[gst-devel] Daily IRC logs

[wim.taymans at chello.be]

[06:30] <omega_movie> of course
[06:30] Action: omega_movie is working on getting rid of that
[06:30] <omega_movie> but /movie first
[06:30] <omega_movie> and please write up ideas for the DEBUG/INFO system into GstDebugColors for now, or create GstDebugDiscussion or something
[06:31] <taaz> for one, debug and info not the same yet
[07:47] Action: taaz builds 2.4.4
[08:04] BBB-zZz (BBB at ucu-105-116.ucu.uu.nl) left irc: [x]chat
[08:21] Action: omega_movie gets 2.4.4
[08:22] Nick change: omega_movie -> omega_
[08:35] ajbusy (me at p51-max6.dun.ihug.co.nz) left irc: Ping timeout for ajbusy[p51-max6.dun.ihug.co.nz]
[08:35] ajbusy (me at p51-max6.dun.ihug.co.nz) joined #gstreamer.
[08:37] <omega_> yo
[08:52] omega_ (omega at omegacs.net) left irc: Ping timeout for omega_[omegacs.net]
[08:54] omega_ (omega at omegacs.net) joined #gstreamer.
[08:54] <omega_> stupid APM
[09:07] <taaz> is there some way to shut off sync and go full video speed?
[09:08] <omega_> dunno, not sure exactly how the clocking works
[09:08] <omega_> you want to play to screen at top speed?
[09:08] <taaz> for timing purposes
[09:08] <omega_> gstclock.c
[09:09] <omega_> ok, incsched test/mpeg2parse3.c is fully correct
[09:09] <taaz> uh... so what's this hard-coded path in mp1vid? ;)
[09:09] <omega_> um, yeah
[09:10] <omega_> if you wanna change it to argv[1], go ahead
[09:10] Action: omega_ just watched the bullet scene in almost 100% speed with audio and sync for the first time <g>
[09:10] <taaz> bullet scene?
[09:11] <omega_> from the matrix
[09:11] <omega_> no changes to mpeg2dec cvs in a long time ;-(
[09:12] <omega_> and /me just listened to the commentary track from the same .vob <g>
[09:12] <omega_> by changing 1 char in mpeg2parse3.c
[09:13] <omega_> it's just a tiny bit too slow ;-(
[09:13] <taaz> yeah... walken got eaten by vmware.  lucky that he got it so good before that.
[09:13] <omega_> yup
[09:14] <omega_> now sampeg may be the next uber-decoder
[09:14] <taaz> someone needs to do the same for ac3dec
[09:14] <omega_> quite
[09:14] <taaz> its got serious issues... 
[09:14] <omega_> yup
[09:14] Action: omega_ is gonna turn off debug and see if that frees up sufficent CPU to watch a DVD on a PIII/500
[09:14] <taaz> most of which can be fixed using similar code to what mpeg2dec uses.  but then someone needs to go in and make it more bullet proof like mpeg2dec
[09:15] <omega_> if I had more time, I'd do it
[09:15] <omega_> but I'd end up rewriting it from scratch
[09:15] <taaz> its not much code... wouldn't be hard to do that
[09:16] <omega_> maybe in a month or so, when gstreamer core settles out a little and I have more time to work on the plugins
[09:16] <omega_> tried mpeg2parse3?
[09:18] <taaz> mp1vid not working too well...
[09:18] <omega_> yeah, that's still the hardest case of any test
[09:20] <omega_> nope, still breaks up
[09:22] Action: omega_ hopes libvo and libao die a slow death <g>
[09:22] <taaz> unlikely
[09:23] <omega_> neat, CVS ac3dec is much worse than what's in gstreamer
[09:23] <omega_> why do you say that?
[09:23] <taaz> vorbis took libao and mutated it
[09:23] <omega_> uh oh
[09:24] <omega_> well, hopefully it'll go away when vorbis sees that the rest of the world is spontaneously converting to gstreamer
[09:24] <taaz> it's not a very good api to start with.  too bad there's no common api for people to use
[09:25] <taaz> did you see pbd's post about wrapper apis on lad?
[09:25] <omega_> that reminds me, we need to get audio-out plugins for other systems like irix and solaris
[09:25] <omega_> checking..
[09:26] <omega_> whoo! 52Mbps used at kernel.org
[09:26] <taaz> yeah, i helped with that
[09:26] <omega_> ditto
[09:27] <omega_> title of pbd's mail?
[09:29] <taaz> Date: Thu, 19 Apr 2001 19:58:33 -0400   Subject: [linux-audio-dev] sound API libraries, servers, etc.                   
[09:30] <omega_> reading
[09:31] <taaz> need to read the thread really...
[09:31] <omega_> yeah
[09:31] <omega_> before or after this message?
[09:31] <taaz> after
[09:31] <omega_> ok, any mention of gstreamer?
[09:31] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[09:31] <omega_> yo
[09:32] <Uraeus> hi omega_
[09:32] <taaz> i dunno... i only skim most of these lad msgs
[09:33] <omega_> what's funny about the first reply is that he seems to think that aRts[d] in GNOME is a done deal
[09:33] <Uraeus> who thinks that?
[09:34] <omega_> From: MOULET Xavier FTRD/DMR/ISS <xavier.moulet at rd.francetelecom.fr>
[09:34] <omega_> but I think aRts is not optimized enough to be said lightweight right now,
[09:34] <omega_> he's got that right..
[09:34] <Uraeus> and btw, Geocrawler sucks, their 15 mintute maintenance is more like 5-6 hours
[09:34] <omega_> really....
[09:35] <omega_> the odd thing is that pbd is fully aware of GStreamer
[09:35] <Uraeus> pbd is a sound server?
[09:35] <omega_> no, pbd is a person
[09:36] <Uraeus> I hate coming in into the middle of a conversation :)
[09:36] <omega_> bleagh, he's still got the wrong idea
[09:36] <omega_> Date: Fri, 20 Apr 2001 10:18:46 -0400
[09:37] <omega_> aes has 32 internal busses for xfering data between apps if they want to
[09:37] <omega_> WHY BUSSES???
[09:37] <omega_> pbd is very very analog-hardware-centric
[09:37] <taaz> what is this "analog" concept you speak of? ;)
[09:37] <omega_> mu
[09:38] <omega_> hah, pbd's forward of stw's response is kinda funny
[09:38] <Uraeus> which mailing list is this?
[09:39] <taaz> i think (though haven't looked at the code) that some of that stuff he's doing is done to squeeze last bit of performance out of the machine.  so 32 busses probably means optimizations done with bit masks all over
[09:39] <omega_> linux-audio-dev
[09:39] <omega_> taaz: most likely, yes
[09:39] <Uraeus> ah
[09:39] <omega_> stw doesn't seem to understand what "in-process client" means
[09:40] <taaz> though i can't imagine that it makes that much difference to use dynamic byte arrays or something instaed... he's the expert though
[09:40] <omega_> out-of-process as pbd means it says that artsd forces a new thread in the client app to talk to the server
[09:40] <omega_> stw thinks in-process means that the client's processig code happens in the server
[09:40] <omega_> neither is strictly true, actually
[09:42] <omega_> Uraeus: I can now play a VOB file with audio, video, and reasonable sync, with about 525MHz of processing power
[09:42] <omega_> unfortunately, I only have 500 ;-(
[09:42] <taaz> whatcha going to do about that? 
[09:43] <omega_> um
[09:43] <omega_> cry
[09:43] <taaz> you could try that rearranging queues to be next to sinks
[09:43] <omega_> no, the problem is actual cycles
[09:44] <omega_> there is mention of gstreamer in this thread, but pbd dismisses it as not a sound server
[09:44] <omega_> which is correct, it's the basis for one <g>
[09:44] <Uraeus> omega_:cool, btw, there is a thread called 'GNOME 2 Sound Architecture and APIs?' on gnome-hackers you guys might find interesting
[09:44] <omega_> yeah
[09:44] <omega_> I haven't read it, I must
[09:45] <omega_> Uraeus: seen color-debug.png ?
[09:46] <omega_> Date: Thu, 12 Apr 2001 12:33:25 -0400 (EDT)
[09:46] <omega_>  From: Elliot Lee <sopwith at redhat.com>
[09:46] <omega_> is interesting
[09:47] <omega_> taaz: as for speeding things up, afaik mpeg2dec *still* doesn't use sse, even though I sent walken a patch for it almost a year ago
[09:47] <Uraeus> omega_: no, I have not seen color-debug.png
[09:47] <omega_> check it out
[09:47] <taaz> hmm?
[09:47] <omega_> it's cool <g>
[09:48] Action: taaz looks at code...
[09:48] <omega_> taaz: sse can squeeze a few percent out of the mpeg2dec turnip
[09:48] <taaz> i thought it was in there...
[09:48] <Uraeus> omega_:URL?
[09:48] <omega_> see -devel
[09:48] <omega_> taaz: nope
[09:48] <taaz> hell, i wrote a motion comp sse patch >1 year ago
[09:48] <Uraeus> omega_: as I said Geocrawler sucks :)
[09:48] <omega_> don't you subscribe??
[09:48] <omega_> gstreamer.net/images/....
[09:49] <Uraeus> omega_:no just read geocrawler
[09:49] <omega_> ick
[09:49] <Uraeus> omega_: if you guys want to move to mail.gnome.org let me know, it has a working archive system at least :)
[09:49] <omega_> heheeh
[09:50] <omega_> taaz: ok, it has some, it uses pavgb sorta
[09:50] <omega_> but I don't like the extreme macroization of motion_comp_mmx.c
[09:51] <omega_> there are structural wins you can get with sse that you can't get just by replacing a few instructions in the mmx code
[09:51] <taaz> omega_: there is some sse code in there
[09:51] <omega_> there's mmxext code there, which isn't quite as far as sse
[09:51] <omega_> afaict
[09:53] <taaz> hmm.. i dunno.    walken was talking about speedups due to prefetching or uncached writes or something which is sse... so he did work on it at one point
[09:53] <omega_> Uraeus: like it?
[09:54] <omega_> I never got prefetching to *not* cause a massive performance hit
[09:54] <omega_> the horizon is a lot farther off than I thought
[09:56] <taaz> eh... ask livid-mpeg2dec list about it.  complain you need to squeeze 525Mhz performance into 500Mhz and maybe walken will help
[09:56] <Uraeus> omega_: hmm, that is actually very nice should ease debugging I guess
[09:57] <omega_> hehehe
[09:57] <omega_> Uraeus: yup, and it'll only get better
[09:57] <Uraeus> omega_: I plan on mailing gnome-hackers about the sound setup, so I need some info on the syncing problems
[09:57] <omega_> taaz: anyone in -mpeg2dec besides walken that is aware of all the issues?
[09:57] <omega_> Uraeus: such as?
[09:58] <Uraeus> omega_: if I understand correctly unless artsd and the kernel starts supporting the stuff Jim Gettys talked about we will never be able to sync video and sound through artsd right?
[09:58] <taaz> omega_: not sure...  there are other app authors that use lib there... and some people that know tech details
[09:59] <omega_> Uraeus: correct
[09:59] <Uraeus> omega_: but if we make a small soundserver atop Gstreamer we will right?
[09:59] <omega_> taaz: I mean deep deep processor things, like prefetch horizon
[09:59] <omega_> Uraeus: how?
[10:00] <Uraeus> omega_: don't know, but Ow3n talked about doing a soundserver for/with GStreamer 
[10:00] <Uraeus> omega_: but you are saying that any soundserver will give us the sync problem?
[10:00] <omega_> right, what does gstreamer have or not have to do with the video sync issue, at this point?
[10:00] <omega_> yes, any separate soundserver will
[10:01] <Uraeus> omega_: could just using Gstreamer without any soundserver be an alternative?
[10:01] <omega_> the problem is that there is no way to guarantee or even really measure (in a generic sense) the latency, in order to account for that in the video end
[10:02] <omega_> Uraeus: the problem is still hardware-sharing
[10:02] steveb (steveb at node1ee31.a2000.nl) joined #gstreamer.
[10:02] <omega_> Ow3n's concept is to ship both audio *and* video to the 'media server' and let it deal with sync internally
[10:02] <omega_> this server just exports an xid for the app to embed as usual
[10:02] <omega_> yo steveb
[10:03] <steveb> morning
[10:03] <omega_> evening...
[10:03] <omega_> oh wait, morning.
[10:03] <steveb> ha!
[10:03] <Uraeus> omega_: and doing that with artsd is not plausible/possible/effecient (the both video and audio part)
[10:03] <Uraeus> ?
[10:03] <omega_> right
[10:03] <omega_> well, the audio part it supposedly does 'right'
[10:04] <omega_> but it can't do video at all
[10:04] <omega_> the other thing is that there's an optimization possible with a gstreamer sound server
[10:04] <omega_> instead of shipping raw audio and video to the server (esp if it's over a network), you ship the compressed stuff
[10:04] <omega_> and the media server does the decode/output
[10:04] <omega_> which for localhost is epsilon difference, but for other scenarios is the make/break point
[10:05] <Uraeus> ok, I guess I have to talk with Owen and see if he is willing to commit to making such a media server for GNOME 2.0 before suggesting it on gnome-hackers
[10:06] <omega_> see gnostream.sourceforge.net
[10:06] <omega_> iirc
[10:06] <omega_> or wiki/GstApplications has it anyway
[10:07] <Uraeus> yes, but in the making is not good enough, I(and GNOME for that matter) would need a more solid commitment
[10:07] <omega_> yup
[10:07] <taaz> how about you just write it first as a gst example ;)
[10:07] <omega_> heh
[10:07] <omega_> we lack a couple plugins still for a simple sound server
[10:07] <omega_> a) a good resampler, b) a good arbitrary mixer
[10:07] <Uraeus> of course the other risk factor here is that a stable release of Gstreamer is out in time
[10:08] <omega_> right
[10:08] <omega_> define 'in time' for gnome 2.0 again? <g>
[10:08] <Uraeus> end of the year :)
[10:08] <omega_> ha
[10:08] <Uraeus> at least aspirations for end of the year :)
[10:08] Action: omega_ takes the crackpipe away from Havoc
[10:09] <Uraeus> well, I guess 'nonone' believes the end of the year date, but setting it do help put priorities straight 
[10:09] <omega_> yeah
[10:10] <omega_> my hope is to have the core reasonably stable within ~3mo
[10:10] <omega_> that includes event system, gobject, and the beginnings of a test suite
[10:10] <omega_> and yes, incsched <g>
[10:11] <taaz> does ridgerun give you deadlines for such things?
[10:12] <omega_> no, but we were discussing some short-term deadlines today
[10:12] <omega_> which means incsched merged by ~monday
[10:12] <omega_> stabilized with autoplug by ~friday
[10:12] <omega_> then the weekend for 0.2.0 release
[10:12] <omega_> a week for gobject
[10:12] <omega_> then /me guesses a week or two for event system
[10:13] <omega_> then we get into a massive code review, documentation, and testing binge
[10:13] <Uraeus> omega_: what about seeking?
[10:13] <omega_> part of the event system
[10:13] <Uraeus> ok
[10:13] <omega_> that's probably week 2 of event system <g>
[10:13] <Uraeus> omega_: did seth from Eazel mail about the issues he was worried about?
[10:13] <omega_> not yet
[10:14] Action: omega_ suggests Uraeus actually subscribe to -devel
[10:14] <Uraeus> guess I should, archive browsing do give me some discussion latency :)
[10:14] <omega_> quite
[10:15] <omega_> he did get things compiling though, after he figure out automake was dying due to lack of memory ;-|
[10:15] <Uraeus> ok, a new subscriber has just joined the list <g>
[10:16] <Uraeus> but I still think we should move of the sucky sourceforge mail system
[10:16] <omega_> or just subscribe an archiver to the list <g>
[10:17] <Uraeus> I mean Gstreamer wouldn't be the first non-pure GNOME project on mail.gnome.org gtk, gimp, fligh-plan etc. is already there
[10:17] <Uraeus> or libxml for that matter
[10:18] <omega_> well, I don't want to move over entirely because there are many sourceforge features we will be using that gnome just doesn't have
[10:18] <Uraeus> no just the mailing list
[10:18] <Uraeus> (s)
[10:18] <Uraeus> hosting the site on gnome.org would suck
[10:18] <omega_> heh
[10:19] <omega_> ergh, /me must get some sleep
[10:19] omega_sonic (omega at omegacs.net) left irc: Read error to omega_sonic[omegacs.net]: EOF from client
[10:19] <omega_> "I'll be back"
[10:20] Nick change: taaz -> taazzzz
[10:20] omega_ (omega at omegacs.net) left irc: killall -9 omega
[10:55] dichro (dichro at CPE-144-132-243-155.nsw.bigpond.net.au) joined #gstreamer.
[10:56] <dichro> any caps-grokking individuals here?
[10:58] <Uraeus> dichro: to late for omega_ and to early for wtay :)
[10:58] <Uraeus> dichro: wtay is the guy you want really
[10:58] Action: dichro nods
[10:58] <dichro> invariably :) any idea when he'll be up?
[10:59] <Uraeus> dichro: give him 2-3 hours
[11:00] <dichro> thanks :)
[11:00] Action: dichro idles
[11:00] <Uraeus> dichro: what are you working on?
[11:01] <dichro> tiny little things - like connecting up a v4l source to a fakesink or fdsink - that don't work unless caps are negotiated
[11:02] <dichro> v4l source has this nasty habit of only resetting various things away from defaults when caps are negotiated
[11:02] <dichro> meaning that if you have a cap-unaware sink, you get the first 30k of each frame :P
[11:04] <Uraeus> ok, but the v4l source core dumps at the moment doesn't it (seem to remember a message on the devel list in that regard)
[11:05] <dichro> bah, humbug. I haven't looked at this in about two weeks :)
[11:06] <dichro> the other thing I was playing with was doing jpeg frame encodes - but the jpegenc stuff seems to have bitrotted over time
[11:06] <Uraeus> ok
[11:07] <Uraeus> so now your cleaning out the bitrot from jpegenc?
[11:07] <dichro> I had a shot at that, but I didn't understand all the gstreamer infrastructure :(
[11:08] <Uraeus> aha, so now you need wtay for some input :)
[11:08] <dichro> I'll probably try again, but it will be easier to write one from scratch and work out things out as I go then to try to debug several different things that I don't understand :) I just don't know if anybody feels proprietary and jealous about the plugin
[11:09] <dichro> and I need a raw video source, too :)
[12:12] ajbusy (me at p51-max6.dun.ihug.co.nz) left irc: Ping timeout for ajbusy[p51-max6.dun.ihug.co.nz]
[12:12] ajbusy (me at p51-max6.dun.ihug.co.nz) joined #gstreamer.
[12:21] dichro (dichro at CPE-144-132-243-155.nsw.bigpond.net.au) left irc: Leaving
[12:30] matth-away (matth at qwest.dsplinux.net) got netsplit.
[12:30] Uraeus (cschalle at c224s9h5.upc.chello.no) got netsplit.
[12:30] steveb (steveb at node1ee31.a2000.nl) got netsplit.
[12:30] taazzzz (dlehn at 66.37.66.32) got netsplit.
[12:30] iGN_ (ign at login1.simplemente.net) got netsplit.
[12:30] Happyfeet (robert at d150-72-75.home.cgocable.net) got netsplit.
[12:30] Gandalf_ (gandalf at tux.rsn.bth.se) got netsplit.
[12:30] seth (seth at null.Stanford.EDU) got netsplit.
[12:41] Happyfeet (robert at d150-72-75.home.cgocable.net) got lost in the net-split.
[12:41] matth-away (matth at qwest.dsplinux.net) got lost in the net-split.
[12:41] Gandalf_ (gandalf at tux.rsn.bth.se) got lost in the net-split.
[12:41] iGN_ (ign at login1.simplemente.net) got lost in the net-split.
[12:41] taazzzz (dlehn at 66.37.66.32) got lost in the net-split.
[12:41] seth (seth at null.Stanford.EDU) got lost in the net-split.
[12:41] Uraeus (cschalle at c224s9h5.upc.chello.no) got lost in the net-split.
[12:41] steveb (steveb at node1ee31.a2000.nl) got lost in the net-split.
[13:36] Nick change: wtay-sleeping -> wtay
[13:36] <wtay> yaw
[13:36] <wtay> er yawn
[13:38] Nick change: ajbusy -> ajmitch
[13:38] <ajmitch> hi wtay
[13:38] <wtay> hi ajmitch, got mail?
[13:38] <ajmitch> wtay: have got mail
[13:38] <wtay> ajmitch: from the cron job?
[13:39] <ajmitch> wtay: i think evolution hasn't helped the situtation
[13:39] <ajmitch> wtay: umm, the first 5 you sent ;)
[13:39] <wtay> today?
[13:40] <ajmitch> ok, i got todays one, evolution is just not marking it as new for some weird reason
[13:40] <ajmitch> evolution is weird at times ;)
[13:40] <wtay> ajmitch: hmm ok, same thing happens here... it might be the header I use...
[13:41] <wtay> To: ajmitch at ihug.co.nz
[13:41] <wtay> From: wim.taymans at chello.be
[13:41] <wtay> Subject: Daily IRC logs
[13:41] <wtay> X-Evolution: 000001ee-0010
[13:41] <wtay> Mime-Version: 1.0
[13:41] <ajmitch> wtay: what prog you use to read it?
[13:41] <wtay> evolution
[13:41] <ajmitch> k
[13:42] Action: wtay is downloading 2.4.4
[13:43] Action: ajmitch goes to grab patch
[13:43] <ajmitch> i've still got a locked mount process due to 2.4.2 loopback bug ;)
[13:45] seth (seth at null.Stanford.EDU) joined #gstreamer.
[13:45] matth-away (matth at qwest.dsplinux.net) joined #gstreamer.
[13:45] steveb (steveb at node1ee31.a2000.nl) joined #gstreamer.
[13:45] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[13:45] Happyfeet (robert at d150-72-75.home.cgocable.net) joined #gstreamer.
[13:45] Gandalf_ (gandalf at tux.rsn.bth.se) joined #gstreamer.
[13:45] iGN_ (ign at login1.simplemente.net) joined #gstreamer.
[13:45] taazzzz (dlehn at 66.37.66.32) joined #gstreamer.
[13:45] <ajmitch> hmm
[13:51] hadess (hadess at pc121-gui14.cable.ntl.com) joined #gstreamer.
[13:52] <hadess> yo
[13:53] <wtay> hi
[13:53] <ajmitch> hi hadess
[13:54] <hadess> hey wtay, ajmitch
[14:00] <ajmitch> night all ;)
[14:00] Nick change: ajmitch -> ajzzzz
[14:00] <wtay> 'night
[14:01] Action: wtay is going to reboot into 2.4.4
[14:01] wtay (wim at cable-195-162-214-190.upc.chello.be) left irc: Read error to wtay[cable-195-162-214-190.upc.chello.be]: EOF from client
[14:46] _gst_newt_ joined #gstreamer.
[14:48] <steveb> 2.4.4 working?
[14:50] dichro (dichro at CPE-144-132-243-155.nsw.bigpond.net.au) joined #gstreamer.
[14:51] <wtay> steveb: yup, no prob, I had to install the ieee1394 CVS sources though
[14:51] <wtay> hi dichro
[14:53] <dichro> hi wtay :)
[15:04] Topic changed on #gstreamer by ChanServ!s at ChanServ: GStreamer: the ultimate multimedia framework
[15:28] steveb (steveb at node1ee31.a2000.nl) left irc: Ping timeout for steveb[node1ee31.a2000.nl]
[15:34] steveb (steveb at node1ee31.a2000.nl) joined #gstreamer.
[15:40] <dichro> wtay: is there a way of creating catch-all caps (ie, ones that match anything)?
[15:41] <wtay> yup, just add a NULL caps
[15:42] <dichro> as in, gst_pad_set_caps(pad, NULL)?
[15:43] <wtay> yes
[15:45] <dichro> hmm. that didn't seem to do what I was looking for... I'm trying to get raw images out of v4l. It defaults to writing out the first 30k of a frame. That default only seems to change at cap negotiation time, if I understand the code correctly.
[15:46] <dichro> the simple sinks (fakesink and fdsink) don't seem to do cap negotiation, hence the earlier question
[15:46] <wtay> dichro: yup, v4lsrc always expects the peer element to tell the output format
[15:47] <dichro> hmm - so what /should/ be happening in this kind of scenario?
[15:47] <wtay> since fdsink never does that, it fails
[15:47] <wtay> dichro: you should tell v4lsrc what format you want (arg) 
[15:47] <wtay> dichro: not implemented though...
[15:47] <dichro>   gtk_object_set(GTK_OBJECT(source), "format", 7, NULL);
[15:47] <dichro>  that's what I have in there at the moment...
[15:48] <wtay> dichro: ok, that's a first step
[15:49] <wtay> hmm v4lsrc needs some work...
[15:49] <dichro> *grin* so does jpegenc :)
[15:50] <wtay> yes :)
[15:50] Action: dichro started trying to fix it but got horribly bogged down. Would a rewrite be rude?
[15:50] alexk3 (alexk3 at 195.96.174.33) joined #gstreamer.
[15:50] <wtay> dichro: not at all, I'm not emotionally attached to anything I write
[15:51] <wtay> yo alexk3
[15:51] <wtay> dichro: which is good and bad at the same time <g>
[15:52] <alexk3> hi to all
[15:52] <dichro> wtay: cool :) I'd love to fix it as it stands, but it scares me. I think I'm much more likely to achieve something useful by aiming at something simpler initially.
[15:55] <wtay> dichro: yup
[15:56] alexk3 (alexk3 at 195.96.174.33) left irc: [x]chat
[16:05] thomas (thomas at adsl-63648.turboline.skynet.be) joined #gstreamer.
[16:05] <thomas> hi all
[16:07] <wtay> hi
[16:07] Action: wtay is running flask in wine (transcoding a vob)
[16:07] <thomas> wtay: I just talked to a friend of mine who sells computers
[16:07] <thomas> wtay: he's going to lend me equipment to make a box like the gbox
[16:07] <thomas> should be cool to work on
[16:07] <wtay> cool
[16:08] <thomas> I think I'll have some catching up to do re: video
[16:08] <thomas> for example is it easy to overlay some text over an image ?
[16:08] <wtay> define easy
[16:08] <thomas> if you have the image map as a gif for example ?
[16:09] <wtay> if you write the code :)
[16:09] <thomas> if you get MJPEG in... how is it treated internally ?
[16:09] <wtay> I'm not sure there are sufficiently capable libraries yet
[16:09] <thomas> for example take a G400-TV card
[16:09] <thomas> but, since a G400 displays your current computer screen if you want...
[16:09] <wtay> MJPEG usually is encapsulated in AVI
[16:09] <thomas> ... you can overlay some video in a window, right ?
[16:10] <wtay> probably
[16:10] <thomas> so when you're watching TV the box can let you know you've got mail or there's someone at the door
[16:11] <wtay> You normally have a handle to the YUV overlay so I would guess so
[16:11] <thomas> ok... that'll be for next weekend.
[16:11] <thomas> btw that aalib plugin is really really cool.  If you stand away from your screen it works pretty well...
[16:11] <thomas> ... but it's hard to find workable MPEG videos.  How can you tell them apart ?
[16:12] <wtay> were are you looking for them?
[16:12] <thomas> just on the net... I had a metal gear preview, an anime clip and some other stuff as well
[16:12] <thomas> but most of them just gave errors
[16:12] <thomas> I couldn't even get aliensong to work
[16:13] <wtay> oh
[16:13] <wtay> does your X server has Xv images?
[16:13] <thomas> I think so.  How can I be sure ?
[16:13] <wtay> xdpyinfo
[16:14] <thomas> ok, what do I look for ?
[16:15] <wtay> in the extensions look for XVideo
[16:15] <thomas> that's there
[16:15] <wtay> ok
[16:15] <wtay> do you have Hermes?
[16:15] <thomas> probably not
[16:15] <thomas> since I removed it from the makefiles
[16:15] <thomas> what is it ?
[16:15] <wtay> a color conversion library
[16:16] <thomas> ok... is it RPM or compile only ?
[16:16] <thomas> and do I need it ?
[16:16] <wtay> gstplay uses it
[16:16] <thomas> had some other issues as well with that.
[16:16] <wtay> so it you want to play an mpeg with gstmediaplay wou'll need it
[16:16] <thomas> I reinstalled glade and glade-devel
[16:16] <thomas> but it's not finding some functions for gstplay
[16:17] <wtay> hmm I have that problem now too
[16:18] <thomas> is it glade's fault or not ?
[16:19] <wtay> dunno yet
[16:19] <wtay> will look at it soon,
[16:28] dichro (dichro at CPE-144-132-243-155.nsw.bigpond.net.au) left irc: dichro
[16:32] steveb (steveb at node1ee31.a2000.nl) left irc: [x]chat
[16:38] steveb (steveb at node1ee31.a2000.nl) joined #gstreamer.
[16:40] <wtay> thomas: what glade symbols are not found on your system?
[16:45] <thomas> sorry wtay, I was working on a site.
[16:45] <thomas> hang on, let's check
[16:45] <thomas> .libs/libgstmediaplay.so: undefined reference to `glade_xml_get_widget'
[16:45] <thomas> .libs/libgstmediaplay.so: undefined reference to `glade_xml_signal_autoconnect_full'
[16:45] <thomas> .libs/libgstmediaplay.so: undefined reference to `glade_xml_new'
[16:45] <thomas> but also /home/thomas/gst/CVS/gstplay/main.c:17: undefined reference to `gnome_init'
[16:45] <thomas> the last one seems like it's my fault
[16:46] <thomas> hmmm....
[16:46] <thomas> maybe I shouldn't have upgraded glade without removing
[16:47] <wtay> hmm strange
[16:50] Aquariophile (root at M148P010.dipool.highway.telekom.at) joined #gstreamer.
[16:50] <Aquariophile> Hello !
[16:50] <thomas> hi
[16:50] <thomas> wtay: I'll retry autogen.sh, maybe I haven't got glade-devel support configured yet
[16:50] <Aquariophile> Anyone here who want to tell me why gstreamer does not let me install it ?
[16:51] <Aquariophile> ./configure seems to work
[16:51] <Aquariophile> make too
[16:51] <Aquariophile> make install not
[16:51] <Aquariophile> error's .... 
[16:51] <Aquariophile> Now I tryed the *.rpm
[16:51] <Aquariophile> does no work too
[16:51] <wtay> Aquariophile: you shouldn't install it for now
[16:51] <thomas> Aquariophile: do a CVS download
[16:51] <wtay> why does make install fail?
[16:51] <thomas> rpm's are old
[16:52] <Aquariophile> make install told me about some errors....
[16:52] <Aquariophile> No idea what exact
[16:52] <thomas> wtay: is a new release scheduled yet ?
[16:52] <Aquariophile> Should I make it GAIN ?
[16:52] <wtay> Aquariophile: can you paste the error here?
[16:52] <Aquariophile> THEN I CANN SHOW YOU ALL
[16:52] <Aquariophile> Sorry
[16:52] Last message repeated 1 time(s).
[16:52] <Aquariophile> My shift
[16:52] <Aquariophile> sorry
[16:52] <thomas> wtay: wtf ? why does automake use 60% mem on a 128 meg system ?
[16:53] <thomas> what does it do that takes so much mem ?
[16:53] <wtay> thomas: a bug
[16:54] <Aquariophile> it makes "make install" now
[16:54] <Aquariophile> moment please
[16:57] <Aquariophile> editor.o: In function `main':
[16:57] <Aquariophile> editor.o(.text+0x34): undefined reference to `gnome_init'
[16:57] <Aquariophile> collect2: ld returned 1 exit status
[16:57] <Aquariophile> make[2]: *** [gsteditor] Error 1
[16:57] <Aquariophile> make[2]: Leaving directory `/root/gstreamer-0.1.1/editor'
[16:57] <Aquariophile> make[1]: *** [install-recursive] Error 1
[16:57] <Aquariophile> make[1]: Leaving directory `/root/gstreamer-0.1.1/editor'
[16:57] <Aquariophile> make: *** [install-recursive] Error 1
[16:57] <Aquariophile> bash-2.04# 
[16:57] <Aquariophile> thomas, wtay ^^^
[16:58] <wtay> Aquariophile: hmm, do you have ximian gnome installed?
[16:58] <Aquariophile> no idea
[16:58] <Aquariophile> I habve SuSe Linux 7.0
[16:58] <Aquariophile> everything was already installed
[16:58] <wtay> Aquariophile: ok
[16:58] <Aquariophile> I am really not good in Linux
[16:59] <Aquariophile> Can you please help me ?
[16:59] <wtay> I'm trying to find out what's wrong...
[16:59] <Aquariophile> thank you
[16:59] <wtay> thomas: you also have the gnome_init problem..
[16:59] <wtay> thomas: any idea what's wrong?
[17:00] <thomas> sec, still doing autogen
[17:00] <thomas> is autogen a standard gnu tool ?
[17:00] <thomas> or is something I'm supposed to fix
[17:00] <thomas> if I feel the need
[17:00] <wtay> standard GNU tool
[17:00] <thomas> ok, clean compile now...
[17:01] <wtay> err, no actually not, something we added I think
[17:01] <thomas> trying it
[17:01] <thomas> if it's something you guys did I could check it if you want
[17:01] <thomas> I don(t think RMS makes standard tools that eat half your ram ;)
[17:01] <wtay> Aquariophile: can you mail me the config.log file you have?
[17:02] <Aquariophile> yes....
[17:02] <Aquariophile> eventually per DCC ?
[17:02] <wtay> thomas: there is a bug in automake I think, richardb has a fix for that
[17:02] <wtay> Aquariophile: use e-mail, wim.taymans at chello.be
[17:02] <Aquariophile> belgium ?
[17:02] <Aquariophile> cool
[17:03] <wtay> yup
[17:03] <Aquariophile> WhatÄ's your native language ?
[17:03] <wtay> dutch
[17:03] <thomas> wtay: gstplay crashes on everest.m1v and aliensong
[17:03] <wtay> thomas: error message?
[17:04] <thomas> none to mention... some assertions failed at start up, but after selecting the clip, it just gives a gnome error box
[17:04] <thomas> "crashed due to a fatal error"
[17:04] <thomas> I'll check the core if there is one
[17:04] <Aquariophile> mail sent 
[17:04] <thomas> no core
[17:04] <Aquariophile> i think it should be available on your server in 30 seconds
[17:07] <Aquariophile> wtay ?
[17:08] <wtay> sorry was distracted
[17:08] <wtay> not yet...
[17:09] <Aquariophile> I do not know the word distracted
[17:09] <Aquariophile> what does it mean
[17:09] <Aquariophile> ?
[17:10] <thomas> it means you've been doing something else
[17:10] <wtay> trying to figure out how to "upgrade" redhat to debian...
[17:10] <Aquariophile> thomas: thank you
[17:11] <Aquariophile> wtay: This could really work ?
[17:11] <Aquariophile> (this upgrade)
[17:12] <wtay> I hope so, I have a hosed redhat PC that seriously needs an upgrade.. but I don't want to use floppies and I don't want to wipe out the existing data...
[17:13] <Aquariophile> it this works, I wont have Suse no longer :-)
[17:13] <wtay> don't like suse?
[17:15] Topic changed on #gstreamer by ChanServ!s at ChanServ: GStreamer: the ultimate multimedia framework
[17:16] <Aquariophile> wtay: no
[17:16] <Aquariophile> nearly every installation won't work
[17:16] <Aquariophile> I think it's a suse related problem
[17:17] <wtay> can you run gnome-config --version
[17:23] <Aquariophile> me ?
[17:24] <Aquariophile> bash-2.04# gnome-config --version
[17:24] <Aquariophile> gnome-libs 1.2.4
[17:26] <wtay> hmm
[17:26] <wtay> Aquariophile: I still didn't receive your mail...
[17:26] <Aquariophile> not ??
[17:26] <Aquariophile> I sent it
[17:26] <Aquariophile> sure
[17:26] <Aquariophile> better idea
[17:26] <Aquariophile> moment
[17:27] <Aquariophile> http://www.salonestetica.at/config.log
[17:28] <wtay> ok thx
[17:28] <Aquariophile> eventually there is a second "make install" visible because i had to run "make install" again to tell you the error 
[17:29] <wtay> looks good..
[17:30] <Aquariophile> but if it does not work, it is not good
[17:30] <wtay> strange
[17:30] <wtay> is gstplay/gstmediaplay built?
[17:30] <Aquariophile> where ?
[17:31] <Aquariophile> under /root/?
[17:31] <wtay>  /root/gstreamer-0.1.1/gstplay/gstmediaplay
[17:33] <Aquariophile> bash-2.04# cd /root/gstreamer-0.1.1/gstplay/gstmediaplay
[17:33] <Aquariophile> bash: cd: /root/gstreamer-0.1.1/gstplay/gstmediaplay: Not a directory
[17:33] <Aquariophile> bash-2.04# 
[17:33] <wtay> it is supposed to be a file
[17:34] hadess (hadess at pc121-gui14.cable.ntl.com) left irc: Read error to hadess[pc121-gui14.cable.ntl.com]: EOF from client
[17:36] <Aquariophile> ok, moment
[17:43] Nick change: taazzzz -> taaz
[18:01] hadess (hadess at pc121-gui14.cable.ntl.com) joined #gstreamer.
[18:12] hadess (hadess at pc121-gui14.cable.ntl.com) left irc: brb
[18:17] hadess (hadess at pc121-gui14.cable.ntl.com) joined #gstreamer.
[18:32] Nick change: wtay -> wtay-away
[18:33] hadess (hadess at pc121-gui14.cable.ntl.com) left irc: brb
[18:50] Nick change: taaz -> taaz-away
[19:11] omega_ (omega at omegacs.net) joined #gstreamer.
[19:13] <steveb> morning
[19:15] <omega_> yo
[19:16] <steveb> omega_: what is the policy of dishing out cvs access?
[19:17] <steveb> s/of/on/
[19:17] <omega_> pretty much anyone who has useful code and is getting on one of the existing maintainer's nerves with repeated changes <g>
[19:17] <steveb> well, wtay hasn't complained yet
[19:18] <omega_> heh
[19:18] <omega_> you're working on the various audio plugins?
[19:19] <steveb> yeah - i have some stuff on float props which I haven't submitted yet
[19:19] <omega_> code or comments?
[19:19] <steveb> code
[19:19] <omega_> cool
[19:19] <omega_> patches to the gst/* code or the plugins only?
[19:20] <steveb> gst/* and tools/*-inspect.c
[19:20] <steveb> oh and my favourite - gst/elements/gstsinesrc
[19:20] <omega_> heh
[19:20] <omega_> well, then.. what's your sf user ide?
[19:20] <omega_> er, id
[19:21] <steveb> try stevebaker_org
[19:22] <omega_> added
[19:23] <steveb> cool, thanks
[19:24] <steveb> now that I have unlimited power to screw things up, do you have any social restrictions you would care to mention
[19:24] Nick change: wtay-away -> wtay
[19:24] <wtay> yo
[19:24] <omega_> yo
[19:24] <omega_> try mpeg2parse3 ?
[19:25] <wtay> omega_: yup, works perfect
[19:25] <omega_> ok, that's the way autoplug should work, for the most part
[19:25] <wtay> can a thread be inside another thread?
[19:25] <omega_> if you want, sure
[19:25] <wtay> autoplug does this 
[19:25] <omega_> I expect it might show a bug or two, but nothing that isn't a bug <g.
[19:25] <wtay> steveb: now that you have cvs access you can apply the patch yourself :)
[19:26] <steveb> wtay: ok
[19:26] <wtay> omega_: not sure what goes wrong yet... I have been kinda distracted... need to upgrade one of my redhat boxes to debian
[19:27] <omega_> hmm
[19:28] <wtay> we should try mpeg2parse4 with 5 threads to be sure...
[19:28] <omega_> oooh, imagine 'splat' playing windoze games via wine with transgaming's direct3d work <G>
[19:28] <wtay> autoplug works when there are two threads so...
[19:28] <omega_> ok, I'll try it
[19:29] <wtay> I just played with wine today.. trying to get some of the multimedia apps working. it failed.
[19:29] <omega_> heh
[19:30] Action: wtay is copying 12GB over the lan.
[19:30] <omega_> ouch, better be 100mbps
[19:30] <wtay> 10mbps :(
[19:30] <steveb> hmm, cvs say "could not chdir to home directory (/home/users/s/st/stevebaker_org) no such file or directory"
[19:30] <omega_> ouch
[19:30] <wtay> uh
[19:30] <omega_> oops
[19:31] <steveb> do you have a shell on that box?
[19:31] <omega_> there's nothing even close to that name on there
[19:31] <omega_> steveadept
[19:31] <omega_> stevecassidy
[19:31] <steveb> how bout sbaker3?
[19:31] <omega_> nope
[19:31] <wtay> heh
[19:31] <wtay> name space issues again :)
[19:32] <omega_> yup
[19:32] <steveb> i created stevebaker_org recently in the vain hope that all of the setup account stuff would work
[19:32] <omega_> hah
[19:32] <omega_> wtay: looks like you're patchman for a while longer <g>
[19:33] <wtay> sbaker3 and sbaker are there...
[19:33] Aquariophile (root at M148P010.dipool.highway.telekom.at) left irc: [x]chat
[19:33] <wtay> steveb is one of those userids yours?
[19:33] <omega_> oh, where?
[19:33] <omega_> oh, they are
[19:33] <steveb> try sbaker3
[19:33] Action: omega_ forgot the * somehow
[19:33] <omega_> sbaker3/src/glamir
[19:34] <omega_> steveb: remove stevebaker_org and add sbaker3?
[19:34] <wtay> steveb: when did you register?
[19:34] <steveb> yeah - abortive project #86364
[19:34] <omega_> heh
[19:34] <steveb> omega_: yes
[19:35] <omega_> added
[19:35] <omega_> actually it's aborted project #19038
[19:35] <steveb> success!
[19:36] <steveb> cheers
[19:39] <wtay> hmm I wonder how we suddenly got second on the multimedia sf category... :(
[19:40] <omega_> what multimedia category?
[19:40] <wtay> http://sourceforge.net/softwaremap/trove_list.php?form_cat=99
[19:41] <wtay> SDLmm..
[19:41] <wtay> oh well..
[19:41] <omega_> by not committing a bunch recently
[19:42] <omega_> sdlmm just got added anyway, they're doing all their "get it into cvs" stuff
[19:42] <wtay> yeah
[19:42] <wtay> our ranking still is the same so..
[19:42] <wtay> give or take 2
[19:43] <omega_> heh, codecs.org is #10 even
[19:43] <omega_> that doesn't bode well for projects #11 and higher
[19:48] hadess (hadess at pc121-gui14.cable.ntl.com) joined #gstreamer.
[19:49] <hadess> yo
[19:49] <omega_> yo
[19:51] <wtay> hi Ha
[19:51] <wtay> hi hadess
[19:51] <wtay> doh
[19:57] <hadess> heh
[20:01] <omega_> I still haven't figured out why the gtk-warnings that say "invalid cast" don't give you file and line number
[20:02] <wtay> omega_: are they supposed to do that?
[20:02] <omega_> ok, 5-thread version works, but spits lots of NULL cothreads
[20:03] <omega_> wtay: you'd think
[20:03] <wtay> could be usefull
[20:03] <omega_> just a bit ,yeah
[20:04] <wtay> they are not macros so its not easy to create them with a line number.
[20:04] <omega_> INFO isn't a macro either, in that sense
[20:04] <omega_> that didn't stop me
[20:06] <wtay> the invalid cast at EOS is something strange actually..
[20:06] <omega_> invalid casts usually are strange <g>
[20:06] <wtay> especially when the pointer is NULL :)
[20:07] <wtay> I think it has to do with the fact that the eos method is called on a ghost pad.
[20:08] <omega_> could be
[20:08] Action: omega_ is reminded again that the editor needs significant work
[20:12] <omega_> wtay: remember that all the debug output in white on red indicates a category with no color assignment yet
[20:12] <omega_> think of what colors you want
[20:12] <wtay> ok
[20:16] Action: wtay is having trouble picking colors :-)
[20:16] <omega_> note that with tricks of bold and reverse, you can use pastel background colors
[20:17] <wtay> hmm
[20:17] <omega_> compile reverse.c (in mail) to see example
[20:17] <wtay> is incsched the place to adjust the colors or HEAD
[20:17] <omega_> incsched for now
[20:17] <wtay> ok
[20:18] <omega_> and if you commit color changes, add them to the wiki page too
[20:19] <omega_> and if you see any black on grey, that means the debug's category is zero, and it should be fixed
[20:19] Action: wtay is just going to pick the next available color
[20:19] <omega_> um, have some plan at least
[20:20] <wtay> is there a plan now?
[20:20] <omega_> afaik, yellow is the only available color
[20:20] <omega_> there is a very minimal plan, yes
[20:20] <omega_> there are several categories that have the same color
[20:20] <omega_> dataflow is green
[20:20] <omega_> plugins are cyan
[20:21] <omega_> autoplug is blue
[20:21] <wtay> caps and negotiation should hava a similar color then..
[20:21] <omega_> yes
[20:21] <omega_> probably to autoplug..
[20:21] <omega_> so blue
[20:21] <wtay> let's do negotiate in underline
[20:22] <omega_> hmm, ok, try it out before you buy it
[20:24] <omega_> mu, odd incsched problems
[20:25] <wtay> ugh, blue is pretty unreadable here..
[20:25] <omega_> oh, one other thing:
[20:25] <omega_> it's set up to leave libgst.la output normal, and plugin output bold
[20:25] <omega_> is this useful?
[20:26] <wtay> yes
[20:26] <omega_> ok, then we can't use bold as an attrib in any of the category colors
[20:27] <wtay> hmm
[20:27] <wtay> how is a plugin supposed to DEBUG something? whith what category I mean
[20:27] <omega_> that's a good question
[20:29] <wtay> there is GST_CAT_PLUGIN_* but that are only 3 cats
[20:29] <omega_> those are for the plugin system anyway, not the plugins themselves
[20:29] <wtay> oh
[20:30] <wtay> then we need another GST_DEBUG_* macros that boldifies everything
[20:30] <omega_> nope
[20:30] <omega_> that happens automatically if you're compiling outside of libgst.la
[20:30] <wtay> oh cool
[20:30] <wtay> good
[20:30] <omega_> see gstinfo.h:39
[20:31] Action: wtay should look at the code first
[20:34] <wtay> so.. caps is underscore blue, nego is reverse blue
[20:34] <omega_> blue on ?
[20:35] <wtay> background color
[20:35] <omega_> ok
[20:35] <omega_> go ahead and commit when you're ready
[20:35] <wtay> props are yellow
[20:35] <wtay> hm
[20:35] <wtay> they should be blue on yellow
[20:36] <omega_> heh
[20:36] <omega_> don't make it unreadable....
[20:40] <wtay> pff yellow is more like orange here
[20:40] <omega_> heh
[20:43] <omega_> oh, and run any gst command with --help
[20:43] <wtay> uh oh
[20:43] <wtay> doh
[20:43] <omega_> ?
[20:45] <wtay> I feel like we're in the '70 again <g>
[20:46] <omega_> psychedelic, baby!
[20:46] <omega_> no all we need are those "surround lighting" things from /. yesterday, and we'll be set!
[20:46] <omega_> er, s/no/now/
[20:46] <wtay> :-)
[20:47] <wtay> topic: "gstreamer: stresses your terminal. too.."
[20:47] <omega_> heh
[20:53] Action: wtay is preparing himself to go play snooker
[20:53] <omega_> uh oh
[20:53] <omega_> commit before you do...
[20:54] <omega_> ah, nevermind <g>
[20:54] <steveb> hmm, is cvs responding for you guys?
[20:55] <omega_> I just updated and shot myself in the head at the same time
[20:55] <omega_> (-A on a branch)
[20:57] <wtay> don't do that
[20:57] <wtay> steveb: yes, crawling as usual though
[20:58] <steveb> i've just started getting "connection refused"
[20:59] <omega_> I haven't had any problems
[20:59] <omega_> beyond my own stupidity, that is
[21:01] <wtay> ok, I'm away now, cya later
[21:01] Nick change: wtay -> wtay-snooker
[21:01] Nick change: steveb -> steveb-dinner
[21:15] <thomas> omega_: gstmediaplay crashes as soon as I open a file.
[21:15] <thomas> omega_: is it only mine or ... ?
[21:15] <omega_> incsched or head?
[21:15] <thomas> head
[21:15] <omega_> doesn't surprise me too much
[21:15] <thomas> works better in incsched ?
[21:15] <omega_> not yet ;-|
[21:16] <thomas> how can I download incsched ?
[21:16] <thomas> or is that a stupid question ?
[21:16] <omega_> cvs -d ..... co gstreamer -r BRANCH-INCSCHED1
[21:16] <omega_> but..
[21:16] <omega_> you'll also want a -d gstreamer.incsched1 at the end of that
[21:16] <omega_> or it'll mangle your existing working copy
[21:17] <thomas> can't I just do the first one in a new directory ?
[21:17] <omega_> yeah
[21:17] <thomas> btw I went to my friend at the hardware store and told him about the gbox idea
[21:17] <thomas> he really liked it and he's going to lend me the hardware I need
[21:17] <thomas> as long as I make him one
[21:17] <thomas> so maybe you should try too
[21:18] <omega_> hehehe
[21:18] <omega_> try where?
[21:18] <thomas> whatever small computer store is in your neighbourhood
[21:18] Action: omega_ doesn't know of a good one
[21:18] <thomas> I just said "the G400-TV works really well" and he included it
[21:18] <omega_> heh, do you have the hardware now?
[21:18] <thomas> not yet
[21:18] <omega_> eta?
[21:18] <thomas> next weekend
[21:18] <omega_> cool
[21:18] <thomas> he said wednesday but ...
[21:19] <thomas> anyway, I was only going for "I want to pay you for it if you let me bring back stuff and exchange it if doesn't do what I want"
[21:19] <thomas> but he liked it so much he wants to do it as a test
[21:19] <thomas> should be a fun weekend project
[21:19] <omega_> heh
[21:19] <omega_> we need a few pieces first, like a source for that mjpeg stuff on that card
[21:20] <thomas> meaning ? source code ? there's no good driver yet ?
[21:20] <omega_> there's a driver, but I dunno if it's v4l2 or something custom
[21:21] <thomas> should check that
[21:21] <omega_> mjpeg.sourceforge.net
[21:21] <thomas> I haven't done a lot of video on linux
[21:21] <omega_> BBB's project
[21:25] Action: omega_ should write a v4l2 source
[21:25] <omega_> since /me had a fair bit to do with the v4l2 API
[21:27] <omega_> but it seems that the v4l2 page is long since dead (no host)
[21:28] <omega_> no, it just moved
[21:29] <thomas> god... CVS is fickly about it's arguments...
[21:29] <thomas> I finally got the right command
[21:29] <omega_> heh
[21:35] <thomas> marvel.sourceforge.net seems to have a g400 driver for v4l
[21:35] <omega_> yup
[21:35] <omega_> though I'd rather use v4l2 for that kind of stuff
[21:36] <thomas> they seem to have that as well
[21:36] <omega_> but if the v4l2 driver for the marvel is not maintained (as the marvel page claims) then we may not have much choice
[21:36] Action: omega_ things that v4l sucks, and v4l2 is done right
[21:36] <omega_> er, thinks
[21:36] <thomas> have no idea yet... will let you know ;)
[21:36] <omega_> that's because I actually tried to use v4l, then helped create v4l2
[21:37] <thomas> heh
[21:37] <thomas> I hope the G450 is the same as the G400 chip-wise
[21:37] <omega_> why?
[21:37] <omega_> you're getting a g450-tv?
[21:37] <thomas> yes
[21:37] <omega_> ooooh
[21:37] <thomas> I should hope so  ;)
[21:37] <thomas> what does it do better ?
[21:37] <omega_> eTV ?
[21:38] <thomas> I saw that one on the site but I'm not sure if that's goin to be the one
[21:38] <thomas> it's listed as a marvel G450 TVIN TVOUT
[21:38] <omega_> it looks like the eTV does mpeg2, not mjpeg
[21:38] <omega_> which is better, but very likely a problem driver-wise
[21:38] <thomas> hmmm...
[21:38] <thomas> maybe I should ask them to change it then
[21:39] <omega_> if you want to get going fast, yeah
[21:39] <thomas> the first thing I want to get working is overlaying messages over the tv signal
[21:40] <omega_> yeah, the g450etv does mpeg2 capture and playback, which is better than mjpeg any da
[21:40] <omega_> any day... but probably isn't supported yet
[21:42] <omega_> it's got dual-head, but only vga + tv
[21:42] <omega_> but that's sane for this kind of device
[21:42] <thomas> I probably wouldn't really use the vga out anyway
[21:42] <thomas> except maybe if the emulators look better on it
[21:43] <omega_> true, my test machine will have a 1024x768 cheapo monitor on it
[21:44] <omega_> looks like the g450etv is DOA at this point for this purpose
[21:45] <thomas> DOA ?
[21:46] <omega_> dead on arrival
[21:46] <omega_> no docs
[21:46] <thomas> the guy in the store set the G400 is hard to get these days
[21:46] <thomas> s/set/said/
[21:46] <omega_> for an app that talks to the v4l marvel: http://zapping.sourceforge.net
[21:46] <omega_> pricewatch.com lists a few sources at $260
[21:46] <omega_> ebay had a couple at ~$200
[21:47] <omega_> ah. and the g450etv doesn't have a tuner
[21:47] <thomas> zapping looks nice
[21:47] <thomas> what ? no tuner ? ok, dumped.
[21:48] <thomas> why did they drop that ?
[21:48] <thomas> hang on, it does...
[21:48] <thomas> according to a zdnet review
[21:48] <thomas> http://www.zdnet.com/products/stories/specs/0,8828,451301,00.html
[21:48] <omega_> hmm, this ebay auction claims it doesn't
[21:49] <omega_> I just bid on a G400-TV
[21:50] <thomas> how much ?
[21:50] <omega_> starts at $120, I bid at $200 max
[21:50] <thomas> hmm... I should probably just ask for a G400-TV for now
[21:51] <thomas> and exchange it later on
[21:51] <omega_> closes in 5.5 days though
[21:51] <thomas> I want to get it *working* next weekend
[21:51] <omega_> um, you'd better define *what* you want working next weekend.....
[21:51] <thomas> ;)
[21:51] <thomas> any app that lets me watch tv and record something
[21:51] <omega_> good luck
[21:51] <thomas> while I think how I'll manage to do the samme thing
[21:51] <thomas> through a web interface
[21:52] <omega_> there are a lot of pieces needed for that
[21:52] <thomas> why ? that difficult ?
[21:52] <thomas> yeah well... I didn't expect it to be easy.
[21:52] <omega_> i.e. not a weekend thing
[21:52] <thomas> If it was it would be in the store already and they wouldn't have let me make it
[21:52] <thomas> well, there are more weekends in a year
[21:52] <omega_> yup
[21:52] <thomas> it should still be a challenge right
[21:52] <thomas> what's the data transfer rate for full screen mjpeg ?
[21:53] Nick change: ajzzzz -> ajmitch
[21:53] <omega_> depends on the compression ratio
[21:53] <thomas> general usage ?
[21:53] <thomas> like 128 kbit mp3
[21:53] <thomas> that ballpark
[21:53] <omega_> there is no such generalization
[21:53] <thomas> nothing that is used the most ?
[21:54] <omega_> raw video at 640x480 is 17.6MB/sec
[21:54] <omega_> so you can compress to 3:1 or 5:1 and get decent quality
[21:54] <omega_> most people compare DV at 5:1 to MJPEG at 3:1, quality-wise, so.....
[21:54] <omega_> I get 13.5MB/sec peak sustained off my laptop disk
[21:55] <omega_> which means I can count on about a third of that in real life
[21:55] <thomas> ok, the MB i suppose is in megabit, not megabyte ?
[21:55] <thomas> just to make sure
[21:55] <omega_> this is a major issue to solve: a general-purpose OS's disk subsystem isn't designed for throughput
[21:55] <omega_> megabyte
[21:55] <omega_> b is bit, B is byte
[21:56] <thomas> hmm... better rethink the project then ;)
[21:56] <thomas> damn that's a lot
[21:56] <omega_> it's still doable, but you have to be aware of these problems <g>
[21:57] <thomas> so that basically means about a gigabyte for each four minutes
[21:57] <thomas> and then compress all night to something like MPEG4 or divx
[21:57] <omega_> the key is either raw device access with a user-space filesystem, or a kernel filesystem that works very well
[21:57] <omega_> yes
[21:57] <omega_> this is why the 450etv would be nice, because it captures mpeg-2 in hardware, which is much smaller
[21:57] <omega_> on the order of 20:1 compression
[21:57] <omega_> <1MB/sec
[21:57] <thomas> that sounds a lot better
[21:58] <thomas> that's about double of a good divx
[21:58] <omega_> yup
[21:58] <omega_> now, one bizarre option would be to come up with algorithm to difference DCT freq blocks
[21:58] <thomas> i suppose matrox only makes X-windows drivers ?
[21:58] <omega_> then you could convert mjpeg into mpeg with zero motion compensation
[21:59] <omega_> how do you mean?
[21:59] <thomas> well, they don't do any v4l stuff I suppose ?
[21:59] <thomas> what do you mean by "difference DCT freq blocks ?"
[22:00] <omega_> an algorithm that takes two sets of dct coeffs, and somehow generates a third that is the difference of the two
[22:00] <omega_> this is tricky because we want the time-domain difference, not freq-domain difference
[22:00] <omega_> so c = diff(a,b)
[22:00] <omega_> where idct(a) + idct(c) = idct(b)
[22:01] <thomas> hmmm... should be doable if you know about fourier transforms
[22:01] <omega_> um, that's the problem
[22:01] <thomas> meaning ?
[22:01] <omega_> ENOCLUE
[22:01] <thomas> ah ok...
[22:01] <thomas> I studied them at university
[22:01] <omega_> ah
[22:01] Action: omega_ nominates thomas
[22:01] <thomas> but it has been some time
[22:01] <thomas> thanks ;)
[22:01] <thomas> but as I said I never did much video
[22:02] <omega_> it's just 8x8 DCT
[22:02] <thomas> though I'd imagine that it wouldn't be that different from the audio
[22:02] <thomas> just that it's two dimensional
[22:02] <omega_> right
[22:02] <thomas> ok so what you're saying is that to get from MJPEG to MPEG
[22:02] <thomas> you have to calculate frames in between the JPEG's ?
[22:03] <thomas> and the JPEG's contain sample data in the frequency domain
[22:03] <omega_> anyway, that would give you light-weight temporal compression, not as good as with motion comp, but can be derived from mjpeg
[22:03] <omega_> partial decode of mjpeg, partial encode of mpeg
[22:03] <omega_> and you still have 100% correct mpeg
[22:03] <thomas> yeah... sounds like my mixing thesis
[22:03] <omega_> but the key is that you can also have a much faster decoder
[22:03] <omega_> ?
[22:04] <thomas> actually, basically all you need...
[22:04] <thomas> is to reverse the dct to get at some sort of subband samples
[22:04] <thomas> average them out inbetween two frames
[22:04] <thomas> and re-dct them
[22:04] <omega_> but the whole point is that if it can be done without the idct/dct, you can save time
[22:04] <thomas> ok, I see what you mean.
[22:05] <thomas> you know, this may seem stupid, but have you tried just averaging the spectral lines ?
[22:05] <omega_> then you're doing a full mjpeg decode, and a full zero-mcomp mpeg encode
[22:05] <omega_> no, I haven't tried any of this
[22:05] <thomas> well, if it is anything like the audio dct's, just averaging should actually work.
[22:05] <omega_> averaging? I want a difference
[22:05] <thomas> or a difference...
[22:05] <omega_> hmm
[22:06] <thomas> does mpeg video also have a two step dct analysis ?
[22:06] <thomas> like mp3 ?
[22:06] <omega_> I don't think so
[22:06] <omega_> I know mpeg video, not audio
[22:06] <omega_> so I can't be sure what you mean
[22:06] <thomas> well, mp3 first does a 32-subband DCT
[22:06] <thomas> equally spaced distance
[22:07] <thomas> then for mp3 does another one to get 500-something spectral lines
[22:07] <omega_> ok, no, video doesn't
[22:07] <omega_> it works only on macroblocks
[22:07] <thomas> so it divides the whole screen into a set of squares
[22:07] <thomas> then does a 8x8 DCT ?
[22:07] <omega_> a macroblock is actually 4 luma and 1 each of the two chroma blocks
[22:07] <omega_> but yeah, it dct's and quantizes those blocks
[22:08] <thomas> ok... does wtay know about fourier transforms ?
[22:08] <omega_> no
[22:08] <omega_> afaik
[22:08] <thomas> or BBB ?
[22:08] <omega_> doubt it, he's only 16 <g>
[22:08] <thomas> heh
[22:09] <thomas> ok, do you know anyone I can talk about it that does this kind of stuff ?
[22:09] <thomas> because I might be wrong but I could be right as well
[22:09] <thomas> though I don't know where to start
[22:09] <omega_> krasic (the other libdv guy) is kinda versed in this stuff
[22:09] <omega_> but only because he had to figure out the 2x4x8 DCT from dv
[22:09] Action: thomas makes mental note
[22:09] <omega_> krasic at acm.org
[22:10] <thomas> but, basically, I could start on trying to "average" two similar JPEG's, right ?
[22:10] <thomas> I mean, calculate the difference
[22:10] <omega_> yeah
[22:10] <thomas> ok, incsched is compiled.
[22:11] <thomas> so now I can try to change the mixer to request pads and add bins on the spot ?
[22:11] <thomas> if I pause the bin ?
[22:11] <omega_> mpeg2parse3 in test/ is the best example at this point
[22:11] <omega_> um, check out mpeg2parse3.c, it'll show you the best way to construct live pipelines
[22:12] <thomas> ok, thanks... I'll try that tomorrow, have to go see a concert
[22:12] <omega_> ok
[22:12] <thomas> hmm... is there a way I can see if the matrox chipset is similar to something that works ?
[22:12] <omega_> the g450etv?
[22:12] <thomas> I suppose not, they wouldn't buy in another chip
[22:12] <thomas> yeah
[22:13] <omega_> only if you can figure out what chips are on it
[22:13] <thomas> sounds like the g450 is better but not supported enouh
[22:13] <omega_> right
[22:13] <thomas> I can't find anything decent anyway
[22:13] <thomas> btw it seems the windows package offers some sort of timeshift program
[22:13] <omega_> that's the idea board for a single-tuner box, though
[22:14] <omega_> cheap (<$200), mpeg2, fast ramdac, tuner, etc.
[22:14] <omega_> and even dualhead so you can have a vga monitor in one room and a bigscreen tv in another
[22:14] <omega_> watch recorded videos on one and live and/or timeshifted tv on the other <g>
[22:15] <ajmitch> omega_: don't tempt me...
[22:15] <thomas> sweet
[22:16] <thomas> and to think there isn't anything on tv worth watching ;)
[22:16] <thomas> or watch tv on the big screen and play old atari vcs games on the monitor
[22:16] <omega_> yeah, though that's pushing for an even more extreme form of convergence
[22:16] <omega_> which I'll work on after this <g>
[22:17] <omega_> that is to push the storage and legacy I/O (the tuner) to a server machine
[22:17] <thomas> yeah well... I'll first see how frustrated I get next weekend
[22:17] <thomas> btw aalib is very cool
[22:17] <omega_> and make the TV smart, but not too smart
[22:17] <omega_> you played a video through it?
[22:17] <thomas> I'm hoping to watch a friends episode in it soon
[22:18] <thomas> yeah
[22:18] <omega_> it doesn't do color though
[22:18] <thomas> just for the fun of it
[22:18] <thomas> I know
[22:18] <omega_> which sucks, and must be fixed
[22:18] <thomas> shouldn't be hard right ?
[22:18] <omega_> not overly
[22:18] <thomas> though I'd think it would be hard to make it look good
[22:18] <thomas> you know, you really only have 2000 color points
[22:18] <thomas> the ascii chars fake pixels well but not color
[22:19] <omega_> yup
[22:19] <thomas> I think you might lose the effect
[22:19] <omega_> well, you assume that the chars are small and plentiful
[22:19] <thomas> does aalib work in higher resolutions ?
[22:19] <omega_> 80x25 won't cut it, but most people who use textmode have higher rez
[22:19] <omega_> 1) create a really big xterm
[22:19] <omega_> 2) play with aasink driver=3
[22:21] <thomas> hmmm... it still opens a window and stops immediately
[22:21] <omega_> with incsched1 or head ?
[22:21] <thomas> head
[22:21] <thomas> i'll try incsched1
[22:21] <omega_> yeah, it works here
[22:21] <omega_> use mpeg2dec
[22:23] <thomas> READY->PLAYING: element "disksrc0" has parent "launch" and sched 0x80729f8
[22:23] <thomas> READY->PLAYING: element "mpeg2dec0" has parent "launch" and sched 0x80729f8
[22:23] <thomas> READY->PLAYING: element "aasink0" has parent "launch" and sched 0x80729f8
[22:23] <thomas> that's where it stops
[22:23] <omega_> oh?
[22:23] <thomas> fyi aasink wasn't in the standard make
[22:23] <omega_> yeah, the configure.in test for it doesn't work all that well
[22:23] <thomas> oh ok maybe I got the wrong mpeg
[22:23] <thomas> everest seems to work
[22:23] <omega_> yeah, you need to use an elementary stream
[22:24] <thomas> hmmm... seems my \n doesn't work when i use stdout
[22:24] <thomas> any simple way to fix that ?
[22:24] <thomas> it just does a line feed, no cr
[22:24] <thomas> meaning no decent output
[22:24] <omega_> eh?
[22:24] <omega_> are you in an xterm?
[22:24] <thomas> ok wait.  i'm in an xterm, but I resized it
[22:24] <thomas> it seems to think it's still 80 wide
[22:25] <thomas> I'll try to start a new one that size
[22:25] <omega_> hmm, create a new one and resize it immediately
[22:25] <thomas> no, still the same.
[22:26] <thomas> weird thing is, if I do normal stuff like man it uses the whole width
[22:26] <omega_> hmm
[22:26] <omega_> ENOCLUE
[22:27] <thomas> oh well... i'll try it tomorrow.
[22:27] <thomas> I've gotta go. thanks for the chat.
[22:27] <omega_> ok, have fun
[22:28] thomas (thomas at adsl-63648.turboline.skynet.be) left irc: Client Exiting
[22:52] <Uraeus> omega_: still around?
[22:54] <omega_> yeah
[22:55] <Uraeus> omega_: mind helping me out with some questions for Jim Gettys about the sound server stuff he was talking about (for my interview with him)?
[22:56] <omega_> hmmm, sure
[22:56] <Uraeus> you could probably formulate more precise questions than me about this
[22:59] <omega_> did you have some in mind?
[23:00] <Uraeus> well, what I basically want is for him to explain the technology he talked about during GUADEC to archieve synching of audio and video, and which Alan refered to in his diary
[23:00] <Uraeus> but I don't know how to formulate the question (and look like I know what I am talking about :)
[23:00] <omega_> o
[23:00] <omega_> er, ok
[23:06] <omega_> um
[23:06] <omega_> I'd really need to see what you have already in order to be of any use
[23:07] <Uraeus> actually nothing related (just a bunch of unrelated X questions) but I try to formulate something first then and get back to you
[23:08] <omega_> ok, that'll help <G>
[23:27] <Uraeus> omega_: hmm, gstreamer CVS doesn't compile for me
[23:27] <Uraeus> gstplugin.c:30:24: gstversion.h: No such file or directory
[23:36] <omega_> you have to run autogen.sh
[23:36] <Uraeus> oh, hmm, usually when you need to do that then there is no configure script 
[23:36] <omega_> or any time configure.in changes
[23:38] <Uraeus> omega_: ok here is a try:
[23:38] <Uraeus> One of the issues discussed during GUADEC was the issue that a stand-alone soundserver makes synchronisation of sound and video impossible. When I spoke to you after the GStreamer presentation you mentioned some technology you had developed for X some years back and which would enable such syncronisation, but if I remember you correctly there was some some pieces still missing in both the kernel and in current sound servers. Could you please 
[23:38] <Uraeus> elaborate on that?
[23:41] <omega_> sounds reasonable
[23:41] <omega_> though it's not 'impossible', just 'really hard'
[23:51] <Uraeus> omega_: gstreamer fails on compilation due to artsdsink.h missing, are we depending on artsd now :)
[23:52] <omega_> no, again a configure.in problem
[23:52] <omega_> just make -k around it for now
[23:52] <omega_> and bug richardb about it <g>
[23:52] <ajmitch> Uraeus: why not depend on artsd? ;)
[23:52] Action: omega_ smacks ajmitch
[23:52] <ajmitch> ouch!
[23:53] <ajmitch> Uraeus: be worried when omega_ starts making it depend on win32 headers as well ;)
[23:53] Action: ajmitch ducks
[23:53] Action: omega_ takes aim....
[23:53] <Uraeus> ajmitch: nah, but I think gstreamer will be renamed esd+ soon :)
[23:53] Action: omega_ falls to the ground twitching
[23:54] <ajmitch> Uraeus: hehe, nice one ;)
[23:57] Action: omega_ goes to help ChiefHighwater move a couple of couches shortly
[23:57] <ajmitch> omega_: twitching subsided?
[23:58] <omega_> barely
[23:59] <ajmitch> oh, so you don't want us to make an esd-compatible soundserver using gstreamer? :(
[00:00] --- Sun Apr 29 2001
[00:00] <Uraeus> you have to allow that I have 5000 lines of Fortran code here already in order to accomplish that
[00:01] <omega_> "600! Clear!" <zap!>
[00:01] <omega_> "800! Clear!" <zap!!>
[00:01] <omega_> "no good!"
[00:01] <Uraeus> hehe
[00:05] Nick change: steveb-dinner -> steveb
[00:07] Action: ajmitch transfers mandrake 8.0 iso to cdwriting machine
[00:11] <Uraeus> ajmitch: no, need for that RH7.1 is out :)
[00:11] <ajmitch> Uraeus: sick sick person ;)
[00:11] <ajmitch> Uraeus: i run debian anyway ;)
[00:12] <Uraeus> ah, you like that hobbyist distro :)
[00:13] <ajmitch> you call RH professional? :)
[00:15] <ajmitch> hmm, iso broken, i might have to get it at the installfest today
[00:16] <Uraeus> of course, RH it is the distribution we Linux fineschmeckers use :)
[00:17] <ajmitch> hah
[00:17] <ajmitch> most people at our LUG use either debian or mandrake ;)
[00:17] <Uraeus> yes, those are popular with Linux wannabees :)
[00:19] <Uraeus> well as I usually say, it is my way or the wrong way :)
[00:40] <seth> Uraeus: you'd be surprised what we long-haired debian-hippies can turn out ;-)
[00:40] <seth> Uraeus: I mean, we do a good job for communists, that is
[00:40] <ajmitch> seth: oh, absolutely
[00:41] <Uraeus> hehe
[00:41] <ajmitch> seth: you have a copy of the Communist manifesto by your bed as well, and a portrait of Lenin on your wall?
[00:42] <seth> ajmitch: And Mao's little red book
[00:42] <seth> ajmitch: of course
[00:42] <seth> ajmitch: Right next to my Debian Maintainer's Guide
[00:42] <ajmitch> seth: what do you maintain?
[00:43] <seth> ajmitch: Nautilus, though not the package maintainer ;-)
[00:43] <ajmitch> seth: hehe
[00:43] <ajmitch> seth: what do you mean, you maintain it but you're not the maintainer? ;)
[00:43] <Uraeus> hmm, I thought you Debian people saw yourself as the staunch defenders of Libertarianism :)
[00:44] <seth> Uraeus: oh wait. you're right.
[00:44] <seth> Quick ajmitch, we need to get our propoganda in line
[00:44] <seth> ajmitch: I hack on Nautilus and GnomeVFS upstream
[00:44] <seth> ajmitch: I'm actually too lazy to go through the Debian maintainer process
[00:44] <Uraeus> cause we RedHat people see outselves as the proud diciples of Margareth Thatcher :)
[00:44] <ajmitch> seth: oh, ok ;)
[00:44] <seth> argh!
[00:44] <ajmitch> the Iron Lady?
[00:45] <Uraeus> the one and only
[00:45] Aquariophile (root at M153P017.dipool.highway.telekom.at) joined #gstreamer.
[00:45] <Aquariophile> Hello
[00:45] <Uraeus> hi
[00:45] <ajmitch> hi
[00:45] <Aquariophile> Anyone who wants to support me ?
[00:45] <Uraeus> depends :)
[00:45] <Aquariophile> gstreamer installing problem
[00:45] <Uraeus> ok, ask away 
[00:46] <Aquariophile> bash-2.04# rpm -ivh gstreamer-0.1.1-1.i386.rpm
[00:46] <ajmitch> seth: you in the UK?
[00:46] <Aquariophile> only packages with major numbers <= 3 are supported by this version of RPM
[00:46] <Aquariophile> error: gstreamer-0.1.1-1.i386.rpm cannot be installed
[00:46] <Aquariophile> bash-2.04# 
[00:47] <seth> ajmitch: not the last time I checked
[00:47] <seth> ajmitch: Maybe its time for me to double check though
[00:47] <Uraeus> ah, you are using a to old version of RMP
[00:47] <Uraeus> err RPM
[00:47] <ajmitch> seth: the evil capitalist US? ;)
[00:47] <Uraeus> Aquariophile: what distro?
[00:47] <Aquariophile> Suse
[00:47] <Aquariophile> and Source does not work too+
[00:47] <ajmitch> Uraeus: gonna bash that one as well? ;)
[00:48] <Uraeus> Aquariophile: try getting the src.rpm and do a rpm -ba , think that will solve it
[00:48] <Uraeus> ajmitch: I only distrobatch people I know :)
[00:48] <Aquariophile> where to get  src.rpm ?
[00:48] <Uraeus> http://gstreamer.net/download.shtml
[00:49] <Uraeus> ajmitch: s/ distrobatch/distrobash/ :)
[00:49] <Aquariophile> gstreamer-0.1.0-1.src.rpm
[00:49] <Aquariophile> this one ?
[00:49] <Uraeus> yes
[00:50] <Uraeus> Aquariophile: you have development libraries installed?
[00:50] <Aquariophile> no ?
[00:50] <ajmitch> Uraeus: so good to know that i'm counted amongst those you know and harass ;)
[00:50] <Aquariophile> no idea
[00:50] <seth> ajmitch: yes, it looks like I'm still in that land of capitalist swine
[00:51] <ajmitch> seth: come join me in the land of the free...
[00:51] <Uraeus> Aquariophile: hmm, hopefully you have, have you compiled anything on the machine before?
[00:51] <Aquariophile> yes
[00:51] <Aquariophile> licq
[00:51] <Aquariophile> and...
[00:51] <Aquariophile> some other things
[00:52] <Aquariophile> opera
[00:52] <Uraeus> ajmitch: yes, it is an exclusive club your in
[00:52] <Aquariophile> gimp
[00:53] <Aquariophile> how to install *.src+rpm ?
[00:54] <Uraeus> Aquariophile: first just do a rpm -Uvh gstreamer*.src.rpm
[00:54] <Uraeus> then go to /usr/src/suse/SPECS
[00:54] <Uraeus> and do a rpm -ba gstreamer.spec
[00:54] <Uraeus> it should then try to make a new RPM for you
[00:54] <Uraeus> if it succeds the new PRM will be in /usr/src/suse/RPMS/i386/
[00:55] <Aquariophile> bash-2.04# rpm -Uvh gstreamer*.src.rpm
[00:55] <Aquariophile> only packages with major numbers <= 3 are supported by this version of RPM
[00:55] <Aquariophile> error: gstreamer-0.1.0-1.src.rpm cannot be installed
[00:55] <Aquariophile> bash-2.04# 
[00:56] <Uraeus> ok, get the gstreamer-0.1.1.tar.gz and do a rpm -ta gstreamer-0.1.1.tar.gz
[00:56] <Uraeus> cause you have a really old version of RPM3, since even the src.rpm fail
[00:56] <Uraeus> would think that Suse have released an update to rpm actaually
[00:57] <Aquariophile> okay
[00:57] <Aquariophile> moment please
[00:58] <ajmitch> Uraeus: when will we get .deb packages? ;)
[00:58] <Uraeus> ajmitch: as soon as you make 'em :)
[00:59] steveb (steveb at node1ee31.a2000.nl) left irc: [x]chat
[00:59] <Aquariophile> what is the best distro ?
[00:59] <Uraeus> hehe, everyone here nows it is RedHat :)
[00:59] <Aquariophile> really ?
[00:59] <Aquariophile> Hm...
[00:59] <ajmitch> Debian, actuallly
[01:00] <Aquariophile> everyone tells me something different.... I here "redhat, mandrake, and most say Debian"
[01:00] <ajmitch> well, anything *but* RH ;)
[01:00] <ajmitch> Mandrake is ok
[01:00] Action: Uraeus looks evily at ajmitch
[01:00] <Uraeus> mandrake is a cheap RH knockoff :)
[01:00] <ajmitch> Uraeus: when was the last time you used it? ;)
[01:00] <ajmitch> RH is a cheap debian knockoff ;)
[01:00] <Uraeus> ajmitch: never, I am just being prejudicial :)
[01:01] <Uraeus> actually I think the 'best' is what you are used to, cause that is the distro you know
[01:02] <Aquariophile>  rpm -ta gstreamer-0.1.1.tar.gz    needs long time
[01:02] <Uraeus> yes
[01:02] <Aquariophile> How long do you think will it take ?
[01:02] <ajmitch> Uraeus: but there are some that just suck, no matter what
[01:02] <ajmitch> Aquariophile: how fast a computer?
[01:02] <Uraeus> ajmitch: yes, like Debian :)
[01:02] <Aquariophile> 64 MB RAM
[01:03] <Aquariophile> atround 300 Mega Herz
[01:03] <Uraeus> ajmitch: strange how all Debian based distros end up going to the dogs (Corel, Stormix etc.)
[01:03] Action: ajmitch drops a few medium sized boulders on Uraeus
[01:03] <ajmitch> Uraeus: they cannot compete with debian itself, it's just so much better ;)
[01:04] <ajmitch> Uraeus: plus it's very easy to switch from Corel or Stormix to debian
[01:04] <ajmitch> Aquariophile: you could be waiting a little while ;)
[01:05] <Aquariophile> as long as you won't go... *g*
[01:05] <Uraeus> ajmitch: yes, switching is easy, the problem is deciding if changing is an upgrade or an downgrade :)
[01:06] <ajmitch> Aquariophile: i'll probably be gone before it finishes ;)
[01:06] <Uraeus> me to, it is 0100 here :)
[01:06] <Uraeus> Aquariophile: but, omega_ should be around and he is our helpdesk supreme
[01:07] <ajmitch> Uraeus: at least with RH you *know* that every new version is a downgrade ;)
[01:07] <Uraeus> ajmitch: hmm, your distribution inferiority complex seems to be shinning through here :)
[01:08] <ajmitch> *my* complex?
[01:08] <Uraeus> hehe
[01:08] <ajmitch> i'll have you know i'm sitting here using debian, wearing a caldera tshirt, and trying to get a mandrake iso ;)
[01:09] <Uraeus> ajmitch, debian and mandrake is ok, but even touching something with that Quisling company Caldera's name on it is a sin, seriously
[01:09] <ajmitch> Uraeus: yeah, but i got it free ;)
[01:10] <Uraeus> hmm, well time for sleep, and please burn that shirt :)
[01:10] Uraeus (cschalle at c224s9h5.upc.chello.no) left #gstreamer.
[01:18] <Aquariophile> I fear it will never stoop
[01:19] <ajmitch> it will, but takes quite a long time
[01:19] <ajmitch> it has to compile all the source
[01:21] <Aquariophile> windows is faster *g*
[01:21] <ajmitch> when have you ever had to compile something on windows?
[01:22] <ajmitch> when have you been able to compile something on windows? ;)
[01:22] <Aquariophile> installing
[01:22] <ajmitch> pardon?
[01:22] <Aquariophile> installing = setup.exe
[01:23] <ajmitch> yeah, installing != compiling
[01:25] <Aquariophile> Could you please try if gstreamer can show this streaming video ?
[01:25] <Aquariophile> mms://194.158.136.76/live1
[01:25] <ajmitch> i doubt it
[01:25] <ajmitch> what format is it in?
[01:27] <Aquariophile> no idea
[01:27] <Aquariophile> it is an austrian live show
[01:27] <ajmitch> gstmediaplay doesn't accept urls like that, i think
[01:27] <Aquariophile> like big brother
[01:27] <ajmitch> i am not a developer so i cannot really say ;)
[01:27] <Aquariophile> but in austria it is called taxiorange
[01:28] <ajmitch> it looks like windows media player format
[01:28] <ajmitch> mms == MS media stream, i guess
[01:29] <Aquariophile> ok
[01:29] <Aquariophile> thank you
[01:30] <Aquariophile> I fear it will never stop to compile
[01:30] <Aquariophile> Oh dear
[01:30] <ajmitch> errors?
[01:30] <Aquariophile> I forgot the ing
[01:30] <Aquariophile> ...nerver stop compiling
[01:30] <Aquariophile> not "to compile"
[01:31] <ajmitch> doesn't matter, i can still understand what you say
[01:31] <ajmitch> ;)
[01:31] <Aquariophile> yes, bbut I hte my bad english
[01:31] <Aquariophile> hte = hate
[01:31] <ajmitch> it's not bad, believe me ;)
[01:31] <ajmitch> i speak worse english at times, i'm sure
[01:32] <Aquariophile> in reality, it is. Yesterday I had to call "Lloyds of london" and my english was bad
[01:33] <Aquariophile> what's your native language ?
[01:33] <ajmitch> english ;)
[01:33] <Aquariophile> oh *g*
[01:33] <ajmitch> i'm from New Zealand
[01:33] <Aquariophile> Just saw you're from newzealand. My uncle was in NZ for 2 years
[01:34] <Aquariophile> yeah, saew it in you DNS
[01:35] Nick change: taaz-away -> taaz
[01:35] <ajmitch> taaz!
[01:35] <taaz> ajmitch!
[01:35] <Aquariophile> Aquariophile: 
[01:35] <Aquariophile> !
[01:36] <ajmitch> taaz: i'll leave you to help Aquariophile, i have to go to an installfest ;)
[01:36] <Aquariophile> :(((
[01:36] <Aquariophile> Please tell me first what I have to do after  rpm -ta gstreamer-0.1.1.tar.gz  
[01:36] <ajmitch> taaz: oh, and explain to Uraeus that RH is *not* the best distro ;)
[01:36] <taaz> ok
[01:37] <Aquariophile> oooooooh Dear !!
[01:37] <Aquariophile> ...
[01:37] Last message repeated 1 time(s).
[01:37] <Aquariophile> make[3]: *** [gsteditor] Error 1
[01:37] <Aquariophile> make[3]: Leaving directory `/usr/src/packages/BUILD/gstreamer-0.1.1/editor'
[01:37] <Aquariophile> make[2]: *** [all-recursive] Error 1
[01:37] <Aquariophile> make[2]: Leaving directory `/usr/src/packages/BUILD/gstreamer-0.1.1/editor'
[01:37] <Aquariophile> make[1]: *** [all-recursive] Error 1
[01:37] <Aquariophile> make[1]: Leaving directory `/usr/src/packages/BUILD/gstreamer-0.1.1'
[01:37] <Aquariophile> make: *** [all-recursive-am] Error 2
[01:37] <Aquariophile> Bad exit status from /var/tmp/rpm-tmp.8083 (%build)
[01:37] <taaz> Aquariophile: what's the problem?  i don't feel like reading the log ;)
[01:37] <Aquariophile> bash-2.04#                                
[01:37] <Aquariophile> paten?
[01:38] <taaz> ?
[01:38] <Aquariophile> paten = what 
[01:38] <Aquariophile> ?
[01:38] <ajmitch> taaz: he is trying to install gstreamer from the source tarball (as per Uraeus' instructions)
[01:38] <taaz> ./autogen.sh; make   what's the problem? ;)
[01:39] <ajmitch> taaz: Uraeus thought it would be easier to use the spec file in the tarball
[01:39] <Aquariophile>  ./autogen.sh ?
[01:39] <Aquariophile> where?
[01:39] <taaz> in the source dir
[01:40] <ajmitch> taaz: i'll leave it in your capable hands - try and comvince him that debian is the best distro ;)
[01:41] <Aquariophile> bash-2.04# ./autogen.sh
[01:41] <Aquariophile> bash: ./autogen.sh: No such file or directory
[01:41] <Aquariophile> bash-2.04# 
[01:41] <ajmitch> unpack the tarball
[01:41] Nick change: ajmitch -> ajbusy
[01:41] <Aquariophile>  already did
[01:48] <Aquariophile> taaz ?
[01:48] <Happyfeet> nakednews has gotta be the best thing to come out of the internet
[01:48] <taaz> yes?
[01:49] <Aquariophile> taaz: what do I have to do ?
[01:49] <Aquariophile> i am in the directory of the tar.gz 
[01:50] <taaz> uh... didn't i tell you already?  run autogen.sh  maybe add configure options, then run make.  what's the problem?
[01:51] <Aquariophile> autogen.sh is not in the directory
[01:53] <taaz> oh... maybe they didnt put that in there
[01:53] <taaz> run ./configure --help, use any options you need and run ./configure, then run make
[01:54] <Aquariophile> I already did it
[01:54] <Aquariophile> BUT:
[01:54] <Aquariophile> "make install" produced errors
[02:04] <Aquariophile> is there an other player, that is able to show streaming mms (window media player) files ??
[02:30] <Aquariophile> taaz?
[02:33] <taaz> ?
[02:34] hadess (hadess at pc121-gui14.cable.ntl.com) left irc: mooooh!
[02:35] Action: omega_ is back
[02:36] <taaz> yo
[02:36] <omega_> yo
[02:40] _gst_newt_ joined #gstreamer.
[02:47] <taaz> there's a 20000 byte shm leak every time i run gst apps.  maybe xvideosink related.  gotta ipcrm a ton of those at some point...
[02:49] <omega_> hmmm
[02:49] <omega_> one per app?
[02:49] <taaz> everytime i run some test app, yeah
[02:49] <taaz> does ipcs show that for you too?
[02:50] <omega_> checking
[02:50] <omega_> yup
[02:50] <omega_> grrrr
[02:51] <omega_> I have a stack of 32KB ones too
[02:51] <omega_> but with a nattch of 2
[02:54] <omega_> is there any way to tell which pid owns a shmseg ?
[02:54] <taaz> i don't understand any of this ;)
[02:54] <omega_> hmm
[02:54] <taaz> you would think they could just ref count the suckers and kill em off when no one uses em
[02:55] <omega_> you'd think
[02:55] <omega_> they do refcount them... nattch
[02:59] <taaz> must be some good reason to keep old ones around i guess
[02:59] <Aquariophile> Where to get rpm version 4 ?
[02:59] <omega_> redhat 7
[03:00] <Aquariophile> no, Suse
[03:00] <Aquariophile> bash-2.04# rpm -ivh gstreamer*.rpm
[03:00] <Aquariophile> only packages with major numbers <= 3 are supported by this version of RPM
[03:00] <Aquariophile> error: gstreamer-0.1.0-1.src.rpm cannot be installed
[03:00] <Aquariophile> only packages with major numbers <= 3 are supported by this version of RPM
[03:00] <Aquariophile> error: gstreamer-0.1.1-1.i386.rpm cannot be installed
[03:00] <Aquariophile> bash-2.04# 
[03:00] <omega_> or get the latest & greatest rpm 3.x
[03:00] <Aquariophile> this is my problem
[03:00] <omega_> the latest 3.x has support for 4.0 packages
[03:00] <Aquariophile> but where ?
[03:00] <omega_> rpm.org
[03:01] <Aquariophile> oh !
[03:01] <Aquariophile> Why do I never have this great ideas ?
[03:01] <Aquariophile> :-(
[03:02] <Aquariophile> to help!
[03:02] <Aquariophile> There are sooooo many !
[03:02] <Aquariophile> ftp://ftp.rpm.org/pub/rpm/dist/rpm-4.0.x
[03:03] <omega_> in your case I'd suggest only the latest 3.x
[03:03] <Aquariophile> but gstreamer want more that 3
[03:03] <omega_> ignore that
[03:04] <omega_> and it's not gstreamer, it's the binary package version
[03:04] <Aquariophile> fact is that I need gstreamer, and it tells me... see above what it tells me
[03:04] <Aquariophile> What do I exactly have to do, to install gstreamer?
[03:04] <omega_> get the latest rpm 3.x
[03:05] <Aquariophile> what's bad with 4.x ?
[03:05] <omega_> if you don't use it already, and you're not planning on upgrading soon to a distrib that uses it (newer suse or whatever) wouldn't tempt it
[03:07] Nick change: taaz -> taaz-food
[03:08] <omega_> ftp://ftp.rpm.org/pub/rpm/dist/rpm-3.0.x/rpm-3.0.6-6x.i386.rpm
[03:09] <Aquariophile> omega_: gfot the list ?
[03:09] <Aquariophile> omega_: Thank you very much
[03:09] <omega_> yeah, all you had to do was post the url
[03:10] <Aquariophile> I install this rpm... then I can install gstreamer*.rpm ?
[03:11] <omega_> theoretically
[03:11] <omega_> though you may need some other rpms from that directory in order to get rpm 3.0.6 installed
[03:11] <Aquariophile> *cry*
[03:11] <Aquariophile> Why can't it be easy ?
[03:11] <omega_> because rpm depends on too much stuff
[03:12] <Aquariophile> bash-2.04# rpm -ivh rpm-3.0.6-6x.i386.rpm
[03:12] <Aquariophile> error: failed dependencies:
[03:12] <Aquariophile>         patch < 2.5 conflicts with rpm-3.0.6-6x
[03:12] <Aquariophile> bash-2.04# 
[03:13] <omega_> this starts to be a suse problem
[03:13] <omega_> patch and rpm should not conflict in the first place, but now you've got a a problem
[03:13] <omega_> most likely there's a single file that overlaps between patch and rpm, so you could --force the rpm install
[03:13] <Aquariophile> *megacry*
[03:14] <omega_> but I wouldn't try it without knowing what that file is
[03:14] Action: omega_ suggests #suse
[03:14] <Aquariophile> #suse is a good idea
[03:14] <Aquariophile> but ... before I have to go online as a user
[03:14] <omega_> it even exists <g>
[03:14] <Aquariophile> (root-banned on #suse)
[03:15] <omega_> dare I ask why?
[03:15] <omega_> oh, right
[03:15] <omega_> root should be banned in all irc
[03:15] <omega_> it's ridiculously unsafe
[03:15] <Aquariophile> ...
[03:15] <omega_> root == "kick me"
[03:15] <Aquariophile> I come back in a minute
[03:15] <omega_> ok
[03:15] <Aquariophile> I am root
[03:15] <Aquariophile> (kick you me ?)
[03:15] <omega_> no, it's a sign for crackers to attack you
[03:16] <Aquariophile> ok...
[03:16] <omega_> and they will
[03:16] Aqua (user at 62.46.9.17) joined #gstreamer.
[03:17] Nick change: Aqua -> Aquariophile_
[03:17] <Aquariophile_> Now I am user
[03:17] <omega_> using a machine as root is just as dangerous, and very much not a recommended practice
[03:17] Aquariophile (root at M153P017.dipool.highway.telekom.at) left irc: [x]chat
[03:17] Nick change: Aquariophile_ -> Aquariophile
[03:21] <Aquariophile> nobody talking in #suse
[03:22] <omega_> hmm, well, I can't help you much with that problem
[03:22] <omega_> but then again, why aren't you just building the tarball?
[03:22] <Aquariophile> I am not able to understand your last question
[03:23] <omega_> why are you working so hard at getting the gstreamer rpm installed?  why aren't you just building it from the tarball?
[03:23] <Aquariophile> bad joke
[03:24] <Aquariophile> very bad joke
[03:24] <Aquariophile> I tryed the tar.gz first
[03:24] <Aquariophile> 3 days ago
[03:24] <omega_> and?
[03:24] <Aquariophile> It does not work
[03:24] <Aquariophile> "make install" and I gett too much errors
[03:24] <omega_> such as?
[03:25] <Aquariophile> good question, did not write it down...
[03:25] <Aquariophile> shell I run it again ?
[03:25] <omega_> yeah, it should build, and if it doesn't, we need to fix it
[03:25] <omega_> but maybe a better question is: what are you trying to do with gstreamer right now?
[03:26] <Aquariophile> I want to view a video
[03:26] <Aquariophile> from a TV-Show
[03:26] <omega_> what format?
[03:26] <Aquariophile> it's 24 hours online a day
[03:26] <Aquariophile> live
[03:26] <Aquariophile> mms://
[03:26] <omega_> won't work
[03:26] <Aquariophile> not ??
[03:26] <omega_> *only* microsoft software can play that
[03:26] <Aquariophile> ;(
[03:26] <omega_> because they refuse to release any specifications
[03:27] <Aquariophile> not nice from them
[03:27] <omega_> nope
[03:27] <omega_> that's why we're working on creating new formats that are completely open
[03:28] <Aquariophile> hm
[03:28] <Aquariophile> there are some linux-softwares that aren't free too
[03:28] <Aquariophile> vmware....
[03:28] <omega_> yeah, that won't help you much, unless you have a copy of windoze
[03:28] <Aquariophile> I really need vmware, but for free, I won't crack it
[03:30] <Aquariophile> everything is so commercial on earth ;(
[03:30] <omega_> yup
[03:36] Nick change: omega_ -> omega_dinner
[03:38] Aquariophile (user at 62.46.9.17) left irc: Ping timeout for Aquariophile[62.46.9.17]
[04:22] seer (seer at gamehendge.org) joined #gstreamer.
[04:24] <seer> hey there!
[04:24] <seer> when is the next "snapshot" going to come out?  I haven't gotten a working CVS version working in quite some time
[04:25] <seer> "Lots of duct tape" seemed to work alright, but like one week after that's CVS rocked
[04:25] <seer> but recently it's been strange
[04:26] <seer> oh well
[04:26] seer (seer at gamehendge.org) left irc: fixing Xtheater for now...
[05:30] Nick change: taaz-food -> taazzzz
[05:46] dichro (dichro at foo.optusnet.com.au) joined #gstreamer.
[05:46] <dichro> morning all
[05:51] <omega_dinner> evening
[05:51] <omega_dinner> er,
[05:51] Nick change: omega_dinner -> omega_
[05:53] Nick change: wtay-snooker -> wtay
[05:53] <wtay> uh?
[05:54] <omega_> yo
[05:54] <wtay> hello
[05:54] Action: wtay can hear the birds sing..
[05:54] Action: omega_ tried to trace the incsched problem
[05:54] <omega_> oops
[05:54] <wtay> can you reproduce the error with 5 threads?
[05:55] <wtay> the == an
[05:55] <omega_> that's what I'm doing
[05:55] <omega_> 4, actually, plus a pipeline
[05:55] <wtay> ok
[05:55] <wtay> what were those 'no cothread..' errors?
[05:56] <omega_> just what they sound like <g>
[05:56] <wtay> right <g>
[05:56] <omega_> the iterator is trying to switch into a queue for some reason
[05:57] <omega_> the chains are getting confused somehow
[05:57] <wtay> hmm
[05:57] <wtay> I have another question...
[05:57] <wtay> can incsched handle unconnected elements?
[05:57] <omega_> theoretically, yes, but not likely
[05:57] <wtay> ok
[05:58] <omega_> rather, it will probably have a cow but not fail
[05:58] <omega_> try it <g>
[05:58] <wtay> not now :)
[06:00] <wtay> I need some sleep.. see you guys later..
[06:00] <omega_> ok
[06:00] Nick change: wtay -> wtay-sleeping
[06:01] <wtay-sleeping> is there anything I can do regarding the autoplugger not working?
[06:01] <omega_> just start making it work like mpeg2parse3
[06:01] <omega_> and write up a description of how on earth the thing works
[06:01] <omega_> cause I certainly have no idea
[06:01] <wtay-sleeping> I have
[06:01] <omega_> exhaustive?
[06:02] <wtay-sleeping> kinda
[06:02] <omega_> I mean actually function-by-function, block-by-block
[06:02] <wtay-sleeping> lemme find it...
[06:02] <wtay-sleeping> docs/random/autoplug1 is a start
[06:02] <wtay-sleeping> ok, not in block by block detail
[06:03] <wtay-sleeping> hmm and it's outdated.. ok I'll do that tomorrow
[06:03] <omega_> it's strange enough that it needs text mirroring the code
[06:03] <omega_> which I will be doing for incsched as well
[06:03] <omega_> and matth may be sanity checks for both of us
[06:04] <wtay-sleeping> yup
[06:04] <omega_> he's got scheduling mostly figured out now
[06:04] <wtay-sleeping> I'll try to write something as complete as possible
[06:04] <omega_> ok, cool
[06:05] <wtay-sleeping> the shortest path algorithm is something I borrowed somewhere, so I don't know how it works (and I don't have my university math books anymore :()
[06:05] <omega_> hehe
[06:05] <omega_> oh well
[06:07] <wtay-sleeping> which reminds me that I need to document caps nego too
[06:07] <omega_> yup <G>
[06:08] <wtay-sleeping> I want all of you to see how much it sucks and how incomplete it is :)
[06:08] <omega_> ehehehe
[06:08] <omega_> --gst-mask=0x81a6
[06:08] <omega_> from that you can tell what I'm doing <g>
[06:08] <wtay-sleeping> uh
[06:09] <wtay-sleeping> clock?
[06:09] <wtay-sleeping> er pads
[06:10] <omega_> this is a test, this is only a test <g>
[06:10] <omega_> bleck
[06:10] <omega_> I practically need to colorize pointers too
[06:11] <wtay-sleeping> anyway, I'm off to bed.
[06:11] <omega_> ok, dream in gstinfo.c colors <g>
[06:11] <wtay-sleeping> I hope not <g>
[06:11] <omega_> ha
[06:15] <dichro> is the data parameter to the negotiation callback just for the function's own callback data, or does it actually carry something useful in it?
[06:15] <omega_> I don't know myself, but I'd presume that it's just for the plugin's own use
[06:15] <omega_> lemme check it out
[06:16] <wtay-sleeping> yup, plugin use
[06:16] <omega_> ok
[06:16] <wtay-sleeping> doesn't work actually
[06:16] <dichro> hmm. weird.
[06:16] <dichro> I'm definitely not setting it to anything, but it's coming back pointing at some object somewhere.
[06:16] <omega_> whoops
[06:17] <wtay-sleeping> well, it's never NULL, it a gpointer*
[06:17] <wtay-sleeping> *data should be NULL, data isn't
[06:17] <omega_> hmm
[06:17] <dichro> looking at the bit in gstpad that calls it, it looks like it /should/ be NULL on the first iteration, and only the plugin should set it... but...
[06:18] <dichro> sorry, in gstcaps.c
[06:18] <wtay-sleeping> it's called in gstpad.c
[06:18] <dichro> um, yes. :)
[06:19] Action: wtay-sleeping drags himself from the keyboard now...
[06:19] <dichro> ah, right. that makes sense, yes :)