From: IRC <wt...@us...> - 2003-09-24 05:43:16
|
******************************************************************* [03:03] Company (~Co...@p5...) left irc: Read error: 104 (Connection reset by peer) [03:04] ChrJW_log (~ch...@p5...) left irc: Connection timed out [03:20] ChrisHJW (~ch...@p5...) joined #gstreamer. [03:20] Company (~Company@pD958BE2C.dip.t-dialin.net) joined #gstreamer. [03:44] robla (~robla@209.210.138.240) left irc: "Client exiting" [04:21] thaytoo (ja...@ad...) joined #gstreamer. [04:30] walters-ny (wa...@ve...) joined #gstreamer. [04:31] <ds-work> ny? [04:31] <Company> no youth probably [04:32] <walters-ny> new york dudes :) [04:32] <Company> d'oh ;) [04:33] <walters-ny> i am at the most awesome coffee shop [04:33] <ds-work> where is it? [04:33] <walters-ny> free wireless, tons of hot chicks, and they're playing Tool [04:33] <walters-ny> ds-work: manhattan, around 83rd and 2nd [04:33] <ds-work> ah, nice area [04:34] <ds-work> pretty down-to-earth, as far as manhattan goes [04:34] <walters-ny> yeah, i like it [04:36] <ds-work> Company: are you awake enough to talk about caps? [04:36] <Company> if it doesn't take long [04:37] <ds-work> Company: I want to change the gstcaps.c and gstprops.c code [04:37] <Company> go ahead :) [04:37] <ds-work> There are two possible models [04:37] <ds-work> GstCaps, which implements a linked list, and has the props code merged in [04:39] <ds-work> or GstCapsChain + GstCaps [04:39] <ds-work> for the latter, many operations (like try_set_caps) would take a CapsChain [04:39] <ds-work> but caps_set and caps_get would take a GstCaps, i.e., an individual link [04:40] <ds-work> the latter case would require a lot of code rework [04:40] <Company> that would be somewhat like returning a gstprops right now [04:40] <ds-work> the former case would be nearly 100% compatible with current code [04:40] <ds-work> not really, since GstProps is 1-1 with GstCaps [04:41] <ds-work> GstProps is _completely_ redundant [04:41] <ds-work> it needs to be merged into GstCaps [04:41] <Company> hm [04:41] <pigeon> Hi all... would anyone here be able to answer me a question with demuxers? [04:41] <Company> silly idea: could we use GObjects and inheritance? [04:42] <ds-work> _after_ that is done, the question becomes: should we carry around extra information (i.e., chain head, reference count, etc.) on each caps link, or separate out GstCaps into a link and a head [04:42] <Company> so that we could use GstCapsChain as a subclass of GstCaps or so? [04:42] <ds-work> possibly, although GstCaps is not an object currently [04:43] <Company> yeah [04:43] <ds-work> I think it should stay that way [04:43] <Company> why? [04:43] <ds-work> speed [04:43] <Company> premature optimization? [04:43] <ds-work> we can have structure inheritance, though [04:43] <ds-work> design optimization [04:43] <ds-work> don't include slow code by design [04:44] <Company> but we dont know if its needed [04:44] <Company> anyway, its just an idea [04:44] <Company> i dont think it'd be slow [04:44] <ds-work> well, not compared to the current code [04:45] <Company> apart from object creation and preallocations, but that could be optimized [04:45] <ds-work> but for the operations I've tested, my new combined caps+props code is up to 10x faster [04:45] <Company> anyway, back to topic [04:46] <ds-work> the basic problem is this. I want to call gst_caps_set() on a caps link. It should fail if the link is not writable, i.e., if the reference count is >1 [04:46] <ds-work> but the reference count is in the chain head, at least logically [04:46] <Company> yeah [04:46] <Company> you need a structure for the head [04:47] <ds-work> one possibility is to have a redundant reference count in all links [04:47] <ds-work> or have a chain head pointer in every link [04:47] <Company> that's stupid, because there might be other reduntant information in the future [04:47] <Company> the caps name is kind of redundant right now, too [04:47] <ds-work> there already is other redundant information [04:47] <ds-work> name is one [04:48] <ds-work> I'd like to get rid of name. nobody uses it [04:48] <Company> i do [04:48] <ds-work> if someone finds a use for it, we can readd it [04:48] <Company> it allows finding out where caps come from, when you debug them :) [04:48] <Company> but apart from that it has no use [04:49] <ds-work> in that case, I think it should be automatic, such as source file and line number [04:50] <Company> maybe [04:50] <Company> but you can remove the name IMO [04:51] <Company> in the current design the "type" information belongs into the gstprops struct really [04:52] <ds-work> assuming that the current design was supposed to be like my plans, which I don't think is the case [04:53] <ds-work> iirc, gstprops was supposed to be a generic run-time structure type [04:53] <ds-work> there were plans for glibification [04:53] <Company> im too tired for those words... [04:54] <Company> a generic run-time structure type ? [04:54] <ds-work> i.e., a C structure whose fields are defined at run time [04:55] <Company> oh, that's possible in a clean way? [04:55] <ds-work> probably not in the way you're thinking [04:55] <Company> well, it would make sense that way [04:55] <ds-work> but a collection of strings and GValues, yes [04:56] <Company> yeah, that's the same like today [04:56] <Company> like props are today basically [04:56] <Company> and then there's the propsentries [04:56] <ds-work> ideally, propsentries would be GValues [04:56] <Company> uh, the propsentries carry the name (stupid idea) [04:57] <Company> yeah [04:57] <ds-work> I've made that possible in the new code, but it doesn't handle comparing GValues that are ints to propsentry ints [04:58] <Company> as long as they are transformable [04:58] <ds-work> shouldn't be a problem [04:59] <ds-work> anyway, I'll make it possible. policy can come later [04:59] <Company> i think the current basic design for caps/props is pretty much ok [04:59] <ds-work> it sounds like your gut instinct was the same as mine, i.e., "< Company> you need a structure for the head" [05:00] <Company> i mean the caps -> props -> propsentries idea [05:00] <Company> but many of the struct members should be shuffled around [05:00] <Company> a propsentry should be a GValue for example [05:01] <Company> the corresponding name should be inside the props structure [05:01] <Company> a props structure should not be refcounted [05:02] <ds-work> yeah [05:02] <Company> but have the mime type (instead of the caps) [05:02] <ds-work> I've basically come to the conclusion that an entire caps chain should be the only thing carrying a reference count [05:02] <ds-work> anything else is unworkable [05:02] <Company> yeah [05:03] <Company> nobody uses props the refcounting way anyway [05:03] <ds-work> uh, actually... [05:04] <ds-work> I've seen caps that have different refcounts than the props it contains [05:04] <Company> well, the props should always have a refcount of 1, no? [05:06] <ds-work> there are cases where props get used by multiple gstcaps [05:06] <Company> you could just post your code to the list, i'll look at it and comment on it [05:06] <ds-work> it's needlessly complicated [05:06] <ds-work> yeah [05:07] <ds-work> one of the side effects of all this is that I'm writing a test suite for caps [05:08] <Company> that's great [05:08] <Company> if you also write a good API documentation for what caps are, then everything is great :) [05:11] <Company> ok, i'm _really_ of to bed now [05:11] <Company> gnight [05:14] <ds-work> later [05:14] Action: ds-work goes home [05:54] walters-ny (wa...@ve...) left irc: "being displaced" [05:55] robla (~ro...@12...) joined #gstreamer. [06:07] Nick change: harshy -> harshy|scaredawa [06:08] Nick change: harshy|scaredawa -> harshy [08:05] chillywilly (danielb@CPE-24-167-193-166.wi.rr.com) left irc: Read error: 104 (Connection reset by peer) [08:10] Nick change: harshy -> harshyZ [08:48] kmaraas (~km...@14...) left irc: Client Quit [08:58] Action: sxpert has one full day of music, according to rb :D [09:10] thaytoo (ja...@ad...) left irc: "later" [09:17] Nick change: sxpert -> oboolo|sxpert [09:19] Action: oboolo|sxpert is away: time to go to work [09:21] wheels (~sc...@ds...) left irc: "work" [09:31] pb_ (~pb...@ds...) joined #gstreamer. [09:33] chillywilly (danielb@CPE-24-167-193-166.wi.rr.com) joined #gstreamer. [09:52] Company (~Company@pD958BE2C.dip.t-dialin.net) left irc: Remote closed the connection [10:27] desrt_ (de...@ko...) joined #gstreamer. [10:27] desrt (de...@ko...) left irc: Read error: 104 (Connection reset by peer) [10:34] Action: not-here|sxpert is back (gone 16:53:33) [10:35] Nick change: not-here|sxpert -> sxpert_work [10:48] kmaraas (~km...@18...) joined #gstreamer. [10:51] desrt_ (de...@ko...) left irc: Remote closed the connection [10:51] desrt (de...@ko...) joined #gstreamer. [10:55] walken (fo...@no...) joined #gstreamer. [11:05] sxpert_work (~sx...@ra...) left irc: "Client exiting" [11:06] desrt (de...@ko...) left irc: Remote closed the connection [11:06] thomasvs (~th...@15...) left irc: Read error: 104 (Connection reset by peer) [11:06] thomasvs (~th...@61...) joined #gstreamer. [11:08] kmaraas (~km...@18...) left irc: Read error: 110 (Connection timed out) [11:09] teuf (~te...@gr...) joined #gstreamer. [11:09] <teuf> hi [11:11] desrt (de...@ko...) joined #gstreamer. [11:12] <teuf> thomasvs: I updated yesterday's patch at http://cfergeau.free.fr/filesrc.diff but it's still ugly :( I guess I'll put it in bugzilla and expect it to be marked as WONTFIX :) [11:16] <thomasvs> teuf: you should ask the list for input, maybe some other people hav eidaes too [11:18] <teuf> thomasvs: can I post to the list without being subscribed ? [11:18] <thomasvs> why aren't you subscribed ? :) [11:19] <teuf> I already have too many mails to read :) [11:19] <thomasvs> it's low volume [11:19] Action: thomasvs chuckles evilly [11:19] <teuf> hehe [11:19] <thomasvs> another ssssssssoul for sssssssatan [11:19] sxpert_work (~sx...@ra...) joined #gstreamer. [11:19] <teuf> as low volume as the linux kernel mailing list ? :) [11:20] <thomasvs> no, really low volume [11:20] <thomasvs> 10 mails a week [11:20] <teuf> ah cool [11:20] <teuf> I'll give it a try then, and unsubscribe when I'm bored [11:20] <thomasvs> or, you become a gst hacker [11:20] <thomasvs> that works too [11:20] <teuf> nah, you won't get me :) [11:21] <teuf> I already have too many things to do, and am not doing anything :-( [11:22] dolphy (~do...@po...) joined #gstreamer. [11:22] apoc_ (~ap...@dy...) joined #gstreamer. [11:25] <dolphy> morning [11:26] <thomasvs> hey [11:28] desrt (de...@ko...) left irc: Remote closed the connection [11:31] kvdb (~kv...@t-...) joined #gstreamer. [11:32] desrt (de...@ko...) joined #gstreamer. [11:34] <kvdb> Hi, I made a nice python script using ffdec_mpegvideo to play an mpeg. It does work, but the video is distorted (playing it in totem works well by the way) and gives errors like: Error while decoding slice, ac-tex damaged at 6 1. Is this an ffmpeg problem or gstreamer? [11:36] <thomasvs> it sounds like an ffmpeg problem [11:37] <thomasvs> have you tried with mpeg2dec ? [11:37] <kvdb> Not yet, what's the plugin name? [11:39] apoc (~ap...@dy...) left irc: Read error: 110 (Connection timed out) [11:40] <kvdb> It doesn't show up as an element in gst-editor so I may not have it installed? [11:40] foser (d0...@22...) joined #gstreamer. [11:41] <thomasvs> yeah [11:41] <thomasvs> plugin name is mpeg2dec [11:41] <thomasvs> what dist ? [11:41] <kvdb> debian (but I can't apt-get anything since I've been messing with pinning and dependencies) [11:42] <kvdb> debian tells me it's an application, not a plugin. I expect it to be wrapped into a plugin for gst then. [11:42] <foser> thomasvs: does a recent mpeg2dec snapshot work for you (not buildtime,but runtime) ? [11:42] <thomasvs> gstreamer-mpeg2dec is an application ? [11:42] <thomasvs> foser: no, it doesn't [11:42] <thomasvs> foser: anything beyond 0.3.1 fails for me [11:42] <thomasvs> we need to ping walken about that [11:43] <kvdb> thomasvs: right, I'll try to install that one now and report back [11:47] <kvdb> thomasvs: hey thanks, this one works on most of my test mpg's. Should I file a bug report at ffmpeg about the problem? [11:51] <foser> thomasvs: i was fooling around with the gsmdec last night and ended up with 'osssink capsnego failed' .. i suppose caps could not be set, but how do i find the exact problem ? [11:52] <thomasvs> kvdb: no, file a bug report with us [11:52] <thomasvs> kvdb: put a link to an mpeg file that works in mpeg2dec but not in our ffmpeg mpeg [11:53] <thomasvs> foser: using debug output, check --help for the right options [11:53] <thomasvs> foser: chances are gsmdec is having old caps [11:53] <foser> thomasvs: i did .. not much help.. what are old caps and whats a proper example of 'new caps' ? [11:55] <thomasvs> foser: gst-inspect osssink tells you what caps it accepts on it's sink pad [11:56] <thomasvs> gst-inspect gsmdec tells you what caps gsmdec can output [11:56] <thomasvs> compare the two [11:57] <thomasvs> foser: do you have a gsm test file ? [11:57] <thomasvs> gst-launch sinesrc ! gsmenc ! gsmdec ! osssink works fine for me [11:58] <foser> hmm so it does for me.. [12:01] walken (fo...@no...) left irc: "l8r" [12:02] <foser> thomasvs: but if i encode to a file and then try to play the file it doesn't work.. or am i missing something here [12:14] <thomasvs> I'm not sure -time to bugzilla it :) [12:21] Action: sxpert_work is away: coworkers call for lunch [12:25] <kvdb> thomasvs: bug filed [12:30] TD (~mi...@m1...) joined #gstreamer. [12:39] <dolphy> thomasvs: ping ? [12:39] robla (~ro...@12...) got netsplit. [12:39] pigeon (~pigeon@203.206.217.82) got netsplit. [12:39] dilinger (ir...@sl...) got netsplit. [12:41] robla (~ro...@12...) returned to #gstreamer. [12:41] dilinger (ir...@sl...) returned to #gstreamer. [12:45] pigeon (~pigeon@203.206.217.82) got lost in the net-split. [13:04] abraar (~foo@AToulouse-105-1-24-31.w81-50.abo.wanadoo.fr) left irc: Read error: 110 (Connection timed out) [13:06] abraar (~foo@AToulouse-105-1-14-172.w80-15.abo.wanadoo.fr) joined #gstreamer. [13:24] ain (~ia...@us...) joined #gstreamer. [13:31] pb_ (~pb...@ds...) left irc: "bbl" [13:44] md` (~illuminat@B6ee5.pppool.de) joined #gstreamer. [13:47] Action: sxpert_work is back (gone 01:25:33) [13:48] <thomasvs> dolphy: pong [13:53] ain (~ia...@us...) left irc: "leaving" [13:53] teuf (~te...@gr...) left irc: "Client exiting" [14:00] <dolphy> thomasvs: want to see some shots of getcontrol ? [14:01] <dolphy> thomasvs: http://dolphy-tech.net/files/getcontrol1.png [14:01] <dolphy> thomasvs: http://dolphy-tech.net/files/getcontrol2.png [14:01] <dolphy> thomasvs: http://dolphy-tech.net/files/getcontrol3.png [14:02] ain (~ia...@us...) joined #gstreamer. [14:02] <swentel> mm nice interface :) [14:02] <thomasvs> dolphy: yeah, getting good :) [14:03] <ain> mornin [14:05] <dolphy> thomasvs: i hope to finish photos browsing tomorrow [14:06] <dolphy> thomasvs: it s day off in spain [14:06] <thomasvs> lucky you [14:07] <dolphy> http://monkey-bubble.tuxfamily.org/ [14:07] <dolphy> a game using gst [14:08] <thomasvs> i knew about the game, but it uses gst ? [14:08] <thomasvs> that's great :) [14:09] <kvdb> I just had a discussion in #pygame about the differences between gst and pygame (using SDL). So what's the main difference? Is gst aiming at being a more complete SDL? I'm confused ;-) [14:09] <kvdb> The 2 projects seem to have quite some overlap [14:09] <thomasvs> SDL is a drawing library [14:09] <thomasvs> gst is a multimedia streaming framework [14:09] <thomasvs> I don't see much overlap [14:09] <ain> damn, thomasvs beat me to it [14:10] <kvdb> uhm, no I was refering to gst <> pygame. They also allow to play sounds and movies and have interaction. [14:11] <thomasvs> I don't know about pygame [14:11] <ain> I would imagine pygame is a case of "play this sound" [14:11] <ain> gstreamer is a media framework [14:12] <kvdb> yes well, i'm comparing technolgies for my project. I'm really expecting much of flash interactivity which is in the pipeline for gst. At first, I thought gstreamer is not really for designing interactive applications, but then I heard about the flash thing. [14:13] <thomasvs> yeah [14:13] <thomasvs> well, the flash library we use doesn't have interactivity yet [14:13] <thomasvs> ds was thinking of adding that though [14:13] <thomasvs> you should ask him about his ETA for that [14:13] <kvdb> yeah, i should [14:13] <thomasvs> not really sure why you want to use flash for games though [14:14] <ain> its the new visual basic of the windows world [14:14] <kvdb> well I'm not really designing a 'game'. It's an educational application. When it's for kids it needs to be fancy. So lots of multimedia elements. But I also need internet connectivity and database support so it's not 'just' a game. [14:17] <kvdb> ain: what's wrong with flash? graphic designers are no programmers. and there's an open source flash thing (swish?) right? [14:18] <swentel> you can't really do a lot with that open source thing iirc [14:20] <TD> swish is not open source [14:20] <TD> or wasn't last time i checked [14:22] TD (~mi...@m1...) left irc: "leaving" [14:22] <kvdb> I saw this in gst wiki: SVG player (free macromedia flash alternative) [14:24] voltron (~vo...@20...) joined #gstreamer. [14:25] <ain> what's wrong with flash? nothing much, but its the "Oh, I wrote a game" langauge of the new millenium [14:25] <ain> even though 99% of the games are exactly the same [14:25] <kvdb> But anyway. I still don't get the difference between the multimedia framework that gst is and pygame. From my point of view, I can use both to say: I want movie a.mpg played in the right upper corner of my application. What 'multimedia kit' would you recommend to me? [14:25] <ain> pygame in that case [14:33] <kvdb> ain: I think pygame would work, but SDL feels like old technology. If I would like an interactive scene in the application. I would have to code it myself using pygame. I think it's nice to have that abstracted away with gst gets svg/flash support. [14:33] <kvdb> WHEN it gets svg/flash support i mean. [14:44] <ain> whats a loop based element I can put into a simple pipeline thats not alsasink? [14:47] <ain> can someone test http://discomachinegun.prettypeople.org/~iain/test-gst.c for me please? [14:48] <ain> run it as ./test-gst file.mp3 osssink [14:48] <ain> (replacing osssink with whatever audiosink you want) [15:07] sublett (~su...@21...) joined #gstreamer. [15:37] <foser> ain: can't resolve it [15:39] <ain> really? [15:39] <ain> aww bugger [15:40] <ain> 62.188.60.41 [15:40] <ain> is the ip [15:41] Action: ain resolves to get a new DNS [15:50] <foser> ain: what should it do ? [15:50] <ain> play the mp3 [15:50] <ain> and stop after 10 seconds [15:51] <foser> hmm no.. it goes terribly slow heer [15:51] <foser> oh wait [15:51] <foser> nah doesnt stop [15:51] <foser> 0.6.3 here [15:52] <ain> it doesn't do anything for me on 0.7 [15:52] <foser> gtg now.. [15:52] <ain> thanks [16:03] sublett (~su...@21...) left irc: "Client exiting" [16:26] TD (~mi...@m4...) joined #gstreamer. [16:38] yippi (~br...@nw...) joined #gstreamer. [16:46] gstreamer (~gh...@ro...) joined #gstreamer. [16:46] md` (~illuminat@B6ee5.pppool.de) left irc: Read error: 110 (Connection timed out) [17:01] gstreamer (~gh...@ro...) left #gstreamer. [17:05] Nick change: harshyZ -> harshyWORK [17:11] swentel (sw...@d5...) left irc: [17:18] kvdb (~kv...@t-...) left irc: "Client exiting" [17:25] foser (d0...@22...) left irc: "[ I want to believe ]" [17:25] foser (d0...@22...) joined #gstreamer. [17:54] walters-ny (wa...@ve...) joined #gstreamer. [17:55] robla (~ro...@12...) left irc: "Client exiting" [18:05] <ds-work> moo [18:05] <thomasvs> moink [18:10] <walters-ny> boing [18:13] <ain> spooge [18:20] <ain> does anyone know how to update the alsa in linux-2.6-test4 to CVS? [18:20] <ain> do you just copy the alsa-kernel dir into the source tree? [18:25] TD (~mi...@m4...) left irc: Read error: 104 (Connection reset by peer) [18:27] TD (~mi...@m1...) joined #gstreamer. [18:46] <walters-ny> so when is the next gst out? [18:47] <ds-work> 0.6.4? [18:47] <ds-work> unlikely soon, since nobody is bothering to merge the bug fixes [18:47] <ds-work> 0.7.1 got lost in the mail [18:48] <ain> at least company's back now [18:48] <taaz> i heard a rumor that someone in ny would do the release for us [18:49] <walters-ny> heh [18:49] <walters-ny> i am not a very good gst hacker honestly [18:50] <walters-ny> i mean i can sort of do it, but my time is going to be much more productively spent hacking rb than gst [18:50] <taaz> uh.. someone just send me "Dr. Lehn" (not quite yet i'm afraid) a query about making a video wall... [18:50] <taaz> where's Zeenix... [18:50] <walters-ny> ok [18:51] <walters-ny> well rb 0.5.4 is going to do the same hack that gst-launch-ext does and look at the extension of the filename and hardcode the decoder [18:51] <taaz> walters-ny: are you a good apt-proxy hacker? fix my setup please ;) [18:51] <walters-ny> taaz: nope :) [18:51] <TD> is spider really that broke? [18:51] <walters-ny> TD: it's just not reliable enough yet [18:52] <TD> are there any docs on what algorithms it uses [18:52] <thomasvs> walters-ny: let's just write an audio-only autoplugger [18:52] <walters-ny> TD: basically AFAIK it just tries all the typefinding functions in the plugins iteratively until it hits something [18:52] <walters-ny> thomasvs: that would be reasonable i think [18:52] <TD> ok [18:52] <walters-ny> thomasvs: but the problem seems to be more with the mp3 typefinding function [18:53] <walters-ny> rather than it deciding it's an mpeg video or something [18:54] <ain> TD: yeah, it just runs all the plugins it can and sticks them all together until it gets the end points connected, then it goes and uses the shortest one [18:54] <TD> that does sound suboptimal [18:54] <TD> but i seem to remember saying that it should use A* before :) [18:54] <TD> time, it all boils down to time :( [18:55] Action: ds-work got a "Dr. Comedi" email the other day [18:55] <TD> maybe if somebody else has some time for autoplug lovin i could point them in the right direction? [18:56] <ds-work> a spider replacement is on my list, at about #4 [18:57] <TD> how are you intending to implement it? [18:57] <ds-work> caps rewrite, interactivity + swfdec hacking, sound server, spider replacement [18:58] <foser> walters-ny: what src plugin is needed for rb radio to work .. httpsrc i assume ? [18:59] <ds-work> trial & error, plus media regression testing :) [18:59] <walters-ny> foser: gnomevfssrc [18:59] <foser> hmm [19:00] <TD> ds-work: have you looked at the a* algorithms? [19:00] <ds-work> TD: no [19:00] <ds-work> TD: it's really not that hard [19:00] <ds-work> TD: we generally only have one possible path [19:00] <TD> yeah. but you don't know what that path is beforehand [19:00] <foser> walters-ny: hmm just locks up for me .. any other special plugins ? [19:01] <TD> a* isn't a complex algorithm but it's good at pathfinding [19:01] <TD> used in games and such to get from A to B, but it's generic, basically traverses nodes on a graph [19:01] <walters-ny> foser: hmm...possibly the site is using an HTTP redirect which gnomevfs doesn't support [19:02] <walters-ny> foser: or alternatively it's just timing out trying to connect to the site [19:02] <ds-work> TD: the algorithm is secondary. The hard part is finding the exact media type at each step [19:02] <foser> walters-ny: i'm using one of the default provided stations .. which one does work for you ? [19:02] <walters-ny> foser: bassdrive works [19:02] <TD> ds-work: mimetype sniffing the stream you mean? [19:03] <foser> walters-ny: and so it does.. duh [19:03] <ds-work> TD: not only that, but also other elements checking the media type along the way [19:03] Action: walters-ny generally only listens to bassdrive, digitially imported, and groove salad [19:03] <ds-work> TD: i.e., mad decoding into raw audio at a specific rate, or specific number of channels [19:04] <TD> ok [19:04] <TD> walters-ny: do you know when you'll implement different streams per station? [19:04] <TD> walters-ny: at the moment i can't use rb to listen to the radio as it always chooses 128k streams [19:05] <walters-ny> TD: you can add different qualities [19:05] <walters-ny> there's not much more we can do [19:05] <taaz> walters-ny: is there a rb text mode client? ;) [19:06] <walters-ny> heh [19:06] <TD> walters-ny: well, being able to choose the default in the preferences and shipping with lower bandwidth defaults would be good [19:06] <walters-ny> prolly yeah [19:08] <TD> walters-ny: btw, did you get that minipatch i sent? [19:08] <walters-ny> minipatch? [19:08] <walters-ny> hm [19:08] <TD> it was like 1 line to HACKING [19:08] <walters-ny> oh...right [19:10] <walters-ny> * commited wa...@rh...--2003/rhythmbox--mainline--1.0--patch-76 [19:10] <walters-ny> TD: thanks for the reminder [19:11] <TD> np. it was more an excuse to fiddle with arch really, though the file was incorrect... [19:25] Nick change: sxpert_work -> not-here|sxpert [19:25] Action: not-here|sxpert is away: time to go home [19:25] walters-ny (wa...@ve...) left irc: "airport" [19:29] md` (~illuminat@B71af.pppool.de) joined #gstreamer. [19:36] dolphy (~do...@po...) left irc: "Network down, IP Packets delivered via UPS" [19:42] Zeenix (~zak@203.135.11.202) joined #gstreamer. [19:43] pb_ (~pb@2002:5160:45ef:0:2e0:7dff:fe74:8b87) joined #gstreamer. [19:43] <Zeenix> hi [19:53] yippi (~br...@nw...) left irc: Remote closed the connection [19:56] <taaz> Zeenix: i got a email for you [19:57] <taaz> someone asking me about gst video wall [19:57] <Zeenix> taaz: send it in, i am now used to of such emails :) [20:05] md` (~illuminat@B71af.pppool.de) left irc: "Client exiting" [20:08] <Zeenix> any know how a gconf client access the gconf database directly (not using the gconfd) ? [20:11] <ds-work> gconftool-2 [20:12] <Zeenix> ds-work: what? that is a client program itself, is'nt it? i need my own client to.. [20:14] <taaz> Zeenix: email addr? [20:16] ds-work (ds...@co...) left irc: "they paved paradise, and put up a slashing dot." [20:19] <Zeenix> taaz: za...@ya... [20:19] <Zeenix> taaz: but always drag anyone to read the howto first [20:20] wheels (~sc...@ds...) joined #gstreamer. [20:23] ds-work (ds...@co...) joined #gstreamer. [20:28] Eldhrin (~ma...@cp...) joined #gstreamer. [20:29] TD (~mi...@m1...) left irc: Read error: 104 (Connection reset by peer) [20:34] Action: Zeenix is impressed by MIT's OCW project [20:34] <Zeenix> and he currently dowloading the lectures of his favourate subject.. [20:35] <Zeenix> s/he/me [20:39] dolphy (~do...@21...) joined #gstreamer. [20:40] sublett (~su...@21...) joined #gstreamer. [21:17] ChristianHJW (~chr...@p5...) joined #gstreamer. [21:19] BBB (~rbultje@213.160.215.2) joined #gstreamer. [21:25] <Zeenix> what the hell is an .idle file? [21:26] <taaz> one that doesn't move very fast [21:26] <taaz> could be for python ide called IDLE [21:27] <Zeenix> actually gconftool's source contained GConfX.idl which is strange from inside, it looks like this: [21:28] <Zeenix> interface ConfigServer { [21:28] <Zeenix> ConfigDatabase get_default_database (); [21:28] <Zeenix> [21:28] <Zeenix> long ping(); [21:28] <Zeenix> void shutdown(); [21:28] <Zeenix> }; [21:28] <taaz> interface definition language [21:29] <taaz> it's a corba thing [21:31] Action: oboolo|sxpert is back (gone 12:11:56) [21:31] Nick change: oboolo|sxpert -> sxpert [21:39] TD (~mi...@m1...) joined #gstreamer. [21:47] madcat (~jrrt@195.174.67.207) joined #gstreamer. [21:48] <madcat> lala [21:53] sublett (~su...@21...) left irc: "Client exiting" [21:56] Eldhrin (~ma...@cp...) left irc: "leaving" [21:58] Company (~Company@pD9E33ED6.dip.t-dialin.net) joined #gstreamer. [22:04] <abraar> BBB: Hey. I was told to talk to you about a problem I have. I'm not able to seek in video files (.mpeg, .avi) with totem 0.99.5 (using gstreamer 0.63 as a backend). Same problem with gst-player. All my GNOME setup is built with jhbuild in case that matters. [22:06] sxpert (~sx...@sx...) left irc: Read error: 104 (Connection reset by peer) [22:06] sxpert (~sx...@sx...) joined #gstreamer. [22:08] <BBB> abraar: it works for me with gst-player... please re-try with gst-player, I have no clue on how totem works [22:09] <BBB> the gst backend in totem is badly maintained anyway [22:11] <abraar> BBB: Just tried with gst-player and I don't even have sound. That's weird. I seem to recall that I used to. [22:11] <abraar> BBB: I have some DEBUG output in the console. I can paste it in /msg if that's OK with you. [22:14] <BBB> debug output doesn't really help... for such problems, I need to be able to reproduce it [22:14] <BBB> and seeking works perfectly well for me :/ [22:15] <abraar> BBB: :/ -- Maybe a plugin I should gst-inspect? Or anything that comes to your mind? [22:17] <BBB> if avi doesn't work, something is badly screwed [22:17] <BBB> avi always works... if it doesn't, something is badly wrong [22:21] <abraar> BBB: Heh, OK. :/ [22:22] <abraar> Thanks for your time. [22:27] <thomasvs> abraar: are you on stable or HEAD ? [22:27] <thomasvs> BBB: and you ? [22:28] <abraar> thomasvs: I'm on the 0.6 branch of gstreamer if that's what you mean. [22:28] <abraar> (gnome24 moduleset in jhbuld) [22:29] <BBB> er... I'm on nothing [22:29] <BBB> I'm doing overlay-work [22:29] <BBB> and it works :) [22:29] <BBB> I have a re-make of xawtv based on gstreamer ;) [22:29] <thomasvs> BBB: you said gst-player works for you [22:29] <thomasvs> BBB: I'm assuming you are on HEAD, not stable [22:29] <BBB> ... last time I tested it [22:29] <BBB> it's on HEAD, yes [22:29] <thomasvs> right [22:30] <thomasvs> so there's the problem :) [22:31] <abraar> thomasvs: So, question is, is it safe for user lambda to go on HEAD? [22:31] <thomasvs> abraar: no, it's not [22:31] <thomasvs> abraar: it should be fixed, whatever it is [22:32] <thomasvs> abraar: I'm writing a testsuite, you can soon run that on your pc and we can compare if our gst's decode the same way [22:32] <abraar> thomasvs: Good news. Where will the testsuite be? Will it go into the 0.6 branch? [22:33] <madcat> thomasvs, how big is the api difference between 0.6 and head? [22:33] <thomasvs> abraar: yeah [22:33] <thomasvs> madcat: not very big [22:33] <thomasvs> madcat: run a diff on the headers [22:33] <madcat> that's nice. [22:33] <Company> debugging is different, the rest is pretty much the same [22:33] <madcat> and is there a planned date for release? [22:34] <Company> after we can watch dvds [22:34] <Company> gnome 2.6 i think [22:34] <madcat> ok [23:06] Galt (ga...@ca...) joined #gstreamer. [23:06] sublett (~su...@21...) joined #gstreamer. [23:10] <desrt> hi, sublett. [23:15] <ds-work> what happened to the 0.7.1 release? [23:15] <BBB> I think uraeus was going to do it [23:15] <BBB> wasn't he? [23:15] <Company> nobody does it [23:15] Action: BBB doesn't know [23:15] <ds-work> thomasvs: btw, we should work on a 0.6.4 release [23:15] <ds-work> ok, then I'll do it :) [23:15] <BBB> oh, I've got a working gst-based xawtv-replacement [23:15] <BBB> :) [23:16] <BBB> it rocks, and it's dead easy [23:16] <ds-work> BBB: does overlay always work? [23:16] <BBB> as in...? [23:17] <thomasvs> 0.7.1 needs fixing, no ? [23:17] <thomasvs> it doesn't do any video for me [23:19] <Company> right [23:19] <Company> ffmpeg or spider is broken [23:19] <Company> or something else ;) [23:19] <ds-work> BBB: does it work for every combination of input devices and video cards? [23:19] <Company> i remember that error message [23:23] <Galt> herm, would this be an appropriate place to ask a stupid question that's probably just me misunderstanding things? :) [23:23] ChristianHJW (~chr...@p5...) left irc: Client Quit [23:23] sublett (~su...@21...) left irc: Read error: 60 (Operation timed out) [23:27] <Company> sure [23:27] <Company> you can go in #flirt or something alternatively for those questions [23:27] <Company> ;) [23:27] <TD> there is a #flirt? [23:27] <TD> oh, too bad :D [23:28] <Company> there is no #flirt, i just checked - just a bad guess of mine [23:28] <Company> he'd have to go to efnet or so [23:29] <BBB> ds-work: no [23:30] <Galt> heh. mostly was wondering... should a GstBin emit an EOS when all the child elements have EOSed? i kind of got the impression they did from the 0.6.1 release notes [23:30] <ds-work> BBB: and the dumb followup question: can you determine when it won't work? [23:30] yippi (~br...@nw...) joined #gstreamer. [23:30] <ain> ** (process:16025): WARNING **: Can't find a decoder for type audio/mpeg [23:30] <ain> is mad not a decoder for audio/mpeg now [23:30] <BBB> ds-work: yes [23:30] <Company> Galt: yes, they're supposed to do that [23:30] <ain> (from media-info-test) [23:31] <BBB> ds-work: and next question: will I implement a fallback? sure [23:31] yippi (~br...@nw...) left irc: Client Quit [23:31] <BBB> but I didn't do that just yet [23:31] <Company> ain: not in stable [23:31] <BBB> but I know how to do it [23:31] <ain> in HEAD [23:31] <Company> ain: in stable it decodes audio/x-mp3 only [23:31] <ds-work> BBB: actually, the next question was: is it in cvs? [23:31] <Galt> company - herm. okay, i thought so. now the question is whether there's something wrong with my gstreamer install, the ruby/gstreamer bindings, or my code [23:31] <ds-work> BBB: feel free to overwrite gst-tv [23:32] <Company> Galt: i'll leave that up to you :) [23:32] <Company> Galt: try running it with --gst-debug=GST_STATES=5 [23:32] <Company> stop [23:32] <Company> try running it with the 0.6 equivalent [23:32] <Company> --gst-mask=0x40 i think [23:33] <ds-work> ugh, why aren't gtk signals documented as to what they do? [23:33] <Company> that'll output state changes [23:33] <Company> ds-work: because gtk isn't that much better in documentation than gst [23:33] <Company> and our signals aren't documented either :p [23:34] <Galt> herm, i wouldn't imagine that flag would necessarily work with a ruby app [23:34] <Company> no idea [23:35] <Company> theres a reason i use C [23:36] <Company> i dont need no freakin bindings for libc/kernel access [23:36] <Galt> i can always call gst_set_debug and see if that shows anything useful :) [23:36] <Company> something like that [23:37] <Company> and file a bug with the ruby people so you can get the command line flags back ;) [23:37] Action: ds-work notes that he never decolourised gst debugging [23:37] <Galt> I'm too lame for C ;) [23:37] <Company> --gst-debug-no-color [23:37] <Company> looks boring [23:38] <ds-work> Company: well, included in that was writing some perl (or python) scripts to colorize debug output and/or convert it to html [23:38] <Galt> Herm, don't think I'm going to tackle this right now, since I should be going home soon to make dinner for the wife. [23:38] <ain> else if (strcmp (mime, "audio/x-mp3") == 0) [23:38] <ain> factory = g_strdup ("mad"); [23:39] <Company> ds-work: how about writing a debug handler, that spits out xml of some sort? [23:39] <ain> I guess someone didn't update media-info.c to handle the audio/mpeg change then [23:39] <Company> go ahead and do it [23:39] <ain> doing so :) [23:39] TD_ (~mi...@m1...) joined #gstreamer. [23:39] <thomasvs> ain: just sync back media-info from nautilus-media [23:39] <Company> or use spider :o [23:39] <thomasvs> ain: just copy source code over [23:39] <thomasvs> ain: media-info is where it will end up, but the main source is in nautilus-media [23:39] TD (~mi...@m1...) left irc: Nick collision from services. [23:40] <Galt> just sticking callbacks on the pipeline to spit out state changes seems to indicate the pipeline is never spitting out EOS and ends up at STATE_PAUSED when the children elements EOS. it's a pain in the ass :) [23:40] Nick change: TD_ -> TD [23:41] <ain> thomasvs: oh, ok [23:42] <BBB> ds-work: I'll commit it during one of the next few days [23:42] <Company> Galt: that's bad [23:42] <ain> cp media-info.[ch] is a command you don't want to run [23:43] <Company> Galt: probably a bug in GStreamer with not propagating the EOS signal correctly [23:43] <Galt> i may give the 0.7.x branch a try, just for shits and giggles. it's not like i'm not already depending on development code [23:44] <Company> we should better find the bug [23:46] <Galt> Regardless, I'll see whether it I can localize where the problem is sometime tommorow. [23:47] <ain> thomasvs: I had to fix come compile warnings for it to compile in GST...want me to commit them to nautilus-media as well? [23:47] <Galt> the gstreamer portion is pretty straightfoward... a GstThread w/ gnomevfssrc >> spider >> osssink (well, eventually i'll have it actually look at the gconf setting for what sink to use, but it's hardcoded right now) [23:54] <BBB> ds-work: do we have a denoiser? [23:54] <ds-work> BBB: no [23:54] <ds-work> BBB: I have some prototype code lying around somewhere [23:55] <ds-work> it worked disturbingly well [23:55] <BBB> commit please ;) [23:55] <ds-work> it was back in the 0.3.3 days [23:55] <ds-work> I'd have to actually find it [23:56] <ds-work> it might be easier to rewrite it :) [23:56] <ain> ohhh a denoiser...that sounds like it could be a useful marlin plugin :) [23:56] <ds-work> it was for video, and assumed NTSC input [23:57] <BBB> dude, we really need to make a system for video norms [23:57] Action: ds-work whistles a happy tune and stares off into space [00:00] --- Wed Sep 24 2003 [00:03] <BBB> norm=GST_PROPS_STRING("ntsc"/"pal"/"secam") [00:03] <BBB> is that so hard? [00:15] tjansen (~tj...@zw...) left irc: Read error: 54 (Connection reset by peer) [00:19] ain (~ia...@us...) left irc: "Lost terminal" [00:25] TD (~mi...@m1...) left irc: "leaving" [00:25] TD (~mi...@m1...) joined #gstreamer. [00:29] ain (~ia...@us...) joined #gstreamer. [00:33] <ain> what was GST_FORMAT_UNITS replaced by? [00:34] <ds-work> ain: in most places, _DEFAULT, iirc [00:37] <ain> ok, thanks [00:43] BBB (~rbultje@213.160.215.2) left irc: "Client exiting" [00:51] dolphy (~do...@21...) left irc: "Network down, IP Packets delivered via UPS" [01:02] Action: wheels is away: sleep [01:24] pb_ (~pb@2002:5160:45ef:0:2e0:7dff:fe74:8b87) left irc: Read error: 110 (Connection timed out) [01:33] <sxpert> the shuffling in rb is totally bogus :( [01:36] foser (d0...@22...) left irc: "[ I want to believe ]" [01:36] TD (~mi...@m1...) left irc: "leaving" [01:50] <Zeenix> bye guys [01:51] Zeenix (~zak@203.135.11.202) left irc: "REDHAT + SUN + ORACLE + IBM, may God help M$" [02:11] Company (~Company@pD9E33ED6.dip.t-dialin.net) left irc: Read error: 104 (Connection reset by peer) [02:12] Company (~Company@pD9E33ED6.dip.t-dialin.net) joined #gstreamer. [02:15] thaytoo (ja...@ad...) joined #gstreamer. [02:38] alley_cat (AlleyCat@pD9E61D53.dip.t-dialin.net) left irc: Read error: 110 (Connection timed out) [02:42] <ds-work> Company: do you if dolphy ever finished his xvideosink changes? [02:42] <Company> no idea [02:43] <Company> im out of the loop since barcelona [02:43] <ds-work> yeah, I think we all feel that way [02:43] walters (wa...@ve...) joined #gstreamer. [02:43] alley_cat (AlleyCat@pD9E61337.dip.t-dialin.net) joined #gstreamer. |