IRC Logs

*******************************************************************
[03:01] <walters> ** ERROR **: file gstoptimalscheduler.c: line 447 (add_to_chain): assertion failed: (group->chain == NULL)
[03:01] <walters> aborting...
[03:01] <walters> unf
[03:01] <walters> that didn't go so well.
[03:02] <walters> Company: what do you make of this?
[03:02] <walters> /build/gstreamer-0.7/bin/gst-launch-0.7 -v gnomevfssrc iradio-mode=1 location=http://neptuner.orbislumen.net:8000/seattletech.ogg ! typefind ! spider ! osssink
[03:03] <walters> i get:
[03:03] <walters> ERROR: from element /pipeline0/spider0/vorbisfile0: The stream is of a different type than handled by this element.
[03:03] <walters> but that's odd, because: /pipeline0/spider0/sink_ident.src: caps = application/ogg
[03:06] <Company> walters: i'm currently fixing another bug, that's eeeevil
[03:07] <Company> walters: does saving to file and then playing work?
[03:10] <walters> Company: doesn't appear to
[03:10] <walters> i get a little bit of sound
[03:10] <walters> then
[03:10] <walters> ** (process:27285): WARNING **: vorbisfile: decoding error
[03:10] Rotty ([email protected]) left irc: Remote closed the connection
[03:10] <Company> probably because you only saved a little bit of file?
[03:11] <Company> does playing with ogg123 work?
[03:11] <walters> Company: i saved 1.1meg
[03:11] <walters> Company: yes
[03:11] <walters> http://web.verbum.org/files/foo.ogg
[03:13] <Company> so it's most probably a vorbisfile bug
[03:13] <walters> do you get that too?
[03:13] <Company> lemme try
[03:13] hyriand ([email protected]) left irc: "Lost terminal"
[03:14] <Company> hm, i forgot that theres a lot of downloads running here atm
[03:19] <Company> walters: i get that too
[03:19] Action: Company assumes some vorbisfile return value is checked wrong
[03:20] <walters> yeah
[03:22] <Company> oggdemux ! vorbisdec works
[03:22] <Company> but that only exists on my local box and can'T seek yetr
[03:23] <Company> walters: if you want to, file a bug and attach that file (the first 300kB should be enough)
[03:24] <walters> do you think your oggdemux/vorbisdec will land soon?
[03:27] <Company> dunno
[03:27] <Company> atm there's 2 issues left
[03:27] <Company> 1 is the fact that it doesn't seek yet, but that should work very soon, the other is the output format
[03:28] <Company> and "consumer float audio" requires figuring out how to do, especially because float2int doesn't do it yet
[03:28] <Company> (i'm hacking on vorbisdec only on my way to uni and back, so it doesn't get 100% attention)
[03:38] <Company> [x] bug fixed
[03:39] <Company> walters: there's a bug wrt typefinding in small buffers, you'll need to update core to fix that when i say "done"
[03:49] <Company> done
[03:49] <walters> cool
[03:51] LeRoutier ([email protected]) joined #gstreamer.
[03:51] <LeRoutier> hello
[03:52] <Company> you're drunk?
[03:52] <Company> or why do you show up at 4 o'clock at night?
[03:53] <LeRoutier> nope, not drunk (not this time)
[03:53] <LeRoutier> i saw light so i entered :)
[03:53] <LeRoutier> this time, light was a commit on gstreamer
[03:54] <Company> ah
[03:54] <Company> walters makes me fix bugs at night
[03:54] <Company> in fact that one was quite nasty
[03:54] Action: walters grins
[03:54] <LeRoutier> i was just in the hope ronald was around so i could answer a question he asked me via bugzilla
[03:55] edmack ([email protected]) joined #gstreamer.
[03:55] <LeRoutier> so, i suppose i'll answer him via bugzilla too
[03:56] <edmack> Hi, my little program is complaining when that there is no eos signal for my filesrc element:
[03:56] <edmack> (a.out:26445): GLib-GObject-WARNING **: gobject.c:1150: invalid signal spec "eos"
[03:56] <edmack> Are there any common causes of this? I'm just putting it in a simple pipeline, connecting it with this line:
[03:56] <edmack> g_object_connect (G_OBJECT (src), "eos", G_CALLBACK (next_music_gst), NULL);
[03:56] <LeRoutier> there is 'ds', 'ds-work' and now a 'ds-gromit'. /me wonders what this last one is
[04:01] <Company> ds logged in from a machine called gromit i suppose
[04:02] <Company> edmack: if you're sure that your element indeed is a filesrc, then that shouldn't happen
[04:02] <edmack> src =  gst_element_factory_make ("filesrc", "src");
[04:02] <edmack> Oh dear :(
[04:02] <edmack> that's before I try to connect the signal
[04:02] water ([email protected]) joined #gstreamer.
[04:17] <edmack> Whenever the location of a filesrc is changed, what must be done to make a spider re-config for another format? currently I change the pipeline state to null, change location, then play, begin iterating
[04:18] Nick change: edmack -> edmack|sleep
[04:19] <Company> remove spider, insert a new spider
[04:19] <Company> (yes, that's broken and we know that)
[04:23] water ([email protected]) left #gstreamer ("The water has parted.").
[04:23] <edmack|sleep> ok, thank you very much (will it be fixed for 1.0?)
[04:24] The_Company ([email protected]) joined #gstreamer.
[04:26] Company ([email protected]) left irc: Nick collision from services.
[04:26] <LeRoutier> got an avi/divx5+mp3 . which pipeline to use to have both video and audio ? with the following, got video working :  ! avidemux ! ffdec_mpeg4 ! xvimagesink
[04:26] Nick change: The_Company -> Company
[04:27] <Company> append avidemux0. ! mad ! osssink
[04:27] <LeRoutier> i don't know how to build on non-linear pipeline (1 to n elements => 1 file to 1 audio sink & 1 video sink)
[04:27] <LeRoutier> k, thx
[04:28] <LeRoutier> so : ! avidemux ! ffdec_mpeg4 ! xvimagesink ! avidemux0. ! mad ! osssink ?
[04:28] <Company> you reference an element with elementname.
[04:28] <Company> no:  ! avidemux ! ffdec_mpeg4 ! xvimagesink avidemux0. ! mad ! osssink
[04:28] <Company> (note the dot in there)
[04:28] <LeRoutier> k, thx. i'll try to put that somewhere in my brain
[04:28] <Company> elements are named 0, 1, ... 
[04:28] sri ([email protected]) joined #gstreamer.
[04:29] <sri> y0, my gstreaming homies.
[04:29] <LeRoutier> works
[04:29] <Company> but that's unsupported, if you do that in an app use avidemux name=demuxer
[04:29] <Company> and then demuxer. ! ...
[04:29] <LeRoutier> so, my avi bug seems to be in gst-player, not in avi demux
[04:31] <LeRoutier> bug report updated for BBB : http://bugzilla.gnome.org/show_bug.cgi?id=131050
[04:32] <LeRoutier> for an unknown reason, gst-player displays funky colors instead of video. so perhaps spiders chooses a bad codec for video. i don't know
[04:32] <LeRoutier> noting this other syntax
[04:33] <Company> you can test with spider
[04:33] <Company> src ! spider ! xvimagesink spider0. ! osssink
[04:33] edmack|sleep ([email protected]) left #gstreamer.
[04:33] <Company> though that probably should be more like this:
[04:34] <Company> src ! spider ! { queue ! xvimagesink } spider0. ! { queue ! osssink }
[04:34] <LeRoutier> how to force spider to tell which plug-in it choosed ?
[04:35] <LeRoutier> hum, with this spider syntax, got a video box which is around 1 cm x 1 cm
[04:36] <Company> you can see it when you use gst-launch -v
[04:36] <LeRoutier> k, thx
[04:36] <Company> and if you get the little box the most likely reason is that the autoplugger didn't find anything
[04:37] <LeRoutier> got a little black box :)
[04:37] <sri> have you guys been seeing "pad" errors from like gstreamer from yesterday or maybe a day older than that?
[04:37] <sri> I keep getting this message from rb that an internal gstreamer error had occured.
[04:38] <LeRoutier> spider tries divxdec. when i used ffdec_mpeg4
[04:40] <Company> yeah, that is an awful error message
[04:40] <LeRoutier> hum, divxdec finds the good window format (framerate=(double)29.969999313354492, width=(int)480, height=(int)272 on its sink)
[04:40] <LeRoutier> but xvimagesink got initialized with 16x16
[04:41] <Company> xvimagesink should be reconnected later on
[04:41] <Company> renegotiated even
[04:42] <sri> Company: I'll see if it's fixed after I recompile gstreamer/gst-plugins 
[04:44] <LeRoutier> Company, it does not seem renegotiated at all
[04:44] <Company> sri: probably not, but it should say something more interesting
[04:44] <Company> sri: i especially dislike the "file a bug" part ;)
[04:45] <Company> LeRoutier: possible if divxdec and xvimagesink can't connect
[04:45] <Company> maybe try ... spider ! { queue ! ffcolorspace ! xvimagesink } instead
[04:45] <LeRoutier> to force colorspace conv, yes
[04:46] <sri> Company: heh.yeah..I wouldn't know what to put in it..no details..that would be a very annoying bug report
[04:47] <LeRoutier> Company, it seems the window first takes the good width/height then is back to 16x16. the problem is that ffcolorspace is negotiated to 16x16
[04:48] <LeRoutier> at start : /pipeline0/thread0/ffmpegcolorspace0.src: caps = video/x-raw-yuv, width=(int)320, height=(int)240, framerate=(double)30, format=(fourcc)I420
[04:48] <LeRoutier> after renego : /pipeline0/thread0/ffmpegcolorspace0.src: caps = video/x-raw-yuv, width=(int)16, height=(int)16, framerate=(double)1, format=(fourcc)I420
[04:49] <Company> weird
[04:49] <LeRoutier> it takes the queue src width+height, not the divxdec sink width+height
[04:49] <Company> you'll need to bug bbb as i don't use closed source here
[04:50] <LeRoutier> queue.src isn't supposed  to have caps of precedent element sink ?
[04:51] ChrisHJW ([email protected]) joined #gstreamer.
[04:51] <LeRoutier> k, i'll be back here "tomorrow". in about 7 hours or more so. in the hope ronald would be around
[04:52] <LeRoutier> for now, time to sleep, it's 5AM after all. good night everyone
[04:53] <LeRoutier> and thx for help Company
[04:53] LeRoutier ([email protected]) left irc: "Leaving"
[04:55] Miko5881 ([email protected]) left irc: "ChatZilla 0.8.31 [Mozilla rv:1.4/20030928]"
[04:59] witless ([email protected]) joined #gstreamer.
[05:01] alley_cat ([email protected]) left irc: "May the Source be with you!"
[05:01] <witless> hi. i'm on debian unstable.  when i run gstreamer and try to play an ogg file, i get these messages streaming on my screen:  
[05:01] <witless> gst-player:1258): GStreamer-WARNING **: push on peer audio_queue:sink of pad src_0:src but peer is not active
[05:01] <witless> any ideas?
[05:28] <walters> witless: how are you trying to play it?
[05:35] <witless> walters: start gst-player from cli, and open an ogg file
[06:24] ChriHJW_log ([email protected]) left irc: Read error: 60 (Operation timed out)
[06:36] ChrisHJW ([email protected]) left irc: Read error: 110 (Connection timed out)
[07:17] hadley ([email protected]) left irc: "Leaving"
[07:18] Marsupilami23 ([email protected]) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!"
[07:19] Marsupilami23 ([email protected]) joined #gstreamer.
[08:15] witless ([email protected]) left irc: "Leaving"
[08:33] Nick change: mathrick|sleep -> mathrick
[08:34] <mathrick> mornin
[08:35] <mathrick> walters: did you realize how horribly wrong you were with your assumptions about that typefind pipeline? ;)
[08:36] <walters> mathrick: eh?
[08:36] <mathrick> <walters> /pipeline0/typefindelement0.sink: caps = application/octet-stream
[08:37] <walters> yeah?
[08:37] markov ([email protected]) joined #gstreamer.
[08:37] <mathrick> walters: _sink_
[08:37] <walters> mathrick: i got the same thing on src
[08:38] <mathrick> walters: it doesn't really matter, I felt for that too, and couldn't understand why I got different results than gst-typefind
[08:38] <mathrick> walters: what's right here is to connect to "have-type" signal of typefind
[08:39] <mathrick> walters: I came to that after reading gst-typefind source :)
[08:40] <mathrick> er, s/felt/fell/
[08:41] <walters> mathrick: in this case it gave the same thing
[08:41] <mathrick> walters: yeah, this was real bug, but in general looking on -v output may be misleading
[08:42] markov ([email protected]) left irc: Client Quit
[09:20] Nick change: harshyOUT -> harshy
[09:22] Nick change: harshy -> harshyZ
[09:35] mathrick ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[09:49] markey ([email protected]) joined #gstreamer.
[10:08] markey ([email protected]) left irc: "bbl"
[10:15] ChrisHJW_log ([email protected]) joined #gstreamer.
[10:15] ChrisHJW_log ([email protected]) left irc: K-lined
[10:21] apoc ([email protected]) joined #gstreamer.
[12:00] sublett ([email protected]) joined #gstreamer.
[12:06] BBB ([email protected]) joined #gstreamer.
[12:06] <BBB> ds: are you there?
[12:06] <BBB> or ds-work or ds-gromit (what is gromit)?
[12:12] Shoragan ([email protected]) joined #gstreamer.
[12:19] Action: BBB just made videobalance 10x faster :)
[12:22] dolphy ([email protected]) joined #gstreamer.
[12:22] <dolphy> yop
[12:24] alley_cat ([email protected]) joined #gstreamer.
[12:25] <dolphy> BBB: hey ronald :)
[12:25] <BBB> hi
[12:25] <BBB> dolphy: videobalance is now fast
[12:25] <dolphy> BBB: do you think ffcolorspace could do colorbalance at the same time ?
[12:25] <dolphy> BBB: that would be even faster
[12:25] <BBB> no
[12:25] <BBB> that's not gst'ish
[12:25] <BBB> and this is fast enough
[12:25] <BBB> it takes 7% CPU for a full pipeline in gst-editor
[12:25] <BBB> that's really a lot
[12:26] <BBB> it's videotestsrc ! videobalance ! ffcolorspace ! ximagesink
[12:26] <BBB> where it does I420->RGB in ffcolorspace
[12:26] <BBB> and all calculations in videobalance
[12:26] <BBB> so 7% is really acceptable
[12:26] <dolphy> ok
[12:26] <dolphy> BBB: i have issues with switch
[12:27] <dolphy> BBB: it blocks when there's nothing to gst_pad_pull
[12:27] <dolphy> BBB: what about making switch chain based ?
[12:27] <jdahlin> hello dolphy
[12:27] <dolphy> jdahlin: hey johan :)
[12:27] <BBB> aha :) I could have known...
[12:27] <BBB> before pulling, do GST_PAD_IS_USABLE(pad)
[12:27] <jdahlin> dolphy, thomas mentioned that you were working for some kind of application to replace the controller app in davedina
[12:27] <BBB> if that gives FALSE, don't pull
[12:28] <dolphy> if (GST_PAD_IS_USABLE (switchpad->sinkpad)) {
[12:28] <dolphy>       data = gst_pad_pull (switchpad->sinkpad);
[12:28] <dolphy> i m doing that already
[12:28] <dolphy> that doesn't work with queue
[12:28] <dolphy> jdahlin: yup it's named getcontrol
[12:28] <dolphy> BBB: that's my only lock now
[12:29] <BBB> dolphy: oh, right, crap...
[12:29] <BBB> ehm
[12:29] <jdahlin> dolphy, I started to hack on something similar, but with an interface like ipod
[12:29] <dolphy> BBB: i have made this pipeline http://getcontrol.sourceforge.net/files/gstplay_pipeline.png
[12:29] <BBB> yes I know the pipeline
[12:29] <dolphy> BBB: and on audio playback it locks
[12:29] <BBB> I don't know how to sovel this
[12:29] <BBB> :(
[12:29] <BBB> maybe ds/company know a way
[12:29] <dolphy> BBB: switch being chain based would solve the issue
[12:29] <dolphy> BBB: setting the same chain function on all sink pads
[12:30] <dolphy> BBB: filling switchpad structure and pushing only the active pad
[12:30] <BBB> no
[12:30] <BBB> multi-input elements cannot be chainbased
[12:30] <BBB> that has scheduling issues
[12:30] <BBB> I've been there with avimux
[12:30] <BBB> don't do it ;)
[12:30] <BBB> you'll make the same mistakes I made
[12:31] <dolphy> hmm
[12:31] <dolphy> have you seen my email about gst_element_release_request_pad
[12:31] <dolphy> ?
[12:31] <BBB> no
[12:31] <BBB> or yes
[12:31] <BBB> I have
[12:32] <dolphy> i implemented that in switch and that really helps :)
[12:33] <dolphy> otherwise switch is trying to pull on lots of pads that are disconnected
[12:33] <BBB> that's odd
[12:33] <dolphy> GST_PAD_IS_USABLE can make it safer but still the pad list is growing for nothing
[12:33] <BBB> those should be unusable
[12:33] <BBB> yes, that's correct
[12:34] <dolphy> so now i m releasing my switch pad when removing visualization bin
[12:35] <dolphy> that works fine
[12:35] <dolphy> BBB: so now ideas about how to solve that blocking gst_pad_pull call ?
[12:36] <dolphy> jdahlin: good :) a lot of projects exist in that domain
[12:36] <dolphy> jdahlin: mine is completely based on directfb
[12:36] <dolphy> jdahlin: and GObjects
[12:36] <jdahlin> dolphy, I'm just using a GtkTreeView right now (and python)
[12:37] <jdahlin> dolphy, but I'm thinking about just using gdk so I can easily port it to whatever
[12:37] <dolphy> jdahlin: i developped my own toolkit
[12:37] <jdahlin> dolphy, seems like a bit overkill :)
[12:37] <dolphy> jdahlin: well not really
[12:37] <dolphy> jdahlin: gdk is far more complex than what i need
[12:38] <dolphy> jdahlin: it was trivial to implement an object hierarchy using directfb calls to draw
[12:38] <dolphy> jdahlin: and that works really well
[12:38] <jdahlin> dolphy, maybe you're right, since applications of that kind dont really need to be integrated with the rest of the system
[12:38] <BBB> dolphy: no idea, really
[12:38] <dolphy> BBB: k so i m stuck once more
[12:38] <BBB> sorry... :(
[12:38] <dolphy> BBB: i ll probably only pull from the active pad
[12:38] <BBB> I don't know much about internals of scheduling et all
[12:38] <dolphy> BBB: and forget others
[12:39] <dolphy> BBB: but that might fill the queues
[12:39] <BBB> it doesn't work anyway... don't use switch for now
[12:39] <BBB> really
[12:39] <BBB> that's easiest
[12:40] <dolphy> well i don't really see how not to use switch atm
[12:41] markey ([email protected]) joined #gstreamer.
[12:42] <dolphy> if i m not using switch then i have a libgstplay that is putting visualization even if there is video signal from spider
[12:42] <dolphy> and as applications like totem and gst-player enables visualization from gconf keys
[12:43] <dolphy> that means you will never see video but only goom :)
[12:43] <dolphy> unless you disable visualization in the settings
[12:44] <BBB> switch needs to know from which chan to pull?
[12:44] <BBB> so you know which to connect if you don't use switch
[12:45] <dolphy> the problem is that when enabling viz i need to connect both the vis bin and spider to the video pipeline
[12:45] <dolphy> and spy for identity handoff to know that there is some video
[12:45] <dolphy> and then disconnect the vis bin
[12:46] <dolphy> if i m not using swith the only solution i see is to use a fakesink after identity
[12:46] <dolphy> and have the vis bin connected to the video pipeline
[12:46] <dolphy> and replug identity when handoff arrives
[12:46] <dolphy> that's not really nice
[12:46] <dolphy> and i might loose some buffer
[12:49] <BBB> I really don't know a solution, as I said
[12:49] <dolphy> ok :)
[12:50] Action: dolphy switch back to getcontrol development then :)
[13:05] <BBB> getcontrol?
[13:05] jdahlin ([email protected]) left #gstreamer ("Leaving").
[13:05] jdahlin ([email protected]) joined #gstreamer.
[13:05] <dolphy> BBB: yup, same category like freevo, or Win XP Multimedia Center
[13:06] <BBB> isn't that what you're running in your living room?
[13:07] <dolphy> yup
[13:07] hyriand ([email protected]) joined #gstreamer.
[13:07] Nick change: markey -> f00d
[13:07] <dolphy> but i haven't touched it since so many times
[13:08] <dolphy> too busy on gst things
[13:11] <BBB> same here for gst-rec... :/
[13:11] <BBB> believe me, I'd love to touch gst-rec
[13:11] <dolphy> hehe i exactly know what you feel
[13:11] <BBB> but well, Gnome is more important, I'm affraid :(
[13:11] <dolphy> true
[13:12] <dolphy> but i can't do anything more now for gnome.. i m completely stuck
[13:12] Action: BBB fixed metadata in AVI, so on to get metadata in matroska working now
[13:12] <BBB> kick company awake ;)
[13:14] <BBB> the gst family image is cute ;)
[13:14] <BBB> I want that too
[13:14] <dolphy> ??
[13:15] <BBB> "part of the gstreamer family"
[13:15] <BBB> on getcontrol.sf.net
[13:15] <dolphy> ah :)
[13:15] <dolphy> that site is completely outdated :)
[13:16] <dolphy> i really have to do a release of the new getcontrol :)
[13:16] <BBB> does it work? :p
[13:16] <dolphy> well the version linked on the website works
[13:16] <dolphy> but it uses gdk pixbufs a lot
[13:17] <dolphy> the new version is completely directfb based with real translucency
[13:17] <dolphy> and has a very flexible plugin system
[13:17] <BBB> do we have dfb plugins in gst?
[13:17] <BBB> :?
[13:17] <dolphy> allowing you to use gstreamer or xine-lib for media playback
[13:17] <dolphy> i will write directfbsink soon
[13:17] <BBB> :p
[13:20] <BBB> why do I get a 16x16 window?
[13:20] <BBB> if I playback AVI files
[13:20] <BBB> divx3 or divx5
[13:21] <dolphy> fixate :)
[13:21] <BBB> why doesn't it update after real size was set?
[13:21] <dolphy> are you talking about gstplay ?
[13:22] <BBB> yes
[13:24] <dolphy> well let me commit the newest one
[13:24] LeRoutier ([email protected]) joined #gstreamer.
[13:24] Action: BBB notices divxdec is broken
[13:25] Action: BBB fixes
[13:25] <LeRoutier> hello BBB
[13:25] <BBB> dolphy: do you have the issue that sound playback (mp3) makes it extremely slow?
[13:26] <LeRoutier> hello everyone
[13:26] <dolphy> BBB: hmm nope
[13:26] <dolphy> BBB: i have an issue with tagged mp3 though
[13:27] <dolphy> BBB: spider is just autopluggin id3tags element and then declare eos
[13:27] <LeRoutier> for the search "mpg slow" in my internal google, i have 1 answer : playing mpg files in gst-player is like 1 fps
[13:28] <BBB> heh :)
[13:28] <dolphy> good mpeg works fine now
[13:28] <dolphy> congrats BBB
[13:28] <BBB> audio playback appears broken here
[13:28] <BBB> appears mad-related
[13:28] <BBB> or so
[13:28] Action: BBB doesn't know
[13:29] <dolphy> i committed latest gstplay
[13:29] <dolphy> and latest switch
[13:30] Action: BBB tries
[13:30] <dolphy> http://getcontrol.sourceforge.net/files/mpeg_scaled.png
[13:30] <BBB> is MPEG video playback fixed too?
[13:30] <BBB> I didn't touch that
[13:30] <BBB> except autoplugging
[13:30] <dolphy> well look at the shot :)
[13:30] <BBB> er
[13:30] <BBB> typefindign
[13:31] <BBB> \o/
[13:31] <LeRoutier> BBB: what was divxdec problem ? 
[13:32] <BBB> 16x16
[13:32] <BBB> dolphy: ok, I fixed it, and you broke it again :p
[13:32] <LeRoutier> oh, so the pb i had yesterday
[13:32] <BBB> I now get 16x16 windows for *all* videos
[13:32] <BBB> (ximagesink)
[13:33] <LeRoutier> well, got 16x16 yesterday only when using spider. not when using an home made pipeline
[13:33] <BBB> time tick 268000
[13:33] Last message repeated 1 time(s).
[13:33] <BBB> LeRoutier: same here
[13:33] <BBB> I fixed that a minute ago
[13:33] <BBB> and dolphy broke it a minute later with his newest gstplay :p
[13:33] <BBB> dolphy: time tick <same num> <- is that bad?
[13:34] <LeRoutier> k, so i just did ctrl+c on my build
[13:35] <LeRoutier> BBB: had the very same 16x16 problem with ffdec_mpeg4 in gst-player. is that fixed too ?
[13:35] <BBB> only in divxdec
[13:35] <BBB> didn't check ffdec_mpeg4
[13:36] <LeRoutier> both codec worked out of spider for the same sample video
[13:38] <dolphy> BBB: time tick the same num means the pipeline is stuck :)
[13:38] <BBB> ugh
[13:38] <BBB> how do I fix that?
[13:38] <dolphy> BBB: what video are you trying ?
[13:39] <dolphy> BBB: i get not issue like yours
[13:39] <BBB> I still get 16x16 windows btw
[13:39] <dolphy> aliensong works fine
[13:39] <dolphy> matrix.avi works fine too
[13:39] <BBB> ximageisnk or xvimagesink?
[13:39] <BBB> I use ximagesink
[13:40] <LeRoutier> xvimagesink here
[13:40] <dolphy> i use ximagesink
[13:41] <BBB> ffdec_mpeg4 is fine
[13:41] Action: BBB retries new gst-register
[13:41] <BBB> I'm trying mpeg, divx3 avi and divx5 avi
[13:41] <LeRoutier> so it's spider or gst-play fault 
[13:41] <BBB> mp3 works
[13:41] <LeRoutier> was a div5 yesterday with mp3 audio
[13:42] <dolphy> ** (process:22212): CRITICAL **: file bytestream.c: line 256 (gst_bytestream_peek): assertion `len > 0' failed
[13:43] <dolphy> do you know why i get that sometimes ?
[13:43] <BBB> avi or matroska sometimes give that for broken files with 0-byte chunks
[13:43] <BBB> it doesn't crash, so leave it for now...
[13:43] <BBB> I'll add a workaround some day later
[13:44] <LeRoutier> still got mpg files running at 1fps or not far in gst-player. known issue ?
[13:44] <BBB> MPEG files don't play at all here
[13:45] Action: BBB does complete fresh checkout
[13:45] <LeRoutier> it might be wonderfull to debug in such conditions :)
[13:46] <dolphy> LeRoutier: getting 1 fps playback is a caps nego issue
[13:46] <dolphy> LeRoutier: which file please?
[13:46] <LeRoutier> BBB: i don't have your latest divxdec & dolphy's gst-play commit here. perhaps it's why we have different behaviour
[13:46] <dolphy> LeRoutier: yeah make a fresh checkout/rebuild please
[13:46] <LeRoutier> dolphy, several p0rn files here
[13:46] <LeRoutier> ok
[13:47] <BBB> omg
[13:47] <LeRoutier> compile in progress
[13:47] <BBB> I just fixed your previous porn files
[13:47] <BBB> d'oh
[13:47] <dolphy> LeRoutier: hehe i tried one porn movie that never played before and now it plays
[13:47] <dolphy> LeRoutier: avi support is better than ever :)
[13:47] Action: BBB likes his new AVI demuxer
[13:47] <BBB> it's really good
[13:48] <BBB> we gotta rewrite asfdemux/qtdemux similarly, one day
[13:48] <LeRoutier> got a large sample of media files, i'll try to find some that won't play with latest version
[13:48] <BBB> especially asfdemux
[13:49] <dolphy> BBB: Pixar Geri's game has issues
[13:49] <dolphy> BBB: it was playing almost fine in 0.6
[13:49] <BBB> avi audio plays
[13:49] <BBB> avi video is borken
[13:49] <BBB> (divx)
[13:49] <BBB> I just fixed that a second ago
[13:50] <BBB> so it's not divxdec
[13:50] <BBB> mpeg doesn't play at all, it stalls after a few seconds
[13:51] <LeRoutier> on the ASF front, there is some time/length computation problem
[13:51] <LeRoutier> i don't know if it comes from gst-player or asfdemux
[13:51] <BBB> asfdemux, I guess
[13:54] <dolphy> can i do a gst_structure_copy of a GstTagList ?
[13:54] Action: BBB make clean && make's
[13:54] <BBB> taglist has its own copy
[13:54] <BBB> but I suppose you can copy it, yes
[13:54] <dolphy> i would like to add the tag list to an asyncqueue
[13:54] <dolphy> and keep it
[13:54] <dolphy> and there doesn't seem to be any ref system on GstTagList
[13:54] <BBB> can someone please please fix video playback using latest CVS & spider & ximagesink? :(
[13:55] <BBB> nope
[13:55] <BBB> structure is unreffed, unfortunately
[13:55] <BBB> :(
[13:57] <dolphy> would be nice to have GstStructure being a Gobject
[13:57] <LeRoutier> got an avi working in gst-player with xvimagesink :)
[13:57] <LeRoutier> a div5+mp3, like the one that failed yesterday
[13:57] <BBB> dolphy: no, takes too much time to initialize
[13:58] <BBB> it's too slow
[13:58] <BBB> hm, then something must be wrong locally
[13:58] <BBB> is that with divxdec or ffdec_mpeg4?
[13:59] <LeRoutier> i suppose it is divxdec has it is primary
[13:59] <BBB> it could be a boxed type
[13:59] <BBB> hm
[13:59] <BBB> then it should work here, too
[13:59] <dolphy> hmm gst_element_found_tags is weird
[13:59] <dolphy> it 's not using the klass->found_tag method
[13:59] <LeRoutier> and the file that failed yesterday works now too
[14:01] <LeRoutier> BBB: i use xvimagesink 
[14:02] <LeRoutier> avi seems to work fine now in gst-player. still have the very slow mpeg 
[14:02] <LeRoutier> dolphy, any way to make gst-player dump the pipeline spider guessed/built on the console ?
[14:03] <dolphy> LeRoutier: to debug please use gst-plugins/examples/gstplay/player
[14:03] <dolphy> not gst-player
[14:03] <LeRoutier> k
[14:03] <dolphy> in gst-plugins/examples/gstplay/player.c there's an xml dumping function for the pipeline
[14:05] <dolphy> k out for lunch
[14:05] Nick change: dolphy -> dolphy_afk
[14:05] <LeRoutier> k, bon app dolphy_afk
[14:06] Nick change: f00d -> afk
[14:06] <BBB> I'm always using player
[14:06] <BBB> not gst-player
[14:06] <LeRoutier> once it is built, i'll try with it
[14:09] <BBB> still nothing
[14:10] <BBB> still 16x16 :(
[14:11] <LeRoutier> with both ximage & xvimagesink ?
[14:11] <BBB> xvimagesink crashes my system
[14:11] <BBB> I'd rather not try
[14:16] <LeRoutier> my avi works with ./player too
[14:16] BBB ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[14:16] ChrisHJW_log ([email protected]) joined #gstreamer.
[14:16] ChrisHJW_log ([email protected]) left irc: K-lined
[14:16] <LeRoutier> trying the mpg
[14:16] BBB ([email protected]) joined #gstreamer.
[14:17] <LeRoutier> <LeRoutier> my avi works with ./player too
[14:17] <LeRoutier> <LeRoutier> trying the mpg
[14:17] <BBB> dammit
[14:17] Action: BBB kicks xv
[14:18] <LeRoutier> mpg works slowly here too
[14:18] <LeRoutier> got several same timetick in a raw
[14:21] <LeRoutier> want me to try ximagesink ?
[14:22] <LeRoutier> works also with ximagesink
[14:23] <BBB> with previous libgstplay, it works fine here
[14:23] <BBB> with the current (1.17), it all fails
[14:23] <BBB> which libgstplay for you?
[14:23] <BBB> gst-libs/gst/play/gstplay.c
[14:24] <BBB> and I get 100% PU use
[14:24] <BBB> s/PU/CPU
[14:24] <BBB> on commandline, it takes little-nothing
[14:25] Uraeus ([email protected]) joined #gstreamer.
[14:25] <Uraeus> good morning
[14:26] <LeRoutier> hello Uraeus
[14:26] <LeRoutier> BBB: got 1.17 too
[14:26] <Uraeus> BBB: why didn't that avi work? unsupported colorspace?
[14:29] <LeRoutier> Uraeus, the "DIB " one in bugzilla ?
[14:29] <Uraeus> the one I submitted to bugzilla two days ago yes
[14:29] <LeRoutier> even mplayer got problems finding how to decode it. but they have fail-over. i don't think gstreamer has it
[14:30] <Uraeus> well it would be cool if we could play it, then we can say we play avi's than even xine fails to play :)
[14:31] <LeRoutier> trying to read it, it makes some "Index entry XXX has invalid stream nr YYY" like i saw a few days ago before BBB somehow fixed avidemux index reading
[14:32] <BBB> Uraeus: we don't handle RGB yet, I'll work on it
[14:32] <BBB> Uraeus: please upload somewhere so I can have a look
[14:32] <Uraeus> BBB: I attached the avi to the bugzilla report
[14:32] <Uraeus> BBB: you are not able to retrieve it?
[14:33] <BBB> oh, ok
[14:33] <BBB> I'll look there
[14:33] <BBB> LeRoutier: if the stream is unrecognized, we don't add it, so then all index entries referring to that stream give that warning
[14:33] <LeRoutier> ok
[14:33] <LeRoutier> my mplayer says that Uraeus file is "RAW BGR8"
[14:33] <BBB> yes, I saw
[14:34] Action: BBB notes that ffmpeg doesn't handle BGR8
[14:34] sublett ([email protected]) left irc: "I like food, food is good!"
[14:34] <Uraeus> BBB: does Hermes? 
[14:36] <BBB> no
[14:37] <BBB> they have PAL8
[14:37] <BBB> what's that?
[14:37] <Uraeus> BBB: so bascially none of our colorspace plugins handles BGR8 ?
[14:37] <BBB> no
[14:37] <Uraeus> unless PAL8 == BGR8
[14:37] Action: BBB will ask ffmpeg to add it
[14:38] <BBB>     PIX_FMT_PAL8,      ///< 8 bit with RGBA palette
[14:38] <Uraeus> if they don't I guess we can add BGR8 to the new gstreamer colorspace plugin that ds made :)
[14:43] <BBB> we don't support it
[14:43] <BBB> neither does ffmpeg
[14:43] <BBB> and I *don't* want colorspace code in gst-plugins
[14:43] <BBB> if we want it, add it to ffmpeg or so
[14:43] <BBB> they'll optimize it for us, it's so much simpler
[14:44] <Uraeus> I guess ffmpeg is becoming what libcolorspace was supposed to be in regards to colorspaces
[14:45] <BBB> ffmpeg is simply a project maintained by mplayer and xine
[14:45] <BBB> it's simple to use for us, too
[14:45] <BBB> we gotta move off doing everything ourselves
[14:45] <BBB> it simply doesn't work that way
[14:46] <Uraeus> um, I don't think we ever did that, gst is one of the projects using most 3rd part libs I think
[14:46] <BBB> yup
[14:46] <BBB> but we started lcs, specialib, liboil, ...
[14:47] <Uraeus> liboil?
[14:47] <BBB> ds' version of specialib
[14:53] <Uraeus> hmm, I am testing out Vega Strike, it is a rather huge CVS checkout ;)
[14:55] <LeRoutier> and it needs gtk1
[14:57] Action: BBB fights with ffmpeg again
[15:06] Rotty ([email protected]) joined #gstreamer.
[15:06] mathrick ([email protected]) joined #gstreamer.
[15:07] Action: BBB thinks his patch is pretty good now
[15:07] sublett ([email protected]) joined #gstreamer.
[15:09] <Uraeus> BBB: your autotools patch?
[15:09] <mathrick> re
[15:10] <BBB> yes
[15:11] hallibaby ([email protected]) joined #gstreamer.
[15:21] <BBB> ah, finished :)O
[15:22] Action: BBB begs ds to fix -fPIC compile of MMX/asm code
[15:22] <LeRoutier> BBB: so we'll soon use a newer snapshot of ffmpeg in gst-plugins ?
[15:24] Nick change: afk -> markey
[15:24] alley_cat ([email protected]) left irc: "May the Source be with you!"
[15:36] <BBB> LeRoutier: I certainly hope so
[15:36] <BBB> LeRoutier: but there's some issues left that I want fixed first
[15:37] <Company> BBB: can i remove the code that assumes memory layout from asf(de)mux? those #if G_BYTE_ORDER == G_BIG_ENDIAN parts?
[15:37] <BBB> Company: I think so, yes
[15:37] <BBB> Company: I assume that's similar to avidemux?
[15:37] <BBB> if so, go ahead
[15:37] <Company> yeah
[15:39] <BBB> (note that asfdemux needs a rewrite similar to avidemux to separate parsing from inerpreting)
[15:39] <BBB> +t
[15:39] <BBB> and I agree that the __atribute__(__packed__); is evil
[15:41] foser ([email protected]) joined #gstreamer.
[15:50] <LeRoutier> hi foser
[15:51] <LeRoutier> seems i've got a crash with player once the avi file has just finished playing
[15:51] <LeRoutier>  in yuv420p_to_rgba32 
[15:52] <BBB> uhm... /me has no clue
[15:53] <BBB> make a bug report + avi file
[15:53] <LeRoutier> ok
[15:53] <BBB> unless it's the 'usual scheduler bug'
[15:53] <LeRoutier> nope don't think so
[15:53] <BBB> (in which case you want to try --gst-scheduler=basicgthread)
[15:53] <LeRoutier> it's in colorspace change fct and at end of stream
[15:56] <LeRoutier> http://bugzilla.gnome.org/show_bug.cgi?id=132468  : uploading avi file asap
[15:57] <LeRoutier> 30MB. upload in progress
[16:02] <LeRoutier> very same bug with basicgthread scheduler
[16:03] <Company> yay, putting network file systems in kernel space rules
[16:04] <Company> unmountable samba share because the other computer shut down :o
[16:04] Company ([email protected]) left irc: Remote closed the connection
[16:05] bab ([email protected]) joined #gstreamer.
[16:07] <bab> hi.. can i ask what license the stuff in gstreamer/common/m4 is under?
[16:07] <bab> i'm wanting to use AS_AC_EXPAND (with modifications) in another project with some other GPLed autoconf macros
[16:08] <dolphy_afk> BBB: i would like to have access to your machine one day :)
[16:08] Nick change: dolphy_afk -> dolphy
[16:09] <dolphy> BBB: so that i can fix that Xv thing
[16:11] <dolphy> BBB: can you try to play a tagged mp3 file with gst-plugins/examples/gstplay/player please ?
[16:24] hallibaby ([email protected]) left irc: "Bye bye"
[16:27] markey ([email protected]) left irc: "leaving"
[16:28] <BBB> do you want it?
[16:31] <BBB> and how? I mean, it's not like remote X works here
[16:34] <dolphy> i don't really care of the x output
[16:34] <dolphy> but i care about adding some verbose infos in get_xv_support and look what's wrong with your card
[16:35] bab ([email protected]) left irc: "using sirc version 2.211+KSIRC/1.2.4"
[16:37] hallibaby ([email protected]) joined #gstreamer.
[16:40] <BBB> dolphy: nothing... but the driver's simply b0rked
[16:40] <BBB> xine crashes too
[16:40] <BBB> so does vlc
[16:40] <dolphy> hmm sweet :)
[16:40] <BBB> no :p
[16:41] <dolphy> BBB: can you run videotestsrc ! xvimagesink with a log level 5
[16:41] <mathrick> BBB: what marvelous card and drivers you have? :)
[16:41] <dolphy> and send that to me
[16:42] <BBB> dolphy: no, because it'll crash my computer and I'm debugging something
[16:42] <dolphy> BBB: ah it makes a complete crash ?
[16:42] <BBB> mathrick: it's a Trident CyberBlade XP/Ai1
[16:42] <BBB> dolphy: yes
[16:42] <dolphy> BBB: wow
[16:42] <LeRoutier> Trident, damn
[16:42] <BBB> (don't ask me how the f*ck this got into X and why the f*ck they didn't fix this)
[16:42] <mathrick> BBB: is that name of chip or card model?
[16:43] <BBB> trident is a vendor
[16:43] <BBB> XP/Ai1 is the card chiptype
[16:43] <dolphy> BBB: well if one day you can do that and send it to me
[16:43] <BBB> sure... is xvinfo or so good enough?
[16:43] <LeRoutier> i thought it WAS a vendor but was burned by the market
[16:43] <mathrick> BBB: wow, never heard of that. And yes, I *know* what Trident is ;)
[16:44] <mathrick> BBB: back in the old days I used Trident 128K VGA cards :)
[16:44] Action: BBB too
[16:44] <BBB> they worked well
[16:44] <dolphy> yeah we all add a trident once
[16:44] <BBB> this card is good
[16:44] <BBB> but the driver sucks
[16:47] Action: mathrick notes that debugging app whose libs have 75MB+ unstripped and takes 75MB+ of mem itself is real fun ;)
[17:04] Company ([email protected]) joined #gstreamer.
[17:08] <dolphy> Company: can you please try playing a tagged mp3 file with gst-plugins/examples/gstplay/player ?
[17:12] markey ([email protected]) joined #gstreamer.
[17:12] <LeRoutier> dolphy, they do not play. only mp3 with no id3 tag plays via player
[17:12] <dolphy> LeRoutier: i know :)
[17:12] <dolphy> LeRoutier: that's why i m asking him to give it a try :)
[17:14] <LeRoutier> well, i got a id3 tagged one that plays but it is the exeception
[17:14] <dolphy> LeRoutier: it's a spider issue
[17:15] <Company> no it's not
[17:15] <Company> if it were a spider issue src ! spider ! osssink wouldn't play either
[17:16] <Company> (unless i'm missing something)
[17:16] <dolphy> well then why south.mp3 plays just fine
[17:16] <dolphy> and only tagged mp3 are not playing
[17:18] <Company> because something is broken somewhere
[17:18] <Company> but that doesn't make it a spider issue by default
[17:18] <Company> luckily you used a big pipeline, so you can now easily find the problem :/
[17:18] <dolphy> /opt/gstreamer/bin/gst-launch-0.7 -t filesrc location=/home/ftp/disk1/music/Naked\ Music\ presents\ -\ Nude\ Dimensions\ volume\ 3/01\ -\ Naked\ Music\ presents\ -\ gaelle\ -\ rain.mp3 ! spider name=autoplug ! { queue ! volume ! osssink } { autoplug. ! queue ! ffcolorspace ! videobalance ! ffcolorspace ! videoscale ! xvimagesink }
[17:18] <dolphy> that works
[17:20] <dolphy> this pipeline is not much different than gstplay's one
[17:23] <LeRoutier> hum. seems that the tagged mp3 i have both contains id3 tags v1 & v2 at the same time (magic done by id3tag)
[17:23] <Company> yay for whoever "fixed" audioconvert
[17:24] <BBB> Company: do you see similar issues as me where audio playback in MPEG video is really slow?
[17:24] <BBB> (just to have someone confirm the issue...)
[17:24] <Company> not the last time i tried
[17:24] <LeRoutier> BBB: got it too, around 1fps for every mpg i have
[17:24] <BBB> yes, that too
[17:24] <BBB> hm
[17:24] <BBB> ..
[17:24] Action: BBB will look at it
[17:25] Action: BBB thinks ffmpeg update is important, too
[17:25] <Company> maybe a timestamp issue
[17:25] <dolphy> Company: how can i fix that tags issue with spider ?
[17:25] <LeRoutier> well, time tick repeats itself a few times
[17:25] <BBB> I'd say timestamps are fine... it looks as if audio plays too slow
[17:25] <BBB> ohwell
[17:25] <BBB> I'll look at it
[17:25] Action: BBB is happy enough that all MPEG autoplug issues appear fixed now
[17:26] <BBB> Company: oh, if you have time, can you please re-do the calculations in gsttypefindfunctions.c?
[17:26] <BBB> I changed them, so the chance calculations need to be redone
[17:26] <BBB> and I can't do them
[17:26] <BBB> gnome calculator bails out
[17:26] <BBB> :/
[17:27] <Company> use gnumeric :)
[17:28] <Company> or perl or C or ... ;)
[17:29] <Company> BBB: i fixed a bug in typefinding yesterday that might affect mpeg video btw, so _maybe_ you don't even need to change the probabilities ;)
[17:29] <BBB> hm...
[17:29] <BBB> I fixed several bugs
[17:30] Action: BBB checks company's logs
[17:30] <Company> it's a bug in the typefind element
[17:30] <Company> btw: do you use spider or typefind for your typefinding?
[17:30] <BBB> no, that's audio
[17:30] <BBB> all my bugs were video-related
[17:30] <BBB> I use gst-typefind for typefinding
[17:31] <BBB> (i.e. to test gsttypefundfunctions.c)
[17:31] <BBB> see http://freedesktop.org/cgi-bin/viewcvs.cgi/gstreamer/gst-plugins/gst/typefind/gsttypefindfunctions.c.diff?r1=1.23&r2=1.24
[17:31] <BBB> er
[17:31] <BBB> no
[17:31] <BBB> http://freedesktop.org/cgi-bin/viewcvs.cgi/gstreamer/gst-plugins/gst/typefind/gsttypefindfunctions.c
[17:31] <BBB> that
[17:31] <Company> i don't mean that fix
[17:31] <BBB> the above three are what I fixed yesterday/tpday
[17:31] <Company> i mean a core fix from tonight
[17:31] <BBB> oh
[17:31] <BBB> hm
[17:31] <BBB> do you have a diff?
[17:32] <BBB> (webcvs link)
[17:32] <Company> it's a problem with using more data than fits into one buffer (which with filesrc is more than 4096 bytes
[17:33] <Company> so i believe that might affect mpeg typefinding :)
[17:33] <BBB> mpeg typefinding uses 2048 byte bufs
[17:33] <BBB> :)
[17:33] <BBB> no, but probabilities were broken because it's a random number
[17:33] <BBB> random as in: it can be 1-100
[17:33] <BBB> same for video
[17:33] <BBB> video can be higher than sys, then video is detected
[17:34] <BBB> but that should *always* be sys
[17:34] <BBB> because video is inside sys
[17:34] <Company> and http://freedesktop.org/cgi-bin/viewcvs.cgi/gstreamer/gstreamer/gst/elements/gstbufferstore.c.diff?r1=1.9&r2=1.10 is the bug :)
[17:34] <Company> how is video muxed into sys btw?
[17:34] <BBB> check my patches... you'll see what was wrong very quickly
[17:35] <BBB> randomly
[17:35] <BBB> MPEG sys chunks are 0x500, 0x1000 or 0x2000 bytes or so
[17:35] <BBB> and each chunk contains audio or video stream, which can be 0-infinite frames
[17:35] <BBB> per chunk
[17:36] <Company> so for video/mpeg to reliably say that it's video it needs to check > 0x2000 bytes?
[17:44] <dolphy> Company: ?
[17:45] <Company> dolphy: no idea - something breaks and I don't know what
[17:45] <dolphy> hmm
[17:46] <Company> but i don'T think it's definitely spider...
[17:46] <BBB> Company: the current code works very well.. I'll leave it at that for now
[17:46] <BBB> until we get new bug reports
[17:47] KA ([email protected]) joined #gstreamer.
[17:48] <Company> BBB: ok
[17:48] <BBB> (unless you see obvious bugs or style issues)
[17:49] <dolphy> filesrc location=/home/ftp/disk1/music/Naked\ Music\ presents\ -\ Nude\ Dimensions\ volume\ 3/01\ -\ Naked\ Music\ presents\ -\ gaelle\ -\ rain.mp3 ! spider name=autoplug ! volume ! tee ! { queue ! osssink } autoplug. ! identity ! ffcolorspace ! { queue ! switch ! videobalance ! ffcolorspace ! videoscale ! xvimagesink }
[17:49] <dolphy> wow :)
[17:49] <dolphy> gstplay pipeline on gst-launch :)
[17:50] <dolphy> cute :)
[17:50] <Uraeus> dolphy: that pipeline works now? :)
[17:51] <KA> Hi. I'm trying to look inside libgstreamer to make some sort of filter/effect. PWG in unfinished yet, and talks about CVS module gst-template. But it seems different from what PWG explains. Is there a simply/std method to create a plug-in I can use now?
[17:51] <dolphy> Uraeus: that one works on gst-launch
[17:51] <dolphy> Uraeus: but for a mysterious reason not in gstplay
[17:52] <BBB> grmbl
[17:52] <BBB> I need to write docs
[17:52] <BBB> you know what?
[17:52] <BBB> I'll work on some PWG docs tonight
[17:53] <Uraeus> KA: current simplest method is taking an existing plugins that is similar to what you want to to and use that as an example
[17:54] <Company> man, i hate "i have no clue, but i'll just do something" people
[17:54] <KA> Company: are you speaking of me?
[17:54] <Company> someone should set up knowledge tests for people before they may do stuff
[17:55] <Company> no, i'm talking about someone committing stuff to audioconvert
[17:55] <Company> KA: you don't seem to add bugs to my code ;)
[17:56] <KA> Company: yes, but I've read only the last sentence, and it was ambigous ;)
[17:57] <Company> sorry :)
[17:57] <KA> well, I can hack the existing code, sure. reading PWG I undestood there should be (in future?) some sort of templating mechs to let developers to write only code for elements
[17:57] <Company> thomasvs was able to make _everything_ wrong in audioconvert
[17:57] <KA> Company: you're welcome, I asked only to be sure :)
[17:57] <Company> KA: i'm not sure how up to date our templates are
[17:58] <Company> KA: but the PWG is way out of date unfortunately
[17:58] <KA> yes, I noticed :)
[17:58] <Company> KA: you might get a good answer from jimmy_dean as he has just started hacking on gst, too
[17:58] <BBB> erm
[17:58] <BBB> there is a template
[17:58] <BBB> and I updated it recently
[17:59] <jimmy_dean> did I hear my name mentioned? :)
[17:59] <BBB> KA: CVS module 'gst-template', dir 'gst-plugin'
[17:59] <KA> BBB: updated 2mins ago
[18:00] <BBB> hm, it fails
[18:00] <BBB> core API changed
[18:00] Action: BBB fix0rs
[18:01] <KA> the template provides also some tool to apply the templating work or is it only a skeleton?
[18:02] <jimmy_dean> KA: everything you need to get started is in it
[18:02] <jimmy_dean> KA: including build scripts, etc...you just make, make install, gst-register and you can use your plugin right away
[18:04] <KA> ./autogen.sh'ing
[18:04] <KA> uh, it needs gsreamer-0.7?
[18:04] <jimmy_dean> yeah
[18:08] <Company> anyone that wants to develop for gstreamer is highly encouragwed to use 0.7
[18:08] <BBB> KA: I've got it fixed, hold on
[18:09] <BBB> comitted
[18:09] <KA> I'm looking for a 0.7 debian repository, ie the one pointed form download section 
[18:09] <BBB> (it builds now)
[18:09] <Company> the reason is that 0.7 is so much better than 0.6 that we all think 0.6 sucks and don't have a clue about its bugs anymore ;)=
[18:11] <KA> well, grestreamer.net/releases/debian/ seems not to be updated :)
[18:13] <Company> 0.7 is the debug version, so there's no debian packages yet
[18:13] <Company> s/debug/unstable/
[18:14] <KA> is 0.7.3 from sourceforge OK?
[18:14] <BBB> sure
[18:14] <Company> uh, that's already quite old
[18:14] <BBB> but I'd suggest current CVS
[18:15] <BBB> in 0.7.x, we only fix CVS bugs
[18:15] <Company> yeah, and we fix bugs a lot ;)
[18:16] <BBB> I'll work on the PWG in train tonight
[18:16] <BBB> where's the "sources"?
[18:17] <Company> docbbook in gstreamer/docs/pwg i think
[18:18] <KA> I hope the rest depending libs in debian unstable are sufficients (libxml and libglib for example) :)
[18:18] <Company> yeah
[18:18] <Company> i use debian on one of my devel machines
[18:19] <KA> BBB: start specifying that PWG refers to 0.7 (or I've not seen it?), I've looked into 0.6 sources for examples' src
[18:20] <BBB> it's only three pages or so in 0.7.x?!?
[18:21] <Company> according to http://gstreamer.net/docs/ it's more
[18:21] <Company> but maybe it's only 3 xml pages, dunno
[18:24] <BBB> no
[18:24] <BBB> the docs build isn't right here
[18:24] hallibaby ([email protected]) left irc: "Bye bye"
[18:25] <BBB> check your docs/pwg/html/index.html
[18:26] <Company> ds: wake up :)
[18:26] <BBB> thomasvs: wake up :)
[18:26] <Company> BBB: there's enough xml files in there
[18:27] <BBB> yes
[18:27] <BBB> but not HTML
[18:28] Action: Company runs --disable-.docs-build because docs building doesn't work here
[18:28] <BBB> d'oh!
[18:28] <BBB> fix it, mention it to thomasvs, I don't care
[18:40] <LeRoutier> bbl. ++
[18:40] LeRoutier ([email protected]) left irc: "Leaving"
[18:59] Action: Company tickles ds, ds-work and ds-gromit
[18:59] <BBB> which version of docbook/xsltproc do I need?!?
[18:59] <BBB> I still don't get docbuild
[18:59] <Company> me neither
[19:00] <Company> especially because thomasvs changes the tools every 2 months
[19:00] <BBB> cut the crap please?...
[19:00] Action: BBB is getting tired of the '<thomasvs> company blabla...' and '<company> thomasvs blabla ...' crap
[19:00] <Company> he really annoys be because he runs around and just does stuff
[19:01] <Company> s/be/me/
[19:01] <BBB> he might not be as good as you in schedulers, autoplugging
[19:01] <Company> most of the time it's obvious to me he doesn't know what he's doing
[19:01] <BBB> but he's doing his best
[19:01] <Company> yes
[19:01] <BBB> and he's doing this fulltime, he *did* that
[19:01] <BBB> give him some credit
[19:01] <Company> but if I don't understand something, I don't touch it
[19:02] <BBB> or help him to improve
[19:02] <BBB> that's what I'm trying to do
[19:03] <Company> it's especially troublesome for me because if i later want to remove the crap he did he comes shouting "that's a regression and most work"
[19:03] <BBB> food, brb
[19:04] <dolphy> you simply have different point of views
[19:04] <jdahlin> gstreamer needs a BDFL
[19:04] <Company> BDFL?
[19:05] <jdahlin> Benedict Dictator For Life
[19:05] <jdahlin> like Linus for the kernel or Guido for python
[19:05] <Company> yes, but there is no such person
[19:05] <Company> if there were one we'd have already declared him leader
[19:06] <dolphy> having a team is not that bad either
[19:06] <Company> yes it is
[19:06] <dolphy> the only issue is to bring the team to work together
[19:06] <Company> there you have the problem
[19:06] <Company> we're as good as the UN in resolving problems
[19:06] <jdahlin> you would have problems with a maintainer too, people will always be pissed
[19:07] <jdahlin> and have different opinions
[19:07] <dolphy> people should just make more efforts
[19:07] <dolphy> that includes you company and thomasvs
[19:07] <Company> more efforts?
[19:07] <dolphy> to understand eachothers goals and try to find common roads
[19:08] <dolphy> i m not sure i understand all your goals yet
[19:08] <dolphy> but i understand thomasvs's ones
[19:08] <dolphy> s/understand/know/
[19:09] <Company> my goal is to make gstreamer a stable, logical framework for handling streams
[19:09] <dolphy> and his one is to make gstreamer a usefull and used framework :)
[19:09] <Company> and gst-plugins a stable, logical collection of code to handle media streams based on gstreamer
[19:10] <Company> those are secondary goals for me
[19:10] <dolphy> see :)
[19:10] <jdahlin> maybe there are different views of available time
[19:11] <Company> if something is useful but illogical i'll go to some effort to make it logical, but if that's not possible, I don't want it
[19:11] <Company> same for something that is used but unstable
[19:11] <dolphy> well yeah but something that's very logical but dead because nobody else want it than you
[19:11] <dolphy> it's not really interesting either
[19:12] <Company> that's true, but it's not hindering the other parts of gstreamer at least
[19:13] <Company> useful and used are primary goals for an app, but not for a lib
[19:13] <Company> thomasvs would obviously make a great app coder
[19:13] <Company> while i suck at it :)
[19:13] <dolphy> do you know how many libs are seating on freshmeat with absolutely no users ?
[19:13] hallibaby ([email protected]) joined #gstreamer.
[19:14] <Company> no idea :)
[19:14] <dolphy> tons of
[19:14] <jdahlin> library and application developers needs to communicate to get a good library though
[19:14] <Company> yeah, probably
[19:14] <dolphy> the day when you will get pissed by gstreamer and you run away
[19:14] <dolphy> then you have an unused lib
[19:14] <dolphy> with no developpers
[19:14] <dolphy> == dead project
[19:14] <dolphy> freshmeat has tons of them too :)
[19:15] <dolphy> that's why it's good to mix both goals
[19:15] <Company> yeah, but i still believe that the reason why most people don't use gst atm is instability
[19:15] <dolphy> having a logical framework that is usefull and used
[19:15] <Company> it's not missing usefullness
[19:15] <Company> and it's not missing use
[19:15] <dolphy> well there are some issues to let people freely using it
[19:16] <dolphy> and i m exactly in the right place to tell you about it :)
[19:16] <dolphy> we are getting things fixed
[19:17] <dolphy> but still..
[19:17] <dolphy> not enough
[19:17] <Company> yeah
[19:17] <dolphy> gnome integration is a key to success for us
[19:17] <dolphy> and gnome integration is a tough task at the moment
[19:17] <Company> yeah, because noone works on it - missing manpower
[19:18] <dolphy> thomasvs, jdahlin and i are working on that but we need support from you
[19:18] <Company> yeah, we're working on it, too
[19:19] <Company> but that's a bit invisible
[19:19] <dolphy> visible points are the one that really hurt
[19:19] <dolphy> such as nautilus media and gnome sound recorder
[19:19] <Company> yeah
[19:19] <jdahlin> or totem?
[19:19] <dolphy> well not yet :)
[19:19] <dolphy> but that will come one day :)
[19:20] <jdahlin> isn't gst-totem important for both gnome and gstreamer?
[19:20] <Company> but if someone breaks the core to get visible results, it's worse than not getting those results
[19:20] <dolphy> yes but it s not hurting now 
[19:20] <jdahlin> dolphy, what's hurting us right now?
[19:20] <Company> the stuff that's supposed to be part of gnome
[19:20] <dolphy> gnome sound recorder and nautilus media are the biggest
[19:21] <Company> gst-totem will hurt us for 2.8
[19:21] <dolphy> because they are in gnome already and are broken
[19:21] <jdahlin> and more specifically, what's broken in them?
[19:21] Action: jdahlin haven't really used them
[19:21] <dolphy> jdahlin: well you better ask thomasvs :)
[19:21] Action: dilinger would like nautilus media to stop crashing so much
[19:21] <dolphy> but anyway we will fix that
[19:21] <Company> rb is broken, too
[19:21] <dolphy> we will also rewrite esd to use gst
[19:22] <Company> though rb is not part of gnome...
[19:22] <Company> rewrite esd to use gst? weee, sounds like a fun project
[19:22] <Company> _not_ :)
[19:22] <dolphy> well shouldn't be too hard
[19:22] <dolphy> thomasvs is investigating that 
[19:22] <Company> especially because deamons need to be stable more than apps
[19:23] <dolphy> gstreamer is quite stable for audio
[19:23] <dolphy> so writing a simple sound server should not be that hard
[19:23] <dolphy> we'll see
[19:23] <Company> yeah, that's true
[19:23] <Company> we're missing sound output plugins most atm
[19:23] <Company> that requires me writing a general gstaudiosink first
[19:24] <desrt> wow.  freenode is a messed up irc network.
[19:24] desrt ([email protected]) left #gstreamer.
[19:24] <Company> but that's not high on my priority list atm because there are more pressing issues (autoplugger, ogg/vorbis seperation, etc)
[19:24] <dolphy> yup
[19:24] <Company> my plan is to fix audio (ie Rhythmbox) first and after that focus on video
[19:24] <dolphy> so if you can't help on the streaminfo thing why not letting thomasvs try something
[19:25] <Company> well, it's not that i removed it because i wanted to piss someone off
[19:25] <dolphy> i know
[19:25] <dolphy> i ve looked at the streaminfo and metadata code in 0.6.x a lot of times
[19:25] <Company> i removed it because after a lot of thinking i came to the conclusion it's not possible because of too much cornercases
[19:25] <dolphy> i know how hard it sucks
[19:26] <dolphy> i think your change is a great change but still we need to bring back streaminfo one way or another
[19:27] <dolphy> we should also think about a way to get spider tell more infos on codec used
[19:27] <Company> but we've discussed this enough times, let's see how this plays out
[19:28] <dolphy> i will need to know that for totem info box
[19:28] <jdahlin> what's the status of metadata for mp3/ogg?
[19:28] <dolphy> jdahlin: tags are working in mp3
[19:28] <Company> jdahlin: that works afaik
[19:28] <Company> mp3/ogg/flac extract tags
[19:28] <jdahlin> cool, we should use that for jamboree
[19:28] <Company> though it's not working the way i'd like it to...
[19:29] <Company> jdahlin: you might want to wait until the way I like is done
[19:29] <jdahlin> Company, when will that be finished?
[19:29] <Company> jdahlin: good question - it requires a nw autoplugger...
[19:30] <Company> dolphy: the video format extraction should probably be some tag like "video-format" or similar
[19:30] <dolphy> yup
[19:30] <dolphy> avidemux should throw that
[19:30] <Company> dolphy: but that requires coming up with a way to do that and implementing it somewhere
[19:31] <Company> dunno if avidemux should throw that
[19:31] <dolphy> demuxer should send tags with codec of stream
[19:31] <Company> IMO the decoder should
[19:31] <dolphy> well that's one way to do it :)
[19:31] <jdahlin> Company, well, something like working metadata is important, so I might just go ahead and add it to jamboree and shout at you as soon as something doesn't work :)
[19:31] <Company> jdahlin: yeah, that's why it's a bad idea to do it right now, i'm already half deaf from walters shoiuting ;)
[19:32] <Company> jdahlin: how does jamboree do it right now?
[19:32] <jdahlin> Company, one ear left for me!
[19:32] <jdahlin> Company, libid3tag for mp3 and stolen parts of vorbiscomment for ogg
[19:33] <Company> jdahlin: the only advantage to that right now is removing the dependancy to those libs
[19:33] <jdahlin> Company, it sort of works today, so we're not in a big hurry
[19:34] <Company> jdahlin: and my goal is to get functions like gboolean has_audio (gchar *file) and GstTagList *get_Tags (gchar *file) working
[19:34] <Company> both of those use some ugly workarounds in Rb atm
[19:34] <jdahlin> What's GstTagList? a GHashTable basically?
[19:35] <Company> jdahlin: a GstStructure, but yes, same idea
[19:35] <dolphy> Company: all tags are strings atm right ?
[19:35] <Company> dolphy: nope
[19:35] <Company> dolphy: every tag is registered with the type it's in
[19:35] <jdahlin> some should be integers and some floats
[19:35] <Company> yeah
[19:35] <Company> they are
[19:35] <mathrick> dolphy: IMO tags with codec should be sent by decoder, this way you can avoid having some central repo of codecs mapping inside demuxers (read: not suck like DS does ;)
[19:35] <Company> tracknumber is an int for example
[19:36] <dolphy> Company: the printtag function seems to dump strings only
[19:36] <dolphy> in gst-launch
[19:36] <Company> dolphy: that's because it uses g_value_strdup_value_contents, which transforms any GValue to a string
[19:37] <Company> which is a nice thing about tags - they're not strings but can easily be made strings
[19:38] <dolphy> goood :)
[19:39] <dolphy> i ll create a hashtable in totem with strings
[19:39] <dolphy> splitting on = i guess
[19:39] <dolphy> ah not needed even
[19:39] <Company> or you could just keep the GstTagList and convert on demand
[19:39] <dolphy> hash_insert (tag_get_nick, stdrup_value_content)
[19:40] <Company> tag_get_name
[19:40] <dolphy> i like having a hash
[19:40] <Company> whatever you like
[19:40] <dolphy> because i just hash_replace and i can receive tags in different signals
[19:40] <dolphy> building a nice hash
[19:40] <Company> or you could use gst_tag_list_merge
[19:40] <dolphy> totem can then get what it needs
[19:41] <Company> which is a lot easier :p
[19:41] <Company> gst_tag_list_insert even
[19:43] <KA> gst-plugins' ./configure takes a long long long time to check all the stuff, and doesn't want to build mad pluging. libmad0{,-dev} are installed
[19:43] <KA> Company: do you get it built on your debian box?
[19:43] <apoc> BBB : Do you know what's the problem with ffmpeg and trm (musicbrainz) ?
[19:44] <apoc> hu i get this with rhythmbox
[19:44] <apoc> (rhythmbox:12004): GStreamer-CRITICAL **: Factory for `id3tag' has no type
[19:45] Action: dolphy still can't understand why gstplay is not playing the mp3 tagged file after finding the tags..
[19:46] <Company> KA: you need libid3tag for mad
[19:46] <KA> is libid3tag related with mad plugin compilation?
[19:46] <KA> indeed
[19:47] <Company> libid3tag and libmad used to be one package
[19:47] <KA> probably the same upstream tarball 
[19:48] <KA> well, food.
[20:01] BBB ([email protected]) left irc: Read error: 110 (Connection timed out)
[20:04] <dolphy> Company: got a tricky question for you
[20:04] <dolphy> Company: i have a problem with switch
[20:04] <dolphy> Company: it's loop based on pull from all the sink pads
[20:05] <dolphy> Company: but that blocks in the case of  spider ! identity ! ffcolorspace ! { queue ! switch ....
[20:05] <dolphy> Company: for audio only media
[20:05] <dolphy> Company: ronald told me you might have an idea
[20:07] <jdahlin> I have a dumb question, what's {} used for in pipelines? 
[20:07] <apoc> jdahlin: thread
[20:08] <jdahlin> oh
[20:13] hallibaby ([email protected]) left irc: "Bye bye"
[20:16] hallibaby ([email protected]) joined #gstreamer.
[20:17] ChrisHJW_log ([email protected]) joined #gstreamer.
[20:17] ChrisHJW_log ([email protected]) left irc: K-lined
[20:17] <Company> dolphy: if the scheduler would support it, you could use gst_pad_switch, but no idea otherwise
[20:18] <Company> s/gst_pad_switch/gst_pad_select/
[20:18] <Company> dolphy: an interesting idea I had last night was to start the threads of the GstPlay pipeline one after another
[20:19] <dolphy> Company: ah ?
[20:19] <dolphy> Company: it's not the case ?
[20:19] <apoc> Company: I think there was a putbits library in the core like getbits ...
[20:19] <Company> dolphy: only start (or connect) the video decoding part if identity grabs a buffer
[20:19] <Company> apoc: it wasn't LGPL
[20:20] <Company> it was mpeg licensed
[20:20] <dolphy> Company: but then identity would be connected to fakesink
[20:20] <Company> i dunno if it still exists in gst-libs/gst
[20:20] <dolphy> Company: and we would loose buffers
[20:21] <Company> dolphy: if there are queues in between, that shouldn't be a problem, because the queues get filled first
[20:22] <dolphy> Company: hmm still that's a kind of workaround
[20:23] <dolphy> Company: the ideal pipeline is the current one imho
[20:23] <dolphy> Company: i just need to find a way so that switch is not pulling from video queue when it's empty
[20:24] <Uraeus> dolphy: could you please ask mathias to package Vega Strike on Freshrpms :)
[20:25] <dolphy> Uraeus: send him a mail :)
[20:25] <dolphy> Company: so you say gst_pad_select might be the solution?
[20:25] <dolphy> Company: is that compliant with a loop based element ?
[20:27] <dolphy> Company: one more thing
[20:27] <Company> dolphy: it's not implemented in opt
[20:27] <dolphy> Company: i don't really get the point of gst_tag_list_foreach
[20:27] Zeenix ([email protected]) joined #gstreamer.
[20:27] <dolphy> Company: it's calling func for each tag
[20:28] <Company> dolphy: yes
[20:28] <dolphy> Company: but then func has to walk through the tag list to get the tags value
[20:28] <Uraeus> hi Zeenix
[20:28] <Zeenix> hello
[20:28] <dolphy> Company: wouldn't it be easier to call func with tag name and tag value ?
[20:29] <Zeenix> Uraeus: are you using Nvidia's graphics card on fedora core 1?
[20:29] Nick change: harshyZ -> harshy
[20:30] <Uraeus> Zeenix: yes
[20:30] <Zeenix> s/on/with
[20:30] <Uraeus> morning harshy
[20:30] <Company> dolphy: i dunno what the reason was to not do this or if i simply forgot it
[20:30] <harshy> morning
[20:30] <Zeenix> Uraeus: and the latest drivers from Nvidia.com?
[20:30] <Uraeus> Zeenix: gotten anywhere with adding that xoverlay interface to the libcaca plugin?
[20:30] <Uraeus> Zeenix: yes
[20:30] <dolphy> Company: ah right
[20:30] <Company> dolphy: you could file a bug about it
[20:31] <dolphy> Company: because the print tag function seems a bit weir
[20:31] <dolphy> weird sorry
[20:31] <Zeenix> Uraeus: I've been in too many troubles latelly
[20:31] <Uraeus> Zeenix: troubles?
[20:31] <Uraeus> Zeenix: nothing to serious I hope
[20:31] <Zeenix> Uraeus: the drivers dont work for me
[20:32] <Uraeus> Zeenix: they don't compile properly with your kernel?
[20:32] <Zeenix> Uraeus: no not so serious, but having severe hardware problem
[20:32] <Company> dolphy: ah right
[20:32] <Company> dolphy: the value mis not included, because there is no value - there might be more than one
[20:32] <Zeenix> Uraeus: compile? I did'nt know they compile. i just run their standard installer & it does all installation
[20:33] <Uraeus> Zeenix: yes, it compiles a small kernel interface stub
[20:33] <Uraeus> Zeenix: no errors during install?
[20:33] <Zeenix> Uraeus: no, it says it installed successfully
[20:33] <Uraeus> Zeenix: what exactly happens? do it spend an eternity getting your login window up?
[20:34] <Zeenix> Uraeus: yes the first time i do startx my screen just goes off untill i press alt+f1
[20:34] <dolphy> Company: so there's no way to do it differently ?
[20:35] <Company> dolphy: i didn't want to expose GValue and GValueList too much
[20:35] <Uraeus> Zeenix: you have updated your XF86Config file?
[20:35] <dolphy> Company: and basically elements are sending a tag list with a single tag inside and print_tag is dumping all the tags of the list on each print_tag call
[20:36] <dolphy> Company: doesn't seems so optimal..
[20:36] <Zeenix> Uraeus: the second time i do startx, it does start X successfully (but 3d graphics almost kill it) & after i log-out, it does'nt allow me to go to X again
[20:38] <Company> dolphy: huh?
[20:38] <Company> dolphy: GstTagList is a list of tags and each tag has a list of values attached to it
[20:39] <dolphy> oh :)
[20:39] <dolphy> right
[20:39] <mathrick> Zeenix: did you comment out Load "GLcore" & Load "dri" as stated in README?
[20:39] <mathrick> Zeenix: and added Load "glx"?
[20:39] <Zeenix> mathrick, yup
[20:40] <mathrick> Zeenix: no fiddling with options?
[20:40] <Zeenix> mathrick: yeah Uraeus just pointed out something imp. that should be added in options
[20:41] <mathrick> Zeenix: hmm? what's that?
[20:41] <Zeenix> mathrick: but i dont think that that would solve the problem as even if X startx successfully, it startx to suck on opengl apps.
[20:42] <Zeenix> mathrick:  Option "IgnoreDisplayDevices" "DFP, TV"
[20:42] Nick change: bluejay|sleep -> bluejay|busy
[20:42] <mathrick> Zeenix: at the end of readme there is troubleshooting section, try various knobs listed in there. Things like disabling XAA, NoHWCursor, make sure no hw accelerated Render is set
[20:47] <Zeenix> mathrick: i'll try that but are you sure that nvidia tnt2 has 'no hw acceleration"?
[20:48] Shoragan ([email protected]) left irc: "Leaving"
[20:50] <mathrick> Zeenix: hw render should be disabled by default, but who knows what have gotten into your XF86Config ;)
[20:51] <mathrick> Zeenix: hw render == hw accelerated X Render extension
[20:53] hadley ([email protected]) joined #gstreamer.
[21:16] shawarma ([email protected]) joined #gstreamer.
[21:16] <shawarma> Hi! Is anybody working on mms-stream-support?
[21:18] jimmy_dean ([email protected]) left irc: "Carrier Signal Lost"
[21:24] <Uraeus> shawarma: yes, mathrick is
[21:25] hallibaby ([email protected]) left irc: "Bye bye"
[21:30] jimmy_dean ([email protected]) joined #gstreamer.
[21:35] <shawarma> Uraeus: Has he just started or? Does anyone know an ETA?
[21:36] jimmy_dean_ ([email protected]) joined #gstreamer.
[21:36] <ds-gromit> moo
[21:36] jimmy_dean ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[21:36] <ds-gromit> gromit is the name of the computer
[21:37] mxpxpod ([email protected]) joined #gstreamer.
[21:38] <mathrick> shawarma: just started
[21:40] Nick change: mathrick -> mathrick|busy
[21:40] <Company> ds-gromit: answer my mail, I need to go on with fixing audioconvert :)
[21:40] <shawarma> mathrick|busy: Need any help?
[21:41] <mathrick|busy> shawarma: could use, although I haven't sorted everything out myself yet
[21:42] wheels ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[21:43] <Company> mathrick|busy: you plan to srite only mmssrc?
[21:43] <Company> or do you want to write mmssink, too?
[21:43] <mathrick|busy> Company: mmssrc is forseeable future, sink is much, much harder
[21:44] jdahlin ([email protected]) left irc: "Leaving"
[21:45] <Company> mathrick|busy: dunno, it'd take time at least :)
[21:47] <mathrick|busy> Company: sending rev-enged protocol is waaay harder than reading it, you should know that ;)
[21:48] <Company> and it's probably stupid, because it's patented...
[21:48] <Company> mathrick|busy: depends, you can at least make it send for your reader :)
[21:49] <mathrick|busy> Company: heh, yep, but it's useless IMO, as micros~1 abandoned it itself :)
[21:49] <Company> mathrick|busy: what does ms use nowadays?
[21:50] <mathrick|busy> Company: rt(s)p
[21:50] <Company> really?
[21:50] <mathrick|busy> Company: rumour is, that they did that in fear of revealing mms specs :)
[21:50] <Company> not msrtp or some such?
[21:50] <mathrick|busy> Company: at least that's what I've found on net
[21:50] <Company> mms is a crappy format from the design view...
[21:52] <mathrick|busy> Company: hmm, have you used it much to be able to tell?
[21:52] Action: mathrick|busy not
[21:53] <Company> mathrick|busy: i've read the reverse engineered specs
[21:54] <mathrick|busy> Company: I'm not that familiar with them yet
[21:54] <mathrick|busy> Company: anything particularly nasty?
[21:55] <Company> mathrick|busy: it just feels patchy
[21:55] <mathrick|busy> Company: ah, your uber lib designer sense... ;)
[21:56] <Company> something like that
[21:56] <mathrick|busy> Company: but I must admit, parts of that protocol are really flooring, like "upgrade your player movie" one ;)
[21:57] <Company> a format that does tcp for every message _but_ one which is sent via udp when using udp
[21:57] <Company> stuff like that...
[21:58] <shawarma> mathrick|busy: right. How do I get access to the CVS?
[21:58] <shawarma> Whoops.
[21:58] <Company> every message looks diferent even though they use the same info
[21:59] <Company> and the fact that they use LE float
[21:59] Nick change: harshy -> harshyLP
[22:00] wheels ([email protected]) joined #gstreamer.
[22:01] <mathrick|busy> Company: I noticed also they never get below something like 3 copies of string data in single packet
[22:07] jimmy_dean_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[22:09] jimmy_dean ([email protected]) joined #gstreamer.
[22:18] ChrisHJW_log ([email protected]) joined #gstreamer.
[22:18] ChrisHJW_log ([email protected]) left irc: K-lined
[22:19] sri ([email protected]) left irc: "Client exiting"
[22:39] mxpxpod ([email protected]) left #gstreamer ("Client exiting").
[23:00] BBB ([email protected]) joined #gstreamer.
[23:00] <BBB> thomasvs: theer?
[23:06] Gldm ([email protected]) joined #gstreamer.
[23:06] <Gldm> hi
[23:08] <BBB> hi
[23:14] <dolphy> BBB: he is skiing :)
[23:14] <dolphy> BBB: i guess he is coming back tonight but might be late in the night
[23:15] <dolphy> BBB: for my switch problem, Company suggested using gst_pad_select but that's not implemented by opt
[23:15] <BBB> nope
[23:15] <BBB> I faced that issue too ;)
[23:15] Action: BBB just documented most interfaces...
[23:16] <BBB> I want to work on the PWG, but can't
[23:16] <BBB> docs aren't building here
[23:16] <BBB> ;)
[23:16] <Gldm> I wish I was skiing :P
[23:16] <BBB> I wish I was in the US
[23:16] <dolphy> BBB: raaah stop discouraging me like that :)
[23:16] <BBB> dolphy: does your company wanna rent me an appartment in NYC?
[23:16] <dolphy> BBB: gni ?
[23:17] <BBB> gni?
[23:17] <Gldm> I'm in NY but still wish I was skiing :P
[23:17] <BBB> oh man! how do I get an appartment in manhattan?
[23:17] <dolphy> BBB: gni? == i m afraid that i don't really understand what you mean can you be more precise please ?
[23:18] <dolphy> BBB: i know a very nice girl who lives there
[23:18] <BBB> long version: I have a room there but want an apprtment
[23:18] <BBB> how do I get one?
[23:19] <dolphy> well apartment in manhattan costs a lot :)
[23:19] <BBB> I know :p
[23:21] <BBB> still I want one ;)
[23:25] Uraeus ([email protected]) left irc: "Client exiting"
[23:29] BBB ([email protected]) left irc: "Client exiting"
[23:34] foser ([email protected]) left irc: "[ I want to believe ]"
[23:35] foser ([email protected]) joined #gstreamer.
[23:35] foser ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[23:56] <KA> what about ladspa-mbeq? does it really equalize or it is a work-in-progress?
[00:00] --- Mon Jan 26 2004
[00:15] alley_cat ([email protected]) joined #gstreamer.
[00:15] <KA> good night[/morning/'noon].
[00:15] KA ([email protected]) left #gstreamer.
[00:17] Action: Misirlou has never heard "good 'noon."
[00:18] ChrisHJW_log ([email protected]) joined #gstreamer.
[00:18] ChrisHJW_log ([email protected]) left irc: K-lined
[00:22] markey ([email protected]) left irc: "g'nite"
[00:22] <thomasvs> dolphy: get to bed, tomorrow is a long working day
[00:27] Nick change: thomasvs -> thomasvz
[00:29] Nick change: mathrick|busy -> mathrick|sleep
[00:29] <mathrick|sleep> g'nite
[00:31] hadley ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[00:32] thread ([email protected]) left #gstreamer ("out").
[00:44] Gldm ([email protected]) left irc: 
[00:49] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[00:58] Shoragan ([email protected]) joined #gstreamer.
[01:07] aeyakovenko ([email protected]) joined #gstreamer.
[01:10] <aeyakovenko> is there v4l support in stable release?
[01:15] aeyakovenko ([email protected]) left #gstreamer.
[01:26] aeyakovenko ([email protected]) joined #gstreamer.
[01:26] <aeyakovenko> is there anybody out there?
[01:34] iain ([email protected]) joined #gstreamer.
[01:36] <aeyakovenko> is there v4l support in stable gstreamer
[01:38] <Zeenix> aeyakovenko: i think there should be
[01:38] <Zeenix> aeyakovenko: try it
[01:40] <aeyakovenko> i only saw gstv4lsrc on the mailing list, but when i try gst-launch-0.6 gstv4lsrc location=/dev/video0 i get an error "ERROR: pipeline could not be constructed: No such element gstv4lsrc"
[01:41] <aeyakovenko> the editor doesn't show one either
[01:44] <Zeenix> no it's name is v4lsrc
[01:44] <aeyakovenko> Zeenix: same thing ERROR: pipeline could not be constructed: No such element v4lsrc
[01:45] <Zeenix> aeyakovenko: have you installed the v4l plugin?
[01:46] Nick change: kmaraas -> km_sleep
[01:46] <aeyakovenko> Zeenix: is it part of the gst-plugins-0.6.4?
[01:48] <Zeenix> aeyakovenko: in the stable version i have (shipped with fedora c1) it is part of gstreamer-plugins-0.6.3-3 package
[01:49] <aeyakovenko> Zeenix: yea, i have 0.6.4 i believe thats the latest stable on Gentoo
[01:49] <Zeenix> aeyakovenko: have you run gst-register?
[01:49] hadley ([email protected]) joined #gstreamer.
[01:49] <aeyakovenko> Zeenix: yep
[01:50] <aeyakovenko> Zeenix: no v4l stuff there
[01:50] <aeyakovenko> Zeenix: is it part of your release?
[01:50] <Zeenix> aeyakovenko: yeah
[01:52] <aeyakovenko> Zeenix: oh ok, i think i know why
[01:54] <aeyakovenko> cool, does anyone know the state of the rtp sink and src?
[01:54] Action: Zeenix hides...
[01:56] <Zeenix> aeyakovenko: no real progress have been made in the direction of RTP/RTSP
[01:56] <aeyakovenko> i worked with Jrtplib, its a pretty complete rtp/rtcp stack but its in c++
[01:56] <aeyakovenko> i am trying to hack together a sip client so i can video chat with my girlfriend that has xp :(
[01:58] Shoragan ([email protected]) left irc: "Leaving"
[01:59] <Zeenix> aeyakovenko: i'll never have a girlfriend who uses xp :)
[02:00] <aeyakovenko> hehe
[02:02] <Zeenix> aeyakovenko: it would be very good if you can help us in making a rtp/rtsp plugin
[02:03] <Zeenix> aeyakovenko: i've recently learned that we have no other choice but to make a plugin from any rtp/rtsp lib
[02:03] <aeyakovenko> id be glad too.  would it be possible to wrap a c++?  
[02:03] <Zeenix> aeyakovenko: yes
[02:04] <Zeenix> aeyakovenko: we dont mind C++ when we dont have any other choice
[02:05] <aeyakovenko> well, is the librtp stuff near complete?
[02:05] <Zeenix> aeyakovenko: librtp?
[02:05] <aeyakovenko> the rtp source sink stuff, i guess thats what the status page says is a dependency for the plugin
[02:06] Rotty ([email protected]) left irc: Remote closed the connection
[02:06] <aeyakovenko> sorry, i just started looking at gstreamer, so i am not to familiar with all the details
[02:07] <aeyakovenko> if you want to take a look at jrtplib its at http://lumumba.luc.ac.be/jori/jrtplib/jrtplib.html  i used it for a sip voice mail server
[02:08] <Zeenix> aeyakovenko: are you it's auther?
[02:08] <Zeenix> s/auther/author
[02:08] <aeyakovenko> Zeenix: no, but the company that i worked on used the libs
[02:08] <aeyakovenko> s/on/for
[02:09] <aeyakovenko> it was a startup me and my buddies started in college, but unfortunately it sank :(
[02:09] <Zeenix> aeyakovenko: is the lib. liscenced under LGPL?
[02:10] <aeyakovenko> yep
[02:10] <aeyakovenko> is gstreamer gpl?
[02:11] <Zeenix> aeyakovenko: i hope you know that LGPL & GPL are different
[02:11] <iain> LGPL is the main core
[02:11] <aeyakovenko> yea i know
[02:11] <Zeenix> aeyakovenko: no gst is lgpl
[02:11] <iain> plugins are "various" mostly LGPL though
[02:12] <aeyakovenko> there shouldnt be a problem with using a lgpl library though
[02:12] <Zeenix> aeyakovenko: if jrtplib is lgpl then please start making the rtp plugin we are missing since a long time
[02:13] thomasvz ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[02:13] <Zeenix> aeyakovenko: i'll try my best to help you
[02:13] <aeyakovenko> ill give it a shot.  i am not to familiar with gstreamer, but the docs look pretty complete
[02:14] <Zeenix> aeyakovenko: trust me, you dont need to know much in order to code for/with gst :)
[02:14] <aeyakovenko> if i am receiving h323 video stream then i need a rtpsrc->h323decoder->videosink, right?
[02:17] <Zeenix> aeyakovenko: most problly yes, i am forgetting much of what we decided about rtp plugins, i need to dig into irc logs
[02:19] <aeyakovenko> does the autoplugin choose the element dynamicaly?  based on what is coming out of the sources?
[02:20] <Zeenix> aeyakovenko: yes but AFAIK not based on the data but the caps on the pads
[02:21] <Zeenix> aeyakovenko: i'll recommend you to read the gstreamer docs
[02:22] <aeyakovenko> Zeenix: i've been reading them, its very cool architecture, just some things i dont quite understand yet
[02:22] <Zeenix> aeyakovenko: but this lib. is rtp only? the lib. on live.com is rtsp too
[02:22] iain ([email protected]) left irc: "And like that *poof* he's gone"
[02:22] <aeyakovenko> Zeenix: i am not familiar with rtsp, what is the difference?
[02:23] <aeyakovenko> oh, ok
[02:23] <Zeenix> aeyakovenko: rtsp is concerned with the session things, whereas rtp is concerned with real streaming 
[02:24] <aeyakovenko> Zeenix: jrtplib doesnt have the sessoin things, it relies on sdp or sip+sdp to do that
[02:28] <aeyakovenko> Zeenix: looks like rtsp uses rtp
[02:29] sublett ([email protected]) left irc: "I like food, food is good!"
[02:29] <aeyakovenko> hehe, i wonder if i can write a siprtp element
[02:31] thomasvz ([email protected]) joined #gstreamer.
[02:32] <aeyakovenko> but i dont think that would fit to well, sip has way to much state junk
[02:32] shawarma ([email protected]) left irc: "BitchX: its not for the tabkey impaired!"
[02:39] <aeyakovenko> are there plugin specific docs?
[02:47] <Company> ds-gromit: moo?
[02:48] <Company> aeyakovenko: you're the 3rd to ask that today, unfortunately the plugin writers guide is way outdated
[02:49] <Company> aeyakovenko: suggested method to get into gstreamer is using the gst-template module and starting from there
[02:50] <aeyakovenko> woo hoo, i made a pipleine from my webcam
[02:50] <Company> :)
[02:51] <Company> mastering gst-launch is an important thing when starting gst hacking
[02:53] Zeenix ([email protected]) left irc: Read error: 110 (Connection timed out)
[02:53] <aeyakovenko> is there a splitter for video streams?
[02:54] <Company> splitter?
[02:55] <Company> to allow duplicating the stream?
[02:55] <Company> in that case it's tee
[02:56] alley_cat ([email protected]) left irc: "May the Source be with you!"


-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn