[gst-devel] Daily IRC logs

[<wim...@ch...>]

[06:32] <dichro> yay! v4lsrc and jpegenc negotiated caps!
[06:32] Action: dichro does a little dance
[06:34] <omega_> oooh
[06:36] <dichro> now if I can just stop it from FPEing...
[06:36] Action: omega_ thinks libjpeg needs a rewrite
[06:36] Action: dichro nods vigorously
[06:36] Action: omega_ would like to take that project up with all the other codec stuff
[06:36] <omega_> if I can get a month solid to focus on that
[06:37] <dichro> I was going to take a hatchet to it and just make it work for my specific case. Given that it currently doesn't work for *any* case, as far as I can tell...
[06:39] <dichro> caps question. For v4l to do anything useful, it needs to negotiate caps with something that has depth and endianness set.
[06:39] <omega_> probably, yes
[06:39] <dichro> is it fair to require that all possible video sinks know about those parameters? jpegenc by default only knows about format and dimensions.
[06:39] <omega_> hmmm
[06:39] <omega_> depth is?
[06:40] <dichro> 24 bpp, for example.
[06:40] <omega_> ah, ok
[06:40] <omega_> there are going to be a significant set of properties for all video
[06:40] <omega_> see dv/dvdec.c for two examples
[06:41] <dichro> is that a canonical list?
[06:41] <omega_> no
[06:42] <dichro> Is there a canonical list? :)
[06:42] <omega_> searching
[06:42] <omega_> no, but there might be a start of one
[06:44] <omega_> http://gstreamer.net/docs/cvs/gst-plugin-writers-guide/cha-basic-types.html
[06:45] <omega_> is where it's supposed to be,  but it's not even hardly started yet
[06:46] <dichro> okay, that's useful, even if out of date.
[06:46] <omega_> the fourcc is the basic unit, everything else comes from that
[06:46] <omega_> which is a reason that the capsnego stuff needs to be done quite differently, if nothing else
[06:47] <dichro> how so?
[06:47] <omega_> because everything that has fourcc will want to settle on that first
[06:47] <omega_> else they may find themselves settling on the red mask before they decide that YUV is the better basic format to use
[06:48] <dichro> and I take it that red mask is not relevant to YUV?
[06:48] <omega_> but then.. caps will probably eventually come with a general protocol/sequence for negotiation anyway
[06:48] <omega_> quite
[06:48] <omega_> YUV doesn't have red
[06:48] Action: dichro is a total graphics ignoramus :)
[06:48] <omega_> www.webartz.com/fourcc for a primer
[06:49] <omega_> on fourcc, at least
[08:14] Nick change: ajbusy -> ajmitch
[08:25] Nick change: ajmitch -> ajbusy
[08:30] Yoshi (ro...@co...) joined #gstreamer.
[08:38] dichro (di...@fo...) left irc: Leaving
[08:44] dichro (di...@fo...) joined #gstreamer.
[08:45] <dichro> woohoo! working jpegenc from v4lsrc! butt-ugly hack job on the code, but very satisfying nonetheless :)
[08:51] steveb (steveb@24.132.238.49) joined #gstreamer.
[09:03] _gst_newt_ joined #gstreamer.
[09:03] <chillywilly> hi ppl
[09:12] matth-away (ma...@qw...) got netsplit.
[09:12] chillywilly (bau...@d8...) got netsplit.
[09:13] matth-away (ma...@qw...) returned to #gstreamer.
[09:13] chillywilly (bau...@d8...) returned to #gstreamer.
[09:23] steveb (steveb@24.132.238.49) left irc: Read error to steveb[24.132.238.49]: EOF from client
[09:24] steveb (st...@no...) joined #gstreamer.
[09:31] dichro (di...@fo...) left irc: Leaving
[09:31] wtay-sleeping (wi...@ca...) got netsplit.
[09:31] taazzzz (dlehn@66.37.66.32) got netsplit.
[09:31] iGN_ (ig...@lo...) got netsplit.
[09:31] Happyfeet (ro...@d1...) got netsplit.
[09:31] Gandalf_ (ga...@tu...) got netsplit.
[09:31] taazzzz (dlehn@66.37.66.32) returned to #gstreamer.
[09:31] iGN_ (ig...@lo...) returned to #gstreamer.
[09:31] Gandalf_ (ga...@tu...) returned to #gstreamer.
[09:31] Happyfeet (ro...@d1...) returned to #gstreamer.
[09:31] wtay-sleeping (wi...@ca...) returned to #gstreamer.
[09:32] wtay-sleeping (wi...@ca...) got netsplit.
[09:32] taazzzz (dlehn@66.37.66.32) got netsplit.
[09:32] iGN_ (ig...@lo...) got netsplit.
[09:32] Happyfeet (ro...@d1...) got netsplit.
[09:32] Gandalf_ (ga...@tu...) got netsplit.
[09:32] matth-away (ma...@qw...) got netsplit.
[09:32] chillywilly (bau...@d8...) got netsplit.
[09:32] seth (se...@nu...) got netsplit.
[09:32] Yoshi (ro...@co...) got netsplit.
[09:32] seth (se...@nu...) returned to #gstreamer.
[09:32] Yoshi (ro...@co...) returned to #gstreamer.
[09:32] chillywilly (bau...@d8...) returned to #gstreamer.
[09:32] matth-away (ma...@qw...) returned to #gstreamer.
[09:32] wtay-sleeping (wi...@ca...) returned to #gstreamer.
[09:32] Happyfeet (ro...@d1...) returned to #gstreamer.
[09:32] Gandalf_ (ga...@tu...) returned to #gstreamer.
[09:32] iGN_ (ig...@lo...) returned to #gstreamer.
[09:32] taazzzz (dlehn@66.37.66.32) returned to #gstreamer.
[09:37] seth (se...@nu...) got netsplit.
[09:37] Yoshi (ro...@co...) got netsplit.
[09:37] seth (se...@nu...) returned to #gstreamer.
[09:37] Yoshi (ro...@co...) returned to #gstreamer.
[09:42] Yoshi (ro...@co...) left irc: Client Exiting
[10:24] omega_ (om...@om...) left irc: sleep
[11:22] chillywilly (bau...@d8...) left irc: Ping timeout for chillywilly[d81.as12.nwbl0.wi.voyager.net]
[11:23] chillywilly (bau...@d5...) joined #gstreamer.
[11:35] Uraeus (csc...@c2...) joined #gstreamer.
[11:36] <Uraeus> morning
[11:38] <Happyfeet> Das Boot is such an incredible movie
[11:42] Nick change: ajbusy -> ajmitch
[11:47] <Uraeus> it is
[11:48] <ajmitch> hi
[11:49] <Uraeus> ajmitch: hi, have you recouperated from last nights flamefest :)
[11:49] <ajmitch> Uraeus: wow there were heaps of people getting debian & loving it at the installfest today ;)
[11:50] <ajmitch> only one redhat user, and that just wasn't going right ;)
[11:51] <Uraeus> hehe, well at least 1 person knows what they are doing in NZ
[11:52] <ajmitch> yeah, i think they may have installed debian in the end... ;)
[11:52] <ajmitch> chillywilly: what distro do you use, btw?
[11:54] <Uraeus> ajmitch: but you know the official GStreamer distro is dsplinux :)
[11:55] <ajmitch> Uraeus: hehe
[11:55] <chillywilly> debian
[11:55] <chillywilly> cause it rocks
[11:55] <chillywilly> and is the most free
[11:56] Action: ajmitch wonders why the linux 2.4.4 patch is so big ;)
[11:58] <Uraeus> hmm, this looks bad, ok here is a deal, I send a Red fedora to anyone here switching to Red Hat :)
[11:59] Action: ajmitch wonders when linus will accept gstreamer into the kernel source...
[11:59] <Uraeus> well, considering the response to korbit, not soon I think :)
[12:00] <chillywilly> Uraeus: why deadcat?
[12:01] <Uraeus> huh?
[12:02] <chillywilly> deadcat == Red Hat
[12:04] <Uraeus> chillywilly: well, it is kept well up to date, it is the one I have gotten used to and Red Hat sponsors lot of great development in the Linux community and they have a strict GPL policy
[12:04] <ajmitch> Uraeus: i gather you are a free software fan then?
[12:05] <Uraeus> I am, and I love it when companies are able to build a business around free software
[12:05] <ajmitch> yup, it's great
[12:05] <Uraeus> that is why I have Caldera
[12:05] <Uraeus> s/have/hate/
[12:06] <ajmitch> hehe
[12:06] <Uraeus> since they are always trying to get ahead by adding proprietary extensions
[12:06] <ajmitch> debian rocks
[12:06] <ajmitch> caldera blows in that respect
[12:06] <ajmitch> well, their distro just sucks ;)
[12:06] Action: chillywilly is a free software zealot too
[12:07] <chillywilly> I have nothing against RedHat they do give back to the community
[12:07] <chillywilly> but debian is the community
[12:07] <chillywilly> ;)
[12:07] <Uraeus> well, I like Debian even if I am not personally inclined to use it, they are strong GNOME supporters too
[12:07] thomas (th...@ad...) joined #gstreamer.
[12:07] <thomas> morning
[12:07] <Uraeus> morning thomas
[12:08] <thomas> hmmm... do you know how the autoplug should work ?
[12:08] <thomas> I mean as it is know as opposed to last week
[12:08] <ajmitch> damn i like how i can just download patch files to update my kernel sources
[12:08] <ajmitch> better than a service pack ;)
[12:09] <Uraeus> thomas: nope not me, wtay knows for sure, but maybe ajmitch or chillywilly also do
[12:09] <ajmitch> Uraeus: hah, we are clueless, remember ;)
[12:10] <Uraeus> ajmitch: finally you admit it :)
[12:11] Action: chillywilly knows nothing of gstreamer just that the ppl are kewl ;), we have bigger fish to fry like GNU Enterprise
[12:12] <Uraeus> thomas: looks like you have to wait for wtay to wake up :)
[12:12] Nick change: Uraeus -> Ura_away
[12:14] <thomas> uraeus: I'll have to wait for myself to wake up as well
[12:14] Action: thomas is going to take a shower and hunt for small animals for food
[12:15] <ajmitch> thomas: no! leave your cat alone! ;)
[12:17] <chillywilly> hehe
[12:24] <ajmitch> why is it that we now get ads in the kernel config? ;)
[12:24] <chillywilly> ads?
[12:25] <ajmitch> goto help for reiserfs in filesystems
[12:25] <chillywilly> I would, but I am too lazy
[12:26] <chillywilly> damn these deals are where you have to buy a hotel accomodation package
[12:26] <chillywilly> which doesn't work when staying months
[12:27] <chillywilly> ;)
[12:27] <chillywilly> but for Tony it would work
[13:32] chillywilly (bau...@d5...) left irc: Philosophers and plow men, each must know his part, to sow a new mentality closer to the heart...
[13:54] ajmitch (me...@p5...) left irc: http://www.freedevelopers.net
[14:02] ajmitch (me...@p3...) joined #gstreamer.
[14:24] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee31.a2000.nl]
[14:33] wtay-sleeping (wi...@ca...) left irc: Read error to wtay-sleeping[cable-195-162-214-190.upc.chello.be]: EOF from client
[14:45] steveb (st...@no...) joined #gstreamer.
[15:10] Nick change: ajmitch -> ajzzzz
[15:42] ajzzzz (me...@p3...) left irc: Ping timeout for ajzzzz[p38-max2.dun.ihug.co.nz]
[15:52] jin (jin@AC9B78CA.ipt.aol.com) joined #gstreamer.
[15:52] jin (jin@AC9B78CA.ipt.aol.com) left #gstreamer.
[16:10] ajzzzz (aj...@p2...) joined #gstreamer.
[16:58] thomas (th...@ad...) left irc: Read error to thomas[adsl-63770.turboline.skynet.be]: EOF from client
[17:11] thomas (th...@ad...) joined #gstreamer.
[18:05] Nick change: matth-away -> matth_
[18:06] <matth_> anyone?
[18:14] <steveb> hello
[18:17] <matth_> hey... it's quiet here
[18:19] <steveb> yeah
[18:19] <steveb> lets make some noise
[18:24] <matth_> "who let the dog out / who / who"
[18:24] <matth_> ;-)
[18:26] <steveb> or not :}
[18:28] <steveb> hey, do you know if C boolean expressions are evaluated lazily?
[18:29] <matth_> what do you mean?
[18:32] <steveb> like (a || b) - if a is true then b won't even be evaluated
[18:35] <matth_> yes: b will not be checked if a is true
[18:35] <steveb> cool
[18:59] dobey (dobey@141.154.95.104) joined #gstreamer.
[18:59] <steveb> yo
[18:59] <dobey> yo
[19:15] Zeenix (programmer@203.135.12.240) joined #gstreamer.
[19:16] <Zeenix> one question. Does gstreamer uses EsounD underneath ?
[19:17] <dobey> it has an esd plugin, so it can
[19:18] <Zeenix> no if i dont want it to do that, will i be given an option
[19:19] <thomas> Zeenix: gstreamer is a programming platform
[19:19] <Zeenix> i mean if i dont want to use EsounD
[19:19] <thomas> it depends on the app that you're writing
[19:19] <dobey> uh, you have to explicitly say "use esound"
[19:19] <Zeenix> yes i know that
[19:19] <thomas> you can also use oss
[19:19] <dobey> or ^H^H^HaRtz^H^
[19:19] <thomas> or write to disk
[19:19] <Zeenix> yes i want to do that i mean using OSS
[19:19] <thomas> well, there's a plugin called osssink
[19:20] <Zeenix> then i think its great
[19:20] <dobey> or you can use videosink
[19:20] <dobey> heh
[19:20] <dobey> haha
[19:20] <Zeenix> is it linked with gnome
[19:20] Action: dobey refrains
[19:20] <Zeenix> can it be used for Text based( only ) Apps.
[19:20] <thomas> Zeenix: what do you mean ?
[19:20] <thomas> Zeenix: it *can* be used for text-based apps...
[19:21] <thomas> ... but you need to run an x-server currently.
[19:21] <thomas> because it uses gtk
[19:21] <thomas> there are two solutions to this :
[19:21] <thomas> a) run a dummy x-server like Xfvb
[19:21] <thomas> b) wait until gstreamer switches to glib which should be in the near future
[19:21] <dobey> open("/dev/dsp", "r")
[19:21] <thomas> I tried a) and it works.
[19:21] <thomas> and I'm living b).
[19:21] <thomas> ;)
[19:23] <Zeenix> untill then i'll be using it on gnome
[19:23] <Zeenix> i'll also try to learn gtk+
[19:23] <thomas> since omega needs it to run embedded on a no-graphics system
[19:23] <thomas> it will rather be sooner than later
[19:23] <thomas> I hope
[19:24] <Zeenix> does it offer sound data compression & decompression
[19:24] <dobey> uh, mp3/ogg
[19:24] <Zeenix> i'm trying to develope a voice chat app. in linux
[19:24] <dobey> voip?
[19:25] <Zeenix> is gstreamer a good choice for that
[19:25] <dobey> look at gnome-o-phone and whatever those other things are called too
[19:25] <dobey> i guess
[19:25] <Zeenix> thomas: what is your opinion
[19:25] <thomas> Zeenix: hard to tell.
[19:25] <Zeenix> whats gnome-o-phone
[19:26] <thomas> if you're planning on doing it across networks (duh)
[19:26] <thomas> then gstreamer needs network plugins
[19:26] <Zeenix> yes accross internet
[19:26] <thomas> which should be easy to do
[19:26] <thomas> do you also want video with that ?
[19:28] <Zeenix> i'll do it myself( using linux sockets )
[19:28] <Zeenix> i'm just asking if it helps for the compression part ?
[19:28] <Zeenix> no video ( not yet ).
[19:28] <Zeenix> after i develope that, i'll be using it for another project
[19:29] <thomas> Zeenix: well then there should not be any real problem
[19:29] <thomas> Zeenix: however maybe gstreamer is a bit overkill.
[19:29] <thomas> Zeenix: but if you plan on doing video with it at some point
[19:29] <thomas> then gstreamer is definitely the way to go
[19:29] <dobey> just use 24kbps encoding at 16bit 11000hz
[19:29] <Zeenix> i asked something about compression( does gstreamer offer that ).
[19:29] <thomas> Zeenix: yes it does
[19:29] <thomas> Zeenix: you can use mp3 or ogg
[19:29] <dobey> heh
[19:29] <thomas> or whatever plugin will be written in the future
[19:30] <Zeenix> i meant real time compression. Can it do that
[19:31] <thomas> Zeenix: what's the difference ? if your pc is fast enough then it can handle real-time compression.
[19:31] <thomas> Zeenix: or are you talking about encoder delay ?
[19:31] <dobey> hrmm
[19:32] <Zeenix> i mean that i dont want files involved. Just memory
[19:32] <dobey> then don't write to disk, just do it in ram
[19:32] <thomas> Zeenix: no problem then.  Just write a pipeline ...
[19:32] <thomas> ... write a small plugin to output data to a socket ...
[19:32] <Zeenix> thanks a lot for so much help thomas
[19:32] <thomas> (... or use the fdsink on a socket file descriptor) ...
[19:33] <thomas> and that's it
[19:33] <thomas> no problem
[19:33] <Zeenix> i mush ask about you ?
[19:33] <thomas> Zeenix: ?
[19:44] steveb (st...@no...) left irc: Ping timeout for steveb[node1ee31.a2000.nl]
[19:47] steveb (st...@no...) joined #gstreamer.
[19:57] Zeenix (programmer@203.135.12.240) left #gstreamer.
[20:09] lsetia (lsetia@203.197.210.66) joined #gstreamer.
[20:18] lsetia (lsetia@203.197.210.66) left irc: Read error to lsetia[203.197.210.66]: EOF from client
[20:34] wtay (wi...@ca...) joined #gstreamer.
[21:09] Nick change: ajzzzz -> aj_uni
[21:19] <thomas> wtay ?
[21:20] <wtay> sorta :)
[21:20] <thomas> busy doing something ?
[21:20] Action: wtay is configuring his new debian system
[21:20] <wtay> pretty busy
[21:21] <wtay> sendmail issues etc..
[21:21] <dobey> la la
[21:21] <thomas> ok, let me know when you're done
[21:21] <wtay> I think I have them solved though (knock on wood)
[21:21] <thomas> don't say that
[21:21] <thomas> ;)
[21:21] <wtay> I might have lost some mails in the process...
[21:36] richardb (ri...@ix...) joined #gstreamer.
[21:36] <richardb> Wooo
[21:37] <thomas> hi
[21:37] <thomas> how do I play mpeg2 videos with gstreamer-launch ?
[21:39] <richardb> Not sure.
[21:42] <richardb> Um: ./gstreamer-launch disksrc location="file" ! mpeg2parse ! mpeg2play ! xvideosink
[21:42] <richardb> I think that'll give you the video.
[21:42] <wtay> hmm no, it wont work just yet :(
[21:42] <richardb> ;-)
[21:45] <richardb> Someone reported that there was a build problem due to artsdsink.h
[21:45] <richardb> missing: anyone else seen this.
[21:45] <richardb> I think it was Uraeus who reported it.
[21:46] <richardb> artsdsink.h is in plugins/artsd/ and I think it should get put into tarballs and dists fine too.
[21:48] <richardb> I'll add an extra check for artsc.h though, in case arts-config is lying. ;-)
[21:48] Action: richardb goes to finish making dinner
[21:48] <thomas> hmmm... anyone know why xvideosink would crash on my machine ?
[21:49] <dobey> it's a mac?
[21:49] <dobey> heh
[21:51] <thomas> dobey: ;)
[21:51] <thomas> it's not a mac, since I'm typing and you can see it
[21:51] <dobey> gah
[21:51] <dobey> this is a mac
[21:51] <dobey> i need a free news server to use :-(
[21:55] <wtay> thomas: crash?
[21:57] <thomas> wtay: yeah, segfault
[21:57] <thomas> should explain why my gstmediaplay didn't want to either ;)
[21:57] <thomas> it's in memcpy in the chain function of xvideosink
[21:58] Ow3n (ow...@ti...) joined #gstreamer.
[21:58] <wtay> what are you playing?
[21:58] <wtay> yo
[21:58] <Ow3n> yo
[21:58] <thomas> hi
[21:58] <thomas> tools/gstreamer-launch disksrc location=/opt/media/mpeg1/everest.m1v ! mp1videoparse ! mpeg_play ! xvideosink
[21:58] <thomas> works fine with aalib
[21:58] <wtay> oh
[21:58] <thomas> incsched version
[21:59] Action: thomas curses at xmms crossfade for crashing again and again
[22:01] <wtay> oh ok, it's normal that it crashed
[22:01] <wtay> wel 'normal'
[22:01] <wtay> gstreamer-launch doesn't provide a toplevel window to embed the xvideosink widget in
[22:01] <thomas> oh ? so i need to give options to xvideosink then ?
[22:01] Action: thomas thinks gstreamer could do with a little more error fallback
[22:02] <wtay> thomas: you can't use xvideosink with launch
[22:02] <thomas> wtay: ok, what can I do then if I want to see video ?
[22:02] <thomas> gstmediaplay just outright crashes
[22:02] <wtay> hmm
[22:02] <thomas> not really, but it shows a dialog saying it did
[22:02] <wtay> any idea where?
[22:02] <thomas> and no core
[22:02] <thomas> well... only when it starts playing.
[22:02] <thomas> if I start it up it's ok.
[22:02] <thomas> When I select a file it crashes.
[22:03] <thomas> when I give it on the command line : same crash.
[22:03] <thomas> I'll try another kind of file.
[22:03] <thomas> hmmm... with an mp3 I get a small window "gstplay" with nothing in it ...
[22:03] <thomas> then and error window...
[22:04] <thomas> "Application lt-gstmediaplay has crashed due to a fatal error (segmentation fault)"
[22:04] <thomas> wtay: does your gstplay work ?
[22:04] <wtay> thomas: sure
[22:04] <wtay> thomas: not in incsched though
[22:04] <thomas> ah ok, maybe that's it.
[22:04] <thomas> i'll try the other one
[22:06] <thomas> no luck, same thing with CVS version
[22:06] <wtay> trying the same here
[22:07] <thomas> why doesn't it give a core file ?
[22:07] <wtay> no idea
[22:07] <wtay> must be a gnome thing
[22:07] <wtay> or maybe libtool
[22:08] <wtay> thomas: no warnings?
[22:08] <thomas> wtay: yeah, a few.
[22:09] <wtay> such as?
[22:09] <thomas> four assertions
[22:09] <wtay> aha
[22:09] <thomas> ** CRITICAL **: file gstbin.c: line 172 (gst_bin_add): assertion `element != NULL' failed.
[22:09] <wtay> oh ok
[22:09] <thomas> but it did those before too, no ?
[22:09] <wtay> tools/gstreamer-inspect colorspace
[22:09] Action: thomas blushes
[22:10] <thomas> it really would help if gstreamer were more verbose about these things
[22:10] <wtay> yeah, I suppose
[22:10] <thomas> is it hard to make gstreamer say : i need this plugin but you haven't got it ?
[22:10] <thomas> I suppose I have to download Hermes then ?
[22:10] <wtay> very easy
[22:10] <wtay> yes
[22:12] <thomas> wtay: what function should be responsible for checking if a plugin exists ?
[22:13] <wtay> well, it all depends actually...
[22:13] <wtay> I would let gstplay report an error
[22:13] <thomas> wtay: but it would be better if internal gst functions would set a global error
[22:14] <thomas> btw: gstplay now shows controls but then crashes
[22:14] <thomas> so that's better
[22:14] <wtay> hmm
[22:14] <wtay> not finding a plugin is not an error
[22:14] <wtay> the app could request another one then
[22:14] <thomas> ok you're right
[22:15] <wtay> any idea what caused the crash now?
[22:15] <thomas> cool... must be some kind of mem leak in gstplay while functions bounce around
[22:15] <thomas> I ran it with debug mask and I got a 10 MB error file
[22:15] <thomas> checking...
[22:16] <thomas> damn... colors are pretty but the output looks worse in plaintext
[22:16] <wtay> hmm, you run esdsink right?
[22:16] <wtay> er, xmms with sd
[22:16] <thomas> how can I tell ?
[22:16] <thomas> ah ok...
[22:16] <thomas> yeah
[22:16] <thomas> turn it off ?
[22:16] <wtay> yes, gstplay uses osssink
[22:17] <thomas> any particular reason ?
[22:17] <wtay> not really
[22:17] <wtay> gstplay.c line 683
[22:17] <thomas> ok still crashes
[22:17] <wtay> eventually all these hardcoded plugins should be config options
[22:18] <thomas> maybe the plugin structure should be cleaned up first ;)
[22:18] <thomas> it's hard to guess what plugins to use based on gstreamer-inspect
[22:18] <wtay> yeah
[22:18] <thomas> here are the last few lines of debug output...
[22:18] <thomas> DEBUG(30961:-1)gst_pad_renegotiate_func:1066: caps compatibility check fail
[22:18] <thomas> DEBUG(30961:-1)gst_pad_renegotiate_func:1103: calling negotiate function on pad queue_video_bin:src data: 0xbf404050
[22:18] <thomas> DEBUG(30961:-1)gst_pad_negotiate_proxy:1212: negotiation proxied from pad (queue_video_bin:src) to pad (queue_video_bin:sink)
[22:18] <thomas> DEBUG(30961:-1)gst_pad_renegotiate_func:1038: negotiating pad colorspace:sink and mpeg_play:src data:0xbf403fb0
[22:18] <thomas> DEBUG(30961:-1)gst_pad_renegotiate_func:1063: checking compatibility with pad mpeg_play:src
[22:19] <thomas> audio works now though
[22:19] <thomas> I mean an mp3
[22:19] <wtay> ok
[22:20] <thomas> hang on... Alien Song works as well
[22:20] <thomas> nice
[22:20] <wtay> ah
[22:20] <wtay> hmm everest fails too here
[22:20] <thomas> is that MPEG1 or MPEG2 ?
[22:20] <wtay> mpeg1
[22:20] <thomas> doesn't finish nicely though.
[22:20] <thomas> Xlib: unexpected async reply (sequence 0x2240)
[22:21] <wtay> nope
[22:21] <wtay> hm, everest needs Xv...dunno why
[22:21] <thomas> ok, if both everest and aliensong are MPEG1...
[22:22] <thomas> why does everest crash gstmediaplay ?
[22:22] <thomas> and why doesn't aliensong not work with tools/gstreamer-launch disksrc location=/opt/media/AlienSong.mpg ! mp1videoparse ! mpeg_play ! aasink driver=2
[22:22] <thomas> when everest does ?
[22:22] <thomas> what makes them different ?
[22:22] <wtay> they seem to disagree about the video properties
[22:22] Ow3n (ow...@ti...) left irc: [x]chat
[22:22] <wtay> AlienSong is a system stream (audio and video) everest is video only (elementary stream)
[22:23] <wtay> for AlienSong you need a demuxer (to split the audio and video) for everest you don't
[22:23] <thomas> wtay: ok.  hmmm.  shouldn't plugins be smarter about that ?
[22:23] <wtay> the demuxer creates dynamic pads (at runtime) and that's why -launch fails (it can't do that)
[22:24] <thomas> ok.
[22:24] <thomas> can launch be adapted to do that ?
[22:24] <wtay> sure
[22:24] <thomas> heh ;)
[22:24] <thomas> time permitting
[22:24] <thomas> wtay: did you study fourier transforms ?
[22:25] <wtay> you should set up some signal handlers to catch the new_pad signal and connect the decoder to the new pad etc..
[22:25] <wtay> yes
[22:25] <wtay> I forgot most of it though <g>
[22:25] <thomas> because omega was telling me yesterday about how to speed up conversion from MJPEG to MPEG2...
[22:25] <thomas> ... and how he'd want an algorithm that would calculate the difference between two DCT blocks...
[22:25] <thomas> which should be possible
[22:26] <thomas> and I thought that, normally, if you have the DCT spectral lines...
[22:26] <thomas> you can just calculate the difference between two frames normally
[22:26] <thomas> and I was wondering what you think about it ?
[22:26] <wtay> what would you do with the diff?
[22:26] <thomas> I'm not sure what omega would want.
[22:26] <thomas> I'm not too smart with video.
[22:27] <thomas> But if it's motion compensation that you need to go from MJPEG to MPEG2
[22:27] <thomas> then that could help, right ?
[22:27] <wtay> I would guess that the DCT for mpeg could be skipped...
[22:27] <thomas> that's what he wanted, skip the IDCT/DCT and don't go to full raw video
[22:27] <wtay> motion comp works on the iDCT of the transform
[22:27] <thomas> what does motion comp do ?
[22:28] <wtay> find similar blocks in the Y and Cr planes
[22:28] <wtay> block = 8x8 pixels
[22:28] <dobey> fod
[22:28] <dobey> err
[22:28] <dobey> food
[22:28] Action: wtay hand dobey a donut
[22:28] <thomas> I should probably read up on it before I make stupid observations ;)
[22:29] <wtay> there are some good articles on the web about it
[22:29] <thomas> If it works like I think it does then the DCT step could be skipped easily
[22:29] <wtay> I'm not too sure about that...
[22:30] <wtay> unless the diff in the DCT coeff are related to the iDCT results
[22:30] <thomas> that would be the idea
[22:30] <wtay> does that assumption hold?
[22:30] <thomas> so that's why fourier transform knowledge would come in handy ;)
[22:30] <thomas> wtay: I have to read up on MPEG video before I can tell
[22:30] <thomas> though I'm sure there must be papers on it, no ?
[22:31] <wtay> yes
[22:32] <wtay> http://icsl.ee.washington.edu/~woobin/ti/intro.html
[22:32] <wtay> hmm, not a good one...
[22:34] <wtay> www.mpeg.org has lots of stuff
[22:35] <thomas> ok, i'll check that.
[22:35] <thomas> are you here tomorrow ?
[22:36] <thomas> free from work i mean ?
[22:36] <wtay> http://wwwam.hhi.de/mpeg-video/papers/sikora/mpeg1_2/mpeg1_2.htm
[22:36] <wtay> no I have to work
[22:36] <wtay> unfortunatly
[22:36] <thomas> and tuesday ?
[22:36] <thomas> I have some issues with the mixer in incsched
[22:36] <thomas> and I'd like to learn why it's not working
[22:37] <wtay> I'm here on tuesday
[22:37] <thomas> is your deb box ok yet ?
[22:37] <wtay> so it seems...
[22:37] <wtay> it's my GFs one
[22:38] <thomas> heh... she wants to work with linux ?
[22:38] <wtay> the sawfish theme selector is not working...
[22:38] <wtay> thomas: she only uses linux :)
[22:38] <thomas> wtay: keep her ! ;)
[22:38] <wtay> I want to manage the box with debian
[22:38] <wtay> oh yes
[22:39] <thomas> wtay: I think some of the autoplug code I copied from another example doesn't work anymore
[22:39] <thomas> is that because you changed caps stuff ?
[22:39] <wtay> thomas: in the mixer?
[22:39] <thomas> yeah
[22:39] <wtay> in CVS?
[22:40] <thomas> in incsched for the moment...
[22:40] <wtay> oh ok
[22:40] <thomas> I want to get the delayed start working
[22:40] <thomas> should i start over based on test/mpeg2parse3 ?
[22:41] <thomas> can this be right : DEBUG(31175:-1)gst_element_add_pad:284: setting parent of pad 'sink' to '(null)'
[22:41] <thomas> the (null) part, can that be or does that mean there's something else wrong ?
[22:44] <wtay> sorry, was away for a moment
[22:45] <wtay> null is normal, the element has no name yet.. so
[22:45] <wtay> brb
[22:46] <aj_uni> damn, no sound...
[22:51] Action: aj_uni waits for alsa drivers to recompile
[23:20] <wtay> back
[23:20] <thomas> good ;)
[23:20] <thomas> you have time to set me on my way ?
[23:21] <wtay> yeah
[23:21] <thomas> ok, here's the last bit of output before the crash
[23:21] <thomas> INFO (31175:-1):gst_bin_add:287: [pipeline] added child "typefind"
[23:21] <thomas> ** CRITICAL **: file gstelement.c: line 739 (gst_element_set_state): assertion `element->sched != NULL' failed.
[23:21] <thomas> DEBUG(31175:-1)gst_bin_iterate:672("pipeline"): entering
[23:21] <thomas> Segmentation fault (core dumped)
[23:21] Action: wtay is distracted by the snooker tournament this week
[23:21] <thomas> I guess you really like snooker ;)
[23:21] <wtay> I do :)
[23:21] <wtay> oh, that's in incsched?
[23:22] <thomas> yes
[23:22] <wtay> it's a bug IMHO...
[23:22] <thomas> my bug or someone else's ?
[23:22] <wtay> typefind is reparented but it doesn't get it's scheduler set to NULL 
[23:23] <wtay> omega's :-)
[23:23] <thomas> hmmm... so I should forget about typefinding for now ?
[23:23] <thomas> just go back to static mp3
[23:23] <wtay> I think I can fix it
[23:24] <wtay> sec..
[23:26] <wtay> hmm I'm looking at gstbin.c line 323
[23:27] <thomas> me too.  don't know what it does though.
[23:27] <thomas> not really I mean.
[23:28] <wtay> gstscheduler.c line 971...
[23:28] <wtay> so..
[23:29] <thomas> ok, there should be only one of those two
[23:29] <wtay> actually it looks good to me, I wonder why that code isn't called..
[23:30] ChiefHighwater (pa...@te...) joined #gstreamer.
[23:31] <wtay> hi
[23:31] <wtay> oh
[23:31] <wtay> mixer.c should use a pipeline as the toplevel bin
[23:32] <wtay> a regular bin doesn't have a scheduler
[23:38] <thomas> wtay: sorry, was distracted.
[23:38] <thomas> ok, I'm checking that now...
[23:41] <thomas> hmmm... does that mean I have to change each use of bin with pipeline ?
[23:41] <thomas> or only gst_pipeline_new instead of gst_bin_new ?
[23:42] <wtay> just the toplevel bin becomes a pipeline
[23:43] <thomas> and still use the gst_bin_add functions ?
[23:44] <wtay> yeah, sure, the pipeline *is* a bin
[23:44] <thomas> ok, then it doesn't work yet
[23:44] <wtay> can you check it in?
[23:44] <thomas> same error
[23:44] <wtay> oh
[23:44] <thomas> yeah, sure.  is it different to check in incsched stuff ?
[23:45] <wtay> no
[23:47] <thomas> ok it's in incsched now
[23:48] <wtay> pff my incsched branch is hosed
[23:49] <thomas> meaning ? you lost it ?
[23:49] <wtay> naah, lot's of errors during cvs update
[23:50] <thomas> hmmm... so that might take a while
[23:51] <wtay> full rebuild
[00:00] --- Mon Apr 30 2001
[00:03] <wtay> ok, trying..
[00:03] <wtay> ok, volenv is not build
[00:04] <dobey> hrmm
[00:06] Ura_away (csc...@c2...) left irc: syntax error - user imploded
[00:06] <wtay> thomas: what's line 140 in mixer.c for?
[00:08] <wtay> uh oh
[00:08] <thomas> wtay: sorry, went for something to eat
[00:08] <thomas> checking
[00:09] <thomas> well, I suppose you must pause the bin before you request a new pad and add a new decoder pipeline to the main bin ?
[00:09] <thomas> hmm... am I allowed to go from null to pause at first pass ?
[00:10] <wtay> dunno, I wouldn't do that
[00:10] <wtay> but!
[00:10] <thomas> yeah ?
[00:10] <wtay> mixer.c line 247
[00:11] <wtay> you can't run typefind if the bin is not insed a pipeline..
[00:11] <thomas> oops, swapped the change probably.
[00:11] <thomas> why did it work before then ?
[00:11] <wtay> s/insed/inside
[00:11] <wtay> before incsched it didn't matter
[00:11] <wtay> I think fixing the typefind function is the best solution..
[00:12] <wtay> lemme try something here
[00:12] <thomas> go ahead
[00:14] <thomas> btw the code's not finished yet anyway, I just want to check if I can delay the second one...
[00:14] <thomas> ... but I couldn't even get it running at first
[00:15] <wtay> woooahh!
[00:15] <thomas> what ?
[00:15] <wtay> hum
[00:15] <wtay> ** CRITICAL **: file gstbin.c: line 274 (gst_bin_add): assertion `gst_object_check_uniqueness (bin->children, GST_ELEMENT_NAME(element)) == TRUE' failed.
[00:16] <thomas> ok, so I probably got my names wrong as well during the sixth or so rewrite
[00:16] <thomas> names should probably be unique, right ?
[00:16] <wtay> so it seems, this is new...
[00:17] <wtay> I'll commit my changes
[00:17] <thomas> ok, I seem to have the second request working now...
[00:17] <thomas> ... but no audio out
[00:18] <thomas> i'll check your changes first
[00:18] <wtay> I get 100% CPU usage, nothing else
[00:18] <thomas> probably the same as me
[00:19] <thomas> maybe the requesting works different than in the non-incshed case
[00:19] <thomas> should I check another example first ?
[00:19] <wtay> it should be the same..
[00:20] <thomas> it is playing, it's just not getting data to the output bin
[00:21] <wtay> you also have the **critical** error?
[00:21] <thomas> no, what ?
[00:21] <thomas> I have a lot of trying to switch to same cothread errors
[00:21] <thomas> all the rest seems fine
[00:21] <thomas> it even exits with eos
[00:22] <thomas> it also found mad for the decoding, so autoplug seems to work
[00:22] <wtay> I have removed the check in gstbin.c and now it just cothread switches without actually doing something
[00:22] <wtay> yes
[00:23] <wtay> hmm, I don't see any osssink...
[00:23] <thomas> no, I used esdsink at first
[00:23] <thomas> changed it back to osssink
[00:23] <thomas> still doesn't work
[00:25] dobey (dobey@141.154.95.104) left irc: Read error to dobey[141.154.95.104]: EOF from client
[00:26] <thomas> hmmm... I only have the cothread errors as long as there's only one channel open
[00:26] <thomas> when the second channel opens, the errors are gone
[00:26] <thomas> I also get an eos signal when the second one opens
[00:27] <thomas> which doesn't seem right anyway
[00:27] <wtay> this is a good case fo omega to look at.
[00:27] <thomas> heh;
[00:27] <thomas> guess I'll have to wait then
[00:27] <wtay> does it also happen with just one channel?
[00:28] <thomas> yeah...
[00:28] <wtay> good, should be a pretty simple case then
[00:28] <thomas> wait... the eos I probably get since I close it manually after a few iterations
[00:28] <thomas> right now I do, do 10000 iterates, open another channel
[00:28] <thomas> in the case of one channel it closes off after this
[00:30] <wtay> ok
[00:34] <richardb> steveb: are you here?  mono2stereo in CVS HEAD doesn't compile.
[00:34] <richardb> It fails due to line 29: #include <gst/meta/audioraw.h>
[00:34] <richardb> Works if this line is removed, I presume this is just a bit of legacy code.
[00:35] <steveb> oh, yeah that shouldn't be there
[00:35] <richardb> float2int.h and int2flaot.h also fail similarly.
[00:35] <steveb> hang on
[00:35] <richardb> ;)
[00:37] <steveb> done
[00:37] Action: steveb hides in shame
[00:38] <richardb> That seems to have fixed it - thanks. :)
[00:39] <richardb> Hmm: the PWG is out of date concerning the audio/raw parameters.
[00:39] <richardb> It doesn't document the "format" parameter, and the possible values if format is "float"
[00:39] <richardb> I don't know what the possible values then are.
[00:40] <richardb> Could someone give a detailed summary / add them into the PWG.
[00:40] <steveb> i could do that
[00:40] <richardb> (I'm happy to do the latter, if someone gives me the info. ;-) )
[00:40] <richardb> That'd be great. :)
[00:40] <steveb> i could give you the info - but first i need to sleep
[00:40] <richardb> I'm off to bed too.
[00:40] Action: richardb yawns
[00:41] Action: steveb is exhausted at the effort of fixing his code
[00:41] <richardb> we need to document the parameters for the audio/raw mime type, especially for the case when format is not "int"
[00:41] <richardb> ;-)
[00:42] <richardb> Goodnight.
[00:42] <steveb> night
[00:42] steveb (st...@no...) left irc: [x]chat
[00:42] richardb (ri...@ix...) left irc: [BX] Hrm... I wonder if I paid this month's electr...EOF From client
[00:44] <thomas> wtay: I'll check with omega tomorrow if he's here.
[00:44] <thomas> I'm going to bed now.
[00:44] <wtay> ok, me too
[00:44] <thomas> later...
[00:45] <wtay> cya
[00:45] <thomas> btw your version of mixer doesn't have the cothread issue
[00:45] <thomas> doesn't play either ;)
[00:46] <thomas> bye
[00:46] thomas (th...@ad...) left irc: Client Exiting
[01:12] Nick change: wtay -> maYam
[01:12] <maYam> hello everyone
[01:13] <maYam> wtay has destroyed my comp..  so i take his
[01:14] <ChiefHighwater> hehe
[01:17] <maYam> and i won't let him go to bed until he fixes everything
[01:17] <maYam> naughty boy..
[01:20] Nick change: maYam -> wtay
[01:22] Nick change: taazzzz -> taaz
[01:24] <wtay> time fo bed, cya
[01:24] Nick change: wtay -> wtay-zZz
[01:55] _gst_newt_ joined #gstreamer.
[02:02] wtay-zZz (wi...@ca...) joined #gstreamer.
[03:18] Topic changed on #gstreamer by ChanServ!s@ChanServ: GStreamer: the ultimate multimedia framework
[04:16] <Happyfeet> hey guys
[04:37] <Happyfeet> what version of autoconf is needed for compiling from cvs ? 
[04:38] <Happyfeet> autogen just gobbles up memory , over 150 megs ,.. and doesnt stop, somthing wrong here
[04:56] <taaz> Happyfeet: read README from cvs code
[04:56] <taaz> it works, just takes lots of mem and time
[05:00] Topic changed on #gstreamer by ChanServ!s@ChanServ: GStreamer: the ultimate multimedia framework
[05:09] Nick change: taaz -> taazzzz