IRC Logs

*******************************************************************
[03:01] BBB ([email protected]) joined #gstreamer.
[03:01] <ds-work> hey
[03:01] <BBB> howdy
[03:02] <BBB> so this is how you normally work? with nobody around? ;)
[03:02] <BBB> at least you're not continuously distracted ;)
[03:02] <BBB> (guess what I'm going to do! *evil grin*)
[03:05] <ds-work> ?
[03:05] <BBB> nevermind :) I'm just tired and still at work so I'm a bit... dunno... weird
[03:05] <ds-work> yeah, things are pretty quiet once Company, dolphy, and thomasvs go to sleep
[03:06] <ds-work> me, taaz, and thaytan, mostly.  We don't talk much
[03:07] <BBB> so it's about time that I change that
[03:07] sub_pop ([email protected]) left irc: "Client exiting"
[03:08] <ds-work> hah, that's what everyone say
[03:08] <ds-work> s
[03:09] <thaytan> ... and the channel goes quiet ;)
[03:10] <BBB> tsk, I'm doing experiments here ;)
[03:10] <BBB> that's not fair
[03:11] <BBB> if I were at home, if I had a decent laptop with a working screen.... then I'd show you something! ;)
[03:11] <thaytan> if I were at home, and my workstation wasn't crashing all day long or my laptop hadn't been stolen... I'd watch :)
[03:12] <ds-work> BBB: what's your eta for a fixed laptop?
[03:12] Nick change: kh_out -> khalek
[03:13] <BBB> was supposed to arrive friday
[03:13] <BBB> I ordered a screen with ultra-fast delivery (one-day, for a #!^^!@^!@^!@ $50!!!) thursday morning
[03:13] <BBB> didn't receive anything yet so far
[03:14] <BBB> the money was - of course - discarded right away
[03:14] <BBB> so.... I'll wait another few days and then fetch the money back
[03:14] <ds-work> grrr
[03:14] <BBB> I'm calling them every day to ask what/when, but that doesn't seem to help
[03:15] <ds-work> has it been shipped yet?
[03:15] <BBB> anyway, my company is buying me a new laptop if I can't get the old one working by the end of this week, it seems, so if the screen doesn't arrive, I'll get myself a new laptop
[03:15] <BBB> ds-work: don't know, I'll ask tomorrow... so far, they only said they "didn't know"
[03:15] <BBB> which is very useful from a customer's point-of-view
[03:16] <BBB> anyway, friday's either a new screen or a new laptop... we'll see how it goes
[03:24] Action: ds-work goes home
[03:35] <BBB> er...
[03:35] <BBB> ok, so apparently Leon has some different ideas on GStreamer & MAS than I originally thought
[03:36] <BBB> http://lists.kde.org/?l=kde-multimedia&m=107758663409264&w=2
[03:36] <BBB> that's pretty f*cked up
[03:50] aldug ([email protected]) left irc: Client Quit
[03:52] bilboed__ ([email protected]) joined #gstreamer.
[03:53] bilboed ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:00] danb ([email protected]) left irc: Read error: 60 (Operation timed out)
[04:11] BBB ([email protected]) left #gstreamer.
[04:21] ds-work_ ([email protected]) joined #gstreamer.
[04:22] ds-work_ ([email protected]) left irc: Client Quit
[04:23] ds ([email protected]) joined #gstreamer.
[04:39] Nick change: jcsston` -> jcsston
[04:52] Dcoder ([email protected]) joined #gstreamer.
[04:56] danb ([email protected]) joined #gstreamer.
[05:00] casaxno ([email protected]) left irc: Read error: 60 (Operation timed out)
[05:08] tehmiller ([email protected]) joined #gstreamer.
[05:58] harshy ([email protected]) joined #gstreamer.
[06:05] sub_pop ([email protected]) joined #gstreamer.
[06:05] 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!"
[06:28] sub_pop ([email protected]) left irc: "Client exiting"
[06:38] Marsupilami23 ([email protected]) joined #gstreamer.
[06:59] Nick change: mathrick|sleep -> mathrick
[06:59] <mathrick> re
[07:01] ChrisHJW ([email protected]) joined #gstreamer.
[07:08] steveb__ ([email protected]) joined #gstreamer.
[07:12] md` ([email protected]) left irc: Read error: 60 (Operation timed out)
[07:27] steveb_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[07:43] ChrisHJW ([email protected]) left irc: Client Quit
[08:17] Dcoder ([email protected]) left irc: Read error: 60 (Operation timed out)
[08:51] Company ([email protected]) joined #gstreamer.
[09:05] markey ([email protected]) joined #gstreamer.
[09:05] dilinger ([email protected]) left irc: Read error: 60 (Operation timed out)
[09:09] md` ([email protected]) joined #gstreamer.
[09:12] jcsston` ([email protected]) joined #gstreamer.
[09:18] markey ([email protected]) left irc: "later"
[09:24] jcsston ([email protected]) left irc: Read error: 60 (Operation timed out)
[09:31] <Company> damn
[09:31] <Company> i need the new bytestream now...
[09:36] <alley_cat> does gst-player still use the gconf keys for audio/videosinks?
[09:38] grub_booter ([email protected]) joined #gstreamer.
[09:39] <Company> it should
[09:40] <Company> i'm pretty sure it does now that i think about it
[09:40] kmaraas ([email protected]) left irc: Client Quit
[09:40] <Company> because i use osssink device=/dev/dsp1 :)
[09:40] <alley_cat> it tries to use ossink here
[09:41] <alley_cat> although gconf has audioscale ! audio/x-raw-int,rate=48000 ! alsasink device=dsp1
[09:43] <Company> it uses osssink if parsing the entry fails
[09:43] <Company> it should be alsasink device=hw:1 i guess?
[09:43] Company ([email protected]) left irc: Remote closed the connection
[09:43] <alley_cat> device=dsp1 works fine in gst-launch, and used to work with gst-player
[09:45] <AlexCONRAD> morning all ! :)
[09:45] Nick change: AlexCONRAD -> KoRnouille
[09:45] <alley_cat> hmm, can't play vorbis files anymore, opt scheduler breaks ...
[09:46] Action: KoRnouille is ready for a day of python coding
[09:48] <alley_cat> ok, totem gives some more useful error message
[09:49] <alley_cat> GConf audio sink not found, using osssinkGConf video sink not found, using xvideosink
[10:02] <alley_cat> i can still play mp3s and mpeg2 videos, but that's about it right now :/
[10:02] markey ([email protected]) joined #gstreamer.
[10:03] sxpert_work ([email protected]) left irc: Client Quit
[10:06] thomasvs ([email protected]) left irc: Read error: 60 (Operation timed out)
[10:09] jcsston` ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:16] <KoRnouille> is a pipeline a BIN ? Or does a BIN has to be in a pipeline, and a pipeline is unique ?
[10:21] dolphy ([email protected]) joined #gstreamer.
[10:22] <KoRnouille> hello dolphy
[10:23] jdahlin ([email protected]) joined #gstreamer.
[10:23] <dolphy> morning
[10:27] steveb_ ([email protected]) joined #gstreamer.
[10:32] sxpert_work ([email protected]) joined #gstreamer.
[10:34] <KoRnouille> what the equivalent in python for gst_element_link_many ?
[10:34] <KoRnouille> I try : gstreamer.gst_element_link_many(filesrc, mpegparser, udpsink) // but I get an AttributeError: 'module' object has no attribute 'gst_element_link_many'
[10:37] thomasvs ([email protected]) joined #gstreamer.
[10:38] <jdahlin> KoRnouille: cvs up?
[10:38] <jdahlin> KoRnouille: err, gstreamer.element_link_many
[10:38] <jdahlin> only with cvs though
[10:38] <jdahlin> brb
[10:38] jdahlin ([email protected]) left irc: "Klienten avslutas"
[10:39] <KoRnouille> humm... doesn't seem to be that syntax
[10:41] <KoRnouille> element = gst_element_factory_make ("mad", "decoder"); <--- "decoder" is the name, right ?
[10:42] <KoRnouille> g_object_set (G_OBJECT (element), "name", "mydecoder", NULL); <--- so setting "mydecoder" will change the name "decoder" to "mydecoder", right ?
[10:44] <thomasvs> KoRnouille: are you doing c or python ?
[10:44] <KoRnouille> python
[10:44] <KoRnouille> but Im reading the C API
[10:44] <KoRnouille> :/
[10:44] <thomasvs> KoRnouille: how about element = Element ("mad", "decoder") in python ?
[10:45] thaytan ([email protected]) left irc: "leaving"
[10:45] <KoRnouille> well, I don't know... I don't have the python docs
[10:45] thaystable ([email protected]) joined #gstreamer.
[10:45] <KoRnouille> but people told me that it was a 1 to 1
[10:46] steveb__ ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:46] <KoRnouille> I used this for making an element : udpsink = gstreamer.gst_element_factory_make("udpsink", "udp_sink")
[10:49] <thomasvs> KoRnouille: no, I'm telling you "how about x" as in "try x"
[10:50] jdahlin ([email protected]) joined #gstreamer.
[10:51] markey ([email protected]) left irc: "bbl"
[10:51] <KoRnouille> im new to that... you're telling me to write this so it works under X ?
[10:55] <thomasvs> huh ?
[10:55] <KoRnouille> im just trying to get that live_server application from videowall written into python
[10:55] <thomasvs> I'm picking up on your questions, you asked stuff about the C api and you're programming in python
[10:55] <thomasvs> so I'm assuming you want to know "How can I create a new element in python with a given name"
[10:55] <thomasvs> so I answered tha timplicit question
[10:56] Shoragan ([email protected]) joined #gstreamer.
[10:56] <thomasvs> ah, I figured out why you think it has something to do with X :)
[10:56] <KoRnouille> lol
[10:56] <thomasvs> when I say "how about x" I really meant "how about (here's a variable that you have to substitute from my previous sentence)"
[10:56] <thomasvs> so, I meant "try element = Element ("mad", "decoder") in python"
[10:57] <thaystable> hi KoRnouille, thomasvs 
[10:57] <thomasvs> hey horse
[10:58] <KoRnouille> okay, let me get clear. I'm a (new) python man. I don't know C. I want to write my application in python. But I don't have examples of how to write gstreamer in python. So because someone told me that python binding were just the C bindings appending "gstreamer.", I'm looking at the C API and try to write my stuff in python
[10:58] <thomasvs> ok, there are a whole bunch of python examples
[10:58] <thomasvs> check gst-python/examples
[10:59] <thomasvs> thaystable: I have some diffs for gst-editor, to add an error handler and so on
[10:59] <thomasvs> thaystable: I am sort of unsure what to do with them currently though
[10:59] <thomasvs> thaystable: I had nice ideas of having the element with the error have a little exclamation mark that would be clickable to get the error dialog, but I'm not a gui guy
[11:00] <thomasvs> thaystable: want me to just commit and let you work on it instead ?
[11:01] pmfp ([email protected]) joined #gstreamer.
[11:18] sysan ([email protected]) joined #gstreamer.
[11:19] thaystab1e ([email protected]) joined #gstreamer.
[11:21] wheels ([email protected]) left irc: "using sirc version 2.211+KSIRC/1.3.9"
[11:29] sysan ([email protected]) left irc: "Miranda IM! Smaller, Faster, Easier. http://miranda-im.org"
[11:31] thaystable ([email protected]) left irc: Read error: 113 (No route to host)
[11:33] <jdahlin> taaz: has examples/identity.py ever worked?
[11:44] pb_ ([email protected]) joined #gstreamer.
[12:41] sysan ([email protected]) joined #gstreamer.
[12:46] sublett ([email protected]) joined #gstreamer.
[12:50] <KoRnouille> is ti better to link, then to add to bin ? or the otherway around ? or it doesn't matter ?
[12:50] <KoRnouille> s/ti/it/
[12:51] <dolphy> should not matter
[12:55] <KoRnouille> ok dolphy
[12:57] Action: bilboed__ is back (gone 09:04:42)
[12:57] Nick change: bilboed__ -> bilboed
[12:57] <bilboed> anybody knows how to remove modules in a cvs ? A friend of mine made a cockup on our server :(
[12:58] <sysan> hi folks i have a problem with a cross compiled gstreamer 0.7.3+ for arm the error is GStreamer-ERROR **: No default scheduler name - do you have a registry ? althought I have run gst-register succesfully
[12:58] <sysan> regarding the cvs I think that you can delete the directory from the repository
[12:59] <bilboed> sysan: I only have remote access, it's the cvs on sourceforge (pitivi is the name of the project)
[13:01] sxpert_work ([email protected]) left irc: Client Quit
[13:02] <thomasvs> sysan: what does gst-inspect say ?
[13:02] ensonic ([email protected]) joined #gstreamer.
[13:03] <grub_booter> bilboed - interesting project - what's the current status? got anything demonstrable?
[13:03] <ensonic> hi all
[13:04] <bilboed> grub_booter: well we've got to present something for the 6th of march... so I guess we HAVE to have something demonstrable
[13:04] <grub_booter> heh
[13:04] <grub_booter> presenting to whom?
[13:05] <bilboed> school :) It's our end of studies project
[13:05] <bilboed> still 12months to go
[13:05] <grub_booter> ah - k
[13:05] <bilboed> for those who heard of the efarm project, it's another eos project
[13:05] foser ([email protected]) joined #gstreamer.
[13:06] markey ([email protected]) joined #gstreamer.
[13:06] <grub_booter> neither efarm nor eos mean much to me i'm afraid
[13:06] <bilboed> efarm: p2p (and other protocols) server, eos : end of studies
[13:07] markey ([email protected]) left irc: Client Quit
[13:07] <KoRnouille> pipeline.set_state(STATE_PLAYING)
[13:07] <KoRnouille> NameError: name 'STATE_PLAYING' is not defined. What's wrong here ? Isn't this the right syntax ?
[13:07] <KoRnouille> I tried adding quotes to it...
[13:09] <thomasvs> KoRnouille: gst.STATE_PLAYING ?
[13:10] <jdahlin> KoRnouille: I'm going to update all the examples later today
[13:10] sysan ([email protected]) left irc: Excess Flood
[13:10] <KoRnouille> jdahlin: they *are* wrong, right ? (well, not wrong, depreciated)
[13:11] <jdahlin> KoRnouille: Well, they should work out of the box, but I'm going to clean them up
[13:14] <KoRnouille> woooow !
[13:14] <KoRnouille> I wrote my 1st python application dealing with API
[13:14] <KoRnouille> :)
[13:14] Action: KoRnouille is happy
[13:14] <KoRnouille> I rewrote the live_server.c to live_server.py
[13:15] <KoRnouille> now, the client... (hummmf, gotta deal with graphics)
[13:17] <KoRnouille> python is really neet
[13:19] harshy ([email protected]) left irc: Read error: 111 (Connection refused)
[13:23] Surphaze ([email protected]) joined #gstreamer.
[13:31] Action: bilboed is away: The power of soap...
[13:44] foser ([email protected]) left irc: "[ I want to believe ]"
[13:47] sysan ([email protected]) joined #gstreamer.
[13:53] Action: bilboed is back (gone 00:21:21)
[13:53] bilboed_ ([email protected]) left irc: Remote closed the connection
[14:00] <bilboed> KoRnouille: once you're done, could you put your python code online ?
[14:03] sxpert_work ([email protected]) joined #gstreamer.
[14:04] Nick change: khalek -> kh_zZz
[14:06] <bilboed> how can I know what plugins are going to be compiled in gst-plugins without running ./configure again ?
[14:07] steveb__ ([email protected]) joined #gstreamer.
[14:08] <sysan> does anyone has tested the gsm plugin on arm platform
[14:09] <bilboed> gstreamer + arm = opensource archos ?
[14:11] <sysan> well it works well on Ipaqs with familiar linux :))
[14:11] <bilboed> wow
[14:16] <thomasvs> ensonic: what are you doing with gstreamer ?
[14:20] Jara[zZ] ([email protected]) joined #gstreamer.
[14:21] n3m0gl081n3 ([email protected]) joined #gstreamer.
[14:26] steveb_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[14:27] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[14:39] markey ([email protected]) joined #gstreamer.
[14:40] Kaetzchen ([email protected]) joined #gstreamer.
[14:40] <ensonic> thomasvs: we are trying to develop a replacement for buzz (www.buzzmachines.com)
[14:41] alley_cat ([email protected]) left irc: Nick collision from services.
[14:41] Nick change: Kaetzchen -> alley_cat
[14:44] thomasvs ([email protected]) left irc: Read error: 113 (No route to host)
[14:47] teuf ([email protected]) joined #gstreamer.
[14:50] thomasvs ([email protected]) joined #gstreamer.
[14:55] <ensonic> thomasvs: were you able to read my reply before your connection was lost?
[15:10] mathrick ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[15:11] <thomasvs> nope
[15:11] mathrick ([email protected]) joined #gstreamer.
[15:11] <thomasvs> post again please :)
[15:14] <ensonic> thomasvs: we are trying to develop a replacement for buzz (www.buzzmachines.com)
[15:15] <ensonic> thomasvs: a little info is already on the project homepage (buzztard.sf.net) , more will come
[15:15] <thomasvs> ah, right
[15:16] <thomasvs> sorry, I forgot :) my mistake
[15:16] <thomasvs> ensonic: but you're doing this on solaris, or am I mistaken ?
[15:17] <ensonic> thomasvs: not really, I have a solaris box here, but in the end we hope that this is platform indep (thats why we would like to choose gst as the base - in the hope it will run on win32 and macos someday)
[15:18] <ensonic> thomasvs: i'll be back in a hour
[15:18] Action: ensonic is away: back in on hour
[15:26] ChriHJW_log ([email protected]) joined #gstreamer.
[15:29] sxpert_work ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[15:30] sxpert_work ([email protected]) joined #gstreamer.
[15:32] ChrisHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[15:33] ibsteveog ([email protected]) joined #gstreamer.
[15:38] steveb_ ([email protected]) joined #gstreamer.
[15:39] bilboed ([email protected]) left irc: Remote closed the connection
[15:40] jaiserca ([email protected]) joined #gstreamer.
[15:44] bilboed ([email protected]) joined #gstreamer.
[15:52] trow ([email protected]) joined #gstreamer.
[15:53] thomasvs ([email protected]) left irc: "Client exiting"
[15:55] steveb__ ([email protected]) left irc: Read error: 110 (Connection timed out)
[15:56] sysan ([email protected]) left #gstreamer.
[16:00] thomasvs ([email protected]) joined #gstreamer.
[16:02] Action: ensonic is back (gone 00:43:58)
[16:04] <ensonic> thomasvs: i am back, currently the issues I have with gst is chaos regarding to audioconvert, audioscale, adder, float2int, and so on - I'll try to come up with a cleanup suggestion, before I ned to find out the rational for having all these separate plugins
[16:05] <ensonic> thomasvs: btw. whats the reson for asking what we are doing? - are you working on the gst-roadmap?
[16:05] <thomasvs> no, curious
[16:06] <thomasvs> the rationale for having them as separate plugins is simple - because they do different things
[16:09] sublett ([email protected]) left irc: Read error: 113 (No route to host)
[16:11] <trow> hi thomasvs
[16:11] <thomasvs> hey jon
[16:11] Shoragan ([email protected]) left irc: Read error: 113 (No route to host)
[16:11] <thomasvs> trow: I'm cleaning out bugzilla and probably doing a new release today/tomorrow
[16:12] markey ([email protected]) left irc: "bbl"
[16:13] <trow> thomasvs: great
[16:14] <thomasvs> 180 bugs to retarget - yay
[16:17] <ensonic> thomasvs: yes of course, but the naming is a bit unconsistent (float2int<->float2intnew, audioconvert->audiochannelconvert or something and 'adder' seems to be broken)
[16:17] steveb__ ([email protected]) joined #gstreamer.
[16:18] <thomasvs> ensonic: raise float2int vs float2intnew on the m/l again please, I agree with you but it's up to benjamin and wingo to decide
[16:18] Shoragan ([email protected]) joined #gstreamer.
[16:19] <ensonic> thomasvs: okay I do, btw do know who is 'responsible' for the adder?
[16:19] <thomasvs> I could be stupid and tell you that I probably am :)
[16:20] <ensonic> thomasvs: try feeding two sinesrc with different freq. in - you will hear just silence (tried with gts-lauch - see mlist)
[16:20] <ensonic> thomasvs: yes, I've seen your name, I already had a look at the source, but get not far enough
[16:22] kwm ([email protected]) joined #gstreamer.
[16:23] <ensonic> thomasvs: I mean, I've read your name in the source ...
[16:26] <thomasvs> ensonic: yeah, I know
[16:26] <thomasvs> I originally wrote it
[16:26] <thomasvs> it's been rewritten by others though
[16:26] mxpxpod ([email protected]) left irc: "Guns kill people as much as spoons made Rosie O'Donnell fat.."
[16:26] <ensonic> thomasvs: any ideas what actually might be broken?
[16:27] <thomasvs> ensonic: caps nego, maybe
[16:27] <teuf> thomasvs: I didn't know about your policy on target milestone for bugs ;)
[16:27] <thomasvs> ensonic: wait, do you mean sinesrc with different freq, or different sample rate ?
[16:28] <thomasvs> teuf: I didn't tell people, which is why I'm telling people now :)
[16:28] <thomasvs> teuf: because retargetting is very boring
[16:28] markey ([email protected]) joined #gstreamer.
[16:35] <ensonic> thomasvs: no caps stuff seems to be fine, I am talking about sinesrc freq=200, sinesec freq=500 -> adder -> someaudiosink
[16:35] markey ([email protected]) left irc: "leaving"
[16:35] steveb_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[16:35] pb_ ([email protected]) left irc: "Client exiting"
[16:35] <thomasvs> ensonic: what pipeline do you use for that ?
[16:38] <ensonic> I've tried severla (if you have access to mail, these have a look that the thread called 'linking two sinesrc to an adder ...')
[16:46] tromey ([email protected]) joined #gstreamer.
[16:46] <thomasvs> tromey: I didn't get to meet you in belgium, didn't know how to find you
[16:46] <tromey> argh, bummer
[16:47] <tromey> I completely spaced on it... I was hanging out with the java folks most of the time
[16:47] <jdahlin> tromey: I saw your talk, it was quite interesting (even though I'm not really a java guy)
[16:47] <tromey> cool
[16:49] <jdahlin> how does eclipse compiled with gcj compare with interpreted versions (sun, ibm)?
[16:49] dilinger ([email protected]) joined #gstreamer.
[16:49] <tromey> it starts up faster and uses less memory.  interactively to me it feels about the same though
[16:50] <tromey> the current JITs, especially IBMs, are really quite good
[16:50] <jdahlin> I thought it was slower, seemed quite slow when I tried it
[16:50] <jdahlin> maybe it's just the linux port that's slow
[16:51] <tromey> yeah, the gtk port is noticeably slower than the windows port
[16:51] <tromey> there is something wrong with the gtk implementation of SWT
[16:52] <jdahlin> tromey: you should make owen look at it :)
[16:56] <tromey> the eclipse guys tried that, but owen reasonably wants a C test case, not 100,000 lines of java code :-)
[17:00] <ensonic> thomasvs: did you found the mail with ths gst-lauch command?
[17:01] <thomasvs> ensonic: still retargetting bugs currently
[17:07] steveb_ ([email protected]) joined #gstreamer.
[17:09] mxpxpod ([email protected]) joined #gstreamer.
[17:14] BBB ([email protected]) joined #gstreamer.
[17:15] <BBB> trow: so when will you be in boston again?
[17:18] apoc_ ([email protected]) joined #gstreamer.
[17:21] <BBB> where is dolphy?
[17:22] <thomasvs> off to bratislava
[17:22] Misirlou ([email protected]) left irc: "asdf"
[17:23] <BBB> hm
[17:23] <BBB> AVI seeking works, right?
[17:23] kmaraas ([email protected]) joined #gstreamer.
[17:23] <BBB> MPEG seeking is broken
[17:23] <thomasvs> we were waiting on you to fix that, no ?
[17:24] <BBB> AVI or MPEG?
[17:24] <thomasvs> MPEG
[17:24] <BBB> I thought AVI worked?
[17:24] <BBB> yes
[17:24] <thomasvs> yeah
[17:24] <BBB> I have a patch for MPEG
[17:24] <thomasvs> :)
[17:24] <thomasvs> yeah
[17:24] <thomasvs> I've seen it in bugzilla
[17:24] <BBB> if it wasn't for the stupid laptop screen...
[17:24] <thomasvs> how about trying it :)
[17:24] <thomasvs> I'm about to do a release
[17:24] <BBB> with no computer?
[17:24] <BBB> ;)
[17:24] Action: BBB is chatting on a win32 thing that doesn't even have cygwin
[17:25] <BBB> did julien solve AVI seeking before or after 0.7.4?
[17:25] <thomasvs> can't you take that monitor and attach it to the laptop ?
[17:25] Action: BBB is closing fixed bugs
[17:25] <thomasvs> BBB: don't !!!
[17:25] steveb__ ([email protected]) left irc: Read error: 110 (Connection timed out)
[17:25] <thomasvs> BBB: or at least, retarget them properly
[17:25] <BBB> it's not mine, I'm at the university
[17:25] <BBB> I know, I read your email
[17:25] <BBB> I will
[17:25] <thomasvs> BBB: you have the laptop no ? so you can attach the monitor you're using to your laptop ?
[17:26] <BBB> it's at home, not here...
[17:26] <BBB> I'd rather just have a real working laptop
[17:26] <BBB> :)
[17:27] <ds> moo
[17:27] Misirlou ([email protected]) joined #gstreamer.
[17:28] <thomasvs> what is xwindowlistener ?
[17:29] <thomasvs> anyone know ?
[17:29] <BBB> it listens to an X window
[17:29] <BBB> for position/size
[17:29] <BBB> it's useful for hardware overlay thingies
[17:29] <BBB> (dxr3, v4l, v4l2)
[17:29] <BBB> it's a support lib plugin, so it is a plugin, but not an element
[17:30] <BBB> http://bugzilla.gnome.org/show_bug.cgi?id=121507
[17:30] <BBB> can we close that as wontfix?
[17:30] <thomasvs> why wont we fix that ?
[17:31] <thomasvs> BBB: http://bugzilla.gnome.org/show_bug.cgi?id=134300
[17:31] <thomasvs> --> can you coment there then ?
[17:31] <thomasvs> any suggestions for release names ?
[17:31] <BBB> thomasvs: because it's a 0.6.x bug which is fixed in 0.7.x
[17:31] <thomasvs> BBB: ok, close it then
[17:31] <BBB> release names: "Welcome to the U.S. of A."?
[17:31] <BBB> :-p
[17:32] <thomasvs> isn't that an illegal name ?
[17:33] <BBB> likely ;)
[17:33] <ds> "Back to the USSA"?
[17:33] <trow> BBB, thomasvs: AVI seeking is still broken for me.  See bug 134550.
[17:34] <thomasvs> it is ? crap
[17:34] Action: thomasvs tests
[17:34] <trow> BBB: I'll probably be in Boston sometime next month.  Don't know for sure yet.
[17:35] <BBB> http://bugzilla.gnome.org/show_bug.cgi?id=131167 <- someone please apply this and close it?
[17:36] apoc ([email protected]) left irc: Connection timed out
[17:36] <BBB> crap @ seeking
[17:36] <BBB> trow: give me a note when you are and I'll book flight or train tickets
[17:36] mathrick ([email protected]) left irc: Remote closed the connection
[17:36] <thomasvs> trow: here it works - does it work with matrix for you ?
[17:42] <Misirlou> thomasvs: What do you mean by illegal?
[17:48] Kaetzchen ([email protected]) joined #gstreamer.
[17:48] alley_cat ([email protected]) left irc: Nick collision from services.
[17:49] Nick change: Kaetzchen -> alley_cat
[17:49] <BBB> http://bugzilla.gnome.org/show_bug.cgi?id=129649 <- close as dunno? there's not much useful info in there
[17:50] <thomasvs> trow: ?
[17:50] <trow> thomasvs: that bug report is only w/ matrix.avi.  Seeking was working semi-OK for me, but then I switched from RH9 to SuSE9 and started seeing the problem.  (The crash I saw on RH9 had a different backtrace: see bug 133006)
[17:50] <trow> s/is only/is/
[17:51] <thomasvs> ah, ok
[17:51] <thomasvs> damn - but both are head cvs right ?
[17:51] <BBB> thomas, can you try the two launch pipelines in http://bugzilla.gnome.org/show_bug.cgi?id=131128 and tell if they work?
[17:51] <BBB> they should
[17:51] <thomasvs> I will do as soon as I fixed milestones
[17:51] <BBB> ty
[17:51] <BBB> or maybe trow? :)
[17:52] <ensonic> thomasvs: 'spring welcome release' (i'm not very creative at thins time anymore)
[17:52] <ensonic> ciao
[17:53] <teuf> BBB: second one works here though the initial window is really small
[17:53] <teuf> 1st one doesn't work
[17:53] <teuf> ** (process:2785): CRITICAL **: attempting to convert unknown formats
[17:54] <BBB> initial window?
[17:54] <BBB> but in the end, it works, right?
[17:54] <teuf> the window
[17:54] <BBB> I mean, the window should resize after a while
[17:54] <teuf> hmm, not sure
[17:54] Action: teuf retries
[17:54] ensonic ([email protected]) left #gstreamer.
[17:54] <teuf> for the aasink one, I don't have aasink installed, so that can't work ;)
[17:54] <BBB> or maybe not.. anyway, videotestsrc allows for 16x16 so it's a valid size
[17:54] <BBB> resizing by hand should work
[17:55] <teuf> yep, it stays at about 16x16, and resizing works
[17:55] <BBB> ok, that's fine then
[17:56] yippi ([email protected]) left irc: Remote closed the connection
[18:00] Action: trow recompiles
[18:01] ibsteveog ([email protected]) left irc: "quitting"
[18:02] ibsteveog ([email protected]) joined #gstreamer.
[18:02] mxpxpod ([email protected]) left irc: "Leaving"
[18:03] <alley_cat> GConf audio sink not found, using osssinkGConf video sink not found, using xvideosink
[18:03] <alley_cat> what could cause this? the keys exist
[18:03] <thomasvs> hang on, putting in a fix
[18:04] <thomasvs> alley_cat: have you recompiled and reinstalled gst-plugins ?
[18:04] <thomasvs> alley_cat: the gconf path got changed
[18:04] <alley_cat> recompiled today
[18:05] <thomasvs> do it again
[18:06] <BBB> thomasvs: http://bugzilla.gnome.org/show_bug.cgi?id=131128 close?
[18:06] <BBB> since we're not at 0.6.x anymore
[18:06] markey ([email protected]) joined #gstreamer.
[18:07] <thomasvs> BBB: we could still use a rgb->yuv colorspace element btw
[18:08] steveb__ ([email protected]) joined #gstreamer.
[18:08] <alley_cat> ah ok, need to set keys with the new paths then
[18:08] <BBB> later... I first want a default gstreamer to play back MPEG, AVI, Quicktime, Ogg and Matroska
[18:08] <BBB> 'default' meaning including ffmpeg and so on
[18:08] <BBB> after that, I'll care about the "without ffmpeg" case
[18:08] <BBB> http://bugzilla.gnome.org/show_bug.cgi?id=126459 <- also for you, thomasvs, I suppose you fixed that already and I can close this?
[18:10] yippi ([email protected]) joined #gstreamer.
[18:10] <alley_cat> k, now i don't get the message anymore, it just segfaults :P
[18:10] mxpxpod ([email protected]) joined #gstreamer.
[18:11] <thomasvs> BBB: no, I just assigned it to myself :)
[18:11] <BBB> eh?
[18:11] <BBB> but it's in CVS already?
[18:11] <thomasvs> leave it open
[18:11] <BBB> gr
[18:11] <thomasvs> I'll close it this week
[18:11] <BBB> just closed it
[18:11] Action: BBB reopens
[18:12] Action: ds remembers to check in some qtdemux fixes
[18:13] <thomasvs> anything specific blocking a release atm ?
[18:13] <thomasvs> ds: adding seeking to it anytime soon ?
[18:13] <BBB> yes, MPEG seeking doesn't work
[18:13] <thomasvs> http://build.fluendo.com:8080/
[18:13] <thomasvs> some pretty colors today
[18:13] <thomasvs> BBB: dude, that's been the way for hte lsat three releases :)
[18:13] <thomasvs> BBB: so I won't block on that for this release
[18:13] Action: ds hasn't been paying attention recently
[18:14] <trow> gstasfdemux.c: In function `gst_asf_demux_add_audio_stream':
[18:14] <trow> gstasfdemux.c:1495: error: `GST_TAG_AUDIO_CODEC' undeclared (first use in this function)
[18:14] <trow> and
[18:14] <trow> gstasfdemux.c: In function `gst_asf_demux_add_video_stream':
[18:14] <trow> gstasfdemux.c:1653: error: `GST_TAG_VIDEO_CODEC' undeclared (first use in this function)
[18:15] <ds> is MPEG still broken?
[18:15] <thomasvs> trow: was added to core, update core
[18:15] <thomasvs> ds: yeah
[18:16] steveb_ ([email protected]) left irc: Read error: 60 (Operation timed out)
[18:16] <BBB> ds: there's a patch from me in bugzilla
[18:16] <BBB> I didn't apply it yet because I wanted to clean it up a bit
[18:17] <BBB> if you guys don't care, just apply it
[18:17] <alley_cat> avidemux is totally broken for me, either hangs or segfaults
[18:18] <BBB> upload some sample files so we can reproduce :)
[18:18] <thomasvs> alley_cat: on what platform ?
[18:18] <thomasvs> BBB: clean it up a bit, we care
[18:18] <BBB> http://build.fluendo.com:8080/ <- pretty
[18:18] <alley_cat> thomasvs: linux, kernel 2.6, nptl glibc
[18:18] <BBB> thomasvs: laptop screen, ?
[18:18] <BBB> :)
[18:19] <thomasvs> alley_cat: platform ?
[18:19] <BBB> actually, I'm not supposed to be here at all, I'm supposed to be studying 24/7 at Cornell...
[18:19] <thomasvs> alley_cat: what processor ?
[18:19] <alley_cat> thomasvs: athlon
[18:19] <BBB> I really have wrong priorities ;)
[18:20] Action: bilboed is away: parti maaaaaaaaaaaaaaaaaanger
[18:20] <alley_cat> hmm, just found an avi for which it works
[18:20] <kwm> avidemux bombs here
[18:20] <kwm> RUNNING pipeline ...
[18:20] <kwm> ERROR: from element /pipeline0/thread1/spider1/src_0: Internal GStreamer error: pad problem.  File a bug.
[18:20] <kwm> Additional debug info:
[18:20] <kwm> gstpad.c(2968): gst_pad_pull: /pipeline0/thread1/spider1/src_0:
[18:20] <kwm> pull on pad src_0:sink but the peer pad mad0:src has no gethandler
[18:20] <thomasvs> alley_cat: ok, so file backtraces
[18:20] <thomasvs> kvm: whatever makes you think that is related to avidemux ? :)
[18:21] <alley_cat> thomasvs: for the hangs should i just susped it in gdb and backtrace?
[18:21] <thomasvs> hangs ?
[18:21] <thomasvs> alley_cat: if it hangs, how do you know if it is avidemux ?
[18:21] <alley_cat> for most files it just sits there eating cpu
[18:21] <thomasvs> dudes
[18:21] Last message repeated 1 time(s).
[18:21] <alley_cat> because remuxing the same file to matroska and playing it with matroskademux works
[18:21] <thomasvs> I'm not a mind reader
[18:21] <thomasvs> what app
[18:21] <kwm> thomasvs: duno I'm not a devven but mpeg/mpg files play
[18:21] <thomasvs> what version
[18:21] <thomasvs> and so on
[18:22] <thomasvs> kwm: ok, so give me the info
[18:23] <BBB> our matroska support is cute, mplayer doesn't even have that ;)
[18:23] iain_ ([email protected]) joined #gstreamer.
[18:24] <ds> but what can one play with matroska?
[18:24] <alley_cat> doh, in gdb the stuff behaves totally different
[18:25] <alley_cat> mplayer can play matroska, but not mux it
[18:25] <alley_cat> ds: half of my anime collection :)
[18:25] <BBB> ds: who csres ;) it's a feature ;)
[18:25] <BBB> alley_cat: their demuxer is ripped from ours *grmbl*
[18:26] <alley_cat> BBB: only the new one, the previous wasn't
[18:26] <BBB> I know... :)
[18:27] <markey> BBB: saw your post on kde-multimedia
[18:27] <markey> nice
[18:27] <BBB> :)
[18:27] <markey> it's good to start communication now
[18:27] <markey> quite exciting isn't it
[18:27] <BBB> I was actually gonna say this as a reply to one of the posts on kde-multimedia
[18:27] <BBB> yup :)
[18:27] <markey> NMM, MAS, GST in the race
[18:28] <BBB> there was one post on kde-mm that c++ wasn't always a good choice... for example, mplayer's C matroska demuxer was much better than their C++ demuxer
[18:28] <BBB> I wanted to add a flame there
[18:28] <BBB> but didn't
[18:28] <BBB> (yet)
[18:29] <markey> hmm
[18:29] <markey> I found that post confusing
[18:29] <markey> IIRC
[18:29] <markey> maybe the poster mixed something up
[18:29] <BBB> markey: I read your post, too... I really do want to help in getting viz. to work, but cannot do that yet :(
[18:29] <markey> eventually, we'll manange, I guess
[18:30] <markey> just had no time for the GST interface
[18:30] <BBB> I guess it'll work in the end
[18:30] <markey> (and was a bit fed up;)
[18:33] <alley_cat> how do i get debug messages from avidemux? i always forget how --gst-debug works
[18:33] <trow> From gst-plugins/gconf:
[18:33] <trow> make: *** No rule to make target `gstreamer.schemas', needed by `gstreamer-0.7.schemas'.  Stop.
[18:34] spyder482 ([email protected]) joined #gstreamer.
[18:36] <thomasvs> trow: you need to autoreconf for that, I think, because it's added to configure.ac how to generate it
[18:36] <BBB> alley_cat: --gst-debug-level=5 2>&1 | grep -i avi
[18:37] <trow> thomasvs: I did autoconf before building.  I'll try again.
[18:37] <BBB> maybe avi has its own debug category
[18:37] <BBB> alley_cat: you cna also | grep -i riff for rifflib debug messages
[18:37] <thomasvs> trow: check if configure generates that file
[18:37] <thomasvs> ds: one for you --> http://bugzilla.gnome.org/process_bug.cgi
[18:37] <thomasvs> argh
[18:37] <thomasvs> http://bugzilla.gnome.org/show_bug.cgi?id=126922
[18:38] <alley_cat> ok, i'm really really confused
[18:38] <alley_cat> with --gst-debug-level=5 it works
[18:38] <trow> thomasvs: There is no mention of it in configure.ac.
[18:39] Action: thomasvs doublechecks locally
[18:39] <thomasvs> trow: I have 1.373 of configure.ac, what version is yours ?
[18:39] <thomasvs> mine has it
[18:40] <trow> Ah... I have 1.372.
[18:40] Action: trow updates
[18:41] <spyder482> BBB: can you explain a small thing to me? :)
[18:42] <alley_cat> hmm, i get the feeling that all problems i have with avidemux are caused by some timing issues
[18:43] <alley_cat> with debug-level 0-2 it hangs, 3 segfaults, 4 and 5 works
[18:43] <thomasvs> alley_cat: you still didn't mention your dist
[18:43] <BBB> spyder: maybe
[18:43] <BBB> be quick, I'm off for lunch in a second
[18:43] <spyder482> BBB: when you catch a new_pad signal from the mpegdemux filter and connect the new pad to the queue, how do you start the movement in the pipeline again?
[18:43] <BBB> spyder: set the queue and all to playing
[18:43] <spyder482> assuming i have 4 threads joined by queues all in a pipeline
[18:43] <alley_cat> thomasvs: sourcemage, source based distro
[18:43] <BBB> I generally put the queue plus connected elements to PAUSED, then wait until the pad is created, link and set it to playing
[18:44] <BBB> where each "set of elements" is in a bin, and I set the bin to playing
[18:44] <thomasvs> alley_cat: what flags are you using to compile ?
[18:44] Action: BBB should post example code some day
[18:44] <spyder482> ah ok
[18:44] <alley_cat> thomasvs: for gstreamer only the default, whih is -O2
[18:44] <BBB> spyder: is it ok if I send you an example (for mpeg->avi transcoding, but matroska works the same) in a few days?
[18:44] <spyder482> BBB: i'm making a "journal" of what i find out that's not really described well in the docs :)
[18:45] <spyder482> BBB: sure
[18:45] <BBB> ok, I'll do that then
[18:45] <alley_cat> for other stuff -march=athlon -mmmx -m3dnow -Os
[18:45] <BBB> if you put a copy of the final journal in CVS, I'll make sure all those notes end up in an updated application development manual that I'm planning
[18:45] <thomasvs> alley_cat: ok, file a backtrace for the segfault at 3
[18:45] <BBB> s/CVS/bugzilla/
[18:45] <alley_cat> i'll try to reproduce the segfault in gdb
[18:46] <spyder482> BBB: i don't have cvs access? :) but i can get it to you when it's finished :)
[18:46] <BBB> I meant bugzilla
[18:46] <BBB> ;)
[18:47] <spyder482> oh... i missed that line ;)
[18:48] Action: BBB first wants to finish at least chapter 4 of the PWG
[18:48] <BBB> and then I'll focus on the application development manual
[18:48] <BBB> and playback issues
[18:49] <BBB> or well, actually, I should focus on my studies, too
[18:49] lupusBE ([email protected]) joined #gstreamer.
[18:50] Action: BBB off
[18:50] <BBB> later
[18:50] BBB ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[18:50] <taaz> wow.  bugzilla emails are out of control.
[18:50] <alley_cat> can't reproduce it with gdb, will try again tomorrow
[18:50] <thomasvs> taaz: I hope so
[18:51] yippi ([email protected]) left irc: "Client exiting"
[18:52] yippi ([email protected]) joined #gstreamer.
[19:12] jaiserca ([email protected]) left irc: "Terminando cliente"
[19:15] Zeenix ([email protected]) joined #gstreamer.
[19:16] <Zeenix> hi
[19:17] <mxpxpod> thomasvs: I just submitted a hug update to gstmm in bugzilla, just so you know :)
[19:24] <trow> Is there an easy way to query if gstreamer knows how to handle a particular mime type?
[19:29] <thomasvs> trow: gst-typefind
[19:29] <thomasvs> trow: or, do you mean, given that you already know the mime type, can gst play it ?
[19:30] <jdahlin> taaz: Have you ever got the identity example to work?
[19:30] <thomasvs> I think, yeah
[19:31] <jdahlin> I can't get it to work
[19:31] <thomasvs> give up :)
[19:34] <trow> thomasvs: yes, I'm looking for something like 'if (gst_knows_how_to_play ("video/mpeg")) { ...'
[19:36] <taaz> jdahlin: sure.  haven't tried recently though.
[19:37] <jdahlin> taaz: so no problem if I break it?
[19:38] <thomasvs> trow: in that case - I put some code in spider that does something like that
[19:38] <thomasvs> trow: though, being able to handle a mime type doesn't guarantee playability for obvious reasons
[19:39] <thomasvs> trow: http://freedesktop.org/cgi-bin/viewcvs.cgi/gstreamer/gstreamer/gst/autoplug/gstspider.c.diff?r1=1.37&r2=1.38
[19:39] <taaz> jdahlin: well... it would be nice if that stuff got to working state again at some point.
[19:39] <thomasvs> trow: cut and paste some code from that diff
[19:39] <taaz> jdahlin: i don't have time to work on it for next couple of days though
[19:40] <taaz> jdahlin: what were you going to break?
[19:40] <thomasvs> later peeps
[19:40] n3m0gl081n3 ([email protected]) left irc: "IceChat IRC Client - Download at www.IceChat.net"
[19:40] Zeenix_ ([email protected]) joined #gstreamer.
[19:40] <jdahlin> taaz: I'd really like identity to work too, seems like a nice example, but it's just that it doesn't work here :(
[19:41] <jdahlin> ah, something related to GstData*
[19:42] <taaz> uh oh ;)
[19:42] <taaz> i have a partial patch for that
[19:42] <taaz> eh, i'll just commit some stuff that probably doesn't break anything ;)
[19:44] <jdahlin> can you send it to me by mail so I can integrate it?
[19:45] smoke ([email protected]) joined #gstreamer.
[19:46] kmaraas ([email protected]) left irc: Read error: 110 (Connection timed out)
[19:48] <taaz> oops.  too late for one commit
[19:49] <trow> Uh oh... rhythmbox once again isn't able to read id3 tags from mp3 files.
[19:50] <trow> What is the gst-launch line to test tag-finding?
[19:50] Nick change: apoc_ -> apoc
[19:51] <apoc> trow: gst-launch -t filesrc location=foo.mp3 ! id3tag ! fakesink silent=true
[19:52] <taaz> jdahlin: i committed struct/taglist foreach support too
[19:54] <taaz> that's all my changes except some minor doc things.
[19:54] <trow> apoc: Thanks.  When I do that, it correctly displays the tags but I also get a lot of messages like:
[19:54] <jdahlin> hmm, can't really get it to work
[19:54] <trow> ** (process:2767): CRITICAL **: how to remove plugins?
[19:54] <trow>  
[19:54] <trow> (process:2767): GStreamer-WARNING **: could not load valid typefind function for feature 'video/x-ms-asf'
[19:55] <trow> I get a message like that for a whole bunch of different mime types.
[19:55] Zeenix ([email protected]) left irc: Read error: 110 (Connection timed out)
[19:57] mathrick ([email protected]) joined #gstreamer.
[19:57] <mathrick> yo
[19:57] <apoc> trow: hm I don't have that ...
[19:57] thomasvs ([email protected]) left irc: Read error: 113 (No route to host)
[19:58] <apoc> trow: with gstreamer HEAD ?
[19:59] <taaz> jdahlin: which?
[19:59] <jdahlin> taaz: so GstData is turned up in the generated code
[19:59] <taaz> jdahlin: you should try player.py on a .ogg to see tags print out
[19:59] Company ([email protected]) joined #gstreamer.
[19:59] <mathrick> damn, it's amazing
[20:00] <mathrick> weather outside is really stunning :)
[20:01] <taaz> jdahlin: eh... there may have to be some hacks related to this.  the code that used to have the boxed GstBuffer as a param now has GstData in some non-OOP sort of mess (IMHO).  which probably means custom hacks for such functions that were not needed previously.
[20:02] <jdahlin> taaz: I'd like to avoid hacks
[20:02] <mathrick> it's a little bit freezing, there's fresh snow, and misty vapor in thick layer, ceiling of which you can clearly see, and which dims streetlamps in beatiful way. Truely sweet :)
[20:02] <taaz> ie, methods like push
[20:02] <jdahlin> taaz: but I think registering GstData as a GBoxed object
[20:02] <jdahlin> just to help bindings
[20:02] <taaz> which is why the examples no longer work
[20:03] <taaz> jdahlin: will that help?  it's more complicated than that i think
[20:04] <jdahlin> taaz: it will help a lot and it's not that complicated, but it requires changes to the core
[20:04] <taaz> you need to be able to ask the data if it's a buffer or event (or something else?)
[20:04] <taaz> since it's boxed you can't just cast to a buffer or something
[20:04] <taaz> which means functions to do magic and convert i guess
[20:05] <jdahlin> how do you do it in C, to convert?
[20:05] <taaz> cast
[20:05] <jdahlin> GstBuffer = (GstBuffer*)data ?
[20:05] <taaz> yeah
[20:05] <jdahlin> err
[20:05] <jdahlin> hmm
[20:06] <mathrick> wow
[20:06] Action: mathrick is amazed seeing wtay-typing here :)
[20:06] <jdahlin> taaz: and can you say a scenario when you need to do that?
[20:06] <taaz> when don't you need to do that?
[20:06] <jdahlin> we could just add a data.get_buffer() or something I guess
[20:06] <jdahlin> and buffer.get_data()
[20:07] <taaz> and data.get_event()
[20:07] <taaz> and event.get_data()?  blah
[20:07] <jdahlin> or just buffer = gst.Buffer(data) / data = gst.Data(buffer)
[20:07] <jdahlin> etc
[20:07] <taaz> Events too
[20:07] <jdahlin> I have to see the API in use before deciding
[20:08] <taaz> look at plugins that process data
[20:08] <taaz> most of them just cast data to buffer now.  i have no idea if that's safe or not.  some handle events with a type check first.
[20:09] <Company> it's supposed to be GObject lite
[20:09] <jdahlin> Company: sounds very suitable for a GBoxed
[20:10] <jdahlin> which makes life a lot easier for language bindings
[20:12] <Company> does GBoxed do deep inheritance?
[20:12] <taaz> it does no inheritance
[20:12] <Company> well, it kinda does, you can derive from GBoxed
[20:13] <taaz> but i dont think you can say Boxed->Data->Event
[20:14] <Company> yeah, that's what i'm asking :)
[20:14] <jdahlin> nope, there is no inheritance
[20:15] <jdahlin> so, we have to add methods to convert between the different types
[20:15] <jdahlin> get_buffer/get_event for gst.Data etc
[20:15] <Company> can't you map them to objects in python?
[20:16] <jdahlin> sure, with help of GBoxed
[20:16] <trow> apoc: Yes, that is w/ HEAD
[20:17] <Company> then you just need to find a way to make a GstEvent a subclass of GstData
[20:17] <taaz> jdahlin: and some support for GST_IS_BUFFER()/EVENT()
[20:18] <Company> and you need to register event, buffer and data as GBoxed (which should probably done anyway)
[20:19] <jdahlin> Company: I'm not sure if it's a good way of treating the objects
[20:19] <taaz> it's doable, just requires low level python hacking.
[20:19] <taaz> create the class(es) by hand rather than using the pygtk support for such things
[20:20] <Company> i'd have made gstdata a regular GObject, but people were afraid that would be slow
[20:21] <jdahlin> would it?
[20:21] <Company> no clue
[20:22] <Company> back then people said it's not an option so i didn't try
[20:22] <taaz> i think omega or wtay did a test long long ago and the answer was yes
[20:22] Action: jdahlin pokes wtay-typing
[20:22] <Company> the problem is the "long long ago"
[20:23] <Company> and the fact that for real usage you don't have that problem because doing de/encoding takes 95% of processing time anyway
[20:23] <ds> shouldn't that be guint64 ago?
[20:23] <Company> no, that would be unsigned long long ago
[20:24] <taaz> if you consider overhead for gobject for, say, a billion buffers during a dvd then it adds up to significant time.  which may be unacceptable for those high end audio people concerned with latency
[20:24] <taaz> not to mix examples ;)
[20:24] <Company> you don't have a billion buffers there ;)
[20:25] <taaz> pointless to argue this without a real test of course
[20:25] <Company> that's like running gnomevfssrc bytesize=1 location=my_one_gig_file ! fakesink
[20:25] <Company> we won't change it right now anyway
[20:26] <jdahlin> but we can make them GBoxed, or something else
[20:27] <Company> OT: what do i call a bytestream where i put the buffers in manually?
[20:28] teuf|la ([email protected]) joined #gstreamer.
[20:29] <Company> bytesplitter sounds stupid
[20:29] <teuf|la> that sounds like a cool toy :)
[20:30] <Company> mathrick: go invent something, you're good at that stuff
[20:30] <mathrick> Company: huh, I don't get your idea :>
[20:33] apoc ([email protected]) left irc: "Leaving"
[20:33] <Company> mathrick: some thingy that you put buffers in and request chunks of those out
[20:34] <Company> mathrick: so you say "give me 9 bytes, then give me 25 bytes, then give me 187 bytes" and so on
[20:35] <Company> and whenever the thingy says "i need more buffers to do that, you push a new one in
[20:36] <Company> when are we going to freeze our API and up or MAJORMINOR to 0.8 btw?
[20:37] <mathrick> Company: hmm, ok, and what is it useful for? after all, all bytes you pull, you have to put in yourself, right?
[20:38] <Company> mathrick: it merges the buffers for me and slices them into the requested sizes
[20:38] <Company> GstByteSlicer
[20:39] <Company> now that sounds cool
[20:39] <mathrick> Company: ahh, ok
[20:39] <mathrick> Company: GstBytePool?
[20:39] kmaraas ([email protected]) joined #gstreamer.
[20:39] bilboed ([email protected]) got netsplit.
[20:39] Jara[zZ] ([email protected]) got netsplit.
[20:39] <Company> i like slicer better
[20:40] <Company> it's not a pool because it discards them once you don't need them anymore
[20:40] <mathrick> Company: how does it know you don't need then any longer?
[20:40] <mathrick> s/then/them/
[20:40] <Company> you tell it
[20:40] <Company> my_thingy_flush (number_of_bytes)
[20:42] <mathrick> Company: GstByteSlicer is cool, but not very explaining
[20:42] <mathrick> Company: hmm, maybe GstBytePocket? :)
[20:43] <mathrick> s/explaining/self-\&/
[20:45] Jara[zZ] ([email protected]) got lost in the net-split.
[20:45] bilboed ([email protected]) got lost in the net-split.
[20:46] jdahlin ([email protected]) left irc: "Klienten avslutas"
[20:46] bilboed ([email protected]) joined #gstreamer.
[20:46] Jara[zZ] ([email protected]) joined #gstreamer.
[20:48] Misirlou ([email protected]) left irc: Remote closed the connection
[20:54] <Company> GstBufferAdaptor
[20:55] jcsston ([email protected]) joined #gstreamer.
[20:55] <Company> GstAdapter sounds even better
[20:56] <Company> though that could be mistaken, but what do i care
[20:58] <mathrick> GstBufferTransformer
[20:58] <mathrick> GstMagicBox ;)
[20:59] Action: ibsteveog mumbles GstTheOneBufferToRuleThemAll
[21:01] BBB ([email protected]) joined #gstreamer.
[21:01] <bitshifter> GstBufSlicer
[21:01] <ibsteveog> does anyone have any ideas about where I could get help with ffmpeg?  I know of their devel list, but I hate to use that first...
[21:01] <mathrick> yo BBB
[21:01] <BBB> howdy
[21:01] <BBB> ibsteveog: like what?
[21:01] <ibsteveog> i've been trying to make mjpegs with ffmpeg and i keep getting files that no one (mplayer or WMP) likes
[21:01] <mathrick> BBB: how comes your username is Administr? some quick hacking into lab's net? ;)
[21:02] <ibsteveog> output_example test.mjpeg
[21:02] <ibsteveog> yields an mjpeg file that everything thinks is just a jpeg with a lot of padding at the end, instead of a 5 second, 125 frame movie
[21:02] <BBB> mathrick: no, just default uni computer
[21:02] <BBB> dunno
[21:02] <BBB> mjpeg is not a format
[21:03] <BBB> you want avi with jpeg/mjpeg as codec
[21:03] <BBB> the "mjpeg" format doesn't exist, contrary to common belief
[21:03] <ibsteveog> i figured i needed a container...  but I dunno how to get ffmpeg to do it
[21:03] <ibsteveog> well, i've found about a dozen different variations of mjpeg
[21:03] <BBB> look in their archives, there was a question 2-3 weeks ago similar to this one
[21:03] <BBB> mjpeg is a codec, not a container
[21:03] <ibsteveog> i looked last week over all of january and february...
[21:04] <BBB> hm
[21:04] <BBB> ok
[21:04] Action: ibsteveog checks again
[21:04] pmfp ([email protected]) left irc: "Leaving"
[21:04] <BBB> ffmpeg -vcodec mjpeg -acodec pcm -format avi, something like that
[21:05] <BBB> if you want a AVFormatContext, then use "avi" as container and CODEC_ID_MJPEG as video codec
[21:06] thomasvs ([email protected]) joined #gstreamer.
[21:07] <ibsteveog> ok... that seems to work with ffmpeg... now how I can get the API to do that :-P  they have zilch api docs and the apiexample is tiny, and has hardcoded values for mpeg... 
[21:10] <BBB> why is ext/ffmpeg/ empty?!?
[21:11] <mathrick> BBB: in gst-plugins?
[21:11] <BBB> no :'(
[21:11] <BBB> ibsteveog: I think guess_format() gives you an AVOutputFormat
[21:11] <BBB> which you can use to write streams
[21:12] <BBB> use "avi"
[21:12] <BBB> then, open an output stream using url_fopen()
[21:12] <Company> BBB: cvs co gst-ffmpeg
[21:12] <BBB> ah
[21:12] Action: BBB looks there
[21:13] <mathrick> BBB: gst-ffmpeg is greatest invention since sliced bread :) It works like charm, no more broken ffmpeg checkouts :)
[21:13] jcsston ([email protected]) left irc: "VDubMOD watchout TCME is coming!"
[21:13] <BBB> pff...
[21:13] <BBB> if nobody uses it, then no wonder we get no bug reports ;)
[21:14] <thomasvs> everyone uses it :)
[21:14] <thomasvs> it's even in jhbuild
[21:14] <alley_cat> there's only one problem with gst-ffmpeg, that's the old ffmpeg version
[21:14] <BBB> I don't see any gst-ffmpeg release or module on gstreamer.net
[21:15] <alley_cat> tried updating it today, but i got some unresolved symbols at the end
[21:15] <thomasvs> BBB: it hasn't been released yet
[21:15] <thomasvs> BBB: yeah, need to put it in the site
[21:15] <BBB> alley_cat: I was working on that but didn't finish it yet
[21:15] <BBB> I want a fully autotoolized ffmpeg, where you can use --disable-../--enable-.. like in gst-plugins ./configure
[21:15] <BBB> ffmpeg had a start there, but it didn't work very well
[21:16] <alley_cat> would be great if it could use an installed ffmpeg
[21:16] <BBB> no way, they don't release often enough
[21:16] <BBB> I'm not gonna wait for that to ever happen
[21:17] <BBB> besides, they don't use versioning, no libtool, no nothing...
[21:17] <BBB> it'd only give issues
[21:17] <BBB> which I don't want
[21:17] <alley_cat> there's a version in the avcodec header
[21:17] <lupusBE> Does anyone know a good site with all the codecses for gstreamer and what they do and there quality?
[21:18] <BBB> alley_cat: binary version
[21:18] <BBB> libtool version
[21:18] <alley_cat> but i guess it's too much work, i know i did maintain some vlc patches for a while to build against installed ffmpeg
[21:19] <BBB> libffmpeg-0.6.so.0.1.1
[21:19] <BBB> those versions
[21:19] <alley_cat> you'll never see that in the official ffmpeg ;)
[21:19] <BBB> that's what I just said ;)
[21:19] Misirlou ([email protected]) joined #gstreamer.
[21:20] <alley_cat> it's not really necessary to be able to use installed one, i just hate building the same code almost every day :P
[21:20] <BBB> so do I, but hey... that's life
[21:20] <BBB> use maintainer mode for ./configure
[21:20] <BBB> ./configure --enable-maintainer-mode
[21:21] <BBB> that helps a lot
[21:21] <alley_cat> i maintain the video section for a distribution, i have to build around 20 programs that include ffmpeg, and almost every day there is some update that needs testing
[21:21] <alley_cat> i think i should just get a faster machine :)
[21:22] <lupusBE> alley_cat, do you have a list with codecses thay I best use with gstreamer so I can play most files?
[21:22] <BBB> we don't have such a list
[21:22] <BBB> we used to, but it got outdated very soon
[21:23] <BBB> I Think thomas was working on updating it (as part of the website), maybe he can tell you more on it
[21:23] <lupusBE> ah that is what happened to it :)
[21:23] <BBB> but we don't have such a list
[21:23] <BBB> I think it's currently at 0.2.1 or so
[21:23] <BBB> ;)
[21:23] <alley_cat> well, i have the dependency list for my package, but that's for gstreamer 0.6
[21:23] <lupusBE> alley_cat url of your src package?
[21:23] <lupusBE> or spec file
[21:24] <lupusBE> plz :)
[21:24] <alley_cat> http://codex.sourcemage.org/test/video-libs/gst-plugins/DEPENDS
[21:25] <alley_cat> that's the depends file, it's a package description for a source based distro
[21:25] <alley_cat> won't really help you i guess ;)
[21:26] <lupusBE> I'm creating a spec file for fedora with all codecses
[21:26] <alley_cat> and 0.7 has lots more than 0.6, so that list is quite outdated for current version
[21:26] <lupusBE> redhat disables a lot
[21:27] <lupusBE> problem is
[21:28] <lupusBE> you have codecses that do the same
[21:28] <lupusBE> and I don't know which one is best
[21:30] <alley_cat> there's not much duplication
[21:35] jimmy_dean ([email protected]) left irc: Read error: 110 (Connection timed out)
[21:55] <Company> thomasvs: when are we going to freeze our API and up or MAJORMINOR to 0.8 btw?
[22:03] <teuf|la> is there anything planned to fix these seeking issues in rhythmbox ?
[22:04] <Company> seeking issues in rb?
[22:04] <teuf|la> pause issues I mean
[22:04] <Company> oh
[22:05] <Company> that's hard to fix
[22:05] <teuf|la> oh :(
[22:05] <Company> i guess the best solution for now is to not set the audiosink to NULL anymore
[22:06] <teuf|la> if rb no longer frees the audio device, people will complain
[22:06] <Company> i know
[22:06] <Company> problem is: the audiosink assumes a new stream starts when it goes READY(or NULL)=>PLAYING
[22:06] <Company> but the timestamp of the first buffer it gets is way into the stream
[22:07] <Company> so it does the correct thing: it waits until that time is reached
[22:07] <Company> correct thing from gstreamer to do would be to send a discont event
[22:07] <Company> but how?
[22:09] <teuf|la> having a way to release the audio device when the sink is paused would be enough for rb I guess ;)
[22:14] <Company> i told you the problem, go find a solution ;)
[22:14] <thomasvs> Company: are you saying it's unfixable for 0.8 ?
[22:15] <thomasvs> because then it makes more sense to revert so we have the old behaviour
[22:15] <thomasvs> I thought it was something that was getting fixed
[22:15] wheels ([email protected]) joined #gstreamer.
[22:18] <Company> there is no "old behaviour" - alsa had that problem since rb started setting the sink to NULL
[22:18] <Company> the other people only noticed that problem when i fixed osssink to correctly sync to timestamps
[22:19] <thomasvs> ok - then the old behaviour is to not make it sync correctly :)
[22:20] <teuf|la> does it make sense to report the content of the TLEN (length) tag when getting a GST_QUERY_TOTAL ?
[22:20] <Company> no
[22:20] <Company> i don't think reporting the TLEN tag is a good idea
[22:20] <teuf|la> that's probably the only sane way to get an accurate length on vbr mp3s
[22:21] <teuf|la> but that can also totally suck on badly tags mp3s
[22:21] <Company> yeah
[22:21] <Company> so rb can decide to use the length tag
[22:21] <teuf|la> ok, I'll do that from rb then ;)
[22:21] <Company> rb should also update the length when it has played the file
[22:21] <Company> it's internal length at least
[22:22] <Company> so it knows how long the file is
[22:22] <Company> thomasvs: that gives you sync errors on some mpegs (which was the reason i fixed it)
[22:23] <wheels> teuf|la: Well, Xing headers are nicer when available.
[22:24] <thomasvs> Company: so in the player it would still be a problem then ? or does the player not set the sink to NULL ?
[22:24] <thomasvs> Company: hm, what would happen if rb reffed the sink, disconnected it, and moved it out of the pipeline ?
[22:24] <teuf|la> wheels: is lame using such headers ?
[22:24] <Company> thomasvs: the problem would be streams where the audio doesn'T start with timestamp 0
[22:25] <teuf|la> I saw code for that in monkeymedia, but didn't look at it too much since they didn't seemt o be available in my mp3s
[22:25] <wheels> teuf|la: IIRC it can.
[22:25] <thomasvs> wheels: why are xing headers better for that ?
[22:25] <wheels> thomasvs: Because they often exist in the case that an ID3v2 tag doesn't.
[22:26] <wheels> thomasvs: And are usually written at encode time rather than being mucked about by various taggers.
[22:26] <teuf|la> hmm, lame seems to write xing headers
[22:26] <Company> thomasvs: either you assume the first timestamp means 0 (in that case you get rb to work correctly) or you sync to the timestamp (in that case mpegs sync correctly)
[22:26] <teuf|la> will look at that later, I'll do a quick fix by using TLEN tags in rb ;)
[22:26] <Company> gstreamer doesn't parse xine headers :)
[22:27] <wheels> TagLib does.  :-)
[22:27] <Company> because i don't know what they are and my mp3s don't have any such headers
[22:27] <Company> TagLib doesn't provide gst elements though :p
[22:27] <thomasvs> Company: ok, so the real issue is that rb sets the sink to READY on paused, causing it to reset the timestamp, right ?
[22:28] <Company> thomasvs: right
[22:28] <thomasvs> and rb does so because ready is the same state where the device is released ?
[22:28] <Company> thomasvs: stop, not right
[22:28] <Company> thomasvs: it sets the sink to NULL
[22:28] Action: thomasvs checks code
[22:29] <thomasvs> oh, we keep the device open even in READY ?
[22:29] <thomasvs> I thought it was paused - why also READY ?
[22:29] <Company> sure
[22:30] <Company> because external devices are opened/closed in NULL<=>READY
[22:30] <thomasvs> yeah, but why was it changed to be that way again ?
[22:30] <thomasvs> ie, what where the use cases saying that was a good idea ?
[22:30] Action: thomasvs should probably check mailing list
[22:30] <Company> so you know the caps of the devices
[22:31] <thomasvs> hm. there must be some better way to do that then.
[22:31] <Company> to do what?
[22:31] <thomasvs> to know the caps of the devices
[22:31] <thomasvs> whatever the reason, if there is no way rb can do what it wants to, it indicates a flaw in our design, not the other way around
[22:32] <Company> it indicates a flaw in our implementation
[22:32] <thomasvs> or, to put it differently, if rb can't do it with us, there's no point for them to use us
[22:32] <thomasvs> how can we implement it differently then ?
[22:32] <Company> to get this bug fixed, we _MUST_ send a discont at the beginning of every stream - mandatory and no way around it
[22:33] <Company> we will eventually make the core auto-send a discont on stream start
[22:33] <Company> but if i just implement it in the core i don't know what will break
[22:34] <thomasvs> "beginning of stream" - you mean paused -> playing ?
[22:34] <Company> no, i mean READY=>PLAYING
[22:34] sxpert_ ([email protected]) joined #gstreamer.
[22:35] Action: ds-work sometimes wishes pads knew more about streams
[22:35] Action: Company wishes we even had the concept of "streams" in our design
[22:36] <ds-work> ugh.  We need to get together again for a week and actually _design_ 1.0
[22:37] <thomasvs> yeah
[22:37] <thomasvs> I'm getting the feeling we don't have enough decent use cases to actually test our design against
[22:37] Action: wheels wonders if there's a guesstimated time frame for 1.0.
[22:37] <thomasvs> I mean, I can't tell my friends rhythmbox can't pause properly :)
[22:38] <thomasvs> wheels: a year from now, I'd say :)
[22:38] <Company> nope
[22:38] tehmiller ([email protected]) left #gstreamer ("So long and thanks for all the fish").
[22:38] <Company> with our current manpower we won't have 1.0 in a year
[22:38] <ds-work> the rb problem is because we _want_ to fix it correctly, but our design won't let us
[22:38] <wheels> thomasvs: Well, this is actually one interesting thing about the Qt bindings -- they usually uncover a few things that haven't apparently ever been tried.
[22:39] <wheels> Is the API completely frozen for 0.8 at this point?
[22:39] <thomasvs> well, to me it seems sort of obvious - if we want to clear timestamps in some state transition, and we want to open/close devices in some state transition, but we have cases where they shouldn't be done at the same time - then we're missing a state :)
[22:39] <Company> gstreamer doesn't properly do NULL=>PLAYING=>NULL=>PLAYING with 50% of elements i'd guess
[22:39] <thomasvs> wheels: I would imagine
[22:39] <ds-work> wheels: yes.  Unless there's something really bad that we missed
[22:39] <thomasvs> wheels: and it would be great if we knew what those things were :)
[22:39] <thomasvs> wheels: for core and libs used, yes
[22:40] <thomasvs> wheels: (i'm still faffing about a bit with media-info atm though, but that's not a real issue)
[22:40] <wheels> thomasvs: Ok, well, at some point I'm going to try to update the Qt bindings to the current API.  I'm sure that will be quite a chore...
[22:40] Nick change: kh_zZz -> kh_out
[22:40] <thomasvs> wheels: well, I'd get on with it if I were you
[22:40] <Company> wheels: not really - the rb API hasn't changed much apart from tags
[22:40] <thomasvs> wheels: it'll be a lot easier to fix bugs that you find because of it before the 0.8 release
[22:40] <wheels> thomasvs: Is there a release schedule yet?
[22:41] <thomasvs> wheels: not a fixed one
[22:41] <thomasvs> wheels: but, I'm still going on the assumption that we do 0.8 around march 10th
[22:41] <wheels> thomasvs: Yeah, that's kind of the idea, but before KDE 3.2 I didn't have any time and just after I needed a break.  :-)
[22:41] <wheels> Oh, that soon...  :-)
[22:41] <thomasvs> heh :)
[22:41] <thomasvs> yeah, we promised gnome :)
[22:41] <Company> gst 0.8 with gnome 2.6
[22:41] <thomasvs> it'll be interesting when we make kde similar promises
[22:41] ds-work ([email protected]) left irc: Read error: 60 (Operation timed out)
[22:41] <Company> gst 0.10 probably with gnome 3.0
[22:42] <Company> (or gnome 2.10?)
[22:42] <wheels> thomasvs: So the GNOME beta1 is this week and the final in 2 weeks?  That seems like a pretty short gap...
[22:42] <thomasvs> wheels: no, the final is a month away I think
[22:42] <thomasvs> wheels: but I want us to be 0.8 quite before the final gnome date
[22:43] <thomasvs> wheels: I'm sure we'll have some stupid bugs left because of jumping to an 8 major
[22:43] ds-work ([email protected]) joined #gstreamer.
[22:43] <thomasvs> wheels: and I'm sure some bugs will still be shaken out before final gnome release too
[22:43] <thomasvs> wheels: btw, did you see leon's reply ? I'm not sure what to think
[22:43] <wheels> thomasvs: Yeah, confused me.
[22:43] <wheels> thomasvs: Well, but not really -- I wouldn't have expected any other answer...
[22:44] <wheels> thomasvs: I don't think he made the connection that "Scott Wheeler" == "the guy he was ranting to at N7y"
[22:44] <teuf|la> thomasvs: btw, what should I do to get a cvs account for gstreamer freedesktop cvs ?
[22:45] <wheels> thomasvs: And it may have just been that -- an atypical rant or something.  We certainly all have our moments.  :-)  But it just wasn't particularly clever to do so to a complete stranger...
[22:46] Action: iain_ has rb play 4 songs in a row perfectly and checks config.log to make sure its actually using gst
[22:46] Action: Company shoots iain_, just because
[22:47] <Company> my rb worked perfectly (apart from that pause thingy) for  ever
[22:47] <Company> MAS development is totally backwards
[22:47] <thomasvs> teuf|la: send me a mail with your desired user name, ssh pubkey, and email address
[22:47] <iain_> I've had nothing but problems with it
[22:47] <teuf|la> thomasvs: k
[22:47] sxpert ([email protected]) left irc: Connection timed out
[22:47] <thomasvs> wheels: well, yeah, maybe
[22:47] <thomasvs> wheels: otoh while in person he is a nice enough guy for real-life-stuff ...
[22:47] <iain_> Company: would you have time to check a gst debug log and see if you can see anything going wrong?
[22:48] <Company> i've never been good at looking at debug logs
[22:48] <thomasvs> wheels: ... ten bad echos and not a good one does add up to either a statistical anomaly or a firm conclusion :)
[22:48] <wheels> thomasvs: Only if he thinks you're important, sadly.
[22:48] <iain_> Company: oh, ok
[22:48] <iain_> Company: I cant see anything going wrong...but its just not working
[22:49] <wheels> thomasvs: That's what really drove me nuts -- I wasn't a donor or an aRts guy so he basically didn't want to talk to me...
[22:50] <Company> mas has no bugzilla, no webcvs, no irc channel, no nothing
[22:50] <Company> it basically does not exist
[22:50] <Company> i have never seen it work
[22:51] <Company> to sy it my way: it's just because of this Leon guy shouting that people think about it at all
[22:54] Action: ds-work has seen it work
[22:54] <ds-work> at the hands of a MAS developer
[22:54] <ds-work> I spent quite a bit of time trying to get the Debian packages to work, but couldn't
[22:54] <wheels> Company: Well, I raised that issue on kde-mm too.
[22:55] <wheels> Company: Leon kind of ignored that part of my mail.  ;-)
[22:57] <Company> people can impress me by doing stuff
[22:57] <Company> people can not impress me by talking
[22:57] Action: ds-work withdraws his last assertion
[22:58] <ds-work> I have mas working locally
[22:58] <Company> that goes for the MAS devs as it does for ramon's RTP stuff
[23:00] <lupusBE> ds-work, trying to package mas?
[23:01] <ds-work> lupusBE: it's been in Debian for over a year
[23:01] <lupusBE> I packaged it for redhat
[23:01] <wheels> NMM is really the only other thing that I'm interested in.  But I'm trying to be fair and balanced and all that.
[23:01] <lupusBE> it had some faults in it
[23:02] <lupusBE> that it looked in the ?/local dirs
[23:02] <lupusBE> but they resolved that
[23:02] <thomasvs> lupusBE: no, that's a matter of imake configuration
[23:02] <thomasvs> isn't it ?
[23:03] <lupusBE> I hate to build from the lasted cvs to use the correct dirs
[23:03] <lupusBE> s/hate/had
[23:03] <thomasvs> the idea alone that someone writes something new in 2000+ and decides to use imake for it is already boggling
[23:03] <wheels> heh
[23:03] <lupusBE> hehe
[23:03] <lupusBE> I packaged imake with autoconf :)
[23:04] <wheels> thomasvs: It's like I've said --  the most redeeming thing about m4 and automake is everything else.  :-)
[23:04] <Company> wheels: I'd love it if NMM got up to speed, but I don't think it's there yet
[23:05] Action: Company wants a gst competitor so he has something to wrap ;)
[23:05] <wheels> Company: Yeah, I have a similar impression of it.
[23:05] <teuf|la> Company: make a library out of mplayer and wrap that, that should keep you busy for a bit ;)
[23:05] <wheels> Company: And it's all done at the same Uni which is good for them in some ways, but makes it hard to interact with as a project.
[23:06] <thomasvs> wheels: hehe, true :)
[23:06] <wheels> Company: But it's at least far enough that I feel like out of fairness I need to give it a good look.
[23:07] <thomasvs> Company: isn't alsa supposed to have a kernel-level mixer btw ?
[23:08] <wheels> thomasvs: Only for cards that do it in hardware IIRC.
[23:08] <wheels> thomasvs: i.e. you can't count on it being there...
[23:08] <Company> yeah, listen to wheels
[23:09] <Company> i can't open anything twice with alsa
[23:09] <teuf|la> isn't this dmix thing for alsa doing that in software if properly configured ?
[23:09] Action: wheels doesn't actually know anything about multimedia but pays attention to people that do often enough that he's good at pretending.  :-)
[23:09] <Company> wheels: the problem that i see is that gst doesn'T have a competitor and all free software projects without competitors have always had big problems in the past :/
[23:09] <thomasvs> I was pretty sure alsa had software mixing at the kernel level
[23:10] <thomasvs> am I that far off ?
[23:10] <Company> dmix is esd for alsalib, yeah
[23:10] <thomasvs> well, isn't that always present ?
[23:10] <Company> thomasvs: i'm pretty sure it doesn't
[23:10] <ds-work> Company: wrap JMF :)
[23:10] <wheels> Company: Well, and it might be a nice one because while it might be a functional competitor it doesn't seem like it's competing with its goals.
[23:10] <Company> no, i don't have dmix and i wouldn't wanna have it - it's more unstable than even alsalib ;)
[23:11] <Company> ds-work: what's JMF?
[23:11] <ds-work> Company: Java Media Framework
[23:11] <ds-work> or wrap Helix
[23:11] <Company> wheels: does it do caps negotiation?
[23:11] <wheels> Company: no idea
[23:11] <wheels> Company: I just pulled down the source for the first time a few days ago.
[23:12] kwm ([email protected]) left irc: "Leaving"
[23:12] <wheels> Company: But I think I might port JuK's Player class to it just to see what the API is like.
[23:12] <Company> ds-work: i'm considering wrapping helix at some point, but my soft toy hasn't finished the license yet to decide if it may read helix to me
[23:12] Action: ds-work wonders what we've gained so far from self-assembling pipelines
[23:13] <Company> wheels: i have never seen that it does negotiate formats, people only connected the right elements and that worked
[23:13] <Zeenix_> this link is broken: http://www.freedesktop.org/~gstreamer/apps/vw/ 
[23:13] <Company> ds-work: rb-player-gst is _very_ short :)
[23:13] <wheels> Company: Which would actually probably be fine in a KDE context since we could just use our mime-magic stuff anyway.
[23:13] <ds-work> Company: ?
[23:14] <Company> ds-work: and we'll gain more when we handle events correcxtly and allow auto-switching of media types
[23:14] <Company> wheels: to detect the AVI formats? ;)
[23:15] <wheels> Company: Ok, so maybe not that.  :-)
[23:15] <Company> wheels: and AVI is easy to detect - try doing it with some weird container like ogg
[23:15] <wheels> Company: I have done it with Ogg.  :-)
[23:17] <Company> wheels: it's probably not that short, is it? (especially for chained bitstreams)
[23:18] <wheels> Company: Well, again, TagLib has a complete implementation of Ogg pages and packets...
[23:18] <Company> hey, someone do a tagliboggdemux :)
[23:19] <wheels> Company: It would probably actually work.  And it's probably a more readable implementation than the one in libogg.
[23:19] Action: wheels admits that there are still a couple of places where I cheated, but I'll fix those at some point...
[23:20] <wheels> Company: But yeah, the Ogg implementation was probably conceptually the hardest code in TagLib.
[23:26] <Zeenix_> who is maintainer for www?
[23:27] sub_pop ([email protected]) joined #gstreamer.
[23:30] Zeenix_ ([email protected]) left irc: "The whole problem with the world is that fools and fanatics are always so certain of themselves, but wiser people so full of 
[23:31] Action: mathrick wishes everybody good night (or whatever preferred time :)
[23:31] Nick change: mathrick -> mathrick|sleep
[23:33] <iain_> Hmm, strange, my sink is receiving buffers constantly, but never gets an EOF to tell it to stop pulling and every time it pulls it gets another buffer
[23:34] <Company> iain_: don't use sinesrc? ;)
[23:34] <iain_> ;p
[23:34] <iain_> I'm using filesrc
[23:35] <iain_> on a 500k mp3
[23:35] <Company> then that's not good
[23:35] thaystab1e ([email protected]) left irc: Remote closed the connection
[23:35] <iain_> hmm, further investigation reveals that it is passing 1 frame at a time
[23:35] thaytoo ([email protected]) joined #gstreamer.
[23:36] <iain_> this may be the problem...
[23:38] mxpxpod ([email protected]) left irc: "Guns kill people as much as spoons made Rosie O'Donnell fat.."
[23:38] jcsston ([email protected]) joined #gstreamer.
[23:38] teemu ([email protected]) left irc: Remote closed the connection
[23:39] teemu ([email protected]) joined #gstreamer.
[23:42] Action: ds-work needs to log out
[23:43] <ds-work> something in my environment is really screwey, and I can't run make distcheck on anything
[23:43] <ds-work> imagine my surprise when I gave up figuring out this problem yesterday, went home, and discovered that the problem had "fixed" itself in my walk home
[23:44] <thomasvs> you're going to get really angry a few times over the next week according to your chakra
[23:44] <thomasvs> I'm trying to fix a similar heisenbug in our buildbot setup
[23:48] ds-work ([email protected]) left irc: "Leaving"
[23:50] lupusBE ([email protected]) left irc: Remote closed the connection
[23:57] teuf|la ([email protected]) left irc: "Client exiting"
[23:57] <iain_> did wingo's big floatint changes get committed?
[23:57] lupusBE ([email protected]) joined #gstreamer.
[00:00] --- Wed Feb 25 2004
[00:02] BBB ([email protected]) left #gstreamer.
[00:03] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[00:05] Action: bilboed is back (gone 05:45:33)
[00:08] ibsteveog ([email protected]) left irc: "quitting"
[00:08] ds-work ([email protected]) joined #gstreamer.
[00:09] grizz ([email protected]) joined #gstreamer.
[00:10] wheels ([email protected]) left irc: Read error: 60 (Operation timed out)
[00:16] Company ([email protected]) left irc: Remote closed the connection
[00:16] <ds-work> changing user IDs is fun
[00:17] wheels ([email protected]) joined #gstreamer.
[00:18] jcsston ([email protected]) joined #gstreamer.
[00:19] chh ([email protected]) joined #gstreamer.
[00:24] mxpxpod ([email protected]) joined #gstreamer.
[00:24] Action: thomasvs finds silly bugs in mulawdec
[00:24] <thomasvs> ds-work: I don't understand your debug registering comments - here's the thing
[00:25] <thomasvs> ds-work: if a lib registers stuff *before* gst_init, it can't because gst and the debugging system isn't inited yet
[00:25] sub_pop ([email protected]) got netsplit.
[00:25] mathrick|sleep ([email protected]) got netsplit.
[00:25] yippi ([email protected]) got netsplit.
[00:25] sxpert_work ([email protected]) got netsplit.
[00:25] ds ([email protected]) got netsplit.
[00:25] <thomasvs> ds-work: if a lib registers stuff *after* gst_init, the options are already parsed
[00:25] <thomasvs> ds-work: the reason it works for all plugins is because the plugin_init functions registering debug stuff get called *during* gst_init
[00:26] <thomasvs> ds-work: am I missing something or are we not talking about the same thing ?
[00:26] <ds-work> thomasvs: but a lib can put stuff in _init that will be called before main()
[00:26] <ds-work> and you can register categories before gst_init() is called
[00:27] <thomasvs> ds-work: ok, but that means a lib has to provide one function to make it register categories, and another to actually init the lib where it can call the gst stuff it needs
[00:27] markey ([email protected]) left irc: "goodnite"
[00:27] <ds-work> no, a lib just needs to register its categories in _init()
[00:29] <thomasvs> ds-work: and what if the lib has to call gstreamer functions in its _init ?
[00:29] <ds-work> if an app is linked to the library, then _init() is called before main(), main calls gst_init(), and the lib's category is magically there
[00:30] <ds-work> I think you're missing something: a function called "_init"
[00:30] <ds-work> void _init(void)
[00:30] yippi ([email protected]) returned to #gstreamer.
[00:30] <ds-work> read the man page for dlopen()
[00:31] ds ([email protected]) got lost in the net-split.
[00:31] sxpert_work ([email protected]) got lost in the net-split.
[00:31] mathrick|sleep ([email protected]) got lost in the net-split.
[00:31] sub_pop ([email protected]) got lost in the net-split.
[00:31] <thomasvs>        Libraries should not export routines named _init or _fini.  Those mech-
[00:31] <thomasvs>        anisms  are  obsolete,  and  may result in undesired behavior.
[00:31] <thomasvs> (this is the first I hear about _init btw :))
[00:32] <ds-work> ok, use the .init section
[00:32] <thomasvs> I thought you were using _init () as shorthand for gst_library_init ()
[00:32] <ds-work> _init doesn't work well for static libs
[00:33] <thomasvs> I must be missing something else.
[00:33] <thomasvs> does this also work for shared linking ?
[00:33] <ds-work> yes
[00:33] <thomasvs> ie, is linking to .so files just some magic that uses dlopen ?
[00:34] <ds-work> yes
[00:34] <thomasvs> and you already fixed something similar were exactly ? (gstplay doesn't have debugging categories)
[00:34] <ds-work> or rather, dlopen() is an interface to the library loading code
[00:35] <ds-work> some app or test was registering a category before calling gst_init()
[00:35] <ds-work> it was failing on Solaris
[00:37] harshy ([email protected]) joined #gstreamer.
[00:38] sxpert_work ([email protected]) joined #gstreamer.
[00:39] ds ([email protected]) joined #gstreamer.
[00:39] sub_pop ([email protected]) joined #gstreamer.
[00:40] Action: ds-work got a Solaris machine today
[00:40] Action: ds-work will get an OS X machine later this week
[00:43] mathrick|sleep ([email protected]) joined #gstreamer.
[00:43] <thomasvs> ds-work: you don't remember which app/test so I can cheat from it ?
[00:45] <ds-work> if I didn't have to wade through 1000's of bug changes, I might find it :)
[00:46] interalia ([email protected]) joined #gstreamer.
[00:47] <interalia> hey
[00:47] <ds-work> oops, it was more complicated than that
[00:47] <interalia> I'm trying to strip the audio out of a .3gp (quicktime-based format) file, using the qtdemux plugin
[00:47] <ds-work> there were multiple bugs
[00:48] <thomasvs> ds-work: well, sorry about the bug changes :)
[00:48] <ds-work> thomasvs: it was related to gtk-scandoc not calling gst_init()
[00:48] <interalia> I've tried: gst-launch filesrc location=Video.3gp ! qtdemux ! filesink location=test.amr
[00:48] <ds-work> so some plugin registered a debug category, and it segfaulted
[00:48] <interalia> but it tells me it can't link element qtdemux0 to filesink0 ...
[00:49] <ds-work> thomasvs: #133414
[00:49] <interalia> anyone give me a hint on maybe how I can manage this?  I'm pretty fuzzy on all this video_%02d stuff I see in examples
[00:49] <thomasvs> ds-work: ah, gthreads not being initalized ?
[00:49] <ds-work> thomasvs: yes, but it was more complicated than just that
[00:49] <ds-work> thomasvs: I *think* I fixed it
[00:49] <ds-work> thomasvs: or maybe I just discovered it :)
[00:50] <thomasvs> uh, that bug report doesn't really tell me what to do with _init or .init :)
[00:50] <ds-work> interalia: filesrc ! qtdemux .audio_00 ! fakesink
[00:50] <ds-work> thomasvs: yeah, I was wondering why you wanted to look at it
[00:51] <interalia> ds-work: thanks for answering :)  that seems kidna better, I'm now getting: syntax error, unexpected IDENTIFIER, expecting '('
[00:52] <interalia> that's whether I do filesrc ! qtdemus .audio_00 ! fakesink or filesink
[00:52] <ds-work> interalia: what shell are you using?
[00:53] <thomasvs> ds-work: are you jerking me around ? :) I'm trying to find an example of where you fixed registration of debug categories using dlopen hacks
[00:53] <ds-work> thomasvs: I don't have an example
[00:54] <ds-work> the bug was merely _related_.  Sorry
[00:54] <interalia> ds-work: bash.
[00:55] <iain_> wow, marlin/gstreamer has decided that my 500k mp3 is actually 26hours long
[00:55] <ds-work> interalia: as long as you have spaces around !, it should be ok
[00:58] <interalia> ds-work: hmm, well I definitely have spaces around the !s.  I'm using gstreamer from debian unstable (0.6.4) but some of the debs are 0.6.2 since I didn't upgrade them all at once.  I'll upgrade to the 0.6.4 gstreamer-tools and see if it makes a diff.
[00:58] <ds-work> interalia: oh, I didn't know you were using 0.6
[00:59] <ds-work> interalia: qtdemux doesn't work in 0.6
[00:59] <interalia> ds-work: bugger.
[00:59] <thomasvs> ds-work: ok, so you're guessing using .init from dlopen would work, but you're not really sure, right ?
[01:01] <ds-work> thomasvs: http://www.schleef.org/~ds/register.c
[01:01] <ds-work> works peachy, although not exactly the way I described :)
[01:02] <thomasvs> heh, wow
[01:02] <thomasvs> constructors ?
[01:02] <thomasvs> o. k.
[01:02] Action: thomasvs will check tomorrow
[01:02] <thomasvs> how high is the chance that this is not really ansi c ?
[01:03] <ds-work> um, __arrtibute__?
[01:03] Action: thomasvs was hoping it wouldn't really show he didn't know what he actually was asking
[01:03] <ds-work> it's not
[01:04] <thomasvs> I mean, will it work on forte ?
[01:04] <ds-work> no
[01:04] <ds-work> iirc, there's a portable way to do it
[01:04] steveb_ ([email protected]) joined #gstreamer.
[01:07] earthw0rm ([email protected]) joined #gstreamer.
[01:08] <earthw0rm> can i use gstreamer to append one ogg file onto another (command line)?
[01:08] <ds-work> ogg files can theoretically be appended using cat
[01:08] <earthw0rm> really?
[01:08] <earthw0rm> weird
[01:09] Action: earthw0rm does man cat
[01:09] <earthw0rm> cat one.ogg two.ogg > three.ogg ?
[01:10] <ds-work> yes
[01:10] jimmy_dean_ ([email protected]) joined #gstreamer.
[01:10] <earthw0rm> i'll be well impressed if that works
[01:10] <thaytoo> it's part of the ogg spec that they can be appended like that
[01:10] <earthw0rm> its a bloody good idea
[01:11] <interalia> ds-work: thanks for your assistance! I might try building 0.7 later
[01:11] <thaytoo> you won't get one file of the total length though in the metadata
[01:11] <thaytoo> but it will play fine
[01:11] <earthw0rm> oh ...
[01:11] <earthw0rm> thats a let-down
[01:12] <thomasvs> earthw0rm: you have a one in 2^32 chance of that failing though
[01:12] <interalia> earthworm: it basically means that oggs can simply have more than one stream, one after the other.  but I think you might find that some players simply play the first stream
[01:12] <earthw0rm> crap
[01:12] Misirlou ([email protected]) left irc: "asdf"
[01:12] <earthw0rm> can gstreamer help me out here then?
[01:13] <interalia> earthworm: yeah, but it's worth a try with whatever your favourite player is anyway ;)
[01:13] <earthw0rm> ok, well i'll do it and try it with totem
[01:13] <earthw0rm> just converting the mp3's to ogg
[01:13] <earthw0rm> (thanks to gstreamer)
[01:16] <earthw0rm> well, playing that seems to have mullered Xine/Totem
[01:17] <earthw0rm> lol
[01:17] <thaytoo> mullered?
[01:17] <earthw0rm> killed it
[01:18] <earthw0rm> gave it a headache
[01:18] <earthw0rm> named after the great maker of deserts, for some odd reason ... but anyway ...
[01:19] wheels ([email protected]) left irc: Read error: 60 (Operation timed out)
[01:19] steveb__ ([email protected]) left irc: Read error: 110 (Connection timed out)
[01:20] <earthw0rm> maybe the bit rates were different or something?
[01:20] <earthw0rm> :/
[01:21] Surphaze ([email protected]) left #gstreamer ("Client exiting").
[01:22] <earthw0rm> would i have to write a gstreamer app to concat files then?
[01:23] <earthw0rm> can't imagine that being easy, whats the gstreamer api like to get something like a media player going, or a utility like i was talking about there?
[01:35] wheels ([email protected]) joined #gstreamer.
[01:40] Shoragan ([email protected]) left irc: "Leaving"
[01:44] harshy ([email protected]) left irc: "I quit for now"
[01:44] harshy ([email protected]) joined #gstreamer.
[01:58] Action: spyder482 still tries to figure out how to mux from avi to mkv
[01:59] iain_ ([email protected]) left irc: "!"
[02:17] ChrisHJW ([email protected]) joined #gstreamer.
[02:18] smoke ([email protected]) got netsplit.
[02:19] sub_pop ([email protected]) left irc: "Client exiting"
[02:24] smoke ([email protected]) got lost in the net-split.
[02:32] chh ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[02:33] ChrisHJW ([email protected]) left irc: Client Quit
[02:44] spyder482 ([email protected]) left irc: "Leaving"
[02:45] sub_pop ([email protected]) joined #gstreamer.
[02:47] spyder482 ([email protected]) joined #gstreamer.
[02:53] earthw0rm ([email protected]) left #gstreamer (""I allege that SCO is full of it." - Linus Torvalds").


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click