <wtay> there's no clean solution for it now I fear [15:34] <shitowax> then I will be obliged to hack something ;) [15:34] <shitowax> there is no keyframe table in avi ? [15:34] <wtay> you can still use metadata, whit a header in the quicktime= dir for now [15:34] <wtay> shitowax: there is, yes [15:35] <wtay> I currently ignore it :) [15:35] <shitowax> ok, then your file are not seekable ... ? [15:35] <wtay> nope [15:36] <wtay> I'm don't know how/when to write the index either... [15:36] <wtay> the encoder has no concept of "last buffer to encode" [15:36] <wtay> and writing it in the state change function seems odd.. [15:36] <shitowax> it's the job of the muxer, isn't it ? [15:37] <wtay> yes, somehow it should send out a buffer with the index at= the end.. [15:37] <wtay> but how do you know when you have encoded the last frame? [15:38] <shitowax> good question, how does your muxer know that the strea= m is finished ? [15:38] <wtay> it doesn't [15:39] <wtay> it doesn't even set the right headers [15:39] <wtay> number_of_frames in the avi header is set to 0 [15:39] <wtay> it's broken [15:39] <shitowax> it can not listen to and eos of a specific pad ? [15:40] <wtay> that would be an option, get the eos signal, set a flag, o= n the next push send the index [15:40] <wtay> do you need to modify the start of the stream too? [15:41] <shitowax> it depends, for the moment, the header is placed at th= e end and is written when the stream is finished [15:42] <shitowax> but, if I want ti create streamable file, I will need = to write it at the beginning [15:43] <shitowax> I don't understand your index problem ... if it's the = number of frame, can't you count them in the muxer ? [15:44] <wtay> you can.. but you cannot write them after finishing the st= ream yet [15:44] <wtay> the muxer has to perform a seek on the downstream element = and send an updated buffer [15:45] <shitowax> could you send me an example of the gstreamer-launch s= yntax you use to encode ? [15:46] <wtay> uhmm.. [15:47] <wtay> -launch disksrc location=3D/opt/data/armageddon.mpg ! mpeg= 1parse video_00! queue ! { mp1videoparse ! mpeg_play ! winenc !video avim= ux ! disksink location=3Dtest.avi } [15:47] <wtay> .. winenc !video_%02d avimux ! .. even [15:48] <shitowax> what is this video_%02d ? [15:48] Action: steveb hides [15:49] <wtay> it's a request pad, you request a new pad from the avi mux= er to hand video to [15:49] <shitowax> can you have more than one video track in an avi file = ? [15:50] <wtay> yup [15:50] <shitowax> and more than one audio track as well ? [15:50] <wtay> yes [15:51] <shitowax> ... hum ... I thought the avi format was much more sim= ple than that =3D) [15:51] <wtay> it very much like qt AFAIK [15:51] <wtay> it also has a lot of garbage in it and extensions and stuf= f [15:52] <shitowax> so basically, you could create a divx with 2 different= sound tracks ...=20 [15:52] <wtay> sure [15:52] <wtay> the maiun thing that is lacking IMO are timestamps in the = muxed stream [15:53] <shitowax> Does this explained why it's not streamable ? [15:53] <wtay> there are "audio" and "video" packets but no indication wh= en to play them [15:53] <wtay> shitowax: it is streamable to some extend... [15:54] <wtay> a well formed avi stream is not streamable though [15:54] <wtay> well, you have to ignore the index I think [15:54] <wtay> hmm, the index could be at the start too actually... [15:54] <shitowax> you mean it's possible to add some sort of private tra= ck to make it streamable ? [15:55] <wtay> that would be possible, yes [15:55] <wtay> but it would still suck [15:56] <shitowax> ;) ... ok then I will return on my encoder ... I'll ke= ep you inform of my progress [15:56] <shitowax> ciao [15:56] <wtay> cya [15:56] shitowax (yann@AAnnecy-101-1-3-163.abo.wanadoo.fr) left #gstreame= r. [16:06] magnesium (mag...@ni...) left irc: = [x]chat [16:38] Nick change: wtay -> wtay-away [16:52] Parapraxis (Sp...@us...) joined #gs= treamer. [16:54] sienap (sy...@ip...) joined #gstreamer. [16:55] <sienap> hej parapraxis!!!!!!!!!!!!!!!!!!!!!!!!!!!!! [16:55] <sienap> please be there ;) [17:01] Action: Uraeus is working through his pile of bills and starts cr= ying [17:02] <sienap> hehehe [17:02] <sienap> :) [17:02] <sienap> damn parapraxis is idle [17:05] <wtay-away> hi [17:05] Nick change: wtay-away -> wtay [17:05] <sienap> wtay! [17:05] <sienap> dude [17:06] <sienap> tell me the magic news of the past few days [17:06] <sienap> :) [17:06] <wtay> nothing special [17:08] <sienap> he [17:08] <Uraeus> s/nothing special/everything we do is magic/ [17:08] <sienap> lol [17:08] <sienap> :) [17:08] <sienap> wtay then tell me the nothing special stuff :) [17:08] <wtay> doing some gobject stuff [17:08] <wtay> nothing special <g> [17:09] <sienap> hehehe [17:09] <sienap> :) [17:09] <sienap> how far is the port now ? [17:09] <wtay> it sorta works [17:10] <sienap> pretty fast done [17:10] <sienap> :) [17:10] <wtay> only audio though [17:10] <sienap> hehe [17:10] <sienap> ;) [17:10] <wtay> we need gtk2.0 and gnome2.0 for video [17:10] <wtay> and I couldn't get those two compiled [17:10] <wtay> well, gtk2.0 went ok [17:11] <steveb> i ran out of hd space when compiling gtk2 :( [17:12] <sienap> >:) [17:13] <wtay> heh [17:13] <sienap> gtk2.0 nice btw ? [17:14] <wtay> it works, nothing special about it AFAICS [17:15] <wtay> but it has cool stuff in it underneat (pango, atk etc) [17:18] <Uraeus> ugh, started the day with 20 000 NOK in my account, pay = half my most urgent bills left with 3000 NOK, and still owe 15 000 NOK on= my creditcards :( [17:19] <wtay> poor guy :) [17:19] <Uraeus> literally [17:21] <Uraeus> shite! just remembered that my houseloan hasn't been ded= ucted yet [17:23] <sienap> wtay wha tabout atk [17:23] <sienap> what does it do ? [17:23] <sienap> ok [17:23] <sienap> you are shitted [17:23] <sienap> :) [17:23] <Uraeus> provides and interface for for accessibility devices [17:23] <wtay> accesibility toolkit, for diabled people like you :-) [17:24] <Uraeus> <g> [17:24] <wtay> mentally disabled even :-) [17:24] <sienap> wtay aah cool so it makes me able to type with my dick ? [17:24] <sienap> verry usefull yeah [17:24] <sienap> :) [17:24] <wtay> yeah :-) [17:25] <Uraeus> mentally disabled? so it makes it easier for C++ hackers= ? [17:25] <sienap> C i hope [17:25] <sienap> :) [17:25] Action: Uraeus notes that sienap didn't get the joke [17:26] <sienap> ehm [17:26] <wtay> right [17:26] <sienap> guess [17:26] <sienap> :) [17:26] <sienap> ooh [17:26] <sienap> now i got it [17:26] <sienap> :) [17:26] <sienap> he=20 [17:26] <wtay> doh [17:26] <Uraeus> <g> [17:26] <wtay> hmm, I can't find a way to get the properties of a gobject= ... :( [17:54] <sienap> btw wtay how is libnle going ? [17:55] <wtay> not started [17:55] <sienap> ah [17:55] <sienap> ok [17:55] <sienap> :) [17:55] <sienap> i wish parapraxis was further with paranormal rewrite [17:56] <sienap> so i can start writing non-sense vis :) [17:58] Action: Uraeus is making himself the most traditional Norwegian d= ish in existence, boiled potatoes [18:08] <Parapraxis> sienap: you're welcome to help :P [18:09] <Parapraxis> sienap: actually, it should be able to do something = kinda soon [18:14] <sienap> aah cool [18:14] <sienap> can you tell me progress stuff please ? [18:14] <sienap> the things you are doing atm are way to difficult for me= >:) [18:16] <sienap> =A0however i think i can do some fancy vis effects [18:16] <sienap> :) [18:16] <sienap> i've got one plasma ready and porting some other to be s= calable [18:16] <sienap> and i am playing with line effects [18:17] <sienap> however i need a own buffer to layer [18:17] <sienap> :) [18:19] <Parapraxis> That will be possible (although the creator of the v= isualization (ie the preset) will be responsible for making it so [18:19] <Parapraxis> sienap: if you want to know the progress read the ch= angelog :) [18:19] <sienap> cvs ? [18:19] <sienap> what is the cvs root [18:20] <Parapraxis> it says on the download page at http://paranormal.sf= .net/download.php [18:21] <Uraeus> wtay? [18:21] <Parapraxis> (but it's paranormal of course) :P [18:21] <Uraeus> wtay: are you here? [18:21] <wtay> yeah [18:21] <sienap> ooh snapshots ? [18:22] <sienap> however what is the current stuff you are working on ? [18:22] <Uraeus> wtay: do we allow stream to a multiplexing "/dev/audio" = with good granularity and latency [18:22] <wtay> Uraeus: dunno [18:22] <sienap> checkout the cvs [18:23] <Uraeus> wtay: I am talking to some of the Sun accessibilty guys = in #gnome and they are wondering about using GStreamer=20 [18:23] <wtay> whatfor? [18:23] <Uraeus> text-to-speech for instance [18:23] <wtay> cool [18:24] <Uraeus> wtay: but they are unsure if we: allow stream to a mult= iplexing "/dev/audio" with good granularity and latency [18:24] <Uraeus> and I coulnd't answer :) [18:24] <wtay> yes [18:24] <Uraeus> I can tell them that its no problem? [18:24] <wtay> there is nothing is gstreamer that would break lowlatency [18:25] <wtay> by design [18:25] <sienap> hehe [18:25] <sienap> :) [18:25] <wtay> of course it depends on the plugins [18:25] <wtay> write a good plugin and you have low latency [18:26] <Uraeus> latency <=3D 100 ms? [18:26] <Uraeus> ? [18:26] <wtay> sure [18:26] <wtay> pass small buffers of data [18:26] <wtay> don't us threads etc.. [18:26] <wtay> s/us/use [18:27] <wtay> set the audiosink to a small fragment size.. [18:27] <wtay> or even dma it right into the audio buffer [18:27] <wtay> of course this is all untested but theoretically possible [18:28] Action: Uraeus feels weird discussing programming techniques with= billh of sun without having any idea what I am really talking about :) [18:28] <Uraeus> billh: the magic is in the plugins, if the plugin in qu= estion is good enough so is gstreamer [18:29] <Uraeus> billh: just pass small buffers of data, dont use threads= and set the audiosink to small fragments and <=3D100 should be possible [18:29] Action: wtay feels weird about low latency without knowing what h= e is talking about [18:29] <Uraeus> <billh> Uraeus: sounds interesting. =20 [18:33] <sienap> parapraxis i hope you'll have a stable plugin architectu= re soon :) and ofcourse a howtodo thingy ;) [18:39] <sienap> is it just me or are all the actuators gone ? [18:39] <sienap> :) [18:39] <Uraeus> wtay: what is multiplexing? [18:40] <wtay> it generally means combining several streams into one and = the reverse process [18:46] <Parapraxis> sienap: it isn't just you :( [18:47] <Parapraxis> sienap: There won't be as many actuators in the new = version [18:47] <Parapraxis> there won't need to be because it will have a micro-= scripting engine [18:47] <sienap> ehm ? [18:49] <sienap> explain [18:49] <sienap> as long as we can get as nice as winamp "Geiss" effects = it is ok with me >:) [18:49] <Parapraxis> it will let the user specify the equations used in t= he transforms [18:50] <sienap> ehm ? [18:51] <Parapraxis> like if you wanted it to rotate, you could do someth= ing like 'theta =3D theta + PI/32' [18:51] <sienap> ooooooh [18:51] <sienap> ic [18:51] <sienap> :) [18:51] <sienap> that is pretty cool [18:51] <Parapraxis> there there won't be any need for many different tra= nsform actuators [18:51] <sienap> hehe yeah [18:51] <sienap> and having some cool presets [18:51] <sienap> :) [18:53] <Parapraxis> Yeah, there will need to be some example presets [18:53] <Parapraxis> I'd also like to have an avs preset converter, to le= ach of the popularity of avs ;) [18:54] Nick change: Parapraxis -> Para[lunch] [18:55] <sienap> i don't know avs that much=20 [18:55] <sienap> the winamp avs that is [18:55] <sienap> :) [18:56] robert (ro...@d1...) joined #gstreamer. [18:57] <sienap> robert :) [18:57] <robert> High sienap [18:58] <sienap> doing something on gstreamer ? [18:58] <robert> watching :) [18:58] <sienap> he [18:59] <sienap> i am off however :) [18:59] sienap (sy...@ip...) left #gstreamer. [18:59] <robert> i'm the tshirt winner, thats my only credit to the proje= ct :) [18:59] <Uraeus> tshirt winner? [19:04] Uraeus (csc...@c3...) left irc: syntax error - = user imploded [19:07] hadess (ha...@pc...) joined #gst= reamer. [19:08] <wtay> yo [19:08] <hadess> hey wtay [19:12] <hadess> how is it going ? [19:13] <wtay> stuggling with gobject... [19:13] <robert> uraeus: back when the project first started, Erik was co= mpiling, I believe the first incarnations of gstreamer on an old 486 or s= ome piece of shit computer.. he had a contest on who could guess how long= it would take [19:14] <hadess> hmm, maybe you guys jumped on gobject a bit too quickly [19:14] <robert> i just hit some random keys on the keypad and emailed th= em off :) [19:14] <hadess> robert: that was an old alpha [19:14] <robert> ahh, yes, your right [19:20] Nick change: Para[lunch] -> Parapraxis [19:21] <hadess> hey Parapraxis [19:31] <Parapraxis> howdee [19:51] <wtay> ok, gotta go.. cya [19:51] Nick change: wtay -> wtay-snooker [19:55] <hadess> doing good, how's pn going ? [19:57] <Parapraxis> pretty slow :( [19:57] <Parapraxis> but I'm almost at the point of drawing something [19:58] <Parapraxis> I'm doing the xmms plugin now (although there isn't = any drawing stuffs in the library yet) [19:58] <Parapraxis> xmms plugin is much simpler than gst so I'm doing it= 1st :) [19:58] <Parapraxis> then I'll use that to test the visuals [19:59] <Parapraxis> should pick up some speed though now that I'm done w= / the tedious/boring abstract classes [20:09] <hadess> ok... [20:22] <Parapraxis> is it valid to assume that if I pass a 32 bit struct= by value it will compile to a single 32 bit push? [20:44] omega_ (om...@om...) joined #gstreamer. [20:51] Nick change: hadess -> hds-busy [21:17] dobey (do...@dr...) joined #gstreamer. [21:25] <Parapraxis> how should an image rendering library communicate it= s images to the application? [21:25] <dobey> uhm [21:25] <dobey> "image rendering library"? [21:25] <Parapraxis> yeah [21:26] <Parapraxis> audio visualization library [21:26] <dobey> that's a very broad library [21:26] <Parapraxis> that's the point ;) [21:26] <omega_> Parapraxis: do you mean in the gstreamer sense or the br= oader sense? [21:27] <dobey> well [21:27] <dobey> you could visualize via midi with lights [21:27] <dobey> ;-) [21:27] <Parapraxis> omega_: the broader sense, keeping in mind speed is = important, but so is flexibility [21:27] <omega_> what model are you thinking of at the moment, as a start= ing point? [21:27] <dobey> well [21:27] hds-busy (ha...@pc...) left irc:= to get drunk [21:28] <Parapraxis> I'm not sure if I should just give the app access to= the image data (in 32 bpp RGB) or write a blitter internal to the librar= y that has support for multiple output formats [21:28] <dobey> VisImage show_image_function(data_to_pass_to_function) [21:28] <omega_> dobey: you don't want to go straight from calc to screen [21:29] <omega_> the idea is for the app to control how the data goes fro= m the viz code to the screen [21:29] <dobey> no [21:29] <dobey> i didn't say that ;-) [21:29] <omega_> that's what your pseudocode says, though <g> [21:29] <Parapraxis> right, the 'application' (which is in most cases act= ually a plugin to something) needs to decide what to do w/ the image [21:29] <dobey> my pseudocode says you should have a function that return= s what you want to display [21:30] <dobey> you should obviously display with gdk-pixbuf if it's an i= mage, or sdl if it's video ;-) [21:30] <omega_> except the function is called show_image <g. [21:30] <dobey> or maybe opengl [21:30] <dobey> fine [21:30] <dobey> s/show/return/ [21:30] <dobey> ;-P [21:31] <Parapraxis> dobey: so you think I should just give the app acces= s to the internally generated 32 bpp RGB? [21:31] <omega_> I think that's the most general solution [21:31] <Parapraxis> I was thinking that if I had an internal blitter, it= could be centrally optimized [21:31] <dobey> Parapraxis: how fast do you want this to be? [21:31] <Parapraxis> dobey: fast ;) [21:32] <omega_> Parapraxis: are your output formats guarateed to always = be 32, 24, 16, 15, or 8 bit RGB ? [21:32] <dobey> Parapraxis: uhm, then you want to use gstreamer or someth= ing, and render it with sdl output or something [21:32] <Parapraxis> dobey: I'm not worrying about the output, that's up = to the application write (which also happens to be me, but not right now) [21:33] <Parapraxis> omega_: nothing is guaranteed ;) [21:33] <dobey> Parapraxis: i don't understand what you are doing then :-= / [21:33] <Parapraxis> omega_: but it would be easy enough to add output fo= rmats [21:33] <omega_> Parapraxis: I'd then say that the viz API should probabl= y handle the common RGB formats, allowing the viz code the opportunity to= optimize internally for the output format [21:34] <dobey> meh [21:34] <omega_> then in a gstreamer plugin or anywhere else, you have to= negotiate between the downstream plugin or screen or whatever, to get th= e right format [21:34] <dobey> i'm gonna go downtown or something [21:34] <Parapraxis> omega_: ick... that means writing 10 different versi= ons of everything [21:34] dobey (do...@dr...) left irc: later [21:34] <omega_> since there are *significant* potential speedups when th= e output format code is merged [21:34] <omega_> Parapraxis: well, don't support them all up front [21:35] <omega_> have a simple interface that allows you to support as ma= ny or as few for any given viz that you want [21:35] <Parapraxis> omega_: I was planning on having everything internal= be 32 bpp rgb [21:35] <omega_> and you can optimize incrementally later [21:35] <omega_> Parapraxis: pseudo-planar? [21:35] <Parapraxis> huh? [21:36] <omega_> you do most/all of your processing on a given color plan= e at a time, right? [21:36] <Parapraxis> omega_: 32 bit rgb allows for huge MMX speed bonus [21:36] <omega_> depending on how the calc is structured, yes [21:37] <Parapraxis> oh, I see [21:37] <omega_> if you're doing very color-plane oriented stuff, then yo= u get better speedup with 3 distinct planes [21:37] <Parapraxis> have 3 separate buffers for r, g, and b? [21:37] <Parapraxis> hmm... [21:37] <omega_> so pseudo-planar would mean that you use standard packed= 32, but consider each pixel to be 8 bits of data that's 32 bits wide [21:37] <Parapraxis> didn't think of that [21:37] <omega_> but that's just a conceptual thing [21:38] <omega_> it has no concrete bearing on how you code it [21:38] <Parapraxis> omega_: it does w/ mmx [21:38] <omega_> right [21:38] <omega_> so the question is, can you make use of planar mode inte= rnally and have any form of significant benefit in doing the final output= format yourself? [21:39] <omega_> I'd tend to think so, since you can pick&choose how to g= o from planar to packed [21:39] <omega_> esp with mmx [21:39] <Parapraxis> omega_: but that would still be treating everything = as 8 bits per R, G, and B, and conversion to 16 bit RGB would still be re= quired [21:39] <omega_> right [21:40] <omega_> but you can do the main calcs much faster in planar, if = you're color-channel oriented [21:40] <omega_> then the packing can be done optimally [21:41] <Parapraxis> it's faster going from color-channel oriented RGB (8= 88) to packed RGB (565) than from 32 bit packed RGB to RGB (565)? [21:41] <omega_> pack to 16 bit would be load(R),load(G),load(B),unpack(r= ),unpack(g),unpack(b),shift,shift,mask,mask,mask,or,or,save(RGB) [21:42] <Parapraxis> but so would the other one [21:42] <omega_> whereas with packed 32 bit you have to pick it apart. N= ot a significant difference [21:42] <omega_> but you get double parallelism.. [21:43] <omega_> I should write it up and time it [21:43] <omega_> and put those routines into libcodec <g> [21:43] <Parapraxis> I think for the transforms (which access all pixels = in a frame) it is the same, but for the rendering effects (which only ace= css a few pixels) it seems like many more cache misses would be happening [21:44] <omega_> possibly, yes [21:44] <Parapraxis> although the rendering effects don't take much time = anyway [21:44] <omega_> but how big is your image, typically? [21:45] <Parapraxis> probably the largest will be 640x480... not sure tho= ugh [21:45] <Parapraxis> depends how fast I can get it to go :) [21:45] <omega_> right [21:45] <omega_> so the question is, are you doing more color-channel-bas= ed ops or pixel-based ops? [21:46] <Parapraxis> I would say pixel, although for the slow effects (wh= ich read/write every pixel) it doesn't make a difference [21:46] <omega_> well, then I'd say you want to use rgb32 internally then [21:47] <omega_> because you can get a pixel in 1 dword rather than 3 sep= arate bytes [21:47] <Parapraxis> hehe... which brings us back to the original questio= n... how to give the image to the app :) [21:47] <omega_> well, in that case, I'd just give it to the app in rgb32 [21:47] <Parapraxis> ok :) [21:48] <Parapraxis> hey omega_... you know about dynamically registering= GObject based classes? [21:48] <Parapraxis> does gst use it, or is it all still static? [21:49] <omega_> I don't think it's static anymore [21:49] <omega_> but I haven't paid too much attention to it yet [21:49] <Parapraxis> omega_: not static, but the g_type_register_static f= unction [21:50] <omega_> btw, if you want the code I'm using to bridge to gtk-1.2= , I'll email it to you [21:50] <omega_> yes, I am using _static, it seems [21:50] <Parapraxis> I don't use gtk yet [21:50] <omega_> I know, you use gobject, right? [21:51] <Parapraxis> yup, I will have an optional gtk dep, which enables = a config widget [21:51] <omega_> no, the gtk bridge I have lets you compile code written = to gobject on top of gtkobject [21:51] <omega_> so you don't have to require glib-2.0 in order to compil= e things [21:51] <Parapraxis> ah, hmmm [21:51] <Parapraxis> ok, yeah, that would be great [21:51] <omega_> until glib-2.0 is commonly avaiable, at least [21:52] <omega_> it'll be in gstreamer CVS on the GOBJECT1 branch sometim= e in the next few hours [21:52] <Parapraxis> oh, I'll just grab it from there then [21:52] <omega_> gobject2gtk.[ch] in gst/ [21:52] <Parapraxis> ok [21:52] <omega_> I'm working on configure.base to handle both gtk and gob= ject right now [22:22] <Parapraxis> is an image buffer's pitch the total # of bytes of t= he width or the # of pixels? [22:22] <omega_> bytes [22:23] <Parapraxis> good :) [22:23] <omega_> the main reason for pitch is that if you have a 640x480 = image on a 1024x768 screen (rgb16), your pitch is 2048, not 1280 [22:24] <omega_> although I've seen cases where pitch was in pixels ;-( [22:46] thomas (th...@ad...) joined #gstreamer. [22:46] thomas (th...@ad...) left irc: Client E= xiting [22:53] <Parapraxis> omega_: oh, I'm using pitch because I'm aligning the= width of each row to 16 byte boundaries for MMX [22:53] <omega_> ok [23:24] <robert> little fluffy clouds, lots of stars at night, and the sk= ies were purple and red and yellow, and the clouds would catch the colour= s everywhere [23:25] <robert> this is the gstreamer theme song, The Orb - Little Fluff= y Clouds, its a techno extravaganza with sounds layered all on top of eac= h other [23:25] <robert> its wild [23:26] steveb (st...@no...) left irc: [x]chat [23:31] <Parapraxis> what happens if you enable profiling in a library? [23:31] <Parapraxis> is there some special way to do it? [23:34] <Parapraxis> ooh... [23:34] Action: Parapraxis will check the gst profiling :) [23:35] <Parapraxis> umm... nm, that didn't help [00:00] --- Sun Jun 17 2001 [00:06] Zeenix (programmer@203.135.48.13) joined #gstreamer. [00:07] <Zeenix> hello [00:07] <Zeenix> have anybody seen Ureaus ? [00:09] <Parapraxis> he was here earlier [00:10] <Zeenix> Parapraxis: never seen you on this channel [00:10] <Parapraxis> Zeenix: oh [00:10] <Parapraxis> :) [00:11] <Zeenix> Paraxxx: your brief into. plz [00:11] <Parapraxis> ? [00:11] <Parapraxis> I'm working on an audio visualization library that w= ill eventually be a gst plugin [00:12] <Zeenix> so you are working for it, i am using gst in my app [00:13] <Parapraxis> cool [00:14] <Zeenix> i am making a voice chat app. [00:15] <Parapraxis> sounds interesting [00:15] <Parapraxis> you wouldn't happen to know how to go about profilin= g a shared library, would you? [00:16] <Zeenix> what ? [00:16] <Parapraxis> never mind ;) [00:18] <Parapraxis> my library is pegging my CPU at 100% and really slow= s everything down, but I suspect it's GLIB [00:18] <Parapraxis> (when in doubt blame glib) [00:19] <Zeenix> what you call glib, glibc.so.xxx ? [00:30] <Zeenix> Parapraxis: where are you ? [00:39] <Parapraxis> New York [00:39] <Parapraxis> what do you mean what do I call glib? [00:40] <Parapraxis> the glib installation should take care of naming the= library [00:40] <Zeenix> by 'i' i meant all [00:41] <Zeenix> what its name on your PC ? [00:41] <Parapraxis> libglib-1.3.so.6.0.0 [00:42] <Parapraxis> actually, glib1.2 is named libglib-1.2.so.0.0.10 [00:42] <Zeenix> you mean its differen from glibc=20 [00:42] <Parapraxis> glibc is the gnu libc >=3D 6.0 I believe (previous v= ersions were just called libc) [00:43] <Parapraxis> glib is a utility library that handles things like l= inked lists, hash tables, etc [00:44] <Zeenix> is glib newer than glibc [00:44] <Zeenix> ? [00:45] <Parapraxis> glib is totally different from glibc [01:07] <Zeenix> is omegs here [01:08] <Zeenix> s/omegs/omega [01:08] <omega_> yeah, sorta [01:09] <Parapraxis> omega_: how does one go about profiling a library (i= f you arent' too busy) :) [01:09] <Zeenix> omega: when i develope a voice chat app using gst & gsm = technologies, do think any mobile phone company might be interested in it= ? [01:09] <omega_> you have to build it with -pg, just like the app. then = you profile an app that uses the library [01:09] <omega_> Zeenix: yup [01:10] <Parapraxis> omega_: that doesn't seem to work [01:10] <omega_> Parapraxis: are you sure you build it with -g and -pg, w= ith no -O ? [01:10] <Parapraxis> I thought -pg overrides -O [01:10] <Parapraxis> not sure I took out the -O's [01:10] <Parapraxis> I'll check [01:12] <Parapraxis> jeez... everything seems to be getting a p# appended= to it's name these days [01:12] <Parapraxis> (ie cvs) [01:12] <omega_> p# ? [01:14] <Parapraxis> cvs 1.11.1p1 [01:14] <omega_> dunno [01:15] <Parapraxis> omega_: does the app that gets profiled need to have= been compiled w/ -pg? [01:15] <omega_> yes. every single file you want results from must be co= mpiled that way [01:16] <Parapraxis> but if I don't care about the app, just the library [01:16] <omega_> rather, every single file must be compiled that way, afa= ik [01:16] <omega_> oh, also you must link with -pg afaik [01:16] <Parapraxis> yeah, I am [01:17] <omega_> else you don't get the routines that actually *do* the c= ounts [01:17] <omega_> if you have [-g] -pg on every compile and link line you = should have no problems [01:17] <omega_> unless you didn't install the glibc-profile package [01:18] <Parapraxis> hmmm... [01:18] <Parapraxis> well I get a gmon.out [01:18] <omega_> that's a start [01:18] <Parapraxis> it has main, atexit, and data_start [01:21] <Parapraxis> yeah -g -pg is on every line [01:21] <omega_> dunno. profiling is very touchy, I haven't had too much= trouble, but where I did, I never really tried too hard [01:23] <Parapraxis> maybe I'll try statically linking [01:23] <omega_> yeah, I think I remember that helping in the past [01:26] <Parapraxis> yup, that worked [01:31] <Parapraxis> just as I suspected... [01:32] <Parapraxis> the 26% of the total execution time is spent in g_ty= pe_instance_is_a [01:33] <omega_> bleagh [01:33] <omega_> can you get a call graph to show what's calling it? [01:33] <Parapraxis> everything calls it [01:34] <Parapraxis> I have tons of g_return_if_fail (PN_IS_<CLASS> )'s [01:34] <Zeenix> when'll wtay's snooker end ? [01:34] <omega_> soon, I hope [01:34] <omega_> Parapraxis: hmmm [01:34] <Parapraxis> I figured I'd have to take them out of the main loop= , but I thought i could leave it in for debugging [01:35] <Parapraxis> guess not though (it runs sooooo slow and isn't doin= g anything) [01:35] <omega_> one thing that bothers me about these macros is that the= re's no guaranteed fast equiv [01:35] <omega_> we should push for a namespace that allows macros that a= re straight casts, so you can choose [01:35] <omega_> maybe just append _FAST [01:35] <Zeenix> omega_ : when you are finished with Paraxxx, dont mind t= o discuss about what i asked earlier [01:36] mattsm (ma...@ad...) joined = #gstreamer. [01:36] <omega_> Zeenix: check with Zaheer, he's working on voip stuff (z= ah...@gr...) [01:36] <Parapraxis> yeah [01:36] <Parapraxis> omega_: but this is just rediculous [01:37] <omega_> yes, this should be fixed [01:37] Action: omega_ looks at the code [01:38] <omega_> G_READ_LOCK (&type_rw_lock); [01:38] <omega_> node =3D lookup_type_node_L (type_instance->g_class->g= _type); [01:38] <omega_> iface =3D lookup_type_node_L (iface_type); [01:38] <omega_> check =3D node && node->is_instantiatable && iface && = type_node_is_a_L (node, iface, TRUE, FALSE); [01:38] <omega_> G_READ_UNLOCK (&type_rw_lock); [01:38] <Parapraxis> yeah [01:39] <omega_> should s/inline// and see which of those takes the most = time [01:39] <Parapraxis> would -O0 do that? [01:39] <omega_> no [01:40] <Parapraxis> is there a gcc flag? [01:40] <omega_> checking [01:40] <Parapraxis> got it [01:40] <Parapraxis> -fno-impliment-inline [01:40] <omega_> ok [01:40] <omega_> recompile glib with that and reprofile [01:42] <Zeenix> must go now, bi & thanks all [01:42] <omega_> l8r [01:42] Zeenix (programmer@203.135.48.13) left irc:=20 [01:45] <Parapraxis> 21.28 1.51 0.73 10489576 0.07 0.07= type_node_is_a_L [01:45] <Parapraxis> 18.95 2.16 0.65 20980022 0.03 0.03= lookup_type_node_L [01:45] <Parapraxis> hehe [01:46] <omega_> yick [01:47] <Parapraxis> the self's are very small, but it doesn't go any dee= per than that [01:47] <Parapraxis> oh, actually, nm [01:58] Apoc (Ap...@pp...) joined #gstreamer. [01:58] <Apoc> yo [01:58] <Parapraxis> yo [01:59] <Parapraxis> omega_: have you run across the same problem w/ gst? [02:00] <Parapraxis> seems like if it happens to me w/ only a few classes= , it would make gst quite unusable [02:00] <omega_> I haven't checked yet [02:01] <Parapraxis> are there glibc profile control functions you can us= e to reset the profiling counters so I can skip the initialization of the= app? [02:02] <omega_> not afaik [02:12] <robert> hi omega [02:12] <omega_> yo [02:21] <omega_> Parapraxis: still there? [02:22] <Parapraxis> yeah [02:23] <omega_> in glib/gobject/gtype.h, comment out line 363 (#define _= G_TYPE_CIT...) [02:23] <omega_> and add: [02:23] <omega_> / experimental!! [02:23] <omega_> #define _G_TYPE_CIT(ip, gt) \ [02:23] <omega_> ((((GTypeInstance*)ip)->g_class->g_type =3D=3D gt) ? TRU= E : g_type_instance_is_a ((GTypeInstance*)ip, gt))) [02:23] <omega_> a really obvious optimization should catch 99-100% of th= e cases you and I have, avoiding that call entirely [02:24] <omega_> I haven't tried to compile that yet, though [02:25] <omega_> similar optimizations can probably be made to mose of th= ose other _G_TYPE_xxx() defines [02:25] <Parapraxis> 1 too many parens [02:25] <Parapraxis> at the end [02:25] <omega_> ok [02:25] <omega_> er [02:26] <omega_> no, one too few at the frint [02:26] <omega_> er, front [02:26] <Parapraxis> oh [02:26] <omega_> there's a big difference, as it turns out <g> [02:26] <omega_> oh, wait [02:26] <omega_> no, you're right [02:27] <Parapraxis> doh [02:27] <omega_> I stripped a open paren from before g_type... in my copy= , so don't need the )) at the end, just ), with the *outer* ) making it t= wo [02:27] <Parapraxis> k [02:28] <omega_> good thing compilers don't typically have a cow when you= have 300 parenthesis on a line [02:28] <Parapraxis> is the a macro that will appear in my code or is it = totally internal to glib? [02:28] <Parapraxis> yeah :) [02:28] <omega_> for fun, compile one of your source files with -E someti= me and see what cpp spits out <g> [02:28] <omega_> that macro will cut&paste into your code [02:28] <Parapraxis> what's -E? [02:28] <omega_> just run cpp [02:28] <Parapraxis> hehe [02:28] <omega_> with all the gcc enviro around it [02:29] <omega_> you get to see some short little line in your code expan= d to many hundreds of characters, mostly parens [02:29] <Parapraxis> yipes! [02:29] <Parapraxis> that's fast :) [02:29] <omega_> it's quite scary until you realize that it all compiles = down to a couple instructions in most cases, because it's all compiler hi= nts (casts) and such [02:29] <omega_> Parapraxis: I fixed it? <g> [02:29] <Parapraxis> yeah :) [02:30] <omega_> cool! [02:30] <Parapraxis> that took them all out of there [02:30] <omega_> whee! [02:30] <omega_> profile again and compare how many times that function r= uns now vs. before [02:30] <Parapraxis> now /me needs to go understand what the hell you did= :) [02:30] <Parapraxis> I did [02:30] <omega_> what % of calls went away? [02:30] <Parapraxis> 0.00 0.81 0.00 5862 0.00 0.00= lookup_type_node_L [02:31] <Parapraxis> the rest aren't even near the top [02:31] <omega_> out of how many before? [02:31] <omega_> whoah [02:31] <Parapraxis> 0.00 0.81 0.00 2496 0.00 0.00= type_node_is_a_L [02:31] <Parapraxis> 0.00 0.81 0.00 1391 0.00 0.00= g_type_instance_is_a [02:31] <omega_> 10-20 =02million=02 before [02:31] <Parapraxis> yeah [02:31] <omega_> so does this qualify me as a gobject guru? [02:31] <Parapraxis> I'm not sure... I'll tell you when I have some idea = what you did :) [02:32] Nick change: taaz-away -> taaz [02:32] <Parapraxis> how can you tell the difference between all those CI= C CCC CLC, etc's? [02:32] <omega_> look up, they're acronyms for what they really do [02:32] <omega_> I haven't figured out the pattern yet though [02:32] <Parapraxis> ahh [02:32] Action: Parapraxis thinks ye do be a gobject guru :) [02:33] <Apoc> bye all ... [02:33] <omega_> taaz: I just removed 26% of Parapraxis's runtime by fixi= ng gobject [02:33] <omega_> Apoc: l8r [02:33] Apoc (Ap...@pp...) left irc:=20 [02:33] <omega_> Parapraxis: you see what I did? [02:33] <taaz> huh? [02:33] <omega_> taaz: the type checking code was using 1/4 of the CPU ti= me in Parapraxis's app [02:34] <taaz> that's pathetic ;) [02:34] <omega_> so I fixed it, brought it from 10-20 =02million=02 calls= of a slow routine down to a couple thousand [02:34] <Parapraxis> not yet :) [02:34] <omega_> Parapraxis: ? [02:34] <Parapraxis> hold on [02:35] <Parapraxis> oh, yeah [02:35] <Parapraxis> it only needs to do the check if it isn't the exact = same class [02:35] <omega_> right [02:35] <Parapraxis> only if it's derived (or wrong) [02:35] <omega_> yup [02:35] <Parapraxis> you need to send that little gem in :) [02:35] <omega_> quite [02:36] Action: Parapraxis wonders how they developed gtk2 w/out running = into this [02:36] <omega_> they probably did, but didn't bother with performance at= all [02:36] <Parapraxis> maybe... but this renders my whole system half-unusa= ble [02:36] <omega_> heh [02:36] Action: omega_ composes mail [02:37] <taaz> you're fixing gobject stuff or user code? [02:37] <omega_> gobject [02:37] <omega_> /#define _G_TYPE_CIT(ip, gt) (g_type_instanc= e_is_a ((GTypeInstance*) ip, gt)) [02:37] <omega_> / experimental!! [02:37] <omega_> #define _G_TYPE_CIT(ip, gt) \ [02:37] <omega_> ((((GTypeInstance*)ip)->g_class->g_type =3D=3D gt) ? TRU= E : g_type_instance_is_a ((GTypeInstance*)ip, gt))) [02:37] <omega_> in glib/gobject/gtype.h [02:38] <taaz> uh... what is that? [02:38] <omega_> type checking. check to see if object *ip is of type gt [02:38] <omega_> before it would always use the slow routine to see if it= 's a derivation, etc., etc. [02:39] <omega_> now it checks brute-force to see if the explicit object = type just happens to match (gt is a guint) [02:39] <omega_> only if it doesn't does it fall back to the slow check [02:39] <omega_> this catches 99.9+% of Parapraxis's cases [02:39] <taaz> what do CIT, ip, and gt stand for? i must be getting dumb= .. i cant read cryptic code anymore [02:40] <Parapraxis> Check Instance Type [02:40] <Parapraxis> Instance Poitner and gtype? [02:40] <omega_> instance pointer, gtype [02:40] <omega_> y [02:41] <taaz> so is this little fix going to speed up all of gobject/gli= b/gtk? ;)=20 [02:42] <omega_> by quite a bit, yes [02:43] <Parapraxis> omega_: the only thing that may be quirky about it i= s multithreading [02:43] <omega_> why? [02:43] <omega_> gt is a static int [02:43] <omega_> the object can't change type [02:43] <Parapraxis> ah, true [02:44] <omega_> the rwlock is just to look into the type structures safe= ly, so the lists don't get changed out from underneath [02:44] <Parapraxis> I see [02:44] <Parapraxis> omega_: what about unregistering dynamic classes? [02:45] <omega_> shouldn't change anything either [02:45] <Parapraxis> ok [02:45] <omega_> the gtype is going to remain the same, and theoretically= the class can't go away with the object still existing anyway [02:45] <omega_> hence g_type_class_ref() [02:46] <Parapraxis> ah [02:46] <Parapraxis> yeah [02:46] Action: Parapraxis is just jealous :P [02:46] <omega_> hehehe [02:54] <Parapraxis> anyone know if pkgconfig will support conditional de= pendencies [02:54] <omega_> meaning? [02:54] <Parapraxis> hehe... I'm not quite sure :) [02:55] <Parapraxis> Right now in my .pc I have "Requires: xml glib-2.0" = (or something like that) [02:55] <Parapraxis> that adds the xml and glib-2.0 cflags [02:55] <Parapraxis> and libs [02:55] <omega_> ah. you'll want to use a .pc.in file and have configure= .sh generate the line as appropriate [02:55] <Parapraxis> but I will eventually have a conditional gtk-2.0 dep [02:56] <Parapraxis> ah, I guess that would work [02:56] <omega_> is that a conditional dep at the time of *your* package = build, or the package the *use* your lib? [02:56] <Parapraxis> omega_: well, it sort of needs to be both [02:56] <omega_> i.e. can you compile your lib to use gtk, then turn it o= ff at a later date somehow? [02:56] <omega_> the only way you can do that is to partition your gtk co= de off into an aux library [02:57] <omega_> then have a paranormal-gtk.pc file as well as paranormal= .pc [02:57] <Parapraxis> what do you mean turn it off? [02:57] <omega_> well, compile paranormal with all the gtk stuff [02:58] <omega_> then build an app that doesn't use the gtk stuff, so you= don't link in paranormal's gtk bits [02:58] <Parapraxis> right, that's what I would like [02:58] <omega_> or do you have to build paranormal *without* gtk in the = first place? [02:58] <Parapraxis> either [02:58] <omega_> ok, then you'll need an aux lib=20 [02:58] <Parapraxis> ok [02:59] <omega_> same idea as gstreamer+gobject: only depend on gtk by wa= y of external plugins that are optional [03:01] <Parapraxis> perhaps I'll only need a separate .pc file, rather t= han a totally separate lib [03:01] <omega_> how would you then decide what to link against? [03:01] <omega_> a single lib will always have the gtk dep, you can't rem= ove it [03:02] <Parapraxis> I see [03:02] <Parapraxis> but it wouldn't matter if the app links to paranorma= l, and paranormal links (w/out the app knowing) to gtk [03:03] <omega_> but that just wouldn't be nice, imo [03:03] <omega_> what if you're building a svgalib app? [03:03] <Parapraxis> yeah, I suppose [03:03] <Parapraxis> would just be a bigger mem footprint [03:03] <taaz> omega_: in 0.2.0 (at least) --gst-mask=3D0 doesn't keep gs= t-register quiet. prints out some INFO lines. is that supposed to be th= e case? [03:03] <omega_> hrm, no [03:04] <taaz> sorry to interrupt your discussion ;) [03:04] Nick change: omega_ -> omega_dinner [03:04] Action: omega_dinner will be back in ~15min [03:18] Apoc (Apoc@212.83.182.32) joined #gstreamer. [03:32] Nick change: omega_dinner -> omega_ [03:37] <Parapraxis> omega_: I think that the type checking was taking up= more that 26% of my cpu usage... [03:37] <Parapraxis> before when I ran the xmms vis plugin it pegged my C= PU usage to 100%, now it's at 0 [03:38] <omega_> eek [03:38] <omega_> and it's still working perfectly? [03:39] <Parapraxis> yeah [03:39] <omega_> wow [03:39] <Parapraxis> well... it doesn't do much :) [03:39] <Parapraxis> but it doesn't crash or anything [03:45] Nick change: wtay-snooker -> wtay [03:45] <wtay> yo [03:45] <omega_> yo [03:46] <wtay> pff I couldn't get the value of a property in gobject :( [03:46] <omega_> oh? [03:46] <wtay> I was trying to get -inspect working [03:47] <Apoc> I just tried the xmms mikmod plugin ... :-) [03:48] <wtay> Apoc: and? [03:48] <Apoc> It works great .. :-) [03:48] <wtay> the more I look at GObject the more bloat I discover.. [03:49] <omega_> yeah, it's not the most elegant [03:49] <omega_> but I think some of it can be fixed [03:49] <omega_> On 16 Jun 2001, Havoc Pennington wrote: [03:49] <omega_> > Seems pretty darn sensible to me. [03:49] <omega_> re: my patch suggestion [03:49] <wtay> yeah, good suggestion [03:50] <wtay> GValue value; [03:50] <wtay> g_object_get (G_OBJECT (element), param->name, &value,= NULL); [03:50] <Apoc> Mikmod should be a xmms plugin or a native gstreamer plugi= n ?? [03:50] <wtay> one would expect to get the value after that with g_value_= get_int(&value) but it just CRITCALs [03:50] <omega_> Apoc: should be a native eventually [03:51] <omega_> wtay: neat [03:51] <omega_> yeah, I have some problems with arguments and signals st= ill [03:51] <omega_> oddly, most of them are with my gobject2gtk bridge [03:51] <omega_> which btw works nicely, for the most part [03:51] <omega_> I can play mp3s with it, at least [03:51] <wtay> cool [03:52] <Apoc> Ok ... I'll look at this ... [03:52] <omega_> and I'm now working on getting the build enviro up to ha= ndling both [03:52] <wtay> so far the "impossible" claims [03:52] <Parapraxis> is anyone working on a shn plugin? [03:52] <omega_> yeah. /me wants to see heroines' manager's face when he= hears about this <g> [03:52] <wtay> shn? [03:52] <omega_> shn ? [03:52] <Parapraxis> shorten, I believe... It's a lossless audio compress= ion format [03:52] <Parapraxis> that's what all my music is [03:52] <wtay> oh [03:53] <wtay> I was going to do a flac one=20 [03:53] <Parapraxis> maybe I'll make one when I get a chance [03:53] <wtay> once I figure out the API, that is [03:55] <Parapraxis> eek [03:55] <Parapraxis> something really wierd just happened w/ cvs :( [03:56] <Parapraxis> importing pnxmms is doing files pnxmms//glib/* (and = there is no pnxmms/glib/... it's just doing the glib source tree) [03:56] <taaz> the flac site has a comparison with other lossless formats= like Shorten. [03:57] <omega_> I wonder if lame can do free-rate lossless mp3s ? [03:57] <taaz> wtay: come on... i bet you could have a flac plugin done i= n 10 minutes ;) [03:58] <wtay> taaz: if I can find a good/clean example of the API usage,= yes [04:01] <wtay> hmm.. maybe something else is wrong with the properties (A= BSTRACT) [04:02] <omega_> dunno [04:02] <omega_> what case are you using to generate this critical? [04:02] <omega_> just -inspect ? [04:02] <wtay> yup [04:03] <wtay> gstreamer-inspect lame [04:08] <Apoc> Oups ! 4 am ... time to go to bed ... ;-) ... Bye all=20 [04:08] <wtay> bye [04:09] Apoc (Apoc@212.83.182.32) left irc:=20 [04:09] <taaz> i was being an idiot the other day... had an LD_LIBRARY_PA= TH entry for dir I install gst into for oms-gst app development. after i= nstalling debs it was doing some really weird things... took me at least = 2 hours to figure it was linking to wrong libs at runtime. duh. [04:10] <omega_> mu [04:11] <wtay> heh, gstreamer started to segfault and plugins would not l= oad anymore... [04:11] <omega_> whoops [04:12] <omega_> don't do that [04:12] <wtay> it turned the gnome compile scripts had set the LD_BIND_NO= W flag... [04:12] <omega_> mu [04:12] <wtay> I discovered it after almost 2 hours.. :) [04:15] <taaz> flac cvs has winamp and xmms plugin source... shouldn't be= too hard to adapt to gstreamer. (no i'm =02not=02 volunteering! <g>) [04:16] gtb456876 (syln@Mix-Lyon-203-1-44.abo.wanadoo.fr) joined #gstream= er. [04:16] <omega_> yo [04:16] <gtb456876> hello ! [04:17] <gtb456876> his there anybody connected ? [04:17] <omega_> to what? [04:17] <wtay> hi [04:18] <gtb456876> ;-) I just wanted to know if you were actually there = ! [04:18] <gtb456876> I come here because I have a very simple question... [04:19] <wtay> taaz: "he who brings up issues and ideas is automatically = volunteering" that's just how it goes :-) [04:20] <gtb456876> I'm trying to read the application development manual= , but I can't make the first application example to work [04:20] <wtay> ah [04:20] <omega_> not surprising, that hasn't been touched in a while ;-( [04:20] <taaz> looks like gtb456876 just volunteered to update the manual= ;) [04:20] <wtay> I updated it as of today :) [04:20] <gtb456876> damn ! [04:20] <omega_> heh [04:21] <gtb456876> is it available on the gstreamer web site ? [04:21] <wtay> checking... [04:21] <gtb456876> tazzz : hummmm... [04:21] <gtb456876> tazzz : hummmm... it's a bit difficult for me ! [04:21] <wtay> it is, yes ! [04:22] <wtay> http://gstreamer.net/docs/cvs/gstreamer-manual/cha-hello.h= tml [04:22] <gtb456876> ok, great ! i'll look at this ! [04:22] <wtay> gtb456876: that manual is built from CVS, so it's fairly u= p to date [04:22] <gtb456876> ok [04:23] <gtb456876> I don't see any difference with the code I tried [04:23] <wtay> gtb456876: the top level element is a pipeline [04:24] <gtb456876> ok, sorry ! ;-) [04:24] <wtay> :) [04:24] <gtb456876> really thank you ! ;-) I will try this ! [04:25] <wtay> gtb456876: feel free to come back if it doesn't work=20 [04:25] <gtb456876> thx ! [04:25] gtb456876 (syln@Mix-Lyon-203-1-44.abo.wanadoo.fr) left irc:=20 [04:27] <wtay> omega_: are you looking at getting a property? [04:27] mattsm (ma...@ad...) left ir= c: Client Exiting [04:29] <omega_> no, I'm verifying that my gtk/gobject stuff works [04:29] <omega_> then I'll commit that and start looking into the gobject= issues [04:29] <wtay> ok [04:33] <wtay> yes! got it [04:33] <omega_> oh? [04:33] <wtay> GValue value =3D { 0, }; [04:33] <wtay> g_value_init (&value, param->value_type); [04:33] <wtay> g_object_get_property (G_OBJECT (element), param->name= , &value); [04:33] <omega_> uh? [04:33] <wtay> to get a property value [04:33] <omega_> what's the zero for, why's it necessary? [04:34] <wtay> you need to fill it with 0 [04:34] <wtay> that was the trick [04:34] <wtay> the GType has to be set to 0 before you can use it... [04:35] <omega_> right, /me wants to know why that's necessary, since it'= s a performance hit, and doesn't have any obvious function [04:35] <taaz> send another patch to gtk list ;) [04:35] <omega_> working on figuring out what it is, first [04:36] <wtay> create a GValue (set it to zero), set the type it is suppo= sed to hold with _init, use it to get a property from an bject [04:37] <omega_> it appears that there's code to handle conversion of typ= es built in [04:37] <omega_> which imo is stupid and never going to be used [04:38] <wtay> I tend to agree.. [04:38] <omega_> so if the type in the gvalue matches the paramspec, it j= ust grabs it [04:38] <omega_> if it's 'transformable' to the paramspec, it transforms = it [04:38] <wtay> worse, if it doesn't match it CRITICALs [04:38] <wtay> or somewhere else that is [04:38] <omega_> um, trying to figure out why it has to be zero (check is= not in the main func) [04:39] <omega_> blagh, there's the same problem we have in gstreamer, bu= t moreso: [04:39] <wtay> bleagh, need to add this stuff to gst_utils.. [04:40] <omega_> a function does checks on its arguments, so we know that= the args are what they say they are [04:40] <omega_> then a function that that function calls *also* does the= same checks [04:41] <Parapraxis> hehe... glib requires pkgconfig 0.7, but the cvs is= only 0.6 :( [04:41] <wtay> only solution is to have internal functions and do the che= cks only on the exported methods [04:41] <omega_> right [04:41] <omega_> but if the internal functions are *also* exported method= s.... [04:42] <omega_> need to have inline sub-methods and such [04:43] <wtay> or static ones.. [04:44] <omega_> yes, static inline [04:48] <wtay> uh? it's getting light outside and it's only 4.48am <g> [04:49] <omega_> oops [04:55] <wtay> great. G_MINFLOAT > 0.0 [04:55] <omega_> uh? [04:55] <wtay> that's screwed up [04:56] <omega_> G_MINFLOAT =3D=3D FLT_MIN [04:56] <wtay> the compression ratio property in lame has G_MINFLOAT to G= _MAXFLOAT range.. the assertion that 0.0 is between those values failed [04:56] <omega_> /usr/include/kpathsea/c-minmax.h:#define FLT_MIN 1e-37 [04:57] <wtay> I know... [04:57] <omega_> that's not MIN, that's EPSILON [04:57] <wtay> (process:7451): GRuntime-CRITICAL **: file gparamspecs.c: = line 1521 (g_param_spec_float): assertion `default_value >=3D minimum && = default_value <=3D maximum' failed [04:57] <omega_> but that's not a stdc header anyway [04:57] <omega_> but that's the only definition of it anywhere [04:58] <wtay> 1 have something else here... [04:58] <omega_> aforementioned header is not included anywhere outside '= kpathsea' [04:58] <wtay> #define FLT_MIN 1.17549435e-38F [04:58] <omega_> what file? [04:58] <wtay> float.h [04:58] <omega_> where? [04:58] <wtay> in /usr/lib/gcc-lib/i386-linux/2.95.4/include [04:58] <omega_> ah [04:59] <omega_> that's invalid [04:59] <omega_> rather, G_MINFLOAT should not be using that valie [04:59] <omega_> regardless of whether gcc is using entirely the wrong na= me for that define [04:59] <wtay> right... [04:59] gtb456876 (syln@Mix-Lyon-203-1-44.abo.wanadoo.fr) joined #gstream= er. [05:00] <gtb456876> the helloworld works now ! [05:00] <wtay> whee! [05:00] <gtb456876> but I had to make a little change... [05:00] <wtay> ah? [05:00] <gtb456876> I had to replace : [05:01] <gtb456876> audiosink=3Dgst_elementfactory_make("audiosink", "pla= y_audio"); [05:01] <wtay> right audiosink with osssink [05:01] <gtb456876> with : [05:01] <gtb456876> audiosink=3Dgst_elementfactory_make("osssink", "play_= audio"); [05:01] <gtb456876> yes, it's that ! ;-) [05:02] <wtay> gtb456876: fixed in CVS :) [05:02] <gtb456876>... [truncated message content] |
From: <wim...@ch...> - 2001-06-18 04:29:53
|
[06:30] <ShrimpX> wow. cool. you don't have a degree from them? [06:30] <omega_> nope [06:30] <ShrimpX> self taught, then? [06:30] <ShrimpX> mostly? [06:30] <omega_> quite [06:30] <ShrimpX> that's impressive [06:30] <omega_> ;-) [06:30] <ShrimpX> you planning on getting a degree at all? [06:30] <omega_> not at this point [06:31] <ShrimpX> how old are you? [06:31] <omega_> 22 [06:31] <ShrimpX> jeez... Now I feel stupid. :( [06:31] <omega_> doh [06:31] <ShrimpX> I'm also 22 [06:32] <ShrimpX> I figured you were 26 or something, with a MS under you= r belt [06:32] <ShrimpX> :) [06:32] <omega_> heh [06:32] <ShrimpX> although you didn't look it [06:32] <omega_> did I sound like one? [06:32] <ShrimpX> knowledge-wise, yes. [06:32] <omega_> cool <g> [06:32] <taaz> hey... some of us are 26 sitting in a school lab working o= n our MS right now [06:32] <ShrimpX> hege [06:32] <ShrimpX> hehe [06:33] <ShrimpX> that's kinda what I see myself doing. :) [06:33] <omega_> s/sitting/stuck sitting/ [06:35] jerwin (je...@66...) join= ed #gstreamer. [06:35] <omega_> yo [06:36] <ShrimpX> ya [06:45] <taaz> omega_: forgot to mention.. that little syncronization iss= ue i told you about before... was totally unrelated to cache issues or wh= atever. [06:45] <omega_> huh? [06:45] <taaz> omega_: on that board with the ppc i'm playing with [06:46] <omega_> hmm, what was it? [06:48] <taaz> dumb copy&paste error ;) had some code someone else wrote= that I copied and modified to do my stuff in various places... had a gl= aring hole in handshaking algorithm we use. specifically, it was reading= back results from the board before waiting for the board to say it was d= one executing a command. [06:49] <taaz> and since it sort of randomly failed depending on timing o= f how long the command took to execute i didnt notice for months untill i= wrote a command that took a while to run [06:49] <omega_> mu [06:50] <taaz> having stared at and worked on the code daily for months i= totally didnt see what was wrong with it ;) oh well.. [06:50] <omega_> whoops [06:56] mattsm (ma...@ad...) joined = #gstreamer. [06:57] <taaz> ha.. CML2 Adventure game... excellent hack ;) [07:00] <taaz> maybe gstreamer could use CML2 stuff to configure plugins = to build. probably wouldnt need the dependency stuff unless we start hav= ing composite plugins that depend on other plugins though [07:02] <taaz> could probably follow along with the next gen kernel build= system too. put in option to build plugins (modules) or hard link every= thing. could probably save some code space if all the dyn linking stuff = was ripped out. kaffe people do this for platforms w/o shlib support [07:03] <taaz> perhaps that just adding too much complexity... [07:03] <omega_> cml2? [07:04] <taaz> http://www.tuxedo.org/~esr/cml2/ [07:05] <taaz> it now has a Zork style interface ;) [07:05] <omega_> question is, how does it interface with autoconf/automak= e? [07:05] <omega_> and of course, it requires python [07:06] <taaz> why do people say it like that? no one complains about re= quiring perl... and perl is evil! evil i say, evil! [07:07] Action: taaz jumps around the room [07:08] Action: taaz likes python and is writing python code right now ;) [07:08] <omega_> so does it interface with autoconf/automake ? [07:09] <taaz> i doubt it.. may have to hack some stuff [07:09] <omega_> who else uses cml2? [07:09] <taaz> i was just thinking out loud [07:09] <omega_> and someone suggested something an apache project is usi= ng as well [07:10] <taaz> yeah. dunno... some other kernel and linux dist people a= re using it. not many app level projects require such things [07:10] <omega_> apache does, we do [07:11] <taaz> exactly... 2 is not alot ;) [07:11] <omega_> it's almost a crowd, though [07:12] <taaz> i'm sure ESR would help out if we need slight mods to his = code... really have to look at the problem and see what's needed before= jumping to the solution ;) [07:12] <omega_> yup [07:13] <taaz> i'm not sure we even have a problem [07:13] <omega_> I have the problem of configuring gstreamer to build on = embedded devices [07:14] <omega_> a big part of that is moving most of the plugins out of = the core package [07:14] <omega_> most/all [07:14] <taaz> perhaps just some improved reporting code in configure wou= ld do it. list all the plugins and if they are going to be built or not = and if not which dependencies are missing. [07:14] <omega_> yes. also need to devise some standards for the configu= re stuff, like HAVE_ vs. USE_ stds, etc. [07:14] <taaz> well, yeah, that's why I initially thought of CML2 [07:15] <omega_> I think most of the problem will be dealt with by moving= plugins out of the core [07:15] <taaz> so we could add symbols for -every- plugin. like being ab= le to exclude disksrc if you dont need it [07:15] <omega_> right, so some degree, yes [07:15] jerwin (je...@66...) left= irc: Ping timeout for jerwin[66-44-59-4.s258.tnt3.lnhva.md.dialup.rcn.co= m] [07:16] <taaz> in that case you could have symbols for each plugin and ea= ch element in the plugin... have some dependency tracking there [07:16] <omega_> back up, start with many separate CVS modules and result= ing tarballs for different plugins [07:17] <taaz> yeah, that will suck too! ;) [07:17] <omega_> yes, but it'll suck in a more compartmentalized way [07:17] <omega_> modular suckage ;-) [07:18] <taaz> take this to what level though? package for every plugin?= that's a big unmanageable mess [07:19] <omega_> no, have to be careful about it [07:19] <omega_> like all the mpeg1 in one module, mpeg2 in another [07:19] <omega_> the mass of audio filters in another [07:19] <taaz> that doesnt work either... where does mpeg2dec go? [07:19] <omega_> video filters in another [07:19] <omega_> in the mpeg2dec package [07:19] <omega_> dunno, this is a non-trivial namespace problem [07:20] <omega_> to start with, I think these plugin packages should at l= east start with gst- instead of gstreamer- [07:20] <taaz> blah... solution is to gain enough momentum so plugins are= maintained upstream by the library authors ;) [07:20] <omega_> to some extent [07:20] <taaz> maybe not... [07:21] <omega_> better to keep things in one place, for the most part, s= o we can do build tests and coordinated releases [07:21] <taaz> there is that little issue of lib naming... we have libgst= foo and libbar plugins... all in usr/lib/gst/ something should be standa= rdized there [07:22] <omega_> yes, we need to pick something [07:22] <omega_> we have to balance plugin libname vs. module name vs. pa= ckage name, etc. [07:22] <omega_> and of course, elementfactory names..... [07:23] <taaz> did you think i was joking about the Intergalactic GStream= er Naming Authority? [07:23] <omega_> just the name part of it [07:23] <taaz> igna.org baby! [07:23] <omega_> not taken, either [07:23] <taaz> yeah... i had a few revisions till i found a free domain ;= ) [07:23] <omega_> mu [07:24] <omega_> why not Pan-galactic? [07:24] <taaz> just trying to be modest [07:24] <omega_> heh [07:25] <omega_> I do want to have a web-based service eventually (maybe = ridgerun-powered) that lets players/devices search for new plugins for un= known media streams [07:26] <omega_> either upload a couple KB of the data stream, or the URL= to it, and it'll tell you the caps, and a list of plugins for your hardw= are/OS that can be downloaded and loaded immediately to play the file [07:26] <taaz> in any case... i'd just put it on gstreamer.net/na/ or som= ething [07:26] <omega_> like real does, without having to restart [07:26] <taaz> for the naming authority at least... [07:26] <omega_> why not gstreamer.net/igna/ ? [07:27] <taaz> you're the leader, you decide ;) [07:27] <omega_> heh [07:27] Action: omega_ creates a wiki node [07:27] <omega_> we need to think about the structure of the wiki too [07:27] <omega_> need a GstIdeas page.... [07:28] <omega_> some stuff has landed elsewhere that should go there [07:28] <taaz> the power of the wiki shall make it so over time [07:28] <taaz> blah.. i really need to uprade wiki sometime [07:29] <taaz> too lazy since had those mods to make it work at gst.net/w= iki/ [07:29] <taaz> should really beg moin people to add that functionality [07:29] <omega_> which? [07:30] <taaz> had to change code to not add cgi-bin/foo/bar/moin.cgi to = the urls in links and so on [07:30] <omega_> ah [07:32] <omega_> http://gstreamer.net/cgi-bin/wiki/moin.cgi/Intergalactic= GStreamerNamingAuthority [07:37] <omega_> http://gstreamer.net/cgi-bin/wiki/moin.cgi/OnlinePluginR= egistry [07:37] <taaz> we gotta support this stuff too: http://www.ipnsig.org/ [07:38] <omega_> of course, if we claim to be the =02i=02gna [07:38] <ShrimpX> so you guys use glade for gui stuff? [07:38] <taaz> naw.. i just mean being able to stream data to pluto or wh= atever ;) [07:38] <omega_> ShrimpX: yup [07:38] <omega_> libglade, actually [07:39] <taaz> that IETF draft there is an interesting read [07:39] <omega_> taaz: right, but that should just be any old STREAM, for= the most part... OS's job [07:41] <omega_> the 'Desiderata of Interplanetary Internetworking' is fu= nny [07:45] <omega_> darn, so I can't really `telnet shell.europa.ju.ss` any = time soon ;-( [07:46] <taaz> unless you have ftl networking that will never be practica= l [07:47] <omega_> just need an ansible, and we'll be set [07:47] <ShrimpX> what's the command array for playing an mp3? [07:48] <omega_> from -launch? [07:48] <ShrimpX> yea [07:48] <omega_> gstreamer-launch disksrc location=3Dfile.mp3 ! mad ! oss= sink [07:48] <ShrimpX> thx [07:49] <ShrimpX> if it can't find mad, what could I be doing wrong? [07:49] <omega_> if you don't have libmad, it won't exist [07:51] <omega_> you can use mp3parse ! mpg123 in the meantime [07:51] <ShrimpX> aah, I see [07:52] <ShrimpX> I get a floating point exception (core dump) [07:52] <omega_> hmmm [07:53] <omega_> well, then grab libmad <g> [07:53] <ShrimpX> hehe ok [07:53] <omega_> if you run redhat or equiv there are rpms for download i= n the 0.2.0 release area on gstreamer.net [07:55] <ShrimpX> I found the tarball. It's installing right now [07:55] <omega_> tarball? [07:55] <omega_> what version> [07:55] <omega_> ? [07:56] <ShrimpX> not on gstreamer.net [07:56] <ShrimpX> ftp://ftp.mars.org/pub/mpeg/ [07:56] <omega_> uh? [07:56] <omega_> I though it was on libmad.sourceforge.net or somesuch ? [07:56] <ShrimpX> hum... what's the latest version? [07:57] <omega_> I think 0.13.2b or something [07:57] <ShrimpX> sounds right [07:57] <omega_> ok [08:01] <ShrimpX> can I compile plugins individually? [08:01] <omega_> should be able to do it with mad [08:01] <omega_> others can't [08:01] <omega_> because configure.sh has to find the libs and cflags for= a lot of the plugins, so you'd have to re-run it after installing a new = dependent lib [08:03] <ShrimpX> the thing is that I have a brand new source dir. I axed= the old one after installing gstreamer. Should that cause problems? (I d= on't want to recompile the whole thing, it takes 45 minutes) [08:04] <omega_> oh, yeah, you have to reconfigure at least [08:04] <ShrimpX> o, ok [08:07] <ShrimpX> sweet!! [08:07] <omega_> oooh, got the autoplugger working, with lots of noise [08:07] <omega_> that means that signals and args are workingn reasonably= , if not silently [08:07] <omega_> ShrimpX: having fun? [08:07] <ShrimpX> yup! [08:08] <ShrimpX> I noticed that the mad plugin is really small... [08:08] <omega_> yup, it just wraps libmad [08:09] <ShrimpX> that's nifty. [08:09] <omega_> even cooler: do you have vorbis installed? [08:09] <omega_> btw, did you regenerate the registry after installing ma= d ? [08:09] <ShrimpX> yea [08:10] <ShrimpX> I don't have vorbis though [08:10] <omega_> darn, ok, well you can still do this: [08:10] <omega_> gstreamer-launch disksrc location=3Dfile.mp3 ! autoplugg= er ! osssink [08:10] <ShrimpX> k [08:11] <ShrimpX> nice!!! [08:11] <omega_> it'll work for any elementary media file that's supporte= d by plugins [08:11] <omega_> so if you have vorbis installed, it'll trivially play .o= gg's [08:11] <omega_> you could even do disksrc ! autoplugger ! vorbisenc ! di= sksink [08:11] <omega_> to [re-]encode all your music files into .ogg [08:12] <ShrimpX> sweet [08:12] <ShrimpX> I have to read some decumentation, understand how the w= hole pipeline functions. [08:12] <omega_> ah, and here's another cool one: [08:12] <omega_> add --gst-mask=3D-1 to the last pipeline you ran [08:12] <ShrimpX> k [08:13] <omega_> and 2>&1 | less if you want to see it all [08:13] <ShrimpX> wow. That's a ton of output. :) [08:14] <omega_> categorized, colorized, and 'easy' to read [08:14] <omega_> you should try it with DEBUG turned on, it triples... [08:14] <omega_> but that's a configure-time option, turned off by defaul= t [08:15] <ShrimpX> what if one would want to write a gui front-end to an a= lready existing plugin? could you get the plugin to spit all the frame in= fo? [08:16] <omega_> gstreamer-inspect [08:16] <omega_> the joys of a real, run-time pluggable object model, som= ething C++ doesn't do without significant help (COM/CORBA, etc.) [08:17] <ShrimpX> cool [08:17] <omega_> see http://gstreamer.net/gsteditor.shtml for some screen= shots [08:38] omega_ (om...@om...) left irc: Read error to omega_[omegacs= .net]: Connection reset by peer [08:52] omega_ (om...@om...) joined #gstreamer. [09:16] steveb (st...@no...) joined #gstreamer. [09:35] <ShrimpX> omega_, you there? [09:40] <omega_> sorta [09:41] <ShrimpX> hehe [09:42] <ShrimpX> I got ogg's running. It's sweet [09:42] <omega_> cool ;-) [09:42] <ShrimpX> didn't know the pipeline name of the vorbis plugin, but= the autoplugger worked wonders. :) [09:43] <omega_> vorbisenc,vorbisdec [09:43] <omega_> just run -inspect with no arguments [09:43] <ShrimpX> ah, ok [10:02] Uraeus (csc...@c3...) joined #gstreamer. [10:03] <Uraeus> hello [10:03] Uraeus (csc...@c3...) left irc: Read error to U= raeus[c33s9h5.upc.chello.no]: EOF from client [10:04] Uraeus (csc...@c3...) joined #gstreamer. [10:05] Uraeus (csc...@c3...) left irc: Read error to U= raeus[c33s9h5.upc.chello.no]: EOF from client [10:07] Uraeus (csc...@c3...) joined #gstreamer. [10:07] Uraeus (csc...@c3...) left irc: Read error to U= raeus[c33s9h5.upc.chello.no]: EOF from client [10:07] Uraeus (csc...@c3...) joined #gstreamer. [10:08] Uraeus (csc...@c3...) left irc: Read error to U= raeus[c33s9h5.upc.chello.no]: EOF from client [10:09] Uraeus (csc...@c3...) joined #gstreamer. [10:09] Uraeus (csc...@c3...) left irc: Read error to U= raeus[c33s9h5.upc.chello.no]: EOF from client [10:11] Uraeus (csc...@c3...) joined #gstreamer. [10:12] <omega_> wb Uraeus, sticking around this time? [10:12] <Uraeus> aargh, the 1.6.4 version of Xchat kept crashing on me [10:12] <omega_> whoops [10:12] <Uraeus> i just updated to 1.7.7 so I hope things works better no= w [10:12] Action: omega_ has no trouble with 1.7.3, except red-carpet keeps= wanting me to upgrade [10:13] <omega_> Uraeus: you have HEAD built right now> [10:13] <omega_> ? [10:13] <Uraeus> of GStreamer? [10:13] <omega_> yeah [10:13] <omega_> what else? <g> [10:13] <Uraeus> no, I don't have a 2.0 setup at all yet [10:14] <omega_> uh? [10:15] <Uraeus> sorry, I misunderstood I have the latest GStreamer CVS y= es [10:15] <omega_> um, sec.. [10:16] <omega_> um, /me has gone insane [10:16] <omega_> no, /me has gone stupid [10:16] <omega_> nevermind [10:16] <omega_> er, or not [10:17] <omega_> does the autoplugger element work for you? [10:17] <omega_> disksrc location=3Dmusic.mp3 ! autoplugger ! osssink [10:19] Uraeus (csc...@c3...) left irc: Read error to U= raeus[c33s9h5.upc.chello.no]: EOF from client [10:19] <omega_> guess not [10:19] <steveb> omega_: works for me [10:19] <omega_> with latest&greatest HEAD ? [10:20] <steveb> hang on [10:20] sienap (sy...@ip...) joined #gstreamer. [10:21] Uraeus (csc...@c3...) joined #gstreamer. [10:21] Action: Uraeus cries [10:21] <omega_> EAGAIN? [10:22] <Uraeus> yeah, now I tried doing a rm -rf on my .xchat directory = in the hopes it helps [10:22] <steveb> hmm, cvs won't respond [10:22] <omega_> works for me [10:22] <Uraeus> the crash error I get is: Gdk-ERROR **: BadDrawable (inv= alid Pixmap or Window parameter) [10:22] <Uraeus> serial 112314 error_code 9 request_code 14 minor_code = 0 [10:22] <omega_> sounds like a gdk/x problem [10:22] <omega_> threading [10:23] <omega_> but xchat doesn't use threads [10:23] <Uraeus> ok, I am running make on my CVS code so I can do the tes= t you wanted [10:23] <omega_> ok [10:23] <omega_> I get some odd results, segfaults in strange places [10:23] <omega_> and I just did an update and rebuild, so I'm confused [10:24] <Parapraxis> finally got my script compiler working... [10:24] Nick change: taaz -> taazzzz [10:24] <Parapraxis> it goes much slower when you can't see the words on = the screen because your eyes are too tired to focus :( [10:25] <omega_> script compiler? [10:25] sienap (sy...@ip...) left irc: Ping timeout for = sienap[ipc379c03c.dial.wxs.nl] [10:26] sienap (sy...@ip...) joined #gstreamer. [10:26] <Uraeus> hi sienap [10:26] <Parapraxis> omega_: yeah, sorta ;) [10:27] <sienap> parapraxis! [10:27] <sienap> hej uraeus! [10:27] <sienap> i just read about the speed improvements [10:27] <sienap> hehe :) [10:28] <Parapraxis> not much of a scripting language (there's no flow co= ntrol), but it outputs the correct pseudo-asm [10:28] Action: omega_ is lost [10:29] <omega_> all: I get a segfault somewhere after: INFO (27581:-1)gs= t_bin_add:292: [launch] added child "disksrc0" [10:29] <Parapraxis> omega_: the transform actuators will allow the user = to enter their own transform equations like 'r =3D r*.99; theta =3D theta= + pi/32;" [10:29] <omega_> ok [10:29] <Parapraxis> oh, you were lost about something else ;) [10:29] <Parapraxis> sienap: yeah, it was lots of fun :) [10:30] <omega_> no, I've got two threads running <g> [10:30] <Parapraxis> now /me is lost [10:30] <Parapraxis> you mean in yer brain? [10:30] <omega_> Parapraxis: keep explaining, and whoever's trying HEAD s= hould look for that segfault [10:30] <omega_> Parapraxis: yup [10:30] <Parapraxis> I gocha [10:30] <omega_> ;-) [10:31] <Parapraxis> well, I need some sleep [10:31] <Parapraxis> night y'all [10:31] <sienap> hehe [10:31] <sienap> ofcourse you don't [10:31] <sienap> ;) [10:31] <sienap> i just woke up [10:31] <Parapraxis> bah [10:31] <sienap> Sun Jun 17 10:35:30 CEST 2001 [10:31] <sienap> :) [10:31] <sienap> sleep well dude [10:31] <sienap> :) [10:31] <Parapraxis> 4;30 here (yes, with a semicolon) [10:31] Parapraxis (Sp...@us...) left irc: = sleeptime [10:32] <sienap> going to do some bad bad stuff tonight :) [10:33] <steveb> omega_: still works for me [10:33] <omega_> hrm [10:34] <steveb> i only built gst/ [10:34] sienap (sy...@ip...) left #gstreamer. [10:34] <omega_> my segfault is in gstparse.c.... [10:35] <omega_> mu [10:36] <steveb> does autoplug interact with gstparse [10:36] <omega_> nope [10:41] Action: omega_ has four builds of gstreamer going on at once [10:41] <omega_> only one of them is on another machine, at another site [10:42] <steveb> doesn't count then [10:42] <omega_> ok, /me has three builds of gstreamer going on at once [10:43] <omega_> and my laptop's fan just kicked on [10:46] <steveb> am i right in thinking that a buffer pool should only po= ol the data, not the GstBuffer structs as well? [10:46] <omega_> right, gstbuffer structs are handled by the gstbuffer.c = code, I think it's a mem_chunk [10:48] <steveb> yep [11:01] <steveb> what happens if a chunk runs out of atoms to alloc - doe= s it grow or barf? [11:01] <omega_> it grows [11:01] <steveb> nice [11:01] <Uraeus> omega_: ok I am ready, what was your test syntax again? [11:02] greg_ (gr...@ho...) joined #gstreamer. [11:05] <Uraeus> omega_: and I still can't do admin on bugs anymore [11:10] <steveb> Uraeus: disksrc location=3Dfoo.mp3 ! autoplugger ! osssi= nk [11:13] <Uraeus> plays fine for me [11:13] <Uraeus> fails on a mp3 with bad data at the begining though [11:18] <omega_> Uraeus: you're set up to do it, if it doesn't work, bug = sf ppl [11:18] <Uraeus> ok [11:19] <steveb> omega_: would it be acceptable to hardcode the number of= atoms per chunk? [11:20] <omega_> it already is, afaik [11:20] <Uraeus> omega_: [root@localhost tools]# gstreamer-launch disksrc= location=3D'/opt/mp3/queen.mp3' ! autoplugger ! osssink [11:21] <Uraeus> INFO (27572:-1) Initializing GStreamer Core Library [11:21] <Uraeus> INFO (27572:-1) CPU features: (c0c1f9ff) MMX 3DNOW MMXEX= T=20 [11:21] <Uraeus> RUNNING pipeline [11:21] <Uraeus> SCHEDULE DUMP FOR MANAGING BIN "launch" [11:21] <Uraeus> schedule has 3 elements in it: osssink0, unnamed_autoplu= gcache, disksrc0,=20 [11:21] <Uraeus> schedule has 1 chains in it [11:21] <Uraeus> 0x80fee90: osssink0, unnamed_autoplugcache, disksrc0,=20 [11:21] <Uraeus> SCHEDULE DUMP FOR MANAGING BIN "launch" [11:21] <Uraeus> schedule has 4 elements in it: unnamed_typefind, osssink= 0, unnamed_autoplugcache, disksrc0,=20 [11:21] <Uraeus> schedule has 2 chains in it [11:21] <Uraeus> 0x8099708: osssink0,=20 [11:21] <Uraeus> 0x80fee90: unnamed_typefind, disksrc0, unnamed_autoplugc= ache,=20 [11:21] <steveb> at the moment I have g_mem_chunk_new ("GstBufferPoolFixe= d", buffer_size,=20 [11:21] <steveb> buffer_size * 32, G_ALLOC_AND_FREE); [11:21] <steveb> so 32 data buffers will be preallocated [11:21] <Uraeus> omega_: some songs play, some songs give me the error ab= ove [11:21] <omega_> steveb: where is that? [11:22] <steveb> when I create my mem_chunk for the default bufferpool [11:22] <omega_> bleagh, /me needs to go to sleep rsn [11:22] <steveb> don't be selfish :) [11:22] <omega_> steveb: huh? the 'defaut' bufferpool is already fully f= unctional [11:22] Action: steveb sighs [11:22] Action: omega_ is confused [11:23] <steveb> previously we had a circular conversation which resulted= in you recommending that I write one using mem_chunks - unless I misunde= rstood [11:24] <omega_> using mem_chunks for the data, yes [11:24] <omega_> more like a bufferpool-factory, actually [11:26] <steveb> ah, yes. thats what I'm writing [11:26] <omega_> ok [11:26] <omega_> btw, I'm not sure you're supposed to give it buffersize*= 32, I think it's just 32 [11:27] <steveb> docs says area_size is in bytes, so that would give me s= pace for 32 data buffers [11:27] <omega_> ok [11:27] <steveb> so is 32 too much? [11:27] <omega_> depends entirely on what you're doing [11:27] <omega_> I'd make it an option right next to buffer_size [11:28] <steveb> ok, thats what I was wondering [11:29] <steveb> the thing is that elements that create bufferpools don't= really know how many to preallocate either - so it should probably becom= e an element arg [11:29] <omega_> possibly, yes [11:30] <Uraeus> omega_: SDL packages are now uploaded btw. [11:30] <omega_> ok, what about gsm ? [11:32] <omega_> Uraeus: is it just me, or does the gnome crowd blow up l= ike this roughly once every year? [11:33] <Uraeus> omega_: yup :) think we have many people with hot latin = blood participating so I guess it is only to be expected :) [11:33] <omega_> doh [11:33] <omega_> ah well, /me needs to go to sleep now [11:33] <omega_> l8r all [11:34] <Uraeus> see ya [11:34] omega_ (om...@om...) left irc: you are getting sleepy, veee= rrryyy sleeeepppyyyy.... [11:41] thomas (th...@ad...) joined #gstreamer. [11:44] Uraeus (csc...@c3...) left irc: Client Exiting [11:52] Action: steveb considers unplugging laptop to avoid death by ligh= tening [11:54] <ShrimpX> hum... I'm trying to execl() gstreamer-launch and it do= esn't wanna work... [11:59] <steveb> ShrimpX: you might have to wake wtay-zZz for help on tha= t one [11:59] <ShrimpX> hehe [12:00] <ShrimpX> I hacked up wmmp3 (a windowmaker applet mp3 player) to = use gstreamer, but gstreamer-launch pukes when execl()'d... Hum. [12:01] <steveb> you'd probably be better off writing your own c app - se= e the helloworld examples [12:02] <ShrimpX> yea, I was thinking about that. [12:10] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee04.a2000.nl] [13:02] steveb (st...@no...) joined #gstreamer. [13:57] Nick change: wtay-zZz -> wtay [13:57] <wtay> yo [13:57] <steveb> yo [13:57] <robert> how is gstreamer coming along ? [13:58] <wtay> robert: we did most of the conversion to GObject now [13:58] <robert> so gstreamer on windows is possible now ? [13:59] <wtay> maybe [13:59] <robert> i assume that means your throwing out some dependency of= a sort [13:59] <wtay> yeah [13:59] <wtay> but we still don't use the glib2.0 threads [14:00] <robert> how far off is .3 ? [14:00] <wtay> pretty far off, we need to implement some new functionalit= y first I think [14:00] <robert> can you put any predictions on a 1.0 release ? [14:00] <wtay> end of this year [14:00] <steveb> robert: use the cvs version - every day is a new release= ;) [14:01] <robert> hihihi [14:01] <wtay> every day? every hour you mean :-) [14:01] <robert> there is a lot of confusion out there about what gstream= er is actually supposed to do [14:01] <wtay> oh? [14:02] <robert> well sometimes i see people saying it'll be the best sou= nd server for gnome the universe has seen [14:02] <wtay> hmm [14:02] <robert> others say it will be a new kind of mpg player of some s= ort [14:02] <wtay> ouch [14:02] <robert> i've seen someone mention using it as a netphone ? [14:03] <steveb> some who should know better think is a streaming media f= ramework [14:03] <robert> gstreamer is useless without the plugins i would gather=20 [14:03] <wtay> robert: what you mention are possible applications built w= ith gstreamer [14:03] <wtay> robert: yup [14:04] <robert> a lot of those plugins i would imagine are garbage right= now=20 [14:04] <wtay> robert: it's really a middleware to connect, manager plugi= ns and their connection/scheduling [14:04] <wtay> robert: not really [14:05] <wtay> robert: they are not full featured maybe, mainly because t= he core doesn't support everything yet [14:06] <robert> are projects like ESD/ARTS/etc in conflict with gstreame= r ? (i'm a bit confused about this ) [14:06] <wtay> robert: not at all [14:06] <wtay> robert: they are sound servers, gstreamer isn't a sound se= rver [14:07] <steveb> wtay: asd isn't using gstreamer is it? [14:07] <wtay> well, aRts tries to do more than just being a sound server= , but... [14:07] <wtay> asd is a sound server, no? [14:07] <steveb> yes - esd replacement [14:07] <wtay> then it's just a gstreamer sink plugin [14:08] <robert> what advantage would a sound server have using gstreamer= over one that isn't ? [14:08] <steveb> yep, asd has a legacy esd interface, and its own which d= oesn't suck [14:09] <steveb> in theory gstreamer would provide all the bits to write = a sound server easily [14:09] <wtay> robert: plugins that could be reused in other apps (like t= he mixer) [14:10] <wtay> robert: when you want to implement a sound server with gst= reamer, you'ld be better of creating a media server instead, where all of= the audio/video stuff in handled in the server [14:12] <robert> hmmm [14:14] <wtay> consider this: [14:14] <wtay> If I would write a sound server, what would I nedd? [14:14] <wtay> 1. some modules to access the various OS flavours of doinf= audio output [14:15] <wtay> 2. a mechanism to load/select those plugins [14:15] <wtay> 3. some way of setting audio formats=20 [14:15] <wtay> 4. a mixer [14:15] <wtay> 5. some way of communicating the audio type to the mixer/a= udio output element [14:16] <wtay> 6. some way of passing data [14:16] <wtay> 7. meybe some threading [14:16] <wtay> 8. filters/equalizers etc.. [14:16] <wtay> 9 some way to manage all those plugins stop/pause/play.. [14:16] <wtay> that's what gstreamer does [14:17] <wtay> but then for *any* media type you can think of [14:17] <steveb> 10: magically resample and convert all sorts of inputs [14:17] <wtay> that too [14:18] <wtay> and find out what sample rate other components want/need/p= refer [14:19] <wtay> most sound server reimplement these components over and ov= er again [14:19] <wtay> what gstreamer gives you is a framework to *reuse* these c= omponents [14:20] <wtay> you can reuse the resampler, for example, if you want to d= o video encoding or an IP phone, or whatever [14:22] <wtay> so the question is: do you want to spend time implementing= a good resampler with gstreamer and know it can be reused? [14:22] <wtay> or do you want to implement is just for this one sound ser= ver purpose [14:24] <wtay> we should create the final missing piece for a sound serve= r (socketsrc/socketsink)... [14:25] <steveb> common case won't be using tcp sockets though [14:25] <wtay> steveb: yup [14:26] <steveb> so how would 2 gstreamer processes connect transparently= ? [14:27] <wtay> steveb: using a socket and some logic to pass a GstCaps [14:27] <wtay> steveb: and maybe some event passing protocol [14:28] <wtay> the local case would use shm [14:29] <wtay> I'm currently thinking about a GstElementBox using CORBA.. [14:29] <steveb> so you've got data, caps & events going down a single st= ream and elements to de/mux the stream? [14:30] <wtay> steveb: yup, the socketsrc/sink [14:30] <wtay> maybe RTP would be better suited [14:30] <steveb> i wouldn't know [14:30] <wtay> several implementation are possible of course [14:30] <wtay> there is no reason to choose only one [14:31] <steveb> how about an element pair for de/mux then a seperate pai= r for the transport of choice [14:31] <wtay> steveb: maybe, yes [14:32] <wtay> I should write up a little random doc about the ElementBox= idea I have... [14:32] <wtay> I'll let you run parts of a pipeline out of process even o= n other machines... [14:32] <wtay> transparently [14:33] <steveb> cool [14:33] <wtay> I'll need that eventually to offload processing to machine= s with specific HW... (the renderfarm) [14:34] <steveb> how would speed/latency be though? [14:34] Nick change: steveb -> corba_troll [14:34] Nick change: corba_troll -> steveb [14:34] <wtay> corba_troll: bad :) [14:34] <wtay> but that's not a real issue in a render farm [14:34] <steveb> true [14:35] <wtay> it'll never make sense to decode the mpeg audio stream on = another machine when you're watching a video :-) [14:36] <steveb> no, i was thinking of out-of-process on the same box tho= ugh [14:37] <wtay> I have no idea how well that'll go with CORBA.. [14:38] <wtay> you'll always have the context switch overhead with OOP... [14:41] Action: wtay wonders how long it'll take before a gstreamer kerne= l module is needed... [14:44] <steveb> i'd like to see Linus apply *that* patch [14:44] <wtay> he will [14:44] <wtay> he was not opposed to wine kernel support [14:44] <steveb> yeah, right after George Bush wins a second term [14:46] <steveb> what advantages would it have [14:46] <wtay> faster context switches [14:47] <wtay> omega did research after that at OGI [14:47] <steveb> ok [14:48] <wtay> it's tied to QoS in some way.. dunno exactly how it goes.. [14:53] <robert> will you guys be able to handle the fame when gstreamer = goes big time ? [14:54] <wtay> sure :) [14:57] <steveb> wtay: don't say 'sure', let you PR agent take care of th= at [14:57] <steveb> :D [14:57] <wtay> hehe [14:57] <steveb> robert: may i ask what your interest in gstreamer is? [15:12] <steveb> wtay: what code profiling techniques do you use? [15:12] <wtay> -pg usually [15:12] <wtay> doesn't work with libtool libs AFAIK, you need to really l= ink an app [15:13] <steveb> what is -pg? [15:13] <wtay> compile option to include profiling stuff [15:14] <wtay> then use gprof to inspect the result [15:14] <steveb> ok, cool [15:15] <steveb> do you usually have -pg on when developing? [15:15] <wtay> no [15:15] <wtay> I seldom do profiling [15:29] sienap (sy...@ip...) joined #gstreamer. [15:29] <wtay> hey [15:30] <sienap> hej [15:37] <steveb> wtay: i've created a simple bufferpool factory but I've = taken it a step further - when getting an instance of GstBufferPool, if a= n instance with the same buffer_size exists that will be returned. The in= stances are stored in a global hash. [15:37] <steveb> do you see any problem with that? [15:38] <wtay> not really... [15:38] <wtay> I hope you didn't touch the GstBufferPool class.. [15:39] <steveb> no, the code is in gstbufferpool.c though [15:39] <wtay> what's wrong with bufferpool? [15:39] <steveb> 1 new public method - GstBufferPool* gst_buffer_pool_get= _default (guint buffer_size, guint pool_size); [15:40] <wtay> that's not usefull [15:40] <steveb> nothing - i have just added a generic factory method whi= ch returns a bufferpool based on g_mem_chunk [15:40] <wtay> elements implementing a bufferpool will want to add more d= ata on the buffers [15:40] <wtay> like a timestamps and stuff [15:41] <steveb> they can do that after they get the buffer [15:41] <wtay> no, because the bufferpool is passed to the other element [15:41] <wtay> or do can they override the create functions? [15:42] <steveb> explain to me the bufferpool/pad relationship [15:42] <wtay> an element can create a bufferpool (for example a videosin= k) [15:42] <wtay> right? [15:42] <wtay> it has buffers the size of the screen [15:42] <steveb> yep [15:43] <wtay> it has a custom create function to allocate a shm segment = (the buffer data) [15:43] <wtay> and a destroy function to free the mem [15:43] <steveb> ok [15:43] <wtay> en element can request a bufferpool from the peer element = and then it'll get a reference to this pool [15:44] <wtay> it then creates a buffer from the pool... [15:44] <wtay> the control is given back to the videosink which allocates= a piece of shm and hands it to the peer element [15:44] <wtay> it'll set some private data on the buffer too so that it k= nows how to deallocate the data in case of a buffer_unref [15:45] <steveb> what if the peer element doesn't have a bufferpool? [15:45] <wtay> so your solution is for the peer element to have a bufferp= ool of its own? [15:46] <wtay> I see... [15:46] <wtay> you don't stricly need a bufferpool then [15:46] <wtay> just a mem_chunck in the element [15:46] <wtay> hmm, yes you do sorry... [15:47] <wtay> sounds good to me then [15:47] <steveb> i think my use case is a bit different. consider sinesrc= ! int2float ! float2int. each element has to create new buffers of diffe= rent sizes than the peer [15:48] <wtay> float2int would have a bufferpool on the sink pad [15:48] <wtay> the others operate in place [15:48] hadess (ha...@pc...) joined #gst= reamer. [15:48] <wtay> yo [15:49] <hadess> hey wtay [15:50] <steveb> this way its trivial for elements to get buffers of any = size from a bufferpool without rewriting the same mem_chunk code [15:50] <hadess> hey steveb [15:51] <wtay> steveb: yup [15:51] <steveb> hi [15:51] <steveb> can I send you what I have done? [15:51] <wtay> sure [15:51] <steveb> or should I just commit :) [15:51] <wtay> just commit ;-) [15:57] <steveb> ok, touched gst.c and gstbufferpool.* [15:57] <steveb> and intfloatconvert uses it [15:58] <wtay> looks perfect [15:59] <steveb> great [15:59] <wtay> maybe a function to unref the default pool would be needed= too [15:59] <wtay> else you'll have a leak [16:00] <wtay> ah, but that's the standard destroy function... [16:00] <wtay> hmm, no.. [16:00] <steveb> there needs to be a pool_destroy added to the bufferpool= api [16:01] <steveb> but then it can only be destroyed if all buffers have be= en handed back [16:01] <wtay> there is a pool destroy, but it won't free your chunk [16:02] <wtay> yup, it needs refcounting... [16:02] <steveb> so there would need to be another func ptr for that [16:03] <robert> steveb: I'm just a guy watching the project [16:03] <steveb> robert: ah [16:03] <wtay> maybe a flag indicating that it's a default pool and free = the data... [16:03] <robert> if you feel like ignoring my questions, thats okay :) ei= ther way i lurk... [16:04] <steveb> its only a leak if buffers of unique sizes keep being cr= eated [16:05] <wtay> different sizes you mean? [16:05] <steveb> oh - i'm wrong [16:05] <wtay> unique in the way that they are all different? [16:05] <steveb> nevermind [16:06] <wtay> every time a buffer is created from the pool, the refcount= increases, when a buffer is handed back it decreases. [16:06] <wtay> when the count is zero, the pool is freed [16:06] <wtay> the destroy function goes away and an pool_unref is added [16:07] <steveb> yep, i can see that should work [16:07] <steveb> should gstbufferpool be a fully fledged GObject? [16:07] <wtay> hmm, mpeg_play could use this new pool too... [16:07] <wtay> steveb: I don't think so [16:07] <wtay> that might be too expensive [16:08] <steveb> rgrep gst_buffer_new * - lots of things could use it :) [16:08] <wtay> make it like GstBuffer [16:08] <wtay> yup [16:17] <wtay> is float2int a mixer and a converter? [16:19] sienap (sy...@ip...) left irc: AAAAAAAAAAH [16:20] <steveb> it doesn't really mix - it just creates interleaved ints= from multiple mono float pads [16:21] <steveb> the idea is that if you're using floats, you're doing it= in a "LAAGA" type environment with no interleaved data [16:41] <wtay> steveb: yup [16:59] Nick change: robert -> Bob [17:00] Nick change: Bob -> robert [17:00] Nick change: robert -> Bob [17:12] <steveb> what is the #ifdef HAVE_ATOMIC_H split about? is it stil= l relavent? [17:14] <wtay> I think so, yes [17:14] <wtay> atimic* functions are faster [17:15] Nick change: taazzzz -> taaz [17:15] <wtay> hi [17:16] <taaz> hey [17:20] hadess (ha...@pc...) left irc: t= roubles [17:41] _gst_newt_ joined #gstreamer. [17:51] Zeenix (programmer@203.135.48.13) joined #gstreamer. [17:51] <Zeenix> hello [17:54] <wtay> hi [17:55] <Zeenix> wtay: how was the snooker, are you good at it too [17:55] <wtay> I won 5-1 :-) [17:56] <Zeenix> congragatulations( dont know the spelling of it ) [17:56] <wtay> I usually win.. [17:57] <Zeenix> can you guide me to a good online intro on embedded linu= x ? [17:58] <wtay> no idea... [17:58] <Zeenix> where is Ureaus these days [17:59] <wtay> cleaning up his appartement :-) [18:00] <Zeenix> thanks God( if exists ), this department is of my mom [18:02] <Zeenix> if she is not home, i wont clean my room for my whole li= fe [18:02] <wtay> I never clean my house [18:03] <wtay> it's incredibly dirty here [18:03] <Zeenix> wtay: what about you video editting software ? [18:03] <wtay> when gstreamer can support it I'll start on it [18:04] <Zeenix> you girlfriend doesnt mind your dirty room [18:04] <Zeenix> ? [18:04] <wtay> not at all :-) [18:06] <Zeenix> how is the gst 0.2.0 going, is it a success ? [18:07] <wtay> not much more trafic here... [18:07] <wtay> if that's an indication of success [18:08] <Zeenix> no , not in terms of that but practically i mean [18:09] <wtay> it sorta works.. [18:09] <Zeenix> wtay: i see doubts in your words [18:09] <wtay> it still needs some work [18:10] <Zeenix> did you updated the manual [18:10] <wtay> yes [18:10] <Zeenix> 100% compatible with 0.2.0 [18:10] <Zeenix> ? [18:10] <wtay> I hope so :) [18:13] <Zeenix> wtay: should i register my project on sourceforge.net ? [18:13] <wtay> yeah [18:13] <Zeenix> are there any benefits ? [18:14] <wtay> sure, you have a website and CVS [18:14] <wtay> and project management [18:15] <Zeenix> wtay: is gstreamer powered by that ? [18:15] <wtay> yup [18:15] dobey (do...@di...) joined= #gstreamer. [18:15] <wtay> hi [18:15] <dobey> hi [18:17] <wtay> hmm, I need food... [18:17] <Zeenix> wtay: will it give us a machine with static IP for runni= ng our server [18:17] Nick change: wtay -> wtay-food [18:17] <dobey> yeah [18:17] <dobey> i'm thinking an ibook2 dvd would be good too ;-) [18:17] <wtay-food> dobey: tell him about SF [18:17] <dobey> who? what? [18:18] <wtay-food> dobey: while I'm eating... :) [18:18] <dobey> tell who? [18:18] <wtay-food> Zeenix [18:18] <dobey> oh [18:18] <dobey> i /ignored him because he never shuts up asking where you= are [18:19] <dobey> in one acronym, and one word though... [18:19] <dobey> SF sucks [18:19] <Zeenix> dobey: i am sorry for that [18:19] <wtay-food> hehe [18:19] <dobey> is that good enough? [18:19] <dobey> besides [18:19] <dobey> i gotta get food too [18:20] <Zeenix> dobey: you must understand that my culture is too differ= ent from yours & so are the values & language [18:28] steveb (st...@no...) left irc: [x]chat [18:29] steveb (st...@no...) joined #gstreamer. [18:51] Zeenix (programmer@203.135.48.13) left irc:=20 [19:17] Nick change: wtay-food -> wtay [19:17] <dobey> whee [19:18] <wtay> ? [19:18] Action: dobey is looking at the ibook2 thingies [19:19] Action: wtay is writing a flac plugin [19:19] <dobey> flac? [19:19] <wtay> flac.sourceforge.net [19:20] <dobey> gah [19:20] Action: dobey saturates his modem [19:26] Uraeus (csc...@c3...) joined #gstreamer. [19:27] <Uraeus> evening [19:27] <wtay> Uraeus: Hyggelig =E5 treffe deg <g> [19:27] <Uraeus> hmm [19:27] <wtay> uhm [19:28] Action: Uraeus scans the list of people who might have told wtay = that sentence :) [19:28] <dobey> hey Uraeus [19:28] <Uraeus> wtay: takker, hyggelig =E5 m=F8te deg =E5 [19:28] <wtay> http://www.sofn.com/cultural/language/language1.html [19:28] <dobey> farsi? [19:28] Action: thomas finally has some pictures of Dave/Dina online [19:28] <thomas> hi Uraeus... [19:28] <Uraeus> hi dobey, no not farsi [19:28] <Uraeus> hi thomas [19:28] <thomas> I've been waiting for your ;) [19:28] <thomas> uhm you that is [19:28] <wtay> Uraeus: michelle is trying to learn norwegian :) [19:29] <Uraeus> cool [19:29] <thomas> Uraeus: did you use/change the spec file in CVS for your= RPM's ? [19:29] <wtay> too bad my keyboard doesn't have those funny chars.. [19:29] <Uraeus> wtay: I can expect to see you two here soon then? [19:29] <dobey> heh [19:29] <wtay> Uraeus: maybe :) [19:30] <Uraeus> thomas: yes [19:30] <dobey> where is mayam anyway? [19:30] <wtay> dobey: wathing TV.... [19:30] Action: dobey blows her a kiss [19:30] <dobey> ;-) [19:30] <Uraeus> thomas: but I haven't updated it with Erik's change requ= ests yet [19:31] <wtay> dobey: se sends you a hug back... [19:31] <thomas> Uraeus: ok, so should I try "make rpm" to make rpm's bas= ed on cvs, or should I do the standard thing (wrap them in a tar.gz and m= ove them to /usr/src/redhat ?) [19:31] <dobey> wtay: hehe :-) [19:31] <thomas> because I have a box doing some stuff with gstreamer and= I want to do a full nfs-based install of everything it needs [19:31] <Uraeus> thomas: never tried the make rpm approach, a maming a ta= r.gz and doing rpm -ta is another alternative :) [19:32] <Uraeus> s/maming/making/ [19:32] <dobey> ooh [19:32] <dobey> you guys are hacking specs now? [19:32] <thomas> Uraeus: ok, I'll try doing it the normal way then ... [19:32] <thomas> ... this will take a long time ;) [19:33] <Uraeus> dobey: I have a almost perfet SPEC file in there now, so= you at Ximian can include GStreamer in Ximian GNOME now without a proble= m :) [19:33] <dobey> heh [19:33] <dobey> we wouldn't have a problem anyway [19:33] <dobey> other than having to ship it's 500000 requirements ;-) [19:33] <dobey> does it build some packages arch and some noarch? [19:34] <Uraeus> dobey: well, I have split each plugin into a separate rp= m so you only need to ship the reqs you want to ship [19:34] <Uraeus> no [19:34] <Uraeus> what could be buildt noarch in there? [19:34] <dobey> gstreamer-devel? [19:35] <dobey> (if you put the static files in a different package) [19:35] <dobey> i don't know [19:35] <Uraeus> it has binaries in it [19:35] <dobey> hence the () ;-) [19:35] <dobey> but i'm having problems with BuildArchitectures: [19:37] <dobey> meh [19:52] Parapraxis (Sp...@us...) joined #gs= treamer. [19:53] <Parapraxis> howdee [19:53] <wtay> hi [19:53] <dobey> la la [19:53] <dobey> either rpm4 is broke [19:54] <dobey> or it's really broke [20:21] <Parapraxis> how do enum's work (under the hood)? Can they be us= ed by asm code? [20:32] <wtay> gobject enums or c enums? [20:33] <Parapraxis> c [20:33] <wtay> they're just ints [20:33] <Parapraxis> are they global? [20:33] <Parapraxis> how does it work if they're in .h's [20:33] <wtay> it's a typedef [20:34] <wtay> so they're know to all who include the .h file [20:34] <wtay> known even [20:34] <Parapraxis> right, but they get put into the symbol table someho= w, right? [20:34] <wtay> nope [20:34] <Parapraxis> so how does gdb recognize them? [20:34] <wtay> it's like #define value1 1 [20:35] <wtay> probably metadata in the debug info [20:35] <Parapraxis> ah [20:35] <wtay> as with all typedefs [20:35] <wtay> AFAIK [20:36] <wtay> I dunno how to acces them from asm [21:07] <dobey> gah [21:07] <dobey> where is hadess? [21:14] <wtay> dunno, I don't manage his schedule.. :-) [21:16] <dobey> heh [21:21] <Uraeus> agh, geez, just fixed about 40 broken links on developer= .gnome.org, that is what I call a boring job :) [21:29] <Bob> i'd do it gladly [21:30] <Uraeus> Bob: well, then you should join the gnome-webmaster mail= inglist [21:30] <wtay> Uraeus: I made a flac decoder plugin :) [21:31] <Uraeus> flac? [21:31] <wtay> flac.sourceforge.net [21:32] <Uraeus> hey, cool, have you told the authors of flac? [21:32] <wtay> not yet... [21:32] <wtay> that's your job :-) [21:32] <Uraeus> heh, ok [21:32] <wtay> unless you are too busy... [21:33] <Uraeus> no, just migrating the GNOME Basic webpage from the old = server to the new one [21:33] <Uraeus> GNOME Basic as in Visual Basic for GNOME [21:33] <wtay> yeah, I heard about it [21:34] <Bob> thats still alive ? [21:34] <Uraeus> yeah, I love it, everytime we post a news story about it= people sends the authors 50 hate mails :) [21:34] <Uraeus> Bob: yes, was lots of CVS commits the other day [21:34] <Uraeus> Bob: moves slowly but surely forward :) [21:34] <Bob> uareus: spelling corrections ? :) [21:34] <wtay> I think its a great project [21:34] <wtay> a stupid language nevertheless [21:34] <Uraeus> Bob: on gnome.org? [21:35] <Bob> the cvs commits :> [21:35] <Uraeus> Bob: not sure I am with you, Ravi Patrap commited some c= ode if I am not mistaken [21:35] <Bob> oky.. just a a little joke [21:36] <Bob> i wish the gbasic effort well but it seems like a project d= oomed from the start [21:37] <Uraeus> well it need hackers who don't mind getting flamed regur= aly at least :) [21:37] <mattsm> people are so retarded [21:38] <dobey> oh, i get flamed regularly [21:46] <Uraeus> hehe funny (http://www.gnome.org/projects/) the gnome-co= re dump directory is listed with a core dump icon in Galeon :) [21:46] <Uraeus> err...s/gnome-core dump/gnome-core/ [21:46] <Uraeus> ;) [21:47] <wtay> hmm, is that galeons fault? [21:48] <Uraeus> wtay: well it obviously parses the word 'core' and assum= es that the gnome-core directory is a core dump :) [21:48] <wtay> yeah, it's the server side software that does that :-) [21:49] <Uraeus> i am sure gnome haters out there would think it fitting = :) [21:49] <dobey> haha [21:49] <dobey> i think it fitting [21:50] <dobey> and i love gnome [21:50] <dobey> it's probably more fitting for galeon though [21:50] <wtay> heh [21:50] <Uraeus> no Galeon bashing dobey, it is my and wtay's favourite b= rowser [21:51] <wtay> yeah, but I can't deny it crashing from time to time.. [21:51] <dobey> i'm not bashing [21:51] <dobey> i'm agreeing with it's tendency to not work [21:51] <dobey> if i wanted to bash it [21:51] <dobey> i'd /j #galeon [21:51] <Uraeus> well, it is pretty stable for me, but there are a few bu= gs here and there, but much fewer than Encompass for instance :) [21:52] <wtay> It's the best all round browser there is, face it [21:52] <dobey> uhm [21:52] <dobey> sigh [21:52] <dobey> mozilla is much better [21:53] <Uraeus> don't agree at all [21:53] <dobey> it actually implements the features correctly [21:53] <wtay> it doesn't fit into my desktop [21:53] <Uraeus> I love Galeon [21:53] <dobey> but don't get me started [21:53] <dobey> i am definately not in the mood for it [21:53] <dobey> and i added a feature yesterday [21:53] <dobey> for all of you feature-kiddies [21:53] <dobey> haha [21:53] <wtay> ? [21:53] <dobey> sigh [21:53] <dobey> don't make me /ignore you [21:54] <dobey> please [21:54] <dobey> :-( [21:54] <dobey> <- the fuckin definition of "jaded" [21:55] <Uraeus> wtay: are you planing on making a FLAC encoder also? [21:56] <wtay> Uraeus: I'm working on it, yes [21:56] <dobey> i'll just say the last 24 hours have not been my best [21:58] <wtay> doh! I forgot to reply to the gtk-devel list confirmation = of subscription [21:59] <Uraeus> wtay: is the FLAC decoder in CVS? [21:59] <wtay> yup [21:59] <Uraeus> dobey: sorry to hear it [22:00] <dobey> yeah [22:01] <dobey> i'll live [22:01] <dobey> i'll be much happier tomorrow [22:04] <wtay> hmm, not me, I have to work :( [22:04] <dobey> i have to work to [22:04] <Uraeus> thanks for reminding me it was monday tommorow wtay, ugh= :( [22:06] <wtay> what's worse, I cannot sleep on sunday nights... [22:06] <dobey> heh [22:06] <dobey> i'm glad tomorrow is monday [22:07] <Uraeus> wtay: yeah, I know those MASH reruns keep me awake too <= G> [22:07] <wtay> Uraeus: no, I just cannot fall asleep... [22:07] <Uraeus> heh, I was kidding [22:08] <Uraeus> I used to be plagued by insomnia as a kid, these days my= problem is staying awake [22:08] <wtay> which means I'm usually a complete wreck on monday morning= and the rest of the week.. [22:08] <wtay> but!... :-) [22:08] <wtay> It'll be a short week.. [22:09] <Uraeus> public holiday? [22:09] <wtay> It's my birthday on thursday and I'll be off for two days = :-) [22:09] Action: Uraeus adds wtays birthday to his dayplanner so he will n= ot forget to congratulate :) [22:10] <wtay> lol [22:10] <wtay> a dayplanner... [22:10] <Uraeus> actually it is a Palm [22:10] <Uraeus> Vx [22:10] <wtay> aha [22:10] <wtay> let's see who reads the logs :-) [22:11] <Uraeus> wtay: noticed your new admin status in #GNOME yesterday?= :) [22:11] <wtay> yeah, cool [22:11] <dobey> dude [22:11] <Uraeus> wtay: yeah, now you can ban dobey from #gnome :) [22:11] <dobey> "evolution" [22:11] <dobey> :-) [22:12] <Uraeus> ? [22:12] <wtay> Uraeus: or YOU.. hehehehehe [22:12] <dobey> anyway [22:12] <dobey> hrmm? [22:12] <dobey> "admin" [22:12] <dobey> whatever [22:12] <Uraeus> hmm [22:12] <wtay> I would never ban dobey <grin> [22:13] <Uraeus> dobey: ok, Op is that better :) [22:13] <dobey> i would just have *bot unban me [22:13] <dobey> and it's not like #gnome is a particularly interesting pl= ace right now anyway [22:14] <Uraeus> actually I think there are suprisingly many good technic= al discussions there despite its size [22:14] <dobey> yes [22:14] <Uraeus> like sopwith and clahey last night [22:14] <dobey> but not right now i'm guessing [22:14] <dobey> because of g-h [22:15] <dobey> and martin's post [22:15] <Uraeus> well, hp and sopwiths discussion an hour ago was interes= ting [22:15] mattsm (ma...@ad...) left ir= c: Client Exiting [22:17] <Uraeus> a little flamewar could be healthy once and again, think= this one was essential to actually get people to synch with eachother ag= ain, and I think martin will continue as release manager [22:17] <dobey> heh well [22:24] <thomas> hey, did anyone check out transcode, which was on freshm= eat a few days ago ? [22:25] <wtay> uh? no [22:25] <thomas> Uraeus: if you're advocating, maybe you should contact t= hem ;) [22:25] <thomas> you want the blurb on it ? [22:25] <thomas> What is transcode?=20 [22:25] <thomas> transcode is yet another linux text-console utility = to encode raw video/audio streams extracted from, for example, DVD or Dig= ital Video sources. It's modular [22:25] <thomas> concept is intended to provide flexibility and easy = user extensibility. However, you need some other packages already working= on your system to [22:25] <thomas> decode/encode your favorite DVDs or vacation Digital= Videos to single CD video files and a platform that supports shared libr= aries. The program is still in beta [22:25] <thomas> state but fairly usable [22:36] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee04.a2000.nl] [22:58] <Uraeus> hmm, ok I look into it [23:06] greg_ (gr...@ho...) left irc: Read error to greg_[home.sen= te.pl]: EOF from client [23:08] <Uraeus> thomas: ok, I mailed the author [23:09] <thomas> Uraeus: ok, cool. It looks like something that should b= e done in gstreamer [23:09] <thomas> and maybe some of his plugin stuff can be converted as w= ell [23:10] <wtay> he uses the standard libs so wa allready have them [23:13] dobey (do...@di...) left i= rc: bbiab [23:18] <Uraeus> time for me to sleep, need to be in at work at 0600 tomm= orow, see ya [23:18] Uraeus (csc...@c3...) left irc: Client Exiting [23:30] <thomas> I'm off too. bye wtay [23:30] <wtay> cya [23:30] thomas (th...@ad...) left irc: Client E= xiting [23:34] dobey (do...@di...) joine= d #gstreamer. [00:00] --- Mon Jun 18 2001 [00:06] dobey (do...@di...) left = irc: later [00:37] ajmitch (aj...@p3...) joined #gstreamer. [00:41] <wtay> yo [00:41] <ajmitch> hi [00:42] <wtay> I gotta sleep now... [00:42] <ajmitch> hehe [00:42] Nick change: wtay -> wtay-zZz [01:02] ryu (username@213.96.225.68) joined #gstreamer. [01:02] <ryu> hi [01:03] ryu (username@213.96.225.68) left irc: Read error to ryu[213.96.2= 25.68]: EOF from client [01:04] ryu (username@213.96.225.68) joined #gstreamer. [01:05] ryu (username@213.96.225.68) left irc: Read error to ryu[213.96.2= 25.68]: EOF from client [01:12] ryu (username@213.96.225.68) joined #gstreamer. [01:12] <ryu> hi [01:13] <ajmitch> hi [01:13] <ryu> I can't get any program working with gstreamer, all the tes= ts in the package fail, and the programs that I've built to learn 0.1.1 d= oesn't work now (although I changed audiosink for afsink) [01:14] <ryu> I get this error always: ** CRITICAL **: file gstelement.c:= line 779 (gst_element_set_state): assertion `element->sched !=3D NULL' f= ailed. [01:14] <ryu> an then a big core :) [01:15] <ryu> (of course I've built the registry again for the new versio= n) [01:15] <ajmitch> you're running 0.1.1? [01:15] <ryu> not, I've just installed 0.2.0 [01:16] <ajmitch> 0.2.0 from cvs or the tarball? [01:16] <ryu> the tarball [01:16] <ryu> I wanted something quite stable to work [01:17] <ajmitch> ok... [01:17] <ryu> Even the simplest example, to play an mp3 doesn't work :( [01:17] <ajmitch> pity i'm not a developer really ;) [01:19] <ryu> and the developers are of course developing, not chating :)= ) [01:19] <ajmitch> or sleeping, at the moment (the ones in europe anyway) [01:19] <ryu> I should also :) [01:20] <ajmitch> hehe [01:21] <ajmitch> omega or wtay would be the ones to ask about this [01:25] <taaz> could use --gst-mask=3D-1 to get more debug output [01:25] omega_ (om...@om...) joined #gstreamer. [01:25] ajmitch (aj...@p3...) left irc: Ping timeout = for ajmitch[p3-max11.dun.ihug.co.nz] [01:25] <ryu> hi omega_ [01:25] <omega_> yo [01:25] <ryu> I've a problem with gstreamer 0.2.0 [01:26] <ryu> I get this error always: ** CRITICAL **: file gstelement.c:= line 779 (gst_element_set_state): assertion `element->sched !=3D NULL' f= ailed. [01:26] <ryu> The code is the same that I used in 0.1.1, only changing au= diosink with afsink (audiosink disappeared :) [01:27] <ryu> In fact, the "mp3" sample included in the sources also does= n't work for me [01:27] <omega_> afsink writes to a disk file afaik [01:28] <ryu> I will change with ossaudio then [01:28] <omega_> osssink [01:29] <ryu> ok [01:29] <ryu> I get the same error [01:29] ajmitch (aj...@p3...) joined #gstreamer. [01:30] <omega_> paste the exact pipeline here [01:30] <ryu> bin =3D gst_bin_new ("bin"); [01:30] <ryu> disksrc =3D gst_elementfactory_make("disksrc","disk_sou= rce"); [01:30] <ryu> gtk_object_set (GTK_OBJECT (disksrc),"location", "cosa.= mp3", NULL); [01:30] <ryu> parse =3D gst_elementfactory_make ("mp3parse", "parse")= ; =20 [01:30] <ryu> decoder =3D gst_elementfactory_make ("mpg123", "decoder= "); [01:30] <ryu> audiosink =3D gst_elementfactory_make ("osssink", "play= _audio"); [01:30] <omega_> oh, first problem... [01:30] <ryu> what? [01:30] <omega_> a Bin has no scheduler, therefore nothing can run [01:30] <omega_> you must have a thread or a pipeline as the outermost co= ntainer [01:30] <ryu> oh [01:30] <ryu> ok [01:32] <ryu> works! [01:32] <ryu> thanks [01:32] <omega_> cool [01:35] <ryu> there is a little problem with the windows codec, either in= gstmediaplay and in my programs, when the plugin is loaded I get this wa= rning: [01:35] <ryu> : error while loading shared libraries: /usr/lib/gst/libwin= codec.so: undefined symbol: gst_colorspace_converter_new [01:35] <omega_> you need libHermes [01:35] <ryu> ok [01:36] <omega_> but file a bug report on sourceforge, since it should ca= tch that case and complain a little more cleanly, IMO [01:36] <ryu> ok [01:37] <ryu> I have libHermes1 [01:38] <ryu> version 1.3.2 [01:38] <ryu> is it a wrong version? [01:38] <omega_> hmmm, dunno [01:38] <omega_> doubt it, it's a pretty stable API [01:38] <omega_> uh, check for /usr/local/lib/gst/libcolorspace.so [01:38] <omega_> or somesuch [01:38] <ryu> I've not [01:39] <ryu> oh [01:39] <omega_> hmm [01:39] <ryu> I've hermes-dev package :DD [01:39] <omega_> ah [01:39] <ryu> sorry [01:39] <ryu> I *need* [01:39] <ryu> :) [01:45] <ryu> the bug is submitted [01:45] <omega_> ok, thanx [02:08] <ryu> nw I have /usr/lib/gst/libcolorspace.so and /usr/lib/gst/li= bgstcolorspace.so, but I get the same problem [02:09] <omega_> did you re-run -register ? [02:09] <ryu> yes [02:09] <omega_> hmmm [02:09] <ryu> I've uninstalled all, rebuild all, and reinstalled all :D [02:09] <omega_> dunno. wouldn't surprise me if it doesn't work at all [02:09] <omega_> try sticking colorspace in the middle of a pipeline befo= re xvideosink [02:09] <ryu> I've not videosink [02:10] <omega_> disksrc location=3Delementary-video.mpg ! mpeg2dec ! col= orspace ! xvideosink [02:10] <omega_> what are you outputting to then? [02:10] <ryu> I've aviencoder [02:10] <omega_> are you decoding or encoding with wincodec? [02:11] <ryu> encoding [02:11] <omega_> oh, well, I tend to doubt that works at all [02:11] <omega_> I'd check with wtay, he would know a lot more about that= than I [02:11] <ryu> I wanted to make a simple app that reads an mpeg-1 and writ= es it in divx format [02:12] <ryu> ok [02:12] <omega_> if you want to do just video for now, you can use -launc= h to test things out [02:12] <omega_> I would very much recommend that you do that to start wi= th, rather than writing a program to do it [02:12] <omega_> saves *lots* of time [02:13] <ryu> launch? [02:13] <ryu> oh, gstreamer-launch [02:15] <ShrimpX> omega_, what does gst_init() do? [02:15] <omega_> all kinds of setup stuff [02:15] <omega_> see gst.c [02:15] <ShrimpX> k [02:21] <ShrimpX> it should work fine if I pass it bogus argc and argv, r= ight? [02:21] <omega_> I think so, yes [02:22] <ShrimpX> cool. I'm modifying an mp3 player to use gstreamer, and= I wasn't sure whether I needed to pass the song name or sonething to gst= _init [02:22] <omega_> no, what you do with gstreamer is solely the domain of t= he elements you use [02:23] <omega_> gstreamer core by definition doesn't care [02:23] <ShrimpX> nice. [02:33] <Parapraxis> omega_: do you know if it's possible to use c enum's= in asm? [02:34] <omega_> not sure, try it <g> [02:34] <Parapraxis> heh ;) [02:34] <omega_> a standalone .S, probably not [02:35] <omega_> extended inline, by way of a "i"(ENUM_NAME) [02:35] <omega_> actually, "ir"(ENUM_NAME), to give ... [truncated message content] |
From: <wim...@ch...> - 2001-06-19 04:30:25
|
[07:07] Nick change: taaz -> taazzzz [07:16] ajmitch (aj...@p5...) joined #gstreamer. [07:16] mattsm (ma...@ad...) left ir= c: Client Exiting [07:49] steveb (st...@no...) joined #gstreamer. [08:00] <Parapraxis> hmmm... [08:00] Action: Parapraxis is suspicious because his scripting engine see= ms to be working on the first try [08:02] <Parapraxis> oh, I just thought of a new way to break it :) [08:08] ajmitch (aj...@p5...) left irc: http://www.fr= eedevelopers.net [08:37] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee4c.a2000.nl] [08:37] steveb (st...@no...) joined #gstreamer. [08:40] <Parapraxis> 'night y'all [08:40] Parapraxis (Sp...@us...) left irc: = sleepses [08:40] <omega_> l8r [08:40] <omega_> doh [08:43] chillywilly (bau...@d1...) joined #gst= reamer. [08:43] <chillywilly> re [08:43] <omega_> yo [08:45] <chillywilly> hey how do you guys do plugin...I wanna implement a= plugin architecture for GEAS (GNUe Application Server) to support object= methods that can be written in multiple langauges [08:45] <chillywilly> plugins [08:45] Action: chillywilly is looking at your API docs [08:46] <omega_> hmmm, how many different languages? [08:46] <chillywilly> GEAS implements the "methods" and rules of business= objects...http://www.gnuenterprise.org we just release 0.0.5 [08:47] <chillywilly> C, python, guile...for a start [08:47] <chillywilly> well we have C and python right now but the whole m= ehtods lib needs reworking [08:48] <chillywilly> as methods need to run in their own thread so as no= t to kill the whole server if a method dies [08:48] <omega_> hmm, we very much rely on gtkobject, and soon gobject, s= o that's what you should be looking at probably [08:48] <omega_> chillywilly: threads do not provide that kind of safety,= only fork() does [08:48] <chillywilly> eh, well the origanl and big GEAS was not too hot o= n OO C [08:48] <chillywilly> guy [08:48] <chillywilly> so it is written in regular old C [08:48] <omega_> good ;-) [08:49] <omega_> C is good [08:49] <chillywilly> but [08:49] <chillywilly> he will be leaving soon [08:49] <chillywilly> I will probably take over as the main GEAS guy [08:53] <chillywilly> GEAS is somewhat modular, but libGDA is getting to = bulky and we are gonna ditch it and write our own db abstraction layer...= anyway I think a lot of things will be rewritten as the current GEAS was = hacked together and not much design went inot it afaik [08:55] <chillywilly> I thought an individual thread recieves a signal? [08:56] <omega_> yes, but if it's a fault, the whole process goes down [08:56] <chillywilly> hmmmm [08:56] <omega_> and if the thread goes postal, the other threads all can= get trashed [08:56] <chillywilly> yeah I suppose....don't know what I was thinking [09:04] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee4c.a2000.nl] [09:04] steveb (st...@no...) joined #gstreamer. [09:27] chillywilly (bau...@d1...) left irc: P= hilosophers and plow men, each must know his part, to sow a new mentality= closer to the heart... [09:35] sbaker (ste...@21...) joined #gstreamer. [09:35] Nick change: sbaker -> steveb______ [09:35] Nick change: steveb______ -> steveb_ [09:59] steveb_ (ste...@21...) left irc: quit has s= teveb, yes [10:08] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee4c.a2000.nl] [10:09] johnix (thomas@61.152.140.57) joined #gstreamer. [10:17] thomas (th...@21...) joined #gstreamer. [12:03] johnix (thomas@61.152.140.57) left irc: Client Exiting [12:42] big_T (th...@fy...) got netsplit. [12:43] omega_ (om...@om...) left irc: [x]chat [12:44] big_T (th...@fy...) returned to #gstreamer. [12:51] big_T (th...@fy...) got netsplit. [12:52] big_T (th...@fy...) returned to #gstreamer. [12:58] big_T (th...@fy...) got netsplit. [12:59] big_T (th...@fy...) returned to #gstreamer. [13:16] richardb (ri...@ix...) joined #gstreamer. [13:16] <richardb> yo [13:17] <thomas> hi [13:18] <richardb> ./gstreamer-launch disksrc location=3Dsome.wav ! parse= wav ! stereo2mono ! chart ! xvideosink [13:18] <richardb> draws a waveform of the wav. :) [13:18] <thomas> richardb: cool ;) [13:19] <thomas> richardb: I had about the same thing with the level plug= in, but not directly to an output plugin [13:19] <richardb> Looks a bit ugly at the moment, but hey. [13:19] <thomas> how does char work ? [13:19] <thomas> chart ... [13:19] <thomas> ? [13:19] <richardb> My plan is that the chart plugin will just take some d= ata and draw it. [13:19] <richardb> And you have a separate plugin to convert audio into d= ata. [13:19] <thomas> ok, let the chart plugin take data from the level plugin= then [13:19] <thomas> I was thinking about creating a raw/envelope mime type f= or that [13:19] <richardb> At the moment, there's just one plugin, but I'll split= it later. [13:20] <thomas> or something like that [13:20] <thomas> then create a second output pad from the level plugin [13:20] <thomas> connect that to something like chart [13:20] <richardb> Yep. [13:20] <richardb> chart will have various arguments to control how it lo= oks. [13:20] <thomas> richardb: ok, I'll check the chart plugin then ;) [13:20] <richardb> And there should be other chart plugins with completel= y different looks. [13:21] <thomas> how does it draw the wav now ? [13:21] <richardb> A waveform. [13:21] <thomas> richardb: oh, so you do a full graph then, of every samp= le ? [13:21] <richardb> Um. It looks a bit like alsaplayer's "monoscope" [13:21] <richardb> But not as neat. [13:21] ajmitch (aj...@p3...) joined #gstreamer. [13:22] <richardb> It picks a chunk of data, 256 samples long, and draws = the sample values [13:22] <richardb> with sample number horizontally, and magnitude vertica= lly. [13:23] <richardb> Draws them as a solid line from the midvalue to the sa= mple value. [13:23] <richardb> Hmm, autoplugger fails to work for the chart stuff. O= r maybe its wav stuff. [13:23] <ajmitch> hi all [13:23] <richardb> ./gstreamer-launch disksrc location=3Dsome.wav ! parse= wav ! autoplugger ! xvideosink works [13:24] <richardb> ajmitch: hi [13:27] <richardb> Anyone know the current state of the autoplugger? [15:03] leviathan (le...@p3...) joined #gstreamer. [16:07] Nick change: ajmitch -> ajzzzz [16:12] thomas (th...@21...) left irc: [x]chat [16:15] big_T (th...@fy...) got netsplit. [16:16] big_T (th...@fy...) returned to #gstreamer. [16:23] dobey (dobey@141.154.95.104) joined #gstreamer. [16:42] big_T (th...@fy...) got netsplit. [16:43] big_T (th...@fy...) returned to #gstreamer. [16:51] steveb (st...@no...) joined #gstreamer. [18:02] Nick change: taazzzz -> taaz [18:08] ryu (username@213.96.225.68) joined #gstreamer. [18:08] ryu (username@213.96.225.68) left irc: Read error to ryu[213.96.2= 25.68]: EOF from client [18:09] ryu (username@213.96.225.68) joined #gstreamer. [18:10] ryu (username@213.96.225.68) left irc: Read error to ryu[213.96.2= 25.68]: EOF from client [18:11] ryu (username@213.96.225.68) joined #gstreamer. [18:12] ryu (username@213.96.225.68) left irc: Read error to ryu[213.96.2= 25.68]: EOF from client [18:13] ryu (username@213.96.225.68) joined #gstreamer. [18:14] ryu (username@213.96.225.68) left irc: Read error to ryu[213.96.2= 25.68]: EOF from client [18:34] mattsm (ma...@ad...) joined = #gstreamer. [19:07] <richardb> Hi, who's awake? [19:08] <dobey> you [19:08] <richardb> Just about [19:08] <richardb> trying to work out why esdsink doesn't like to be auto= plugged. [19:09] <richardb> think its because it specifies its possible values of = the channels parameter as a list (1,2), which isn't compatible with a sou= rce which specifies them as a range (1-2) [19:11] Apoc (Ap...@pp...) joined #gstreamer. [19:11] <Apoc> Hi all ... [19:11] <richardb> hi [19:13] <Apoc> Does someone know (well) buffer in gstreamer ?? [19:17] <steveb> Apoc: can you elaborate? [19:19] <Apoc> I try to make a libmikmod plugin ... [19:20] <Apoc> In this lib there is a wrapper to read data from memory or= file or something else ... [19:21] <Apoc> You just have to write function that make fread, ftell, ft= ell, fseek ... [19:22] <Apoc> My question is about fseek ... Buffer come from the disksr= c element ... But there are many piece of buffer right ?? [19:24] Nick change: wtay-zZz -> wtay [19:24] <wtay> yo [19:24] <wtay> Apoc: you *need* fseek? [19:24] <Apoc> yo wtay [19:26] <Apoc> Yeah .. I need to write a function that do the same job as= fseek but with buffer an not with physical file=20 [19:26] <wtay> Apoc: then I would suggest you use pullregion for now [19:27] <Apoc> wtay : How to use it? [19:28] <wtay> gst_pad_pullregion(sinkpad, GST_REGION_OFFSET_LEN, offset= , size); [19:30] <Apoc> Is this the offset of the current buffer or of the file if= we use a disksrc element ? [19:30] <richardb> wtay: got time to answer a couple of questions about t= he autoplugger and things? [19:31] <wtay> Apoc: offset in the source file [19:31] <wtay> richardb: sure [19:31] <wtay> Apoc: avidemux does a little bit of pullregion stuff, so d= oes quicktimedemux [19:32] <richardb> wtay: firstly, I find that ./gstreamer-launch disksrc = location=3Dfile ! autoplugger ! esdsink fails, but ! osssink works. [19:32] <richardb> I was wondering why, and it seems to be a caps negotia= tion thing. [19:32] <richardb> esdsink give channels as a list containing 1 and 2 [19:33] <Apoc> wtay : this the offset from the begining ... [19:33] <richardb> osssink gives channels as an integer range from 1-2 [19:33] <wtay> yeah, they are not compatible [19:33] <wtay> Apoc: yes [19:33] <richardb> should these be compatible? [19:33] <wtay> I think so, yes [19:33] <wtay> an int range should be compatible with a list of those dis= crete values [19:34] <richardb> I was thinking the easiest way to make it so would be = to collapse lists to ranges in gst_props_new [19:34] <richardb> and remove duplicates, and merge overlapping ranges, t= oo. [19:34] <wtay> richardb: that's not always possible, but I can see your p= oint [19:34] <richardb> when isn't it possible? [19:34] <wtay> for a float region this obviously doesn't work [19:35] <Apoc> wtay : I'll try ... thanks [19:35] <wtay> but for ints it makes sense [19:35] <richardb> overlapping float regions could be merged, and values = inside an existing region could be removed. [19:35] <wtay> and it'll be faster too [19:35] <richardb> but more can be done for ints. [19:35] <richardb> yep. [19:35] <richardb> second part of question: [19:36] <richardb> if caps aren't compatible, is the autoplugger doomed, = or can it use negotiation to try and find particular values for caps whic= h are compatible. [19:36] <richardb> or is that just planned for future...? [19:36] <wtay> it's doomed [19:37] <wtay> we need a capsnego rewrite soon [19:37] <richardb> Change the protocol from plugins' point of view? [19:37] <richardb> or just a rewrite of the implementation. [19:37] <wtay> I thinkt eh implementation only [19:38] <richardb> good. :) [19:38] <wtay> we need an algo that can find compatible props for both el= ements [19:39] <richardb> It needs to take the two sets of caps, find their unio= n, and try and do negotiation between them if the union is not null. [19:39] <wtay> es [19:39] <wtay> yes [19:39] <richardb> rather than just checking that one contains the other. [19:40] <richardb> right. [19:40] <richardb> I'll try and work out a patch to gstprops.c to simplif= y lists, for now. :) [19:40] <wtay> cool [19:41] <richardb> icky problem. :) [19:41] <Apoc> bye all ;-) [19:41] <wtay> kinda [19:41] <wtay> Apoc: cya [19:42] <richardb> not sure whether to build up the list simplifying each= time, or to simplify afterwards... ho hum. [19:42] Apoc (Ap...@pp...) left irc:=20 [19:43] <wtay> currently you can even have different types inside a list = too [19:43] <wtay> you can create a list with ints, floats, ranges and string= s=20 [19:43] <wtay> all in the same list [19:43] <richardb> yeah: I was going to give up simplifying if list didn'= t contain just ints and int ranges, or floats and float ranges. [19:44] <richardb> lists can't contain lists, though, it seems. [19:44] <wtay> ok, that makes sense [19:44] <wtay> yeah, no lists inside lists :) [19:44] <richardb> that's a start. :) [19:44] <wtay> it'll help :) [19:45] <wtay> I wonder if we should even allow floats and ints at the sa= me time in the same list... [19:47] <richardb> it seems invalid. [19:47] <richardb> there's no real way to complain and return an error. [19:48] <richardb> Could define it as not allowed, and put an assert in. [19:49] <wtay> something like that [19:50] <wtay> g_warn and only keep a valid subset [19:50] <richardb> actually, that'll help lots. :) [19:51] <wtay> that's why I bring up the issue :-) [19:55] <richardb> is there a GST_WARN macro, or similar? [19:55] <richardb> Can't see one. [19:56] <wtay> there isn't :( [19:59] <wtay> I gotta take a bath now.. [19:59] Nick change: wtay -> wtay-away [20:31] Nick change: wtay-away -> wtay [21:20] Nick change: wtay -> wtay-taxes [21:22] ShrimpX_ (marius@Photon.pdx.netroedge.com) joined #gstreamer. [21:30] Nick change: wtay-taxes -> wtay [21:30] <wtay> yo [21:31] <taaz> 8 cothreads? [21:31] <taaz> gotta fix flac... [21:32] <richardb> making slow progress with gstprop simplification stuff. [21:32] <wtay> I tried... [21:32] <richardb> I need to have dynamically allocated GstPropsEntry ite= ms. [21:33] <richardb> No easy way to copy them at the moment, though, so I'm= writing a method to do so. :( [21:33] <richardb> with a big switch for all the different types... [21:33] <wtay> bleagh [21:33] <richardb> yup [21:34] <richardb> Perhaps we should make GstPropsEntry into a proper obj= ect. [21:34] <richardb> Then we could just give it a clone method. [21:34] <wtay> that would be too slow I think... [21:34] <richardb> I think I must be doing something wrong. [21:34] Action: richardb reviews the code [21:36] <richardb> Hmm. I don't need to copy after all. [21:36] Action: richardb phews [21:36] sienap (sy...@ip...) joined #gstreamer. [21:36] <wtay> heh [21:37] <wtay> hi sienap [21:37] <sienap> hej wim! [21:37] <sienap> any news ? [21:37] <wtay> I made a FLAC plugin... [21:38] <sienap> a FLAC ? [21:38] <wtay> flac.sourceforge.net [21:38] <sienap> is HEAD the gobject port ? [21:38] <wtay> not yet [21:38] <sienap> like in any change i'll be able to compile HEAD out of t= he box ? [21:38] <sienap> =A0he ok [21:38] <sienap> then i cvs check it out [21:39] <dobey> hrmm [21:40] <sienap> wtay sounds cool [21:41] <sienap> the FLAC thingy [21:42] omega_ (om...@om...) joined #gstreamer. [21:42] <steveb> yo [21:42] <richardb> evening [21:42] <sienap> oemga_! [21:43] <wtay> hello [21:43] <omega_> yo [21:44] <richardb> omega_: new chart plugin. [21:44] <wtay> bleagh, the flac encoder only works when lowering the nr o= f cothreads to 8 :(=20 [21:44] <wtay> and I can't find a reason in libFLAC... [21:44] <richardb> omega_: ugly and slow, though. :( [21:44] <omega_> new chart plugin? [21:45] <richardb> currently takes audio data and draws a chart of it. [21:45] <omega_> define 'chart' ? [21:45] <sienap> chart plugin ? [21:46] <richardb> chart =3D diagram displaying data [21:46] <richardb> I called it chart because it will draw various differe= nt types of diagram, selected by option. [21:46] <richardb> and I couldn't think of a better name. [21:46] <richardb> would you like me to rename it? [21:47] <richardb> my plan was to have one plugin which took abstract dat= a frames, and other plugins to generate those data frames from different = sources, or by different processes. [21:48] <taaz> wtay: what's the stack size when at only 8 cothreads? [21:48] <richardb> so, could have something which plugs into a level mete= r and records the last 200 levels, and then send that to the chart plugin= to get a chart of the last 200 volume levels [21:48] <wtay> taaz: 2MB/8 [21:49] <richardb> anyone think of a better name that "chart" for what I = just described? [21:49] <omega_> I'll have to look more at what it does before I can poss= ibly come up with another name [21:49] <richardb> <g> [21:55] thomas (th...@ad...) joined #gstreamer. [21:56] <thomas> hi [21:56] <wtay> yo [22:02] <ShrimpX_> what does it mean if 'last' tells you that there was a= user 'date' that logged into your machine, on terminal '}' although ther= e is no user 'date' on your system? [22:02] <omega_> mangled wtmp [22:02] <ShrimpX_> hum... [22:02] <omega_> wtay: we need to have a plan for progress on GOBJECT1, r= egarding the gtk/gnome dependencies [22:04] <wtay> omega_: right [22:05] <wtay> omega_: I think I will use gobject with the old gtk+ [22:05] <omega_> you can't, afaik [22:06] <wtay> we need a check for gtk2.0.. [22:06] <omega_> yes, and figure out how much of the gnome stuff we can r= emove from the test programs, and how to handle gnome2.0 for the editor a= nd player [22:06] <wtay> if glib2.0 && gtk2.0->build gui stuff [22:06] <wtay> if !glib2.0 && gtk1.2 -> build all [22:07] <wtay> omega_: I didn't manage to compile gnome2.0 yet [22:07] <omega_> how much of it do you need? [22:07] <omega_> just libgnome and friends? [22:07] <wtay> I think so [22:07] <omega_> I'll give it a try from HEAD [22:09] <wtay> richardb: heh, a little leak in chart I think... [22:09] <richardb> quite probably... [22:09] <richardb> you mean *big*, don't you. [22:10] <wtay> pretty big...=20 [22:10] <wtay> and I know where [22:11] <wtay> line 382, where you ref the output buffer [22:11] <richardb> Ah. That happens automatically, does it? [22:12] <wtay> yup [22:12] <richardb> I'll fix in a minute then. [22:13] <richardb> done [22:13] <wtay> lukily we both have 16 bit video :) [22:14] <richardb> Is there an element which can convert between depths y= et? [22:14] <wtay> colorspace can sorta do that [22:16] <wtay> on a good day... [22:18] <richardb> Hmm. I think it should be possible to modify the buil= d system so that it does a non-recursive make, with automake 1.5 [22:19] <richardb> That would mean that it build dependencies would work = better: I just did make in tools/ having modified libgst.la, but it doesn= 't relink. :( [22:19] Action: wtay tries to install quicktime4 with wine... [22:19] <sienap> ___________ [22:19] <sienap> [root@Oegaah svgasink]# ldconfig -v |grep vga [22:19] <sienap> libvga.so.1 -> libvga.so.1.4.1 [22:19] <sienap> [root@Oegaah svgasink]# gcc -o vgatest vgatest.c -lvga [22:19] <sienap> collect2: ld returned 1 exit status [22:19] <sienap> /usr/bin/ld: cannot find -lvga [22:19] <sienap> that is [22:19] <sienap> :) [22:19] <sienap> _____ [22:19] <sienap> hmmm someone any clue ? [22:20] <dobey> you don't have libvga.so [22:20] <sienap> hmm maybe making a link yeah [22:20] <sienap> good suggestion [22:20] <sienap> thanks dobey! [22:26] <omega_> wtay, richardb: I have ORBit2/configure.in, it doesn't l= ike my automake 1.4p3 afaict [22:28] <richardb> a patched automake 1.4p3? [22:28] <omega_> yes [22:28] <omega_> ./configure: line 579: syntax error near unexpected toke= n `AM_CONFIG_HEADER(config.h)' [22:28] <omega_> ./configure: line 579: `AM_CONFIG_HEADER(config.h)' [22:28] <richardb> does it work with an unpatched one? [22:28] <richardb> that's an autoconf thing, I think. [22:29] <omega_> I haven't changed autoconf from rh7.0 std [22:29] <richardb> no, sorry. [22:29] <richardb> Um. [22:29] <richardb> I'll try with CVS automake... [22:29] <richardb> where does ORBit2 come from? [22:30] <wtay> wine segfaults on quicktime.exe.. what a surprise... [22:30] <omega_> this is ORBit2 HEAD from gnome cvs [22:30] <omega_> wait, it works now [22:30] <omega_> it's very fragile, I didn't have a dependency installed [22:30] <omega_> (linc, whatever that is) [22:30] <richardb> ah. (no idea) [22:31] <omega_> something about making networked servers easier [22:31] <richardb> How much has the configure.base in GOBJECT been change= d? [22:31] Action: omega_ needs to hook up wtih gleblanc and work on buildin= g snapshots of gnome 2.0 [22:31] <omega_> richardb: almost not at all [22:31] <richardb> I want to go through it and modify the way it does LIB= S and CFLAGS (to use the AM_LIBS and AM_CFLAGS) at some point. [22:31] <omega_> see viewcvs [22:32] <richardb> Good. [22:32] <richardb> Just a few pkg-config lines. [22:32] omega_omicron (om...@om...) joined #gstreamer. [22:32] omega_omicron (om...@om...) left irc: [x]chat [22:32] <omega_> grrrr, xchat annoys me [22:34] steveb (st...@no...) left irc: [x]chat [22:34] <taaz> richardb: if you want to further refine the ALSA check you= could try to figure out how to get the 0.9.0 check to not have double fl= ags (ie -lasound -lasound) [22:34] Action: taaz find it easy to pile work on other people ;) [22:35] Action: richardb sighs [22:36] <richardb> Anyone know if a g_list_free() invalidates other point= ers into the GList. [22:36] <richardb> I'm guessing it does. [22:36] <taaz> that fix I put in earlier today was nasty... i'm not sure = if the ALSA checks should be saving flags like that. Kinda sucked before= having =02every=02 lib link with -lasound [22:36] <richardb> s/g_list_free/g_list_remove/ [22:37] <wtay> richardb: it only frees the chain [22:37] <wtay> richardb: doesn't touch the data pointer [22:37] <richardb> Sorry. I'll rephrase [22:37] <taaz> what's a good way to test if lame plugin is working? [22:37] <richardb> If I remove an entry from a glist, do glist pointers i= nto the same list (at different locations) become invalid? [22:38] <wtay> taaz: disksrc location=3D... ! parsewav ! lame ! disksink = location=3Dbar.mp3 [22:38] <wtay> richardb: I would assume only the onew pointing to the rem= oved link change [22:40] <wtay> richardb: but you should not rely on that [22:40] <richardb> hmm [22:40] <richardb> yep, I think I won't. [22:43] greg_ (gr...@ho...) joined #gstreamer. [22:44] <richardb> whee. I can now remove duplicate integer properties. [22:44] <richardb> just a simple step to do merging of lists and things. [22:45] <wtay> neat [22:46] <greg_> hi ! just a shot: have you seen this ? http://linuxtoday.= com/news_story.php3?ltsn=3D2001-06-18-005-20-NW-CY-GN&tbovrmode=3D1 [22:47] <sienap> looking [22:51] <sienap> my god [22:51] <sienap> wtf [22:52] <wtay> gnome is dead!! that's why it didn't compile! :-) [22:53] Parapraxis (Sp...@us...) joined #gstre= amer. [22:54] <dobey> meh [22:54] <wtay> yo [22:54] <greg_> wtay: ;-)))))) I think they want to catch up at what kde = offers to user as quickly as possible, so they cut off the things that st= ops them. From UI point of view bonobo may not be "the best thing since s= liced bread". [22:54] <greg_> From the User of UI point of view of course... not a prog= rammer. [22:54] <dobey> bonobo is a component system, it can be used without a ui [22:54] <dobey> or with any [22:56] Action: richardb begins the process of copyright assignment for a= n automake patch [22:56] <greg_> it interesting if people from KDE will to use it too... a= nd *when* (what release?) [22:56] <sienap> parapraxis! [22:56] <sienap> damn martin bualig gone ? [22:56] <sienap> :((( [22:56] <omega_> sienap: sorta [22:57] Action: omega_ discussing gnome2 build insanities in #gnome [22:57] <dobey> oh [22:57] <dobey> have fun [22:58] <taaz> richardb: is that the thing that says you $1? I never got= mine ;) (of course I'd donate it back... but they never asked...) [22:58] <taaz> s/you/you get/ [22:59] greg_ (gr...@ho...) left irc: [x]chat [22:59] <richardb> taaz: didn't hear anything about $1, but so would I. ;= -) [23:00] greg_ (gr...@ho...) joined #gstreamer. [23:07] <taaz> show of hands: who thinks a releases/current symlink to la= test dir is a good plan? [23:08] <omega_> um, real files go where? [23:08] <taaz> so I can use releases/current/debian for apt [23:08] <taaz> hmm? current -> 0.2.0 for now [23:09] <omega_> ah, ok, that sounds good [23:09] <omega_> so releases/0.2.0/debian has the debs [23:10] <taaz> yes [23:11] <taaz> well, not just yet... give me 5 mins ;) [23:11] <taaz> you do know the web site is a mess ;) [23:12] <taaz> what's with the 'new' dir? [23:15] <taaz> I was also thinking of breaking off an -x package too. fo= r xvideosink. so people who use sdl or framebuffer or whatever wont have= to grab that. i'm seriously overthinking this whole thing ;) [23:17] greg_ (gr...@ho...) left irc: Ping timeout for greg_[home.= sente.pl] [23:17] <richardb> whee: props simplification works. [23:17] <omega_> richardb:=20 [23:17] <omega_> ? [23:18] <richardb> omega_: I've been doing a little work on the props stu= ff: making it so that a list of integer entries gets merged into a range,= if possible. [23:18] <richardb> This stops some places where caps are compatible from = being considered as incompatible. [23:18] <omega_> ah [23:19] <richardb> It's a chunky patch to gstprops.c [23:20] <richardb> is there any problem with committing to that, at the m= oment? [23:20] <richardb> I ask since I don't usually do core hacking... [23:20] <omega_> shouldn't be a problem as long as you're sure it works a= nd/or have tested it [23:20] greg_ (gr...@ho...) joined #gstreamer. [23:20] <richardb> Will commit in a bit then. Have only begun tests so f= ar... [23:25] sienap (sy...@ip...) left irc: My damn controlli= ng terminal disappeared! [23:29] <taaz> doh... just found bug in all the plugin debs [23:29] <taaz> minor.... will fix [23:32] ChiefHighwater (pa...@su...) joined #gstre= amer. [23:37] <wtay> yo [23:38] <ChiefHighwater> ello [23:42] <Bob> any thought of changing the name of gstreamer ?=20 [23:42] <Bob> gstreamer incites a lot of hatred in some people because it= infers that its a Gnome application because the G [23:43] <dobey> blah blah [23:43] <dobey> let them hate [23:43] <dobey> :-) [23:43] <Bob> just a thought, i couldn't care really but thats the kind o= f shit that people will vent=20 [23:43] <richardb> Havn't noticed that happening lately. Think most of t= he relevant people have gone away or got it out of their systems by now. [23:43] <taaz> which people? [23:43] <Bob> KDE people [23:44] <taaz> tell G is for GNU [23:44] <ChiefHighwater> Bob: haven't really seen that in here 8-] [23:44] <richardb> yeah, they've gone away or got it out of their systems= . ;-) [23:44] <Bob> if it doesn't have a K in the front, then it must be poison [23:44] <richardb> [gk]streamer [23:44] <richardb> hehe [23:44] <ChiefHighwater> richardb: hehe [23:44] <Bob> hehe [23:44] <omega_> best of both worlds: 'geek-streamer' [23:45] <richardb> yay [23:45] <ChiefHighwater> hehe [23:45] <richardb> will it do teleportation. [23:45] <Bob> personally I'm just sick of seeing G and K prepended to eve= ry application in the known universe :) [23:45] <richardb> ? [23:45] <taaz> GangstaStreamer [23:45] <dobey> Kstreamer [23:45] <dobey> so [23:45] <taaz> http://gstreamer.net/releases/current/ [23:45] <taaz> ok, debs are up. [23:45] <dobey> blah blah [23:46] <Bob> hehe [23:46] <taaz> someone test it for me ;) [23:46] Action: richardb edits sources.list [23:46] <dobey> i'm tired of hearing people bitch about it [23:46] <Bob> then there is the matter of how to pronounce GNU .. hehhe..= some people say gah-nu.. that sounds silly to me [23:47] <dobey> uhm [23:47] <dobey> it is guh-nu [23:47] <taaz> removeing plugins has a reg update bug... doing -register = in the wrong spot. will fix soon [23:47] Action: omega_ doesn't really care about names so much as the thi= ng actually working [23:47] <Bob> why 'guh' ..=20 [23:47] <dobey> that's how gah-nus are pronounced [23:47] <dobey> use the dictionary [23:47] <Bob> 'guh-nome' ... i prefer nome [23:47] <dobey> it's your friend [23:49] leviathan (le...@p3...) left #gstreamer. [23:49] <richardb> fireworks are going off here. College balls. :) [23:49] <Bob> mmmm , college girls , beer, .. ahhh :) [23:49] <dobey> sigh [23:49] <dobey> i hate college [23:50] ChiefHighwater (pa...@su...) left irc:=20 [23:50] <wtay> apt-get install gstmediaplay [23:50] <wtay> Reading Package Lists... Done [23:50] <wtay> Building Dependency Tree... Done [23:50] <wtay> The following extra packages will be installed: [23:50] <wtay> gstreamer-colorspace gstreamer-common gstreamer-elements= gstreamer-gnomevfs gstreamer-oss gstreamer-runtime libgst0 [23:50] <wtay> libgstmediaplay0=20 [23:50] <wtay> whoa! [23:50] <dobey> eww [23:51] <Bob> is that the process .deb takes infecting the system ? [23:51] <taaz> wtay: yup.. those are all the statically defined elements = in the gstplay.c [23:52] <wtay> cool [23:53] <wtay> whoa! mp3 plays [23:53] <wtay> apt-get install gstreamer-all-plugins [23:53] <wtay> Reading Package Lists... Done [23:53] <wtay> Building Dependency Tree... Done [23:53] <wtay> The following extra packages will be installed: [23:53] <wtay> gstreamer-aa gstreamer-alsa gstreamer-arts gstreamer-art= sd gstreamer-audiofile gstreamer-cdparanoia gstreamer-esd [23:53] <wtay> gstreamer-gsm gstreamer-jpeg gstreamer-lame gstreamer-mp= eg2dec gstreamer-mpg123 gstreamer-sdl gstreamer-vorbis=20 [23:54] <dobey> apt-get remove my-life [23:54] thomas (th...@ad...) left irc: Read err= or to thomas[adsl-63515.turboline.skynet.be]: EOF from client [23:54] <wtay> yay, got mpeg1 [23:56] <wtay> I like it, good job :) [23:56] <taaz> the mpeg2dec and lame plugins are kinda screwy since they = are broken unless you have non-deb stuff installed. [23:57] <taaz> I'll do the mpeg2dec packaging real-soon-now. maybe will = fixup the lame stuff so it builds shared libs into the debs and put those= up too [23:57] <Parapraxis> wtay: is gst making it into debian unstable soon, or= you just working w/ debs? [23:58] <Parapraxis> (ie for some other reason) [23:59] <wtay> Parapraxis: we have .debs now! [23:59] <Parapraxis> ... [00:00] --- Tue Jun 19 2001 [00:00] <wtay> Parapraxis: ? [00:00] Action: wtay jumps around the table [00:00] <taaz> Parapraxis: yeah, it'll probably be in unstable after I fi= x a few things. md...@de... offered to sponsor the packages untill I = become a real developer. [00:01] <taaz> If anyone wants to help figure out why lintian doesnt like= almost all the plugins due to a PIC problem, let me know ;) [00:01] <Parapraxis> taaz: yeah, I remebered that, but wasn't sure if it = was happening soon or just in a month or two [00:01] <Parapraxis> btw... where do you guys want gstparanormal to live?= (ie what source tree)? [00:02] <Parapraxis> I'm probably going to make a (extremely) primitive v= ersion in a couple days [00:02] <wtay> Parapraxis: I would suggest to put into gst-cvs for now... [00:02] <Parapraxis> ok [00:03] <richardb> taaz: debs work, apart from one problem... [00:03] <wtay> Parapraxis: is there something like a libparanormal we can= link against or do we need to put the complete paranormal source into cv= s? [00:03] <richardb> error loading plugin /usr/lib/gst/libgst_arts.so, reas= on: /usr/lib/gst/libgst_arts.so: undefined symbol: __tiQ23Gst19ArtsStereo= Sink_skel [00:03] <Parapraxis> wtay: paranormal is now a library [00:04] <richardb> when doing gstreamer-register [00:04] <wtay> hmm, yeah, here too [00:04] <wtay> Parapraxis: cool [00:05] <Parapraxis> wtay: It probably won't be ready for end-users any t= ime soon though (and shouldn't be put in a release or anything) [00:06] <Parapraxis> wtay: If that's a problem I can use the pn cvs for n= ow [00:06] <wtay> taaz: where's gstreamer-inspect? [00:06] <wtay> Parapraxis: no prob for me [00:06] <Parapraxis> ok, cool [00:06] <Parapraxis> wasm [00:07] <taaz> richardb: ok, you caught me... i didn't even try arts ;) [00:07] <Parapraxis> wasn't sure if you wanted 100% guaranteed non-functi= oning plugins :) [00:07] <wtay> richardb: it's a C++ issue [00:07] <wtay> Parapraxis: we have plenty of them :-) [00:07] <taaz> wtay: in gstreamer-tools [00:08] <wtay> cool [00:08] Action: Parapraxis wonders if gstreamer will have the most debs o= f any project in existence [00:08] <richardb> taaz: I think gstreamer-register doesn't really need a= package all of its own. [00:08] <taaz> Parapraxis: webmin is getting up there [00:09] <Parapraxis> taaz: yeah, but gst is only version 0.2 ;) [00:09] dobey (dobey@141.154.95.104) left irc: home [00:09] <taaz> richardb: where should it go then? [00:10] <richardb> taaz: well, it's completely essential, so I'd suggest = libgst0 [00:10] <taaz> richardb: i did that because the manual suggested to use a= -runtime package for support binaries that a lib package needs [00:10] <richardb> Ah. [00:10] <richardb> Fair enough then. [00:10] <taaz> yeah, it does seem rather silly in this case [00:10] <richardb> gstreamer-compprep should go in same package as gstrea= mer-register. [00:11] <wtay> I got to sleep now :( [00:11] <wtay> cya [00:11] Nick change: wtay -> wtay-zZz [00:11] <richardb> and be called by all the plugins whenever gstreamer-re= gister is [00:11] <richardb> wtay-zZz: night [00:11] <taaz> hmm... probably. suppose it needs to be re-run each plugi= n install/remove [00:12] <richardb> yep. we need to make gstreamer-register, etc, quieter= too. [00:12] <omega_> taaz: I would like to find a way to do incremental regis= try updates [00:12] <richardb> but, other than that, great job. :) [00:12] <omega_> it shouldn't be hard [00:13] <richardb> Speaking of which, I'm wondering if the check that the= registry is up to date is such a good idea. [00:13] <richardb> With a decently packaged system, the registry will alw= ays be up-to-date. [00:14] <richardb> Perhaps it should be a developer only option, instead. [00:14] <omega_> how do we know that for certain? [00:14] <omega_> yeah, possibly [00:14] <taaz> i mentioned that -register INFO output to omega_... i have= n't looked to see what causes it yet though [00:14] <richardb> if the registry isn't up to date, it's a packaging err= or. [00:14] <richardb> It's the same situation as ldconfig, I think. [00:14] <omega_> I suppose, yeah [00:15] <taaz> richardb: it's going to get more complicated when users ca= n add plugins to a GST_PLUGIN_PATH or whatever [00:15] <richardb> taaz: yeah, true. [00:15] <taaz> could assume system regs are up-to-date but users ones may= not be [00:15] <richardb> you typed it quicker than me. [00:15] <taaz> but in the user case, it could just be updated automagical= ly [00:15] <omega_> right, need to have that separation [00:15] <richardb> :) [00:16] <richardb> system ones are updated by packaging system, basically= , so we rely on that. [00:16] <richardb> taaz: gstreamer-register.c:199 [00:16] <richardb> gst_info_enable_category(GST_CAT_PLUGIN_LOADING); [00:17] <omega_> richardb: that can get turned off right away with --gst-= mask=3D0 [00:17] <richardb> that's what causes the output, I think. [00:17] <omega_> since _init is where that happens, which is *after* that= line [00:17] <richardb> Oh, yeah. [00:17] <richardb> Ignore me. :) [00:18] <omega_> but I still get the first two messages, which is wrong [00:18] Action: taaz sighs [00:19] <taaz> its alot of effort to keep that deb diff small [00:19] <taaz> if I just rebuild twice from the same dir the diff is like= 3M [00:19] <omega_> huh? [00:19] <taaz> but from newly untarred source its really small [00:19] <taaz> make clean doesn't actually clean jack [00:19] <omega_> it must ignore .cvsignore [00:20] <omega_> if the diff isn't ignoring files in .cvsignore, it'll ge= t Makefile and Makefile.in and all that [00:20] <wtay-zZz> make distclean? [00:20] <taaz> or maybe i'm doing it wrong... i'm not sure how to return = a tree to tarball size [00:20] <taaz> hey.. you are asleep [00:20] <omega_> make distclean will remove the majority of that stuff, i= ncluding configure.sh [00:20] <richardb> Hmm, autoplug still doesn't work with esdsink. :( [00:20] <omega_> taaz: he talks in his sleep sometimes <g> [00:20] <taaz> distclean removes too much i think... [00:20] <richardb> But my patch to gstprops.c does help, and doesn't seem= to break anything. [00:20] <richardb> committing. [00:21] <taaz> but then there is the doc problem. rebuilding those chang= es a whole load of files [00:21] <taaz> err... building... I think there is too much stuff in the= tarball and make clean should be removing alot of it but isnt [00:21] <omega_> taaz: get a faster machine <g> [00:22] <taaz> like the pdf, ps, html versions of the docs... why are tho= se in the 0.2.0 tarball? [00:22] <omega_> dunn [00:23] <omega_> er, dunno [00:23] <richardb> not sure either. I thought they weren't going to be. [00:23] <richardb> rules aren't quite right for make dist, perhaps. [00:24] <taaz> oh yeah, now I remember the problem ;) due to the fact th= at those docs are in the tarball and not cleaned i have a unfixable error= . after building the source it rebuilds those docs, some of which are bi= nary. the debian tools =02fail=02 to build things when there are binary = files that differ. I'm not sure how to fix that. [00:24] <taaz> it says "unrepresentable differences in source tree" or so= mething [00:25] <omega_> distclean will get you as close to clean HEAD as possibl= e [00:26] <taaz> i think distclean doesnt work too well in docs dir. still= have generated stuff left (i think) [00:26] <omega_> ok, then that needs fixing [00:26] <taaz> maybe for 0.2.1 we should try to fix this just to make me = happy ;) [00:26] <taaz> those doc makefiles are rather hairy though [00:34] Nick change: ajzzzz -> ajmitch [00:38] <Parapraxis> what does pcm stand for? [00:39] <taaz> pulse coded modulation [00:40] <Parapraxis> ok [00:40] <taaz> err.. maybe just pulse code modulation [00:49] <omega_> to solve the -register noise problem, we can do a couple= things [00:50] <omega_> first question I have is: what happens if you call gst_i= nit twice? [00:50] <omega_> I can't see any code that would cause it to only run onc= e... [00:50] <omega_> meaning, no static gboolean _gst_initialized =3D FALSE a= nywhere [00:51] <omega_> but I guess that can be worked around for this particula= r issue [00:51] <richardb> such code should probably be added [00:51] <omega_> yup [00:52] <omega_> I'm not convinced the structure of the _init stuff is an= ywhere near correct for that [00:52] <richardb> need to be aware that init might get called from a dif= ferent thread before it finishes. [00:52] <omega_> does gtk_init assume that could happen? [00:52] <omega_> if it doesn't, I wouldn't have gst_init do it either [00:52] <richardb> ENOCLUE [00:53] <omega_> though we do know that we're very non-thread-safe for mo= st of the global structures [00:53] <omega_> like plugin lists, etc. [00:56] gtb (gtb@Mix-Lyon-106-2-210.abo.wanadoo.fr) joined #gstreamer. [00:57] <richardb> hmm. when autoplugging, esdsink's newcaps function is = being called with empty caps [00:57] <richardb> I'm sure that's not meant to happen. [00:57] <gtb> (hi!) [00:57] <richardb> hi [00:58] <gtb> is there any up to date documumentation about autoplugging = on the web ? [01:00] <gtb> because the application example in the application developm= ent manual is marked "out of date" [01:01] <omega_> for simple stuff you just stick an 'autoplugger' element= in the middle [01:01] <omega_> we're working on updating that element to handle more co= mplex cases [01:02] <gtb> i create the autoplugger exactly as another element ? [01:03] <omega_> yup [01:03] <gtb> ok [01:03] <omega_> but it only works for a->b transforms, not a->b,c,d stuf= f like audio/video mpegs [01:03] <ShrimpX_> the autoplugger rocks [01:03] <gtb> I'll try this ! [01:04] <gtb> does it work with the 0.2.0 version of gstreamer ? [01:05] <omega_> should, yes [01:05] <gtb> ok, thx ! [01:10] <gtb> ok, it works ! [01:12] <taaz> any idea what to do about this arts symbol bug? [01:12] <omega_> not atm, no [01:17] <taaz> hmm... where are those skel functions supposed to be? [01:18] <omega_> MCOP stuff [01:19] <taaz> uh... so is "MCOP stuff" broken or is this our bug? ;) [01:20] <omega_> dunno [01:24] <taaz> heh... i'm going to take a wild guess and say that no one = at all is uses the arts plugin [01:24] <omega_> that would be correct [01:24] <taaz> the makefile doesnt even build the gst_artsio.cc file [01:24] <taaz> so of course it isnt linked in ;) [01:24] <omega_> oops [01:25] <taaz> i'm not sure how to fix that... its a generated file so it= shouldnt be in _SOURCES but it being in _DEPENDENCIES just makes sure it= s generated not actually built [01:25] <taaz> richardb, mr auto* master, how shall we fix this? [01:27] Action: richardb is distracted [01:29] <richardb> back. [01:31] <richardb> ah, right. [01:31] <taaz> maybe using BUILT_SOURCES? not sure if those are excluded= from dist or not [01:31] <richardb> generated files shoudln't be in sources; the source of= them should be. [01:31] <richardb> you can sometimes get that to work nicely., [01:32] <richardb> is gst_artsio.cc desired in the dist? [01:32] <taaz> no [01:32] <richardb> what's the problem at the monet? [01:33] <richardb> gst_artsio.cc isn't in the _SOURCES. [01:33] <richardb> gst_artsio.idl is, instead. [01:34] <taaz> problem is gst_arsio.cc is generated right now but not bui= lt and linked in. but we dont want that in the dist [01:34] <richardb> It's linked in for me. I think. [01:34] <taaz> its not even built here [01:34] <richardb> maybe it's because I'm using a newer automake. [01:35] <taaz> for 0.2.0 at least... [01:35] <richardb> could you do grep _la_OBJECTS Makefile.in in the arts = dir [01:35] <richardb> ? [01:36] <richardb> better: grep _la_OBJECTS\ =3D Makefile.in [01:37] <taaz> libgst_arts_la_OBJECTS =3D gst_arts.lo gst_artsio.cc gst_= artsio_impl.lo [01:37] <richardb> am_libgst_arts_la_OBJECTS =3D gst_arts.lo gst_artsio.l= o gst_artsio_impl.lo [01:37] <richardb> libgst_arts_la_OBJECTS =3D $(am_libgst_arts_la_OBJECTS= ) [01:37] <richardb> Yeah, problem which has been fixed. [01:38] <richardb> Which automake do you have? [01:38] <taaz> ii automake 1.4-p4-1 A tool for generating GN= U Standards-complian [01:38] mwc (ma...@ly...) joined #gstreamer. [01:38] <taaz> latest unstable deb [01:39] gtb (gtb@Mix-Lyon-106-2-210.abo.wanadoo.fr) left irc: using =02si= rc=02 version 2.211+=034KSIRC=03/1.0 [01:40] <taaz> what to do until 1.5 deb comes out... hmm [01:41] <richardb> A later 1.4-p release may fix it. [01:41] <richardb> Don't bother with CVS automake unless you want to manu= ally apply a patch I have. [01:41] <richardb> And a few other hassles too. [01:41] <richardb> 1.4-p4 is latest. [01:42] <taaz> so, is there a mutation of that Makefile.am that would fix= it for now? [01:42] <taaz> ie, a hack ;) [01:43] <richardb> I'm wondering... [01:44] <richardb> libgst_arts_la_SOURCES =3D gst_arts.c gst_artsio.cc gs= t_artsio_impl.cc [01:45] <richardb> and put a dist-hook target in which does rm -f ${build= _dir}/gstartsio.cc [01:45] <richardb> Might work. [01:53] ajmitch (aj...@p3...) left irc: Ping timeout = for ajmitch[p36-max1.dun.ihug.co.nz] [01:54] ajmitch (aj...@p3...) joined #gstreamer. [01:56] greg_ (gr...@ho...) left irc: Read error to greg_[home.sen= te.pl]: EOF from client [02:13] <mwc> Hi. I am trying to compile from cvs. I get gstwindec.cc:27= :22: creators.h: No such file or directory [02:14] <omega_> hmmm, not surprising. the wincodec stuff is not maintai= ned very well, because no one cares about it too much [02:15] Nick change: omega_ -> omega_dinner [02:17] <mwc> worked a few days ago. Strange. I wonder where creators.h= went. [02:45] Nick change: omega_dinner -> omega_ [02:53] <taaz> richardb: did that dist-hook thing... appears to work [02:53] <richardb> Cool. :) [02:54] <taaz> i also added CLEANFILES =3D gst_artsio.... to clean up th= e mcop mess [02:54] <taaz> i suppose that's the right way to do it? [02:54] <richardb> Ugly though - make sure you put a comment saying what = the cleaner solution is, but that it only works with automake 1.4f or gre= ater [02:54] <richardb> yes, probably. [02:54] <richardb> Havn't actually compiled it, so I'm not sure what the = mess is... ;-) [02:55] <taaz> idl -> h, cc, mcoptype, mcopclass [02:56] <taaz> Extension Expansion Technology i bet the call it [02:56] <richardb> ugh. CLEANFILES should do it, though, yes. [02:57] <taaz> its rather confusing having numerous *CLEAN* rules that yo= u just have to guess when to use them [02:57] <taaz> err.. s/rules/vars/ [03:01] ajmitch (aj...@p3...) left irc: Ping timeout = for ajmitch[p36-max1.dun.ihug.co.nz] [03:05] ShrimpX_ (marius@Photon.pdx.netroedge.com) left irc: ircII EPIC4-= 1.0 -- Are we there yet? [03:08] rowenc (rowenc@Roaming-209-53.PLU.edu) joined #gstreamer. [03:17] ajmitch (aj...@p4...) joined #gstreamer. [03:43] <taaz> omega_: you around? [03:43] <omega_> yeah [03:43] <taaz> question about web pages... [03:43] <omega_> uh oh [03:43] Nick change: ShrimpXawy -> ShrimpX [03:44] <ShrimpX> hehe my signoff msg from work is still on the screen [03:44] <taaz> the downloads.shtml page has RPMs and DEBs headers that ar= e broken...=20 [03:44] <taaz> but i'm not sure how this should be organized [03:44] <omega_> yes, I didn't finish that [03:44] <omega_> I was gonna put a table in each of rpm and deb [03:44] <taaz> seems to be some duplication between downloads.hstml and t= he releases/index... page [03:44] <omega_> with the two columns being the gst packages and the pack= ages that those depend on [03:44] <omega_> taaz: yes [03:44] <omega_> if you can think of a better way to do it, please, have = at it [03:45] <taaz> and i'm not sure if/how the current symlink should be used= instead of hardlinks to specific releases [03:46] <taaz> any reason the download.html file is there? [03:46] <omega_> nope [03:47] <taaz> what is the 'new' dir for? [03:47] <omega_> that's I think where we were constructing a new website = that would go in CVS [03:49] <taaz> too much effort to duplicate info between download.shtml a= nd rel/x.y.z/index.shtml... should the dl page just have list of release = info pages? [03:50] <omega_> I dunno, try something <g> [03:50] <taaz> sorry for asking so many questions... i just didn't want t= o trample on some master plan [03:50] <omega_> you think I have a master plan? [03:50] Action: omega_ falls and the floor and laughs his head off [03:50] <taaz> heh.. [03:50] <omega_> "Just Do It." [03:51] <taaz> does that include moving some of the new/ stuff to the unn= ew dirs? like all the releases in one spot instead of all over the place= ? [03:51] <omega_> it'd be best if we can move towards the new/ dir and act= ually get that in CVS, then publish it [03:52] <taaz> ha.. so you do have a master plan! [03:52] <omega_> um, not really [03:53] <omega_> at least, not as far as what actually goes *in* the webs= ite [03:53] <taaz> well, looks like richardb owns many new/ files... maybe he= has one [03:53] <omega_> however, you should talk with br...@ri... before = mangling things too much, because he is working on a redesign [03:53] <omega_> yes, he did a lot of that stuff [03:53] <taaz> i don't really care that much... just trying to organize t= hings [03:54] <omega_> ok, but if you're gonna organize things, try to coordina= te somewhat <g> [04:03] <richardb> I have no master plan. [04:11] <omega_> taaz, richardb: check with brock, see if he has any plan= s besides a visual rework [04:12] <taaz> not so fast amigo... i'm not going to become web site main= tainer ;) [04:12] <omega_> I didn't say that, just that you should check with brock= before you go changing too much stuff [04:13] <omega_> richardb: how do I convert a standalone .sgml to html ? [04:13] <taaz> so are the files in /etc/gstreamer supposed to be there? = seems more like /var info [04:13] <omega_> taaz: dunno what the emerging standards/thoughts are on = that [04:13] <richardb> taaz: that may well be a better place. [04:14] <omega_> I would guess /var/cache would make sense possibly [04:14] <omega_> richardb: help! <g> [04:14] <richardb> omega_: umm... [04:14] <taaz> sgml2html? [04:14] Action: taaz has that command [04:14] <omega_> ah, yes, that looks right [04:14] <omega_> doh [04:15] <omega_> except it fails [04:15] <richardb> I don't have that. [04:15] <richardb> Um, are these docbook sgml files [04:15] <omega_> it seems to completely miss the dtd/dsl [04:15] <omega_> right, I have a .sgml file in gnome cvs I need to read [04:16] <richardb> If you've got nautilus installed, that should do it. ;= -) [04:16] <omega_> erm [04:16] <richardb> There's a gnomedb2html2 script which does conversion t= oo. [04:16] <richardb> Or use jade, and a big tool chain. [04:16] Action: omega_ waits for nautilus to swap half his harddrive into= ram [04:17] <omega_> it loads it as text [04:17] <richardb> hmm. [04:17] <richardb> do you have the gnome-db2html2 script? [04:17] <omega_> I can read the raw sgml, but I have better things to do = <g> [04:17] <omega_> yes [04:17] <omega_> it didn't do anything [04:18] <richardb> I have a script which uses it to generate pages. [04:18] <richardb> how fast do you pick up emails? [04:19] <omega_> almost instantenous [04:19] <omega_> er, however you spell that [04:19] <richardb> script sent. [04:19] <omega_> got it [04:20] <richardb> No guarantees, but it works for me. [04:20] <omega_> um, it assumes that gnome-db2html2 actually does somethi= ng [04:20] <omega_> it has zero effect on my machine [04:20] <richardb> Oh. Hangon then. [04:20] <ShrimpX> what's the best way to start a skeleton for a coding pr= oject? (creating makefiles, ./configure, etc) [04:21] <omega_> um, copy from somewhere else <G> [04:21] <richardb> omega_: sent you a tarball which has no dependencies a= nd should build nicely [04:21] <ShrimpX> hehe [04:22] <omega_> n/c [04:22] <omega_> how am I supposed to invoke it>? [04:22] <richardb> ./gnome-db2html2 file.sgml [04:23] <richardb> gives you the front page [04:23] <omega_> ok, it has no effect [04:23] <richardb> uh? [04:23] <richardb> Odd. [04:23] <richardb> What doc is it? [04:23] <omega_> sending, what addr? [04:23] <richardb> ri...@ta... [04:23] <omega_> sent [04:24] <omega_> your script gives me a dir with a zero-byte index.html [04:24] <richardb> Give's me nothing. [04:24] <richardb> I have a plan though... [04:24] <omega_> hrm, I see a copy online [04:28] <richardb> No idea why those tools don't work. [04:33] <richardb> Got it to build using the proper build system though: [04:33] <richardb> put it in gst/docs/manual, and change manualname in Ma= kefile.am to writing-bindings. [04:36] <omega_> I'm not too worried about that specific case, I have wha= t I need now online [04:36] <richardb> bedtime for me now. [04:36] <omega_> ok, l8r [04:36] <richardb> gstreamer-launch doesn't support tee yet, does it? [04:36] <omega_> nope, no syntax for it yet [04:37] <richardb> Right. Can't see my scope at the same time as hearing= it, then. :( [04:37] <richardb> night. [04:37] Nick change: richardb -> richardb-away [04:43] ajmitch (aj...@p4...) left irc: Ping timeout= for ajmitch[p48-max11.dun.ihug.co.nz] [05:00] ChiefHighwater (pa...@su...) joined #gstre= amer. [05:00] <omega_> yo [05:01] <ChiefHighwater> Ello 8-] [05:08] mwc (ma...@ly...) left irc: Ping timeout for mwc[ly= chee.ntu.edu.au] [05:23] mwc (ma...@ly...) joined #gstreamer. [05:26] <Parapraxis> got my first real actuator working :) [05:26] <mwc> hey is anyone arround? [05:27] <omega_> maybe <g> [05:28] <Parapraxis> omega_: is it not possible right now to have a vis p= lugin and hear music w/ gst? [05:28] <mwc> I am having real problems getting to any international web = sites. I was wondering if it is a higher problem [05:28] <omega_> it's possible, just not with -launch [05:28] <omega_> mwc: such as? [05:28] <Parapraxis> ah [05:29] <mwc> sf, slashdot, cnn, etc.etc.etc [05:29] Action: omega_ just loaded /., but he's in the states [05:30] <Parapraxis> same here [05:30] <mwc> Yes thanks. I think either my provider or an upstream prov= ider has bgp problems. [05:43] mwc (ma...@ly...) left irc: Ping timeout for mwc[ly= chee.ntu.edu.au] [05:47] mwc (ma...@ly...) joined #gstreamer. [05:58] Nick change: ChiefHighwater -> CHW_away [06:07] mattsm (ma...@ad...) left ir= c: Client Exiting [06:17] <ShrimpX> is it more efficient to implement object functions as p= ointers to functions (struct members) or simply functions which take the = object as a parameter? [06:18] <omega_> straight functions [06:18] <omega_> a function pointer always incurs an extra couple cycles = overhead [06:18] <ShrimpX> hum [06:18] <omega_> because the function pointer still has to get the object= pointer on the stack anyway, you just don [06:18] <omega_> don't see it in C++ [06:19] <ShrimpX> aah... ok. [06:20] <ShrimpX> what about privacy? can you enforce provacy on object m= embers in c? [06:20] <taaz> the question you should be asking is what is the most effi= cient way for =02you=02 to write the code correctly. worry about cycles = later. blah blah... [06:20] <taaz> yeah, put a comment there "dont touch these" [06:20] <ShrimpX> haha [06:20] <taaz> seriously [06:21] <omega_> ShrimpX: you can do that either by hiding the header, or= , um,=20 [06:21] <omega_> hiding the header [06:21] <taaz> or have a pointer to a private struct that isnt defined in= the public headers [06:21] <omega_> at the expense of an indirect [06:22] <ShrimpX> does gstreamer use any privacy stuff? [06:22] <omega_> um, no [06:22] <omega_> best case would be to have the public members up front, = and a private version that has that struct at the front [06:22] <omega_> only give out the pointer and the public header, just ma= ke sure the pointer is big enough for the public+private [06:23] <omega_> but that's a lot of work, and policy is a lot simpler: "= don't touch!" [06:24] <ShrimpX> cool. [06:24] <taaz> can do stuff like _foo or foo_ to make it more clear to us= ers too [06:25] <omega_> yup [06:25] <omega_> whoah. #gnome just netsplit. that's entertaining, actu= ally [06:25] <ShrimpX> on this server? [06:25] <omega_> no, irc.gnome.org [06:26] <omega_> just lost 100 members of the channel [06:26] <ShrimpX> hehe |
From: <wim...@ch...> - 2001-06-20 04:30:21
|
[06:30] <ShrimpX> s/ne/net [06:56] omega_ (om...@om...) left irc: Read error to omega_[omegacs= .net]: Connection reset by peer [06:57] omega_ (om...@om...) joined #gstreamer. [06:57] omega_omicron (om...@om...) joined #gstreamer. [06:57] omega_omicron (om...@om...) left irc: [x]chat [06:59] ajmitch (aj...@p6...) joined #gstreamer. [07:22] Action: omega_ helps OctobrX get his laptop set up for wavelan in= 2.4, and firewire capture [07:28] big_T (th...@fy...) left irc: Ping timeout for big_= T[fysgr763.sn.umu.se] [07:29] big_T (th...@fy...) joined #gstreamer. [07:46] Nick change: CHW_away -> ChiefHighwater [07:50] johnix (thomas@61.152.140.57) joined #gstreamer. [07:56] steveb (st...@no...) joined #gstreamer. [08:07] <Parapraxis> anyone know why gdb might ctl-z itself? [08:08] <omega_> hrm? like suspend to shell ? [08:08] <Parapraxis> yeah [08:08] <omega_> never seen that [08:09] <Parapraxis> then when I resume it, it goes until I start my vis = plugin (by playing a song in xmms) and it segfaults, but gdb tells me the= wrong line (ie the stack says it's in a different function) [08:09] <Parapraxis> debugging with a broken debugger =3D not fun [08:09] <Parapraxis> omega_: do you work on gst 24 hours a day? [08:10] <omega_> Parapraxis: no, just 18 <g> [08:10] <ajmitch> heh [08:10] <Parapraxis> I would laugh if I thought you were kidding ;) [08:10] <omega_> doh [08:11] <ChiefHighwater> I would submit that omega_ even dreams code...so= 24, yes 8-] [08:12] <omega_> details [08:13] <Parapraxis> ugh... and I can't break the program gdb is running,= it just suspends gdb :( [08:14] <Parapraxis> hah... making the terminal window smaller fixed it [08:14] <omega_> uh? [08:15] <Parapraxis> now it hits a 'press enter to continue because the s= creen is full' when listing all the 'no debugging symbols found' and that= seems to make it not suspend [08:34] OctobrX (oc...@en...) joined #gstreamer. [08:34] <omega_> doh [08:34] <OctobrX> omega_: heh [08:34] Action: OctobrX ducks [08:34] <omega_> OctobrX: so you're a gstreamer convert now? [08:34] <OctobrX> ummm, (mic on... screeching)... sure! (holds up two pe= ace signs) [08:35] <omega_> "I am not a crook!" [08:35] <OctobrX> hehe [08:35] <omega_> erm, did it finish playing yet? [09:04] <ShrimpX> so, always pass structs by reference, right? [09:07] <omega_> pointer [09:08] <omega_> reference is a C++-ism [09:08] <omega_> only hides what's really going on, makes it hard to keep= track of [09:08] <omega_> IMO [09:09] <johnix> Hi guys [09:09] <ShrimpX> hehe [09:10] <ShrimpX> I'm beginning to think that I'm starting off on the wro= ng foot by learning c++ in school [09:10] <omega_> yeah, would agree. but you don't have much choice ;-( [09:11] <ShrimpX> I'm learning C at home though. OO and all. :) [09:11] <ShrimpX> the gstreamer code is a very good reference. [09:11] <ShrimpX> err.. pointer. [09:11] <ShrimpX> :) [09:11] <omega_> doh [09:14] ChiefHighwater (pa...@su...) left irc:=20 [09:15] Action: Parapraxis gives up [09:15] <Parapraxis> later y'al [09:15] Parapraxis (Sp...@us...) left irc: Cli= ent Exiting [09:22] Nick change: taaz -> taazzzz [09:43] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee02.a2000.nl] [09:43] steveb (st...@no...) joined #gstreamer. [09:50] <ShrimpX> what does a "conflicting types [09:50] <ShrimpX> comile error mean? [09:50] <omega_> what context? [09:50] <ShrimpX> can I paste? [09:50] <omega_> past the compiler error [09:50] <ShrimpX> obj.h:20: two or more data types in declaration of `_pe= rson_initialize' [09:50] <ShrimpX> obj.c:5: conflicting types for `_person_initialize' [09:50] <ShrimpX> obj.h:20: previous declaration of `_person_initialize' [09:51] <omega_> means that your prototype and implementation don't match [09:51] <ShrimpX> hum [09:51] <omega_> compare those two lines, there's some difference between= them [09:51] <omega_> like int vs long or something [09:51] <omega_> or mismatched pointer types [09:53] rowenc (rowenc@Roaming-209-53.PLU.edu) left irc: I am quitting [09:58] <ShrimpX> wow, that was weird. That error was because of a missed= ';' at the end of a struct. (I gotta make that a reflex) [09:58] <omega_> yup [09:58] <omega_> missing ';'s can have bizarre consequences [09:58] <omega_> just like missing \'s in multi-line macros [10:00] <ShrimpX> yea, I really gotta make sure I pay attention to those. [10:02] <omega_> wanna see something that's *=02really=02* easy to break? [10:02] <omega_> http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/codecs/li= bcodec/libcodec/mcomp/mcomp.c?rev=3D1.2&content-type=3Dtext/vnd.viewcvs-m= arkup [10:02] <steveb> is there a way to tell the length of some media without = actually running the pipeline? [10:03] <omega_> steveb: depends on the source plugin [10:03] <omega_> going to READY will cause the disksrc at least to open u= p and discover the length of a file [10:03] <omega_> but without pushing data there's no way to tell how 'lon= g' it is in a media-specific way [10:03] <omega_> ShrimpX: even worse, try parsing the above code with thi= s: http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/~checkout~/codecs/libco= dec/gen.pl?rev=3D1.4&content-type=3Dtext/plain [10:05] <steveb> does READY state start pushing data so elements can get = any header info? [10:06] omega_ (om...@om...) left irc: Ping timeout for omega_[omeg= acs.net] [10:06] omega_ (om...@om...) joined #gstreamer. [10:07] <ShrimpX> what is that? [10:07] <omega_> steveb: no [10:07] <omega_> ShrimpX: which? the first or the second? [10:07] <ShrimpX> the result [10:07] <omega_> um, lots more code <g. [10:09] <omega_> http://temple-baptist.com/~omega/mcomp.h is the result f= rom the mcomp directory [10:09] <ShrimpX> aagh! [10:10] <omega_> ShrimpX: you're at school??? [10:10] <ShrimpX> nope. why? [10:10] <omega_> 131.252.244.168 - ... [10:10] <ShrimpX> I live on campus. [10:10] <omega_> oh [10:10] <ShrimpX> cheap rent and dsl. :) works for me [10:10] <omega_> dsl? [10:10] <ShrimpX> yup [10:10] <omega_> I thought you were on aracnet.. [10:11] <omega_> or is that someone else? [10:11] <omega_> yeah, someone else [10:11] <ShrimpX> probably... not sure [10:11] <omega_> um, you should be on resnet which would be ethernet [10:11] <ShrimpX> yes [10:11] <omega_> 131.252.x.y is pdx's class B [10:11] <omega_> so not DSL [10:11] <omega_> psu is on Internet-2 [10:11] <ShrimpX> aah. [10:11] <omega_> has a 100Mbps connection to UW [10:12] <ShrimpX> really? [10:12] <omega_> via OC3 from psu campus to Pittock Block [10:12] <ShrimpX> nifty [10:12] <omega_> I've 'fondled' the machine that switches those packets [10:12] <ShrimpX> I know I'm on 100mbps to the psu network... [10:12] <ShrimpX> wow, cool. [10:12] <omega_> doubt you have direct i2 routing though [10:12] <omega_> I think the rest is a T3 to i1 [10:13] <omega_> but psu is the core of the original internet in pdx [10:13] <ShrimpX> really? [10:13] <ShrimpX> people keep telling me that they are buying service fro= m someone else... [10:13] <omega_> was involved way back with Alan Beattie was starting up = his stuff locally [10:14] thomas (th...@21...) joined #gstreamer. [10:14] <ShrimpX> you're not really 22, are you? :) [10:14] <omega_> almost 23 [10:14] <omega_> well, sorta [10:14] <ShrimpX> hehe [10:14] <omega_> 22.~7 [10:14] <ShrimpX> 22.78878588449 [10:14] <omega_> do a traceroute to www.washington.edu [10:15] <ShrimpX> k [10:15] <omega_> you might go through I2 routes [10:15] <ShrimpX> want me to paste it? [10:15] <omega_> sure [10:15] <johnix> what's i1 and i2? [10:15] <omega_> internet, internet-2 [10:15] <ShrimpX> 1 131.252.244.10 (131.252.244.10) 0.471 ms 0.322 ms= 0.363 ms [10:15] <ShrimpX> 2 garm.dmz.pdx.edu (131.252.3.12) 1.077 ms 1.353 ms= 0.868 ms [10:15] <ShrimpX> 3 ptld-car1-gw.nero.net (207.98.64.17) 1.255 ms 1.1= 94 ms 1.089 ms [10:15] <ShrimpX> 4 ptld-core2-gw.nero.net (207.98.64.177) 1.703 ms 1= .425 ms 1.328 ms [10:15] <ShrimpX> 5 eugn-core2-gw.nero.net (207.98.64.13) 4.665 ms 20= .532 ms 4.638 ms [10:15] <ShrimpX> 6 0-car.oregon-gigapop.net (207.98.64.22) 5.112 ms = 5.284 ms 4.852 ms [10:15] <ShrimpX> 7 pos-5-0.core0.eug.oregon-gigapop.net (198.32.163.25= ) 5.091 ms 4.913 ms 5.116 ms [10:15] <ShrimpX> 8 eug-snva.oregon-gigapop.net (198.32.163.10) 17.941= ms 17.408 ms 17.444 ms [10:15] <ShrimpX> 9 sttl-scrm.abilene.ucaid.edu (198.32.8.9) 39.919 ms= 41.534 ms 34.850 ms [10:15] <ShrimpX> 10 westincar2-S5-0-0.pnw-gigapop.net (198.48.91.77) 3= 5.395 ms 35.283 ms 34.935 ms [10:16] <ShrimpX> 11 uwbr1-GE0-0.cac.washington.edu (198.32.170.12) 34.= 988 ms 35.104 ms 34.964 ms [10:16] <ShrimpX> 12 zinc-V24.cac.washington.edu (140.142.154.10) 35.82= 6 ms 35.524 ms 35.702 ms [10:16] <ShrimpX> 13 www3.cac.washington.edu (140.142.3.35) 35.141 ms = 35.918 ms 34.922 ms [10:16] <omega_> yup, i2 [10:16] <omega_> see line 9 [10:16] <omega_> but not through the Pittock Block [10:16] <omega_> still routed through nero [10:16] <omega_> ;-( [10:17] <ShrimpX> heh [10:17] <omega_> or wait... [10:17] <omega_> there's a lot of politicking going between osu and psu [10:17] <johnix> There's technically no difference between i1 and i2 host= s, is there? [10:18] <omega_> johnix: no, just the routing clusters around them [10:18] <omega_> and the width of the pipes [10:18] <johnix> except for net bandwidth [10:18] <johnix> what are routing clusters? [10:18] <omega_> a term I just invented <g> [10:18] <ShrimpX> hehe [10:18] <omega_> i2 is its own backbond [10:18] <omega_> er, backbone [10:19] <omega_> and it peers with the rest of the internet via the unive= rsities it connects [10:19] <omega_> you're not gonna get routed onto the i2 unless you're go= ing to a university from a university, or a related research site [10:19] <johnix> i1 is too, isn't it? [10:19] <omega_> i1 is a whole stack of backbones [10:19] thomas (th...@21...) left irc: Ping ti= meout for thomas[212-100-172-175.adsl.easynet.be] [10:20] <omega_> ShrimpX: traceroute to anakin.cse.ogi.edu [10:20] <ShrimpX> k [10:20] <johnix> mmh... I don't really get it [10:20] <omega_> johnix: i2 is a separate adminstrative domain [10:20] <johnix> maybe I should take some "advanced routing" course... [10:20] <ShrimpX> 1 131.252.244.10 (131.252.244.10) 0.486 ms 0.405 ms= 0.267 ms [10:20] <ShrimpX> 2 garm.dmz.pdx.edu (131.252.3.12) 1.160 ms 1.179 ms= 0.854 ms [10:20] <ShrimpX> 3 ptld-car1-gw.nero.net (207.98.64.17) 1.708 ms 1.0= 58 ms 1.009 ms [10:20] <ShrimpX> 4 ptld-core2-gw.nero.net (207.98.64.177) 1.572 ms 1= .471 ms 1.386 ms [10:20] <ShrimpX> 5 eugn-core2-gw.nero.net (207.98.64.13) 4.532 ms 5.= 739 ms 4.433 ms [10:20] <ShrimpX> 6 0-car.oregon-gigapop.net (207.98.64.22) 5.164 ms = 4.660 ms 4.974 ms [10:20] <ShrimpX> 7 pos-5-0.core0.eug.oregon-gigapop.net (198.32.163.25= ) 37.704 ms 4.795 ms 4.933 ms [10:20] <ShrimpX> 8 eug-snva.oregon-gigapop.net (198.32.163.10) 17.767= ms 17.294 ms 17.383 ms [10:20] <ShrimpX> 9 sttl-scrm.abilene.ucaid.edu (198.32.8.9) 34.756 ms= 36.452 ms 35.336 ms [10:20] <ShrimpX> 10 westincar2-S5-0-0.pnw-gigapop.net (198.48.91.77) 3= 4.671 ms 42.059 ms 42.986 ms [10:21] <ShrimpX> 11 westincore1-ge-0-0-0-0.pnw-gigapop.net (198.32.170.= 104) 43.589 ms 36.798 ms 35.755 ms [10:21] <ShrimpX> 12 pdxcore1-so-0-1-0.pnw-gigapop.net (198.48.91.42) 3= 8.407 ms 40.042 ms 38.458 ms [10:21] <ShrimpX> 13 pren-pnw-GE0-0.pnw-gigapop.net (198.32.40.4) 38.59= 8 ms 39.149 ms 38.284 ms [10:21] <ShrimpX> 14 ogi2-gw.i2.ogi.edu (129.95.255.1) 39.400 ms 39.19= 7 ms 39.353 ms [10:21] <ShrimpX> 15 129.95.10.1 (129.95.10.1) 40.213 ms 40.297 ms 40= .219 ms [10:21] <ShrimpX> 16 129.95.10.54 (129.95.10.54) 40.175 ms 39.691 ms = 39.353 ms [10:21] <ShrimpX> 17 anakin.cse.ogi.edu (129.95.50.58) 39.079 ms 39.56= 5 ms 39.145 ms [10:21] <omega_> the routing is policed by the AUP, which states that onl= y non-commercial traffic can flow there [10:21] <ShrimpX> omega_, you took a routing course at psu, right? [10:21] <omega_> oh man, that sucks [10:21] <ShrimpX> why? [10:21] <omega_> ShrimpX: took it at OCATE, from Binkley, who's a PSU guy [10:21] <omega_> lines 3 through 12 should not happen [10:21] <ShrimpX> oops [10:22] <omega_> line 13 is the Cisco 12000 at the Pittock Block [10:22] thomas (th...@21...) joined #gstreamer. [10:22] <omega_> there's supposed to be a direct link from the basement o= f Fourth Avenue Building to Pittock Block, fiber straight from the dmz to= the pren router [10:22] <omega_> and from pren there's 100MBps to the uw gigapop [10:23] Action: omega_ should sync with Binkley on how this is all struct= ured these days [10:23] Action: ShrimpX shoudl too [10:23] <omega_> you know Binkley=20 [10:23] <omega_> ? [10:23] <ShrimpX> nope [10:24] <ShrimpX> but I should learn about routing [10:24] <ShrimpX> :) [10:24] <omega_> heh [10:25] <omega_> btw, the actual code involved in that mcomp.h file: http= ://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/codecs/libcodec/libcodec/mcomp= /average.c?rev=3D1.2&content-type=3Dtext/vnd.viewcvs-markup [10:25] <omega_> one of four [10:26] <omega_> or http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/codecs= /libcodec/libcodec/yuv/yuv420-rgb555-mmx.c?rev=3D1.3&content-type=3Dtext/= vnd.viewcvs-markup for something different <g> [10:28] <ShrimpX> that looks complicated [10:28] <omega_> actually it's not too bad.. [10:28] <omega_> but it took me a month to get to that point [10:29] <omega_> are you going to be taking any courses where you need to= write a parser any time soon? [10:29] <omega_> or know someone who is? [10:29] <omega_> cause I need to get http://cvs.sourceforge.net/cgi-bin/v= iewcvs.cgi/codecs/ewocc/ewocc.c?rev=3D1.1.1.1&content-type=3Dtext/vnd.vie= wcvs-markup rewritten correctly [10:29] <ShrimpX> I'll be taking one in the winter, I think [10:29] thomas (th...@21...) left irc: Ping ti= meout for thomas[212-100-172-175.adsl.easynet.be] [10:29] <omega_> I could pretty easily write the BNF for it, but I can't = get lex&yacc to cooperate [10:30] <ShrimpX> hum [10:30] <omega_> in fact, I think I have a start at the BNF for that some= where [10:30] <omega_> it parses http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi= /codecs/ewocc/mpeg2.parse?rev=3D1.2&content-type=3Dtext/vnd.viewcvs-marku= p [10:31] <ShrimpX> I don't really get how the persing works... Does it gen= erate C code out of that? [10:31] <omega_> that one will, yes [10:31] <omega_> ewocc =3D Erik Walthinsen's Optimized Codec Compiler [10:31] <ShrimpX> why do you need to write it in that pseudo code first? [10:32] <omega_> because massive optimizations can be made from that [10:32] <ShrimpX> aah... [10:32] <omega_> look at sequence_header [10:32] <omega_> instead of reading N bits at a time, read a large chunk = and do the bitmasking [10:32] <omega_> reading N bits takes 10+ cycles, masking can take 1 if y= ou do it right [10:32] <ShrimpX> wow [10:33] <omega_> hier2.c is the other half, optimized reading of vlc coef= fs [10:34] <omega_> that'll be rewritten to asm, as per mpeg1_dct_coeff.c [10:35] <ShrimpX> that's gcc asm, right? [10:35] <omega_> yup [10:36] <omega_> at some point soon I'm going to write the *entire* libdv= decoder in asm [10:36] <ShrimpX> nifty stuff... I wish I could wrap my mind around this = whole concept. C is easy to read, but the concept behind it is not as eas= y... [10:36] <omega_> do enough of it and it becomes second nature [10:37] <omega_> asm is starting to make considerable sense to me now, fi= nally [10:37] <ShrimpX> can you write usable asm by hand? [10:37] <omega_> you're lookin at it <g> [10:37] <ShrimpX> hehe. cool [10:37] <omega_> of course, that's about 50% of my total inline asm outpu= t ever [10:37] <omega_> but that'll change when I get to libdv [10:38] <ShrimpX> that'll be interesting [10:38] <omega_> quite [10:38] <omega_> the parsing for DV is very very nasty [10:38] <omega_> I need to go to OGI and spend a day having Buck explain = it to me [10:38] <ShrimpX> who's Buck? [10:38] <omega_> the other author of libdv [10:38] <omega_> a grad student at OGI [10:39] <omega_> and my ex-manager at a failed startup <g> [10:39] <ShrimpX> I noticed that the OGI people work on all kinds of cool= projects [10:39] <ShrimpX> heh [10:39] <omega_> see temple-baptist.com/~omega/ogimabot2/ [10:40] mwc (ma...@ly...) left irc: [x]chat [10:40] <ShrimpX> cool! [10:41] <omega_> a group of students should be (hopefully) building a new= and improved one this summer [10:41] <omega_> and I hope to help out, since I built that one, and have= done a lot of research on how to build a better one [10:41] <ShrimpX> were you a student there? [10:41] <omega_> nope, staff [10:41] <ShrimpX> programmer? [10:41] <omega_> yup [10:41] <omega_> and everything else... [10:42] <ShrimpX> what were yuor credentials when you were hired? [10:42] <omega_> um... [10:42] <omega_> worked at Sequent for 2.5years before that [10:42] <omega_> testing and test development for the clustered volume ma= nager [10:43] <ShrimpX> how old were you when you started there? (/me thinks th= at Omega was born when he was -10) [10:43] <omega_> last box I worked on before being laid off (randomly) wa= s http://www.omegacs.net/~omega/STiNG/ [10:43] <omega_> ShrimpX: was 17 [10:43] <ShrimpX> holy shite [10:43] <omega_> started during Christmas break my senior year [10:43] <ShrimpX> I was busy dropping out of high school and running away= from home at 17 :) [10:44] <omega_> doh [10:44] <omega_> the sqnt hardware is really fun to play with [10:44] <ShrimpX> I bet [10:44] <ShrimpX> I can't get to omegacs.net... [10:44] <omega_> that machine was set up to read 330MB/sec from disk [10:45] <omega_> www. ? [10:45] <omega_> works from another dsl node [10:45] <ShrimpX> heh [10:45] <johnix> omega_: random access? [10:45] <omega_> johnix: linear [10:45] <omega_> random could be higher, with that many disks [10:46] <johnix> yow! [10:47] <omega_> ShrimpX: remove the www. [10:47] <omega_> fixing... [10:47] <ShrimpX> I'm doing it by ip. It works [10:48] <ShrimpX> eehhh... geek porn [10:49] <omega_> just imagine =02using=02 it... ;-) [10:50] <ShrimpX> heh [10:50] <omega_> ok, just added www.omegacs.net, it should work in anothe= r 48hrs at most (DNS trickle-down) [10:50] <ShrimpX> cool [10:50] <omega_> recent models have up to 64 PIII Xeon 700 2MB's [10:51] <ShrimpX> those racks look like the ELI racks [10:51] <omega_> see sequent.com [10:51] <omega_> ELI =3D=3D electic lightwave ? [10:52] <ShrimpX> no, not sure... It's a space provider and ISP [10:52] <ShrimpX> they provide rack space and a T1 [10:52] <omega_> eli does fiber, and isp for businesses [10:52] <ShrimpX> yup [10:53] <omega_> you seen their racks where? [10:53] <omega_> FAB ? [10:53] <ShrimpX> they're hosting some of our stuff [10:53] <omega_> our=3D=3Dpsu ? [10:53] <ShrimpX> some HP servers that we maintain [10:53] <ShrimpX> Edge (my work) [10:53] <omega_> ah [10:54] <ShrimpX> www.netroedge.com (design agency) I do www stuff (perl/= database programming) [10:54] <omega_> ah [10:58] <ShrimpX> I gotta get some sleep. I gotta wake up in a couple of = hours... [10:58] <omega_> heh [10:58] <omega_> yeah, /me should too [10:58] <ShrimpX> omega_, thanks a lot for your help [10:58] <omega_> no prob [10:58] <ShrimpX> I'll bug you later. :) [10:58] <omega_> heh, cya [10:58] Nick change: ShrimpX -> ShrimpXzZz [10:59] Nick change: richardb-away -> richardb [10:59] <omega_> yo [10:59] <richardb> Morning. :) [10:59] Action: omega_ is having difficulties with the gobject2gtk port [11:00] <johnix> richardb: where are you from? [11:00] <johnix> to get up at 5 pm ? [11:00] <richardb> UK [11:00] <omega_> johnix: where are you for this to be 5pm??? [11:00] <richardb> I'ts 10am here. [11:01] <johnix> Shanghai, PR China [11:01] <omega_> it's 2am here [11:01] <johnix> omega_: and you're in America? [11:01] <richardb> omega_: could libcodec help with drawing waveform diag= rams? [11:01] <omega_> ShrimpX is smart, he's going to sleep already [11:01] <omega_> johnix: west coast [11:01] <omega_> richardb: possibly, but not yet [11:01] <omega_> no algos in it that are relevant atm [11:02] <richardb> omega_: right. bitstream stuff looks like it might be= , but havn't taken a close look. [11:02] <richardb> I'll wait then. [11:02] <omega_> why would bitstream be useful? you're taking pcm data, r= ight? [11:03] <richardb> probably wouldn't be useful. [11:03] Action: omega_ found the problem! yay! [11:05] <omega_> still one more at least [11:11] <steveb> omega_: do you think there should be an easy way for app= /elements to discover the length of a given media? [11:11] <steveb> without running the pipeline [11:11] <omega_> define 'length' [11:11] <steveb> number of samples/frames before eos [11:12] <omega_> that's a very format-specific thing, you must read some = data before the parser can figure that out [11:12] <steveb> when its easy to discover that is (file on disk with len= gth in header) [11:12] <omega_> someone can construct a meta-element like the autoplugge= r to do that safely [11:13] <steveb> so you temporarily run transport ! decoder ! lengthfinde= r then reconnect transport ! decoder to its original state? [11:14] <steveb> or no decoder - just transport ! lengthfinder [11:14] <omega_> transport ! autoplugcache ! parser [11:14] <omega_> and wait for a signal from the parser [11:14] <omega_> the parser should make an effort to get the length right= away [11:15] <omega_> but there are many issues lurking there [11:15] <steveb> so then how does the app get the length from the parser? [11:16] <omega_> via a signal or somesuch [11:19] <steveb> ok. the case I am thinking of is an editor app which has= a list of media resources to use (mostly audio files on disk but it coul= d be anything with no sinks). The user would at least have to know the le= ngth of each resource [11:19] <omega_> of course [11:19] thomas (th...@21...) joined #gstreamer. [11:26] <richardb> omega_: you mentioned that after GObject, you're going= to be doing a testsuite... [11:26] <omega_> heh, yeah [11:26] <richardb> ... is part of that going to be a "lint" tool for chec= king the behaviour of plugins? [11:26] <omega_> yup [11:26] <richardb> great. [11:27] <omega_> and a test framework for plugin developers to build test= suites for their plugins [11:27] <richardb> even better. [11:27] <richardb> I'm not sure of what correct behaviour for plugins is,= in many cases... [11:27] <omega_> we'll probably fill a CD with test media [11:27] <richardb> eg, when newpads can be created. [11:27] <omega_> richardb: part of that is defining what correct is [11:28] <richardb> Indeed. Writing testsuites often helps clear the mind= on that sort of issue... [11:28] <omega_> and I expect the testsuite to quickly grow to more lines= of code than the core and plugins combined [11:28] <omega_> and I'm gonna need a lot of help coming up with as many = unit and functional tests as possible [11:28] <richardb> It should be at least double, IMO. [11:28] <omega_> I'll be starting on that next Monday [11:30] <steveb> so gobject merge will happen before then? [11:30] <omega_> should, yes [11:31] <omega_> I'm closing on the last few bugs now in the shim [11:31] <omega_> we still need to deal with the port of the player and ed= itor and stuff to gnome-2, but that can wait [11:31] <steveb> so how do libs like flac end up using so much stack? [11:31] <omega_> because they're poorly written [11:32] <steveb> specifically [11:32] <omega_> should never allocate large amounts of data on the stack [11:32] <omega_> always malloc anything over about 8KB [11:32] <steveb> ok [11:36] <steveb> i've been thinking of adding a location arg to bin so th= at serialised pipelines can be easily run [11:36] <omega_> serialized pipelines? [11:36] <steveb> *.gst [11:36] <omega_> no, what we need is ghost parameters [11:37] <steveb> what do you mean [11:37] <omega_> ghost the location parameter to the parent bin, just lik= e pads [11:38] <richardb> It would also be good if ./gstreamer-launch could acce= pt a parameter to tell it to run the pipeline stored in a .xml file, or s= omething like that. [11:38] <omega_> yup [11:38] <steveb> richardb: thats what i'm talking about -launch bin locat= ion=3D~/bla.gst [11:38] <richardb> Oh, right. [11:38] <omega_> hrm, ok, not quite how I'd do it [11:39] <richardb> Surely that should be done by a special bin, like the = autoplugger. [11:39] <omega_> no, we have an API call for that [11:39] <omega_> see examples/xml/runxml.c [11:40] <richardb> gst_xml_get_element() gets the element with the approp= riate name from the xml file? [11:40] <omega_> yup [11:41] <omega_> just need to enhance -launch to understand that it's bei= ng passed a .gst filename, and load/run that [11:41] <omega_> even better.... [11:41] <omega_> treat it just like any other element, but one that's loa= ded from disk [11:41] <steveb> what if you could have nested bins which load their own = .gst file [11:41] <omega_> the compound-element idea [11:41] <steveb> yes! [11:42] greg__ (gr...@ho...) joined #gstreamer. [11:42] <richardb> I might want to do: -launch disksrc ! xml loc=3Dfile != osssink [11:42] <omega_> -launch dvdsrc ! dvdplay.gst [11:42] <steveb> maybe a custom bin element would be good for this [11:42] <omega_> nope [11:42] <omega_> no nead [11:42] <omega_> er, need [11:43] <richardb> I suppose -launch would just recognise an element name= which ended in .gst as an element to load from disk. [11:43] <omega_> just have gstparse.c check for files if the element does= n't exist [11:43] <richardb> or that. [11:43] <omega_> and load that as any other element into the pipeline [11:43] <omega_> need to make sure we can save ghostpads and ghostparams,= of course [11:43] <omega_> need the editor to make this complete [11:44] <steveb> explain how ghostparms would work [11:44] <omega_> just like ghostpads [11:44] <richardb> -launch should also have an option to save the complet= ed bin/chain rather than run it. [11:44] <omega_> ghost a param from a disksrc to its parent bin [11:44] <omega_> then bin suddenly gets a param that really is the disksr= c's param [11:44] <omega_> richardb: easy enough [11:46] <steveb> so would params in elements inside the .gst be settable = from -launch? [11:46] <omega_> yes, if they're ghosted to the parent container [11:47] <omega_> though I'd like to enhance the syntax so you can set arb= itary parameters like: [11:47] <omega_> disksrc0:location=3D".." [11:47] <omega_> though that overloads the : operator, which is used to s= eparate element:pad names [11:47] <steveb> bin0:thread0:disksrc0:location [11:48] <steveb> how about . notation [11:48] <omega_> steveb: see gst_object_get_path_string [11:50] <steveb> / [11:54] johnix (thomas@61.152.140.57) left irc: Client Exiting [12:06] <richardb> Would a -o initial parameter to -launch be appropriate= for saving the pipeline? [12:07] <omega_> probably [12:07] <richardb> eg, -launch -o myfile.gst disksrc blahblahblah [12:07] <omega_> yup [12:07] greg__ (gr...@ho...) left irc: Ping timeout for greg__[hom= e.sente.pl] [12:16] greg__ (gr...@ho...) joined #gstreamer. [12:16] greg__ (gr...@ho...) left irc: [x]chat [12:22] bln (bj...@h1...) joined #gstreamer. [12:33] Nick change: richardb -> richardb-work [12:35] Action: omega_ goes to sleep, finally [12:35] omega_ (om...@om...) left irc: [x]chat [12:45] bln (bj...@h1...) left irc: bbl [13:07] sienap (sy...@ip...) joined #gstreamer. [13:07] <sienap> dudes :) [14:16] Nick change: ajmitch -> ajzzzz [14:29] bln (bj...@h1...) joined #gstreamer. [14:29] <sienap> hej [14:29] <bln> tjena [14:35] thomas (th...@21...) left irc: Ping ti= meout for thomas[212-100-172-175.adsl.easynet.be] [15:19] bln (bj...@h1...) left #gstreamer. [15:27] sienap (sy...@ip...) left irc: My damn controlli= ng terminal disappeared! [15:49] dobey (do...@dr...) joined #gstreamer. [16:50] Nick change: taazzzz -> taaz [17:34] Uraeus (csc...@c3...) joined #gstreamer. [17:35] <Uraeus> hi [17:35] bln (bj...@h1...) joined #gstreamer. [17:45] marius_ (marius@Photon.pdx.netroedge.com) joined #gstreamer. [17:45] Nick change: marius_ -> ShrimpX_ [17:57] Uraeus (csc...@c3...) left irc: Client Exiting [18:05] o2 (oc...@en...) joined #gstreamer. [18:05] dobey-ppc (do...@dr...) joined #gstreamer. [18:05] steveb_ (st...@no...) joined #gstreamer. [18:05] Bob (ro...@d1...) got netsplit. [18:05] dobey (do...@dr...) got netsplit. [18:05] steveb (st...@no...) got netsplit. [18:05] OctobrX (oc...@en...) got netsplit. [18:07] Nick change: steveb_ -> steveb [18:07] Possible future nick collision: steveb [18:16] Bob (ro...@d1...) got lost in the net-s= plit. [18:16] OctobrX (oc...@en...) got lost in the net-split. [18:16] dobey (do...@dr...) got lost in the net-split. [18:32] steveb (st...@no...) left irc: Ping timeout for ste= veb[node1ee02.a2000.nl] [18:59] bln (bj...@h1...) left irc: Leaving [19:04] ChiefHighwater (pa...@te...) joined #gstreamer. [19:36] Bob (ro...@d1...) joined #gstreamer. [19:56] Nick change: wtay-zZz -> wtay [19:56] <wtay> hellouw [19:57] <ChiefHighwater> ello [20:03] Nick change: dobey-ppc -> dobey [20:11] <wtay> yo [20:20] <ShrimpX_> wtay, (newbie question) why are threads better for thi= s project rather than processes? [20:21] mattsm (ma...@ad...) joined = #gstreamer. [20:21] <wtay> ShrimpX_: threads are easier to use when a lot of data is = shared [20:21] <ShrimpX_> aren't they more expensive? [20:21] Zeenix (programmer@203.135.48.13) joined #gstreamer. [20:21] <wtay> ShrimpXnot really [20:21] <wtay> yo Zeenix [20:21] <Zeenix> hello wtay [20:22] <ShrimpX_> there's a lot of talk on linux-kernel right now as to = why threads suck, and "if you use threads it means that you are not able = to put together a good process model design", that type of stuff.... I wa= s wondering. [20:23] <wtay> ShrimpX_: don't beleive it [20:23] <wtay> deblieve even [20:23] <ShrimpX_> hehe [20:23] <Zeenix> wtay: i registered my project at FS & they approved it [20:23] <wtay> believe even :-) [20:24] <wtay> Zeenix: cool [20:24] <Zeenix> wtay: but there is a prob. [20:25] <wtay> ? [20:25] <Zeenix> i gave my name in the field given for project's full nam= e in the form [20:25] <wtay> rotfl ! [20:26] <wtay> so now we have a project called "zeenix" :) [20:26] <Zeenix> wtay: no its my full name: Zeeshan Ali Khattak [20:26] <wtay> oops [20:27] <Zeenix> wtay: but its unix name is zchat [20:28] <wtay> =20 [20:28] <wtay> Project: Zeeshan Ali Khattak [20:28] <Zeenix> yes [20:28] <wtay> cool [20:28] <wtay> I think you can still change it [20:28] <Zeenix> wtay: what should i do now [20:28] <ShrimpX_> hey, I use Zeeshan Ali Khattak! [20:29] <wtay> yeah, developed by zchat. [20:29] <ShrimpX_> c in written. [20:29] <Zeenix> its ok you can make a joke of it, i dont mind, i myself = loughed on it for hours [20:30] <wtay> Zeenix: you should change it :) [20:30] <Zeenix> wtay: i am trying to do that at the moment [20:32] <Zeenix> wtay: they were talking of logging-in from shell, whats = that ? [20:33] <Zeenix> wtay: its not the 'bash' shell they are talking about, i= s it ? [20:33] <wtay> Zeenix: ssh za...@sh... [20:34] <Zeenix> wtay: never got a chance to get onto internet from Linux [20:36] <ShrimpX_> that sucks [20:36] <Zeenix> wtay: was ssh on you debian box or did you got that from= somewhere [20:37] <wtay> Zeenix: It was with debian, yeah [20:38] <Zeenix> wtay: any idea about the name of rpm, a gues only ? [20:38] steveb (st...@no...) joined #gstreamer. [20:39] <wtay> Zeenix: somthing with ssh in it.. [20:39] <wtay> ssh client [20:42] <Zeenix> wtay: my friend here want to see your pic. , what was th= e URL, i forgot=20 [20:44] <wtay> uhm.. no no, last time I did that you wanted to date me :-= ! [20:45] <Zeenix> wtay: no i didnt [20:47] <wtay> uhmm..ok: http://gstreamer.net//Gstreamer_team_at_Guadec.j= pg [20:48] <dobey> hahaha [20:48] <dobey> does that mean mayam is free? [20:48] <wtay> dobey: not at all, it means I'm in a good mood [20:48] <dobey> damn [20:49] Action: dobey winks in mayam's general direction [20:49] <Zeenix> debey: havent you acccepted my appologies yet ? [20:50] <wtay> appologies? for what? [20:50] sienap (sy...@ip...) joined #gstreamer. [20:50] <sienap> sjonge jonge [20:50] <sienap> hi all [20:50] <Zeenix> wtay: ask dobey [20:50] <wtay> sienap: hallo [20:51] <sienap> goeiendag meneer wim :) [20:51] <sienap> ik vandaag nog beetje rond gekloot met svga [20:51] <sienap> svgalib dus [20:51] <sienap> maarja [20:51] <sienap> nu zit ik sdlsink source te lezen [20:51] <sienap> en snap ik er lekker helemaal NIKS van :) [20:51] <Zeenix> wtay: dobey seems to be very upset on asking where from = he is [20:51] <wtay> sienap: welgemeende edele groeten, mijnheer... [20:51] <sienap> zo hoort het ;) [20:51] <Zeenix> speak english [20:52] <sienap> zeenix you are not missing a thing [20:52] <sienap> :) [20:52] <wtay> sienap: hoezo, nieks snappen [20:52] <wtay> ? [20:52] <Zeenix> else have a private chat [20:52] <wtay> sienap: gewoon negeren die zijknieks :) [20:52] <sienap> wtay issut een eikol ? [20:52] <dobey> i am [20:53] <wtay> sienap: nogal van het zware kaliber, ja [20:53] <sienap> haha :) [20:53] <Zeenix> dobey: i promise, i'll never ask this again [20:53] <dobey> missing something that is [20:53] <sienap> hij doet dus ook niks nuttigs ? [20:53] <wtay> sienap: niet echt nee.. [20:54] <wtay> sienap: en dat terwijl ik een opendivx spul aan't schrijve= n ben... [20:54] <sienap> naja ik ook niet :) [20:54] <sienap> aah netjes [20:54] <sienap> opendivxdec ? [20:54] <Zeenix> ka la pa dha mmuuu ejrkjwerkj dskfjkj [20:54] <sienap> ik snap echt werkelijk niks van sdlsink echt klote [20:54] <wtay> yup [20:54] <sienap> wtay issut niet zo dat jij ops in dit channel kan krijge= n :) [20:54] <ChiefHighwater> Zeenix:umf gup rapft um nortgesung [20:54] <wtay> hmm, tuurlijk [20:55] <sienap> hehe [20:55] <sienap> :) [20:55] <sienap> schop die gozer dan [20:55] <sienap> maar goed [20:55] <Zeenix> CHw: madar chod kay bachay [20:55] <sienap> balen dat ik dom ben :) [20:55] <sienap> ik ben trouwens over he :) [20:55] <sienap> zonder herkansingen [20:55] <sienap> echt goed ;) [20:55] <sienap> niet verwacht enzo [20:55] <wtay> cool [20:55] <sienap> had ouders al voorbereid van ik ga zakken papie en mamie [20:55] <sienap> die mentor van me vond dat oko niet leuk [20:56] <sienap> dat ik over was :) [20:56] <sienap> had hij vast ook nie verwacht met m'n 30% aanwezigheid [20:56] <sienap> en heb 2 van 14 toetsen boek opengehad [20:56] <sienap> :) [20:56] <sienap> is nu wel beetje de man dacht ik zo [20:56] <ChiefHighwater> Zeenix: sienap ubermango zuit fregue dov wtay [20:56] <wtay> ChiefHighwater: yeah, yeah [20:56] <Zeenix> give me a brake [20:56] <sienap> sjonge jonge jonge [20:57] <sienap> al dat geweld [20:57] <sienap> :) [20:57] <sienap> maar goed [20:57] <wtay> noegabollen! [20:57] <sienap> ehm ... belgen hebben ook wel interresante scheldwoorden= :) [20:57] <wtay> hmm... no makefiles [20:57] <sienap> naja [20:57] <wtay> that sucks [20:57] <sienap> maar goed svgasink wilt niet lekker opschieten [20:57] <sienap> want tja ik ben dom [20:57] <sienap> :) [20:57] <sienap> ik snap er echt niks van . [20:57] <wtay> sdlsink? [20:58] <Zeenix> sienap: are you having a very serious prob. writting eng= lish ? [20:59] <sienap> ja dat heb ik [20:59] <wtay> sienap: what do you want to know about sdlvideosink? [20:59] <sienap> wtay vidsink [20:59] <sienap> wtay mwha [20:59] <sienap> ik snap echt geen hond van al die stuff :) [20:59] <sienap> hoe je data krijgt enzo [20:59] <sienap> naja gewoon niks eigenlijk [21:00] <sienap> maarja [21:00] <ChiefHighwater> Zeenix: you'll find sienap has many very serious= problems [21:00] <sienap> zo eens random gaan hacken [21:00] <wtay> sienap: start from the beginning? [21:00] <sienap> wtay he :) [21:00] <sienap> snapt ook niet hoe sdlsink naar screen blaat [21:00] <sienap> maarja [21:00] <wtay> sienap: I can guide trough the code if you want... [21:00] <sienap> dat is eigenlijk enige wat ik hoef te weten [21:00] <sienap> rest hack ik wel neer [21:00] <sienap> wtay dat is nouw eens lief [21:00] <sienap> moment [21:00] <sienap> hebbut uitgeprint liggen [21:00] <sienap> wtay shall we query ? [21:00] <wtay> heh [21:01] <sienap> i am not in a english mood anyawy :) [21:01] <wtay> no, do english and make a guide :-) [21:01] <sienap> chiefhighwater one of those problems is my serious large= dick :) [21:01] <taaz> wtay: that gint you commited isnt really the best way to d= o things... why not just use size_t since that's what fwrite returns? [21:01] <wtay> taaz: fair enough.. [21:01] <sienap> wtay kan dit echt niet in het nederlands [21:01] <sienap> dit ga ik niet verwoord krijgen [21:01] <sienap> :) [21:02] <wtay> taaz: it's better than the guint16 is used to be :-) [21:02] <wtay> sienap: allez dan... Ik zal wel in the engels antwoorden a= ls you don't mind [21:02] <Zeenix> wtay: dont you think sienap is misbehaving [21:02] <sienap> wtay neuh :) [21:02] <sienap> zeenix no he doesn't [21:03] <sienap> zeenix actually we think you do [21:03] <sienap> zeenix shall we assbutt kick you under the nuts ? >:) [21:03] <wtay> hmm, shut up=20 [21:03] <sienap> ik ? [21:03] <wtay> yes [21:03] <sienap> ooh [21:03] <sienap> sorry hoor [21:03] <wtay> no insulting here [21:03] <sienap> ehm ok ok my excuse :) [21:03] <wtay> my time is short.. [21:04] <sienap> ooh pardon :) [21:04] <sienap> ehm however [21:04] <wtay> 3 more hours to do something productive this evening [21:04] <sienap> in the sdlvideosink there are some parts at the begining= that has enum { somestuff } what is this doing ? [21:04] <wtay> no offense of course :) [21:04] <sienap> like in SIGNAL_FRAME_DISPLAYED and stuff [21:04] <wtay> ok, the first one defines some signals.. the second one ar= guments [21:05] <wtay> you know what an enum is? [21:05] <sienap> always has to be in that orde r? [21:05] <sienap> ehm no [21:05] <taaz> wtay: play find-the-big-stack-function-in-flac! [21:05] <sienap> what is enum doing ? [21:05] <wtay> taaz: yeah [21:05] <wtay> sienap: it's just a declaration [21:05] <taaz> i was looking... couldn't find it though. some seriously = large allocations though [21:05] <wtay> sienap: like #define SIGNAL_ 0 [21:05] <wtay> sienap: starting from 0 to ... [21:05] <sienap> aah ic [21:05] <wtay> taaz: weird how the stack overflow could happen there... [21:06] <wtay> sienap: so.. ARG_HEIGHT =3D=3D 2 [21:06] <taaz> those Frames and Subframes are really big. all malloced t= hough [21:06] <wtay> for example.. [21:06] <sienap> think i got the point [21:06] <wtay> taaz: there was one thing on the stack I removed... but it= didn't help.. [21:06] <wtay> sienap: ok, next? [21:06] <sienap> all is standard 0 ? [21:07] <wtay> sienap: starts from 0, goes up by 1 [21:07] <sienap> gst_padtemplate_factory just sets some properties about = the element huh ? [21:07] <wtay> sienap: you can also give it a type so the compiler can ch= eck it etc... [21:07] <wtay> sienap: that is what the pads will accept [21:07] <wtay> sienap: for example, this plugin has a sink pad (input)=20 [21:08] <taaz> wtay: i actually haven't even tried it yet ;) been workin= g on the debs. just did a scan for anything suspicious [21:08] <sienap> aah i understand [21:08] <wtay> sienap: it takes video/raw in the YUV format=20 [21:08] <wtay> taaz: I recompiled with various changes.. to no avail [21:08] <sienap> hmm how am i going to do that with svga 16bits colorspac= e ? [21:08] <wtay> sienap: sec.. [21:08] <sienap> he ok [21:09] Nick change: taaz -> taaz-meeting [21:09] <wtay> sienap: look in xvideosink, line 567 [21:09] <wtay> sienap: that's for format definition of RGB images [21:09] <sienap> one mom [21:10] <wtay> also dvdec line 92... [21:10] <sienap> so i can just request it ? [21:10] <sienap> like in if i want it in YUV or RGB or BGR [21:10] <sienap> aah that is nice [21:10] <wtay> sienap: what do you mean 'request it'? [21:11] <sienap> in what format i want the stream [21:11] <wtay> sienap: yes [21:11] <sienap> for RGB i can use svgagl really nice [21:11] <sienap> that is rather cool :) [21:11] <wtay> the better sinks of course accept many formats... [21:11] <sienap> he :) [21:11] <wtay> sienap: gstreamer will insert a converter if the format do= esn't match [21:11] <sienap> aah cool cool :) [21:12] <sienap> RGB is perfect for vgagl [21:12] <wtay> ok [21:12] <sienap> ok after that there are bunch of static voids [21:12] <sienap> ehm one mom raeding [21:12] <wtay> the best way would be to query for the vga capabilities an= d generate a list of those definitions [21:12] <sienap> ooh that are declaration for upcoming functions d0h [21:12] <wtay> yup [21:13] <sienap> ok one mom [21:13] <wtay> sienap: note the 'static ' keyword.. you know what that is= ? [21:13] <sienap> i am not taking too much time hopefully [21:13] <sienap> wtay not exacly [21:13] <sienap> :) [21:13] <wtay> sienap: it means that the function will not be exported in= the shared lib.. [21:13] <wtay> actually, not exported from this .c file [21:14] <wtay> so other plugins will not see this function at all... [21:14] <sienap> aah i though so [21:14] <sienap> so extern does [21:14] <sienap> yeah i got the point [21:14] <wtay> or without the static [21:14] <sienap> already though about it that way [21:14] <sienap> cool [21:14] <wtay> sienap: a plugin only has one external symbol: plugin_desc [21:14] <sienap> ok what about the gst_sdlvideosink_get_type=20 [21:15] <wtay> sienap: that's the gtk+ object stuff [21:15] <sienap> ehm ok [21:15] <sienap> something special to note on ? [21:15] <wtay> it creates a new type in gtk [21:15] <wtay> not really, be sure to adjust the name [21:15] <sienap> aah ok [21:15] <wtay> and the pointers to the class_init and _init functions [21:16] <wtay> sienap: you know about OO programming, right? [21:16] <wtay> ... about classes and instances...? [21:16] <sienap> you really think so ? [21:16] <sienap> ofcourse i don't :) [21:16] <wtay> right.. [21:16] <sienap> aah i will learn by time [21:16] <sienap> :) [21:17] <ShrimpX_> heh [21:17] <sienap> he :) [21:17] <sienap> wtay you are 10 years older than me :) [21:17] <wtay> sienap: the class is the 'template' for the objects you ar= e going to create [21:17] <ShrimpX_> you already program OO by using structs and such... [21:17] <Zeenix> sienap: this means you are at my age [21:17] <ShrimpX_> in part [21:17] omega_ (om...@om...) joined #gstreamer. [21:17] <wtay> yo [21:17] <omega_> yo [21:17] <ShrimpX_> omega_ wakes up [21:17] <Zeenix> i know about OO programming, thats no excuse [21:18] <wtay> sienap: so... [21:18] <sienap> Zeenix that will declare the lameness of both of us :) [21:18] <sienap> the age thingy that is [21:18] <ShrimpX_> hehe [21:18] <sienap> zeenix you are a genius now don't bother me please [21:18] <ShrimpX_> age don't matter [21:18] <sienap> wtay next function [21:19] <wtay> sienap: in the class init, you declare the properties and = the signals intances are going to have [21:19] <sienap> yeah i know about that a little [21:19] <sienap> gtk has it as well [21:19] <wtay> ok, good [21:19] <sienap> ok ok=20 [21:19] <wtay> sienap: so some syntactic fluf to declare the properties a= nd signals... [21:19] <sienap> _class_init... what todo ? [21:20] <wtay> sienap: cut and paste, adjust=20 [21:20] <wtay> sienap: you want to have signals and properties to your pl= ugin? [21:20] <wtay> gstelement_class->change_state =3D gst_sdlvideosink_change= _state; [21:20] <sienap> wtay guess so :) [21:21] <wtay> that's where we override the class method change_state to = our own implementation [21:21] <wtay> inheritance it is... [21:21] <sienap> a bit lame.. watching vid in console launching in X [21:21] <sienap> can't wait till gobject :) [21:21] <sienap> ok next function the newcaps [21:22] <wtay> sienap: first we go down to the _init method [21:22] <sienap> ehm ? [21:22] <wtay> that's the method that will be called when we create a new= plugin instance [21:22] <sienap> what function we are talking right now ? [21:23] <wtay> sienap: gst_sdlvideosink_init [21:23] <wtay> sienap: skip the newcaps function for now [21:23] <sienap> ooh koie [21:23] <sienap> ok found :) [21:23] <wtay> see it? [21:23] <sienap> yeah [21:23] <sienap> i've got it on paper [21:23] <sienap> so i can read / make notes easily [21:23] <wtay> first we create a new pad (from the template) [21:23] <wtay> gst_pad_new_from_template [21:24] <wtay> see how it goes? [21:24] <sienap> mom [21:24] <sienap> raeding it [21:24] <sienap> ehm [21:24] <sienap> guess so :) [21:24] <wtay> it'll create a pad with the properties of the template we = declared a bit higher.. [21:25] <wtay> this is where the plugin will get it's data from [21:25] <sienap> ehm=20 [21:25] <wtay> no? [21:25] <sienap> not really :) [21:25] <sienap> what are those -> things doing ? [21:26] <sienap> like sdlvideosink->sinkpad [21:26] <sienap> and where did it get the name sdlvideosink from .. [21:26] <wtay> sdlvideosink is a pointer to a struct, the -> accesses a v= ariable in it [21:26] <sienap> ooh ic.. [21:26] <wtay> sienap: it's the argument of the function _init [21:26] <sienap> why not just . ? [21:26] <sienap> sdlvideosink.sinkpad [21:26] <steveb> sienap: BBB chose that name [21:26] <wtay> sienap: cause it's a pointer [21:26] <sienap> steveb what name ? [21:26] <sienap> ooh ic.. [21:27] <steveb> sdlvideosink - he wrote it [21:27] <sienap> ooh yeah i kno [21:27] <sienap> w [21:27] <sienap> ooh those structures are in the header file .. [21:27] <sienap> sdlvideosink.h or something :) [21:27] <wtay> right [21:27] <sienap> that declares :) [21:28] <wtay> sienap: ready now? [21:28] <sienap> yes [21:28] <wtay> ok, then we add the pad to the plugin... with gst_element_= add_pad [21:28] <sienap> okie [21:29] <wtay> now our element has a sink pad that accepts video/raw [21:29] <sienap> all looks like gtk actually :) [21:29] <wtay> yeah [21:29] <wtay> then we set some other stuff in the pad... [21:29] <sienap> how to make the sickpad accepting rgb ? [21:29] <steveb> would this be good for gstreamer? http://www.videolan.or= g/libdvdcss/index.html [21:29] <wtay> steveb: yes, downloaded it [21:29] <steveb> the site is a bit /.ed at the moment [21:30] <omega_> is that what taaz was talking about? [21:30] <wtay> sienap: change the padtemplate used to generate the pad [21:30] <dobey> whee [21:30] <sienap> wtay mom looking into that [21:30] <omega_> taaz was talking about a two-part library, that dlopen()= 's the crypt part [21:31] <sienap> GST_padtemplate_factory right ? [21:31] <wtay> omega_: I don't think so.. [21:31] <wtay> sienap: yup [21:31] <sienap> so "video/rgb" should give me rgb data ? [21:31] <wtay> sienap: see how we use that definition to create the pad [21:31] <wtay> sienap: no, look at dvdec for example [21:31] <sienap> mom [21:32] <wtay> sienap: it's video/raw, format=3DRGB, ... [21:32] <sienap> what dir it is ? [21:32] <wtay> plugins/dv [21:32] <wtay> dvdec.c [21:33] <wtay> lin 92 [21:33] <sienap> what is dv anyway ? [21:33] <wtay> sienap: the DV format (used in modern camcorders) [21:33] <sienap> ooh got the point [21:33] <wtay> sienap: usually with firewire and such [21:33] <sienap> about the rgb thingy [21:34] <sienap> it will always give RGB huh ? [21:34] <sienap> with a converter or something [21:34] <sienap> really handy / nice [21:34] <sienap> :) [21:34] <wtay> sienap: yes [21:34] <sienap> btw can i also set the scale reso [21:34] <sienap> like 320x200 [21:34] <sienap> or the other [21:34] <sienap> :) [21:35] <wtay> sienap: then just set the width/height to something fixed [21:35] <sienap> ooh also in the padtemplate factory [21:35] <wtay> sienap: a plugin that cannot generate the format will not = be able to talk to your plugin [21:35] <sienap> rgb format you mean ? [21:35] <sienap> then a converter will be added huh ? [21:35] <wtay> whatever you define in the template [21:36] <wtay> sienap: yes [21:36] <sienap> also for resolution a converter will be added ? automati= cly ? [21:36] <wtay> sienap: semiautomaticlly, yes [21:36] <sienap> for fixed width just "width",320 or something.. [21:36] <wtay> sienap: yes [21:36] <sienap> do i have to do things my self to get those converters t= oo work ? [21:36] Zeenix (programmer@203.135.48.13) left irc: Ping timeout for Zeen= ix[203.135.48.13] [21:37] <wtay> sienap: not in the plugin [21:37] <wtay> sienap: continue? [21:38] <sienap> ehm ok [21:38] <sienap> i am worried about the format convertion [21:38] <sienap> rgb / and scale [21:38] <wtay> sienap: don't worry about it yet [21:39] <sienap> however don't want to take too much time from you so go = on.. [21:39] <wtay> sienap: then you need to set a chain function on the pad [21:39] <sienap> which does ? [21:39] <wtay> sienap: this function will be called when data is sent to = the pad [21:39] <sienap> think i am getting the point like a normal gstreamer pad= the plugins has their own internal pad as well ? [21:40] <sienap> or something.. [21:40] <wtay> sienap: pads always belong to a plugin [21:40] <wtay> element to be exact [21:40] <sienap> aah ic [21:40] <wtay> sienap: se how gst_pad_set_chain_function (sdlvideosink->s= inkpad, gst_sdlvideosink_chain); sets the chain function to this pad? [21:41] <wtay> sienap: and if you look a few lines down, you'll see the c= hain function [21:41] <sienap> ehm i think so [21:41] <sienap> yeah isee it [21:41] steveb (st...@no...) left irc: [x]chat [21:41] <wtay> sienap: that function will be called when data enters the = plugin (on this pad) [21:41] <sienap> so the chain function is probably the draw function as w= ell ? [21:42] <sienap> yeah i got myself that far [21:42] <wtay> sienap: it does the drawing, yes [21:42] <wtay> sienap: it's gets a buffer as the input [21:42] <wtay> sienap: that's the buffer with the image in it [21:42] <sienap> one mom [21:43] <sienap> man toilet brb :) [21:43] <wtay> omega_: is there a way to measure stack usage of a plugin? [21:44] <omega_> only during a switch [21:44] <omega_> compare sp to top_sp [21:45] <wtay> omega_: any way of doing that inside a plugin (when it cal= ls a function of some other library)? [21:46] <omega_> atm it's only possible during switch [21:46] <wtay> ok [21:46] <omega_> you'd have to set up a SIGALRM to check at arbitrary tim= es [21:46] <wtay> SIGALRM? [21:47] <omega_> set a signal handler that compares the sp to the current= top_sp [21:47] <wtay> oh [21:47] <omega_> set the alarm to a very short time [21:47] <omega_> keep track of the current cothread and stack size [21:48] <omega_> wtay: I want to get GOBJECT1 stable and merged ASAP [21:48] <wtay> omega_: good [21:48] <omega_> and thus I need your help testing and debugging [21:48] <wtay> sure [21:48] <omega_> and preparing it with the dependencies [21:49] <omega_> need to remove gnome/gtk deps from everything that doesn= 't actually need it [21:49] <omega_> and set up Makefile.am's to only build stuff for now if = we're using gtk-1.2 [21:49] <wtay> like -launch I'm afraid.. [21:49] <wtay> yes [21:49] <sienap> back [21:49] <sienap> i can tell you one thing.. that was some major peement := ) [21:49] <sienap> ok where were we ... [21:49] <sienap> the chain function [21:49] <wtay> omega_: 30mins? [21:49] <sienap> go ahead and explain wim.. [21:50] <omega_> wtay: ? [21:50] <wtay> omega_: and then I'll start working on gobject.. [21:50] <omega_> ok [21:50] <wtay> we're writing a plugin dev manual here :) [21:50] <sienap> hehe [21:50] <omega_> ah [21:50] <wtay> sienap: so chain... lot's of stuff to draw the buffer usin= g SDL [21:51] <sienap> one mom reading over [21:51] <wtay> sienap: the first part is to wait for the clock [21:51] <sienap> there is an GstSDLVideoSink *sdlvideosink; thingy.. what= kind of type is that ? [21:52] <wtay> sienap: that's your object as defined in the .h file [21:52] <sienap> ooh mom [21:52] <wtay> sienap: ..as defined by the _class_init and _init function [21:52] Parapraxis (Sp...@us...) joined #gstre= amer. [21:52] <sienap> aah true [21:52] <sienap> yeah ic [21:52] <wtay> hi [21:53] <wtay> sienap: you following? [21:53] <sienap> ehm yes [21:53] <sienap> hej parapaxel! [21:54] <wtay> sienap: then there is some stuff to get/set properties... [21:54] <wtay> sienap: some stuff to open/close the SDL videosink (show/h= ide it) [21:55] <sienap> i know about the SDL open and close stuff [21:55] <sienap> :) [21:55] <sienap> i made that plasma thingy in SDL last time [21:55] <wtay> sienap: then the plugin init method to register this plugi= n... [21:55] <Parapraxis> hey sienap [21:55] <sienap> locksurface that is [21:55] <sienap> Parapraxis=02:=02 did some cool stuff today ? [21:56] <sienap> wtay ehm where ? [21:56] <wtay> plugin_init [21:56] <omega_> wtay: do you have a functioning divx/3ivx player? [21:56] <sienap> ehm what function ? [21:56] <Parapraxis> sienap: got my first real actuator working (a scope = that uses end-user provided equations) [21:56] <wtay> omega_: divx, yes, 3ivx, no [21:56] <sienap> ooh wait [21:56] <sienap> wtay mis comunication [21:57] <sienap> :) [21:57] <omega_> wtay: emailing you a movie to see if you can view it [21:57] <sienap> para hell yaeh :) [21:57] <Parapraxis> sienap: it's not committed yet, but you should be ab= le to check it out later today (I have also done an xmms plugin) [21:57] <wtay> omega_: ok, what's it supposed to be? [21:57] <ShrimpX_> pr0n [21:57] <sienap> parapraxis nice nice :) [21:57] <omega_> wtay: MPEG-4, very simplistic (I-only, etc.) [21:57] <wtay> omega_: hmm, avi? [21:57] <omega_> yup [21:57] <wtay> should work [21:58] <wtay> omega_: not too big I may hope... [21:58] <Parapraxis> sienap: still have to work out some wierd bug with e= xponents in the scripts [21:58] <omega_> wtay: 300KB [21:58] <wtay> ok [21:58] <sienap> parapraxis succes :) [21:58] <sienap> wtay ehm sorry we went wrong :) [21:58] <sienap> i was still at the chain function [21:59] <sienap> what is up with the "jitter " ? [21:59] <wtay> sienap: that's for the timing stuff. [21:59] <sienap> explain [21:59] <wtay> sienap: buffers have a timestamp indicating at what time t= hey should be displayed [21:59] <wtay> sienap: we use a clock to wait for that time [22:00] <wtay> sienap: but only if the timestamp is not too far off... [22:00] <sienap> aah ic [22:00] <sienap> what if the sink is too slow ? [22:00] <wtay> sienap: nothing.. [22:00] <sienap> just a general question [22:01] <sienap> what will happen if my svgasink will draw around 10 fram= es/sec and the stream does 25 fps [22:01] <wtay> sienap: sdlvideosink doesn't drop frames [22:01] <sienap> so what will happen ? [22:01] <wtay> sienap: you can drop frames if the timestamp is too far of= f [22:01] <wtay> sienap: it'll slow down to a crawl :) [22:01] <sienap> it will or ? [22:01] <sienap> aah shittie [22:01] <sienap> so no QoS in sdlsink ? [22:01] <wtay> nope [22:02] <wtay> no good QoS in gstreamer even [22:02] <sienap> QoS will be implented system wide later ? [22:02] <ShrimpX_> what's QoS? [22:02] <sienap> in a way that sinks won't have to watch it ? [22:02] <wtay> sienap: more or less, yes [22:02] <omega_> wtay: luck? [22:02] <wtay> omega_: nope.. [22:02] <sienap> aah nice [22:02] <wtay> ** WARNING **: avidecoder: could not autoplug [22:02] <wtay> Not a JPEG file: starts with 0x00 0x00 [22:03] <omega_> hrm [22:03] <sienap> wtay so i won't really have to watch it ? since it we'll= be fixed later anyway he okie [22:03] <wtay> sienap: yup [22:03] <sienap> nice nice [22:03] <sienap> :) [22:03] <sienap> ok the drawing we let for what it is for now [22:04] <wtay> omega_: it's mjpeg [22:04] <sienap> ehm what about the set_arg ? [22:04] <wtay> sienap: that's general gtk stuff to set properties in the = plugin [22:04] <sienap> aah so just takeover and change names ? [22:04] <omega_> wtay: how do you know that? [22:04] <wtay> INFO (23024: 4)gst_avi_demux_strh:236: gst_avi_demux: fcc= _handler 0x67706a6d (mjpg) [22:05] <wtay> INFO (23024: 4)gst_avi_demux_strf_vids:264: gst_avi_demux:= compression 0x47504a4d (MJPG) [22:05] <omega_> hrm. ok, well, the data is actually mpeg4 [22:05] <omega_> I'll report that problem though [22:05] <omega_> what fourcc should it be? [22:05] <wtay> that's a bug then [22:05] <wtay> MP31 or something [22:05] <omega_> it's ISO, not divx M$ v3 [22:06] <omega_> you probably need 3ivx or opendivx to play it [22:06] <wtay> omega_: avidecoder currently uses jpegdec for it.. [22:07] <wtay> omega_: I can hack it to use something else... [22:07] <sienap> wtay where does sdlvideosink->frames_displayed; get's hi= s frames-displayed information from ? [22:07] <omega_> wtay: it is definitely mpeg-4 [22:07] <wtay> omega_: you think opendivx can handle it? [22:08] <wtay> sienap: ? [22:08] <omega_> if opendivx is really ISO, yes [22:08] <wtay> sienap: unused [22:08] <sienap> ooh :) [22:08] <wtay> omega_: it should.. [22:08] <sienap> that declares [22:08] <sienap> and frame_time .. ? [22:08] <sienap> also unused ? [22:08] <sienap> what is it btw :) [22:08] <sienap> (i need to make logner sentences) [22:08] <wtay> unused [22:09] <wtay> aka, not implemented :) [22:09] <omega_> sienap: yes, please use longer sentences, too many lines= scroll my xchat waaay off the screen [22:09] <sienap> hehe [22:09] <sienap> :) [22:09] <sienap> omega sorry it is something i am trying to get learn off= for months already :) [22:09] <wtay> sienap: remove you enter key from the keyboard.. [22:09] <sienap> LOL :) [22:10] Uraeus (csc...@c3...) joined #gstreamer. [22:10] <Uraeus> hi [22:10] <omega_> yo [22:10] <sienap> maybe i should hang a paper on my monitor that says don'= t press enter every 2 seconds.. [22:10] <wtay> hi [22:10] <sienap> however i understand the open / close thingy.. should be= a vgainit and vga(TEXT) thingy for svgasink [22:10] <sienap> ok what about the gstelementstatereturn ? [22:11] <wtay> sienap: that's the state change function [22:11] <sienap> explain that a little please (if i am not taking too muc= h time) [22:11] Nick change: o2 -> OctobrX [22:11] <wtay> sienap: an element will go through different states as exp= lained in the manual [22:11] <wtay> null->ready->paused->playing [22:12] <wtay> sienap: on NULL, we close the videosink, else we open it [22:12] <wtay> omega_: I'm looking at opendivx now.. [22:13] <sienap> aah i understand [22:14] <sienap> what about the ready state ? [22:14] Nick change: taaz-meeting -> taaz [22:14] <wtay> sienap: just make the plugin ready to accept data [22:15] <omega_> taaz: is the vlc libdvdcss thing what you were talking a= bout? [22:15] <sienap> oh okie [22:15] <Uraeus> omega_, wtay: I never got around to following this up, b= ut I think we should aim at inclussion in the upcomming Sun System Archit= ecture process, what do you think? [22:15] <omega_> Uraeus: sure [22:15] <wtay> yes [22:15] <Uraeus> ok, I mail nils from Sun and ask what and how [22:16] <Uraeus> wtay: did I remember to forward the FLAC mail to you? [22:16] <wtay> Uraeus: last tile I checked they need design docs to start= the process.. [22:16] <wtay> Uraeus: yes [22:16] <Uraeus> do we currently have any design docs? [22:16] <wtay> sienap: and the last thing: the newcaps function... [22:16] <Uraeus> of a type that might be usefull? [22:16] <wtay> Uraeus: not really.. [22:17] <wtay> Uraeus: some exist though.. [22:17] <taaz> omega_: no, that's vlc stuff. the libdvdread has IFO pars= ing and uses libcss... though i was arguing earlier that we should cooper= ate and just have one css code base. libdvdread may convert to use libdv= dcss [22:17] <sienap> wtay mom i am going over everything one time :) [22:18] <omega_> taaz: hmm, the css lib shouldn't do IFO ... [truncated message content] |