[gst-devel] Daily IRC logs

wim.taymans at chello.be wim.taymans at chello.be
Sat Mar 31 06:27:23 CEST 2001

[06:31] <PartyZan> What's is Beowolf and other?
[06:31] <_maestro_> have fun guys, I m going to sleep.
[06:31] _maestro_ (Banane at HSE-Sherbrooke-ppp78458.qc.sympatico.ca) left #gstreamer.
[06:31] <PartyZan> And what time?
[06:31] <jmv> Beowolf is a cluster of computers for Linux
[06:32] jmv (jmv at HSE-Montreal-ppp333276.sympatico.ca) left irc: [x]chat
[06:32] <PartyZan> Could you send me any link about it?
[06:35] <PartyZan> Hey,Omega, did you hear about project of Ajmitch?
[06:37] <omega-packing> which one?
[06:38] <PartyZan> MP3-encoding on-the-fly
[06:38] <PartyZan> And he have any another?
[06:38] <PartyZan> And he has(sorry) any another?
[06:38] <omega-packing> yeah, he's got others, I don't know much about them
[06:39] <PartyZan> OK
[06:48] Nick change: taaz-working -> taaz
[06:49] <omega-packing> can you implement what toolman suggested ?
[06:49] <taaz> who me ;)
[06:49] <omega-packing> um, yeah, mr wiki
[06:50] <taaz> um.. hang on... let me sort through my email.  which suggestion is it?
[06:50] <omega-packing> the rewrite idea for wiki urls
[06:50] <PartyZan> I'm goind to sleep,buy:)
[06:50] PartyZan (root at ip95-221.dialup.wplus.net) left irc: [x]chat
[06:56] <taaz> ok, let me try it...
[07:00] <taaz> funky... it works
[07:01] <taaz> hmm.  has to have trailing / though
[07:01] <omega-packing> ooh
[07:03] <taaz> i'd rather have main wiki in /wiki/ and data stuffs in /wiki-????/ though
[07:03] <taaz> no need to stress out my shift key ;)
[07:04] <omega-packing> agreed
[07:05] <taaz> i'm going to break things for a sec here... hold on
[07:13] <taaz> hmm..
[07:15] <taaz> well, gst.net/wiki/SomeNode works but all the links in it are to cgi-bin/...
[07:39] <omega-packing> I'm gonna log out for a while, my cd burner has just made two frisbees in a row
[07:39] Action: omega-packing is minimizing the number of potential sources for failure
[07:39] omega-packing (omega at omegacs.net) left irc: [x]chat
[09:41] steveb (chatzilla at 212186169160.chello.com) joined #gstreamer.
[09:59] steveb (chatzilla at 212186169160.chello.com) left irc: ChatZilla 0.8 [Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; 0.8.1) Gecko/20010323]
[11:30] _gst_newt_ joined #gstreamer.
[11:55] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[11:55] Uraeus (cschalle at c224s9h5.upc.chello.no) left irc: syntax error - user imploded
[11:55] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[12:05] ajmitch (me at p40-max4.dun.ihug.co.nz) joined #gstreamer.
[12:10] <Uraeus> hi ajmitch
[12:10] <ajmitch> hi Uraeus
[12:12] <taaz> heh.. my logo entry posted to list ;)
[12:13] Action: taaz waves to _gst_newt_ 
[12:13] Action: ajmitch will have to look ;)
[12:13] <ajmitch> taaz: why do you wave to the bot? ;)
[12:14] <taaz> _gst_newt_ is in my logo!
[12:14] <ajmitch> oh?
[12:14] <Uraeus> hmm, personally I would like a logo which translates easily into an Icon for gstmediaplay
[12:14] <Uraeus> or could easily be used to illustrate posts on Gnotices/Slashdot etc.
[12:14] <taaz> mine's in svg
[12:15] <taaz> http://crib.lehn.org:8080/~dlehn/projects/gstreamer/logo/
[12:16] <Uraeus> taaz: I have to admit I am having trouble deciding if I like your logo or not; It do give the project a mascot, but I miss the sense of 'speed' that the RidgeRun logos have
[12:24] <taaz> i added some fake motionblur to "streamer" ... looks a little faster now ;)
[12:27] <taaz> bed time
[12:27] Nick change: taaz -> taazzzz
[12:35] holger (holger at Q.convergence.de) joined #gstreamer.
[12:37] <ajmitch> hi
[13:10] _gst_newt_ joined #gstreamer.
[13:33] Uraeus (cschalle at c224s9h5.upc.chello.no) left irc: syntax error - user imploded
[13:34] Nick change: ajmitch -> aj_zzzz
[14:35] iGN_ (ign at joined #gstreamer.
[14:42] sienap (synap at ipc379c060.dial.wxs.nl) joined #gstreamer.
[15:21] zaheer (zaheer at odin.cs.ucl.ac.uk) joined #gstreamer.
[15:36] sienap (synap at ipc379c060.dial.wxs.nl) left irc: sienap has no reason
[16:37] vektor (vektor at HSE-Kitchener-ppp194641.sympatico.ca) joined #gstreamer.
[17:19] omega-waking (omega at omegacs.net) joined #gstreamer.
[17:27] <matth> yo
[17:27] <matth> when do you take off?
[17:27] <matth> s/when/omega: when
[17:35] <omega-waking> in 2:15
[17:44] BBB (BBB at joined #gstreamer.
[17:44] <BBB> hey!
[17:44] <BBB> I like the second logo :P
[17:45] <zaheer> omega-waking: not long be4 your flight :)
[17:46] <BBB> what was the cafe called again where we were gonna meat?
[17:46] <zaheer> why dont you post that info to the list.....
[17:46] <zaheer> BBB: s/meat/meet i hope :)
[17:47] Action: zaheer hopes to meet omega at the airport..
[17:47] <BBB> oops
[17:47] <BBB> meet
[17:47] <BBB> :)
[17:48] Action: BBB will send his preference to the list
[17:48] <BBB> hold on
[17:48] <zaheer> i think steveb and omega have agreed on a place for our meeting
[17:48] <zaheer> tho i dont know where
[17:49] <BBB> they told me yesterday
[17:49] <BBB> let's check the logs
[17:49] Action: zaheer is feeling very cold, the aircon is pretty strong
[17:49] <BBB> http://www.apollofirst.nl/
[17:49] <BBB> that's the place
[17:50] omega-waking (omega at omegacs.net) left irc: Ping timeout for omega-waking[omegacs.net]
[17:55] <zaheer> i hear amsterdam is pretty cold at the moment
[18:00] <BBB> yep
[18:02] Nick change: wtay-sleeping -> wtay
[18:06] <zaheer> yo wtay
[18:07] <zaheer> BBB: its there at 12pm?
[18:08] omega-europe (omega at omegacs.net) joined #gstreamer.
[18:08] <zaheer> yo omega
[18:09] <zaheer> is the meet at apollofirst?
[18:09] <omega-europe> yes
[18:09] Action: omega-europe leaves now for airport
[18:09] Action: omega-europe is away: europe
[18:09] <zaheer> ok ill see you at the airport..
[18:10] Nick change: taazzzz -> taaz
[18:10] <BBB> :)
[18:10] <BBB> have fun guys, see you tomorrow at 12 in apollofirst
[18:11] <BBB> I sent a routedescription to the list :)
[18:12] <zaheer> ok thx bbb :)
[18:12] greg_ (greg at home.sente.pl) joined #gstreamer.
[18:13] <greg_> hi ! what's up ?
[18:13] Action: zaheer is gonna go to the office now and pickup the whiteboard and pens
[18:13] <wtay> yo guys
[18:15] <zaheer> wtay: i had a prob with my stereo2mono capsnego last night, can i talk to you later about it?
[18:16] <BBB> zaheer: no problem
[18:16] Action: BBB is gone - work to do
[18:16] Nick change: BBB -> BBB-[away]
[18:16] Action: BBB-[away] is away: work to do
[18:17] <wtay> zaheer: sure
[18:20] <zaheer> ok thanks
[18:20] <zaheer> ill see you guys laters
[18:20] zaheer (zaheer at odin.cs.ucl.ac.uk) left irc: on way to office
[18:20] <wtay> ok
[18:24] Nick change: wtay -> wtay-eating
[18:25] Nick change: taaz -> taaz-away
[18:35] steveb (steveb at node1ee02.a2000.nl) joined #gstreamer.
[18:49] <steveb> anyone know what time we're meeting tomorrow?
[18:58] holger (holger at Q.convergence.de) got netsplit.
[18:58] omega-europe (omega at omegacs.net) got netsplit.
[18:58] steveb (steveb at node1ee02.a2000.nl) got netsplit.
[18:58] taaz-away (dlehn at got netsplit.
[19:09] holger (holger at Q.convergence.de) got lost in the net-split.
[19:09] taaz-away (dlehn at got lost in the net-split.
[19:09] omega-europe (omega at omegacs.net) got lost in the net-split.
[19:09] steveb (steveb at node1ee02.a2000.nl) got lost in the net-split.
[19:12] holger (holger at Q.convergence.de) joined #gstreamer.
[19:12] taaz (dlehn at joined #gstreamer.
[19:30] greg_ (greg at home.sente.pl) left irc: Ping timeout for greg_[home.sente.pl]
[19:36] greg_ (greg at home.sente.pl) joined #gstreamer.
[19:39] BBB-[away] (BBB at left irc: Ping timeout for BBB-[away][]
[19:47] Nick change: wtay-eating -> wtay
[20:30] Nick change: aj_zzzz -> ajmitch
[20:32] dobey (dobey at dreadnought.ximian.com) joined #gstreamer.
[20:32] <wtay> yo
[20:32] <wtay> ximian? <g>
[20:32] <dobey> yeah
[20:32] <wtay> what are you working on?
[20:32] <dobey> how is endian stuff coming along?
[20:32] <dobey> too much
[20:32] <ajmitch> hi
[20:32] <wtay> heh
[20:33] <wtay> endianness is largly untested...
[20:33] <wtay> we have a PPC guy pointing us to the problems..
[20:34] <dobey> yeah, i just rebuilt from a couple days ago, and the mp3 seems to work better, but not good enough
[20:34] <dobey> also, gstmediaplay doesn't like to do anything for me
[20:34] <wtay> hmm
[20:34] <wtay> mpeg1 is reported to work on PPC
[20:35] <dobey> it locks up for me
[20:35] <wtay> oh
[20:35] <dobey> hrmm
[20:35] <dobey> are you using smpeg?
[20:35] <wtay> no, plain mpeg_play
[20:35] <dobey> oh
[20:37] <wtay> but the videosink and mpg123 endianness largly cause the real problems
[20:37] <dobey> yeah
[20:37] <dobey> smpeg works, which is why i asked, maybe you should use it?
[20:38] <wtay> it does too much, it's not really suited for the gstreamer architecture...
[20:39] <dobey> too much? and i can use xmms plugins under gstreamer, sheesh, well, i can't get them to work
[20:40] <dobey> ogg/vorbis also locks up gstmediaplay
[20:40] <dobey> mp3 in gstmediaplay does nothing
[20:40] <wtay> hmm, looks like there is some other problem then
[20:40] <wtay> did you run gstreamer-register?
[20:40] <dobey> yes
[20:40] <wtay> CVS?
[20:41] <dobey> yep
[20:41] <wtay> hmm
[20:41] <dobey> hrmm, gst uses synaesthesia?
[20:41] <dobey> brb
[20:42] <wtay> dobey: quite poorly I should add...
[20:42] <ajmitch> wtay: what could problem be? hadess has this stuff working on PPC, no?
[20:42] <wtay> ajmitch: yep
[20:42] <wtay> need to find out why it hangs...
[20:43] <ajmitch> wtay: ogg playback worked on i386 for me a couple of days ago
[20:43] <ajmitch> and yes, this computer does act like an i386 for me at times ;)
[20:43] <wtay> ajmitch: gstmediaplay supports ogg/avi/mpeg1/mp3/mpeg2/flx on my machine..
[20:44] <dobey> back
[20:44] <wtay> dobey: any idea where it hangs?
[20:44] <dobey> when i try to open a file
[20:44] <wtay> can you run gstmediaplay with --gst-debug-mask=-1 flags?
[20:44] <dobey> click 'ok' and lock
[20:44] <dobey> sure
[20:45] <wtay> dobey: did you try to run gstmediaplay <filename>
[20:45] <dobey> yes
[20:45] <ajmitch> wtay: sound hardware in use, maybe?
[20:45] <dobey> and it locked
[20:45] <dobey> DEBUG(26427:-1)gst_thread_signal_thread:339: signaling thread
[20:45] <dobey> DEBUG(26427:-1)gst_thread_main_loop:322: thread "thread_play_audio" waiting
[20:45] <dobey> DEBUG(26427:-1)gst_thread_wait_thread:348: waiting for thread
[20:46] <wtay> final output?
[20:46] <dobey> yeah
[20:46] <wtay> sec..
[20:47] <wtay> can you capture the complete log and mail it to me (wim.taymans at chello.be)?
[20:48] <dobey> yeah, sure
[20:48] <wtay> cool
[20:49] <dobey> how would i do that then? > doesn't capture debug output
[20:49] <wtay> 2>&1 >/tmp/log
[20:50] <wtay> redirect stderr to stdout
[20:51] <dobey> didn't work
[20:51] <wtay> sec...
[20:53] <wtay> gstmediaplay --gst-debug-mask=-1 <filename> 2>/tmp/test
[20:53] <wtay> sorry
[20:55] holger (holger at Q.convergence.de) left irc: Client Exiting
[20:55] <dobey> there you go
[20:57] <wtay> looking at the log...
[20:59] <wtay> hmm a deadlock...
[20:59] <dobey> yep
[21:00] <wtay> looks like it is blocking somewhere in osssink
[21:00] <dobey> can't i use esd?
[21:01] <wtay> hmm try to disable any programs using oss...
[21:02] <ajmitch> lsof /dev/dsp should tell you what programs are using the sound driver
[21:02] <dobey> nothing is using oss
[21:02] <wtay> fuser -av /dev/dsp
[21:03] <dobey> none
[21:04] <wtay> do you (on the console) lines like this:
[21:04] <wtay> osssink: attempting to open sound device
[21:04] <wtay> osssink: Capabilities 00003b00
[21:04] <wtay> osssink:   Full duplex
[21:04] <wtay> osssink:   Realtime
[21:04] <wtay> osssink:   Has coprocessor
[21:04] <wtay> osssink: opened audio with fd=5
[21:04] <wtay> osssink: negotiate
[21:04] <wtay> etc..
[21:04] <dobey> gg 2> gst.log
[21:04] <dobey> gst_play: realize
[21:04] <dobey> osssink: attempting to open sound device
[21:04] <dobey> osssink: Capabilities 00000100
[21:04] <dobey> osssink:   Full duplex
[21:04] <dobey> osssink: Formats 00000030
[21:04] <dobey> osssink:   S16_LE
[21:04] <dobey> osssink:   S16_BE
[21:04] <dobey> osssink: opened audio with fd=7
[21:05] <wtay> ok, that looks good..
[21:05] omega-europe (omega at omegacs.net) joined #gstreamer.
[21:05] <ajmitch> myabe run strace on gstmediaplay to see what it is locking on?
[21:05] steveb (steveb at node1ee02.a2000.nl) joined #gstreamer.
[21:05] <ajmitch> hey omega-europe, steveb
[21:05] <wtay> yo omega-europe steveb
[21:06] <dobey> rt_sigprocmask(0x2, 0, 0x7ffff4fc, 0x8) = 0
[21:06] <dobey> write(9, "\17\357\24\300\0\0\0\0\0\0\0\0\17\374\251\30\20\10\203"..., 148) = 148rt_sigprocmask(0x2, 0, 0x7ffff448, 0x8) = 0
[21:06] <dobey> rt_sigsuspend(0x7ffff448, 0x8 <unfinished ...>
[21:06] <dobey> --- SIGRT_0 (Real-time signal 0) ---
[21:06] <dobey> <... rt_sigsuspend resumed> )           = 32
[21:06] <dobey> sigsuspend([]
[21:06] <wtay> not very helpfull...
[21:07] <dobey> i'm guessing it has something to do with the osssink:   Realtime not existing for me
[21:07] <wtay> look like a thread issue again...
[21:07] <dobey> yep
[21:09] <steveb> anyone know what time we're meeting tomorrow?
[21:10] <dobey> 9:00 PM EST?
[21:10] <steveb> erm
[21:11] <dobey> err
[21:11] <dobey> wait
[21:11] <dobey> n/m
[21:11] Action: dobey confused
[21:11] <wtay> dobey: flying over to europe then? <g>
[21:11] <dobey> it happens when you work on 30+ projects simultaneously
[21:11] <steveb> at Amsterdam tomorrow
[21:11] <dobey> no, i'm not going to guadec
[21:12] <steveb> wtay: what time will you arrive?
[21:12] <wtay> steveb: not
[21:12] <steveb> oh, not coming?
[21:12] <wtay> steveb: nope
[21:16] <wtay> dobey: trying to find a good testcase...
[21:16] <dobey> ok
[21:17] <wtay> gstreamer-launch disksrc location=<some.ogg> ! vorbisdec ! osssink
[21:17] <wtay> can you try this?
[21:18] <wtay> no threads in this case
[21:20] <dobey> that works
[21:20] <wtay> hmm
[21:20] <dobey> yeah
[21:20] <wtay> thread issues again...
[21:21] <dobey> what should i put there for an .mpeg movie?
[21:21] <dobey> i'll test that too
[21:21] <wtay> can't do that yet on the command line...
[21:21] <dobey> oh
[21:21] <wtay> yeah, no dynamic pads for -launch yet...
[21:23] <wtay> what happens here is that right after the place where it locks in your case, the main thread is running again (after setting the state on osssink)
[21:23] <dobey> hrmm, so fix soon?
[21:23] <wtay> somehow this doesn't happen with you.. dunno why yet
[21:25] <wtay> hmm, which reminds me of a bug in hadess's PPC audio card driver...
[21:26] <wtay> dobey: does the ogg file play well?
[21:26] <dobey> yeah
[21:26] <dobey> mp3's don't though
[21:26] <dobey> can i do mp3 from the command line with -launch?
[21:26] <wtay> yes, sec..
[21:27] <wtay> gstreamer-launch disksrc location=some.mp3 ! mp3parse ! mpg123 ! osssink
[21:27] <wtay> mp3 layer 3 bitstreams fail in a horrible way on PPC
[21:28] <dobey> that didn't work too well
[21:28] <dobey> too bad i can't use the xmms mp3 player
[21:28] <dobey> in gst
[21:29] <dobey> brb, more wokr
[21:29] <dobey> wor
[21:29] <dobey> k
[21:29] <wtay> ok
[21:34] <wtay> omega-europe: ETA for incsched1?
[21:46] <wtay> dobey: got a little fix checked in for you to try..
[21:46] <dobey> in cvs?
[21:46] <wtay> yup
[21:46] <dobey> ok
[22:05] <wtay> dobey: ?
[22:06] <dobey> yeah?
[22:07] <wtay> better?
[22:07] <dobey> installing now
[22:07] <wtay> ok <g>
[22:09] <dobey> still locks
[22:09] <wtay> grr
[22:10] <dobey> and mp3 loads now, and locks instead of doing nothing
[22:11] <wtay> care for another log mail? this time also add --gst-info-mask=-1
[22:11] <ajmitch> hmm, i'm still here? ;)
[22:12] <wtay> ajmitch: so? <g>
[22:12] <ajmitch> wtay: just wondering at how i haven't been kicked off for flooding (pasted about 50 lines in #freedevelopers)
[22:12] <wtay> hehe
[22:13] <ajmitch> 62 lines actually, quite a lot ;)
[22:15] <wtay> ajmitch: you won't get kicked of #gstreamer by flooding #freedevelopers :)
[22:16] <dobey> wtay: can i dcc the log to you?
[22:16] steveb (steveb at node1ee02.a2000.nl) left irc: [x]chat
[22:16] <ajmitch> wtay: but i could get kicked off OPN for sending that much ;)
[22:17] Action: ajmitch watches ppl in #freedevelopers drop offline from ping timeout
[22:17] <wtay> dobey: nope, just email me (i'm behind a FW)
[22:17] <dobey> ok
[22:17] <ajmitch> bye ;)
[22:17] ajmitch (me at p40-max4.dun.ihug.co.nz) left irc: http://www.freedevelopers.net
[22:20] <wtay> dobey: could you add --gst-info-mask=-1 too?
[22:20] <dobey> i did
[22:20] <wtay> did it go to the console then?
[22:20] <dobey> nope
[22:20] <dobey> gstmediaplay --gst-debug-mask=-1 --gst-info-mask=-1 /home/dobey/10-melange.ogg 2> gst.log
[22:20] <dobey> i'm guessing that's what all the lines that begin with INFO are
[22:21] <wtay> ermm, I don't have the INFO lines in the log here...
[22:22] <wtay> it is another logfile though...
[22:22] <dobey> INFO: Initializing GStreamer Core Library
[22:22] <dobey> INFO: setting DEBUG categories to 0xFFFFFFFF
[22:22] <dobey> INFO: CPU features: (00000000) NONE
[22:23] <dobey> i'm guessing that's useless though
[22:23] <wtay> there should be way more INFO lines in the log
[22:24] <wtay> sec...
[22:25] <dobey> nope
[22:25] <wtay> hmm
[22:26] <dobey> weird
[22:26] <wtay> very
[22:26] greg_ (greg at home.sente.pl) left irc: Ping timeout for greg_[home.sente.pl]
[22:26] <wtay> INFO: Initializing GStreamer Core Library
[22:26] <wtay> INFO: setting DEBUG categories to 0xFFFFFFFF
[22:26] <wtay> INFO: setting INFO categories to 0xFFFFFFFF
[22:26] <wtay> INFO: CPU features: (c1c7f9ff) MMX 3DNOW MMXEXT
[22:27] greg_ (greg at home.sente.pl) joined #gstreamer.
[22:29] <dobey> hold up
[22:30] greg_ (greg at home.sente.pl) left irc: Ping timeout for greg_[home.sente.pl]
[22:32] <dobey> there you go
[22:33] <wtay> much better :)
[22:34] <wtay> hmm
[22:34] <wtay> but no info with gstmediaplay?
[22:34] <dobey> hrmm?
[22:35] <wtay> for -register it correctly sets up the INFO logging
[22:36] <dobey> and?
[22:36] <wtay> using the same --gst-info-mask=-1 flag for gstmediaplay doesn't work?
[22:41] <taaz> hmm.  thread fix commit?  i'll have to try it since i still get lock ups in gstmediaplay but they don't seem to happen when debug output is on.
[22:42] <dobey> yeah, that's what i used
[22:43] <wtay> dobey: then there is a bug in the arg parsing in gst.c (gst_init_check) but it doesn't happen here...
[22:43] <dobey> ahh
[22:44] <wtay> dobey: are you interested in debugging it?
[22:46] <dobey> maybe later
[22:46] <wtay> ok
[22:47] <wtay> dobey: omega-europe is working on a completely rewritten thread system
[22:47] <dobey> ok
[22:48] <wtay> although I'm not convinced the thread sync is the only issue here
[22:49] <dobey> yeah
[23:34] greg_ (greg at home.sente.pl) joined #gstreamer.
[23:51] zaheer (zaheer at tnt-16-171.easynet.co.uk) joined #gstreamer.
[23:51] <zaheer> yo
[23:51] <wtay> hi
[23:51] <zaheer> i have a bin of:
[23:52] <zaheer> disksrc -> mpg123 -> stereo2mono -> osssink
[23:52] <zaheer> basically helloworld.c modified to add stereo2mono
[23:52] <wtay> add mp3parse too...
[23:52] <zaheer> yes i forgot that :)
[23:52] <wtay> ok
[23:53] <zaheer> all i did was add stereo2mono in the bin in the right place, with helloworld.c
[23:53] <wtay> good
[23:53] <zaheer> let me give you the error i get
[23:54] <zaheer> the only nego function that gets called is for the src pad...
[23:54] <zaheer> i have them printf'ing
[23:55] <zaheer> isnt it the sink pad's nego that should be called first?
[23:55] <wtay> depends...
[23:56] <wtay> it should...
[23:56] <zaheer> i guess i should send you the stereo2mono.c file, no doubt its an obvious problem that i have done due to my lack of total understanding :)
[23:57] <wtay> ok, go ahead, I noticed from the logs that it looks like you didn't add the pad to the elements... we'll see
[23:57] <zaheer> what happens is the things dont negotiate properly....tho the mp3 is definitely a stereo mp3
[23:57] <zaheer> i fixed that error :)
[23:57] <zaheer> i made a booboo on the padtemplates :)
[23:58] <wtay> hehe, can you send me the plugin?
[23:58] <wtay> wim.taymans at chello.be
[23:59] <zaheer> ok ill send it to you now
[00:00] --- Sat Mar 31 2001
[00:01] <zaheer> sent..
[00:02] <zaheer> i probably have a memory leak too...cant remember if i fixed it with not unreffing a caps object when i should
[00:02] <wtay> zaheer: caps refcounting is largely untested too...
[00:03] <zaheer> :)
[00:04] <zaheer> i guess we should try and cause it to break :)
[00:05] <wtay> that's too easy :)
[00:05] <wtay> let's try to make the s2m work instead :)
[00:05] <zaheer> yah definitely
[00:05] <zaheer> once thats done i will get the ulaw converter done, coz its nearly the same in terms of caps nego
[00:05] <wtay> ok I spotted a bug
[00:06] <dobey> hopefully it'll help me :-P
[00:06] ChiefHighwater (paul at temple-baptist.com) joined #gstreamer.
[00:06] <wtay> in _init you set the caps of the pads... why?
[00:06] <wtay> line 205
[00:06] <zaheer> coz you did in converter2.c
[00:06] <wtay> hehe, well....
[00:06] <wtay> ok what happens is this:
[00:07] <zaheer> thats only reason i set the caps in init :)
[00:07] <wtay> the element has decided on the caps, so it sets them.
[00:07] <wtay> the audiosink is connected to the s2m plugin (which has caps) and tries to adjust itself
[00:08] <wtay> it fails bacause the caps are incomplete (it doesn"t state the rate and stuff) so it calls the src pad negofunction of s2m to give something else
[00:08] <zaheer> aaah
[00:08] <ChiefHighwater> Ello
[00:09] <wtay> I'm going to insert the s2m plugin in the helloworld example too here... sec..
[00:09] <zaheer> so i should only set caps in nego function?
[00:09] <wtay> hi ChiefHighwater
[00:09] <zaheer> hiya chw :)
[00:09] <ChiefHighwater> so, how long until the "pah teh"?
[00:09] <zaheer> a few hrs till my flight :)
[00:09] <ChiefHighwater> hehe
[00:10] <wtay> pah teh?
[00:10] <zaheer> but till the party is about 11 1/2 hrs :)
[00:10] <zaheer> my flight is in 7 hrs...
[00:10] <taaz> someone will take notes if you all discuss interesting things right? ;)
[00:10] <zaheer> taaz: i have my laptop and will be noting things down
[00:11] <zaheer> if the airline lets me, i will also have a whiteboard there....
[00:11] <wtay> taaz: _gst_newt_ will :)
[00:11] <taaz> bring a digital camera to take whiteboard pics <g>
[00:11] <ChiefHighwater> wtay:pah = par ; teh = ty  in urban slang here in the US
[00:12] <wtay> taaz: doh!! 
[00:12] <zaheer> taaz: i can take a real camera :)
[00:12] <wtay> taaz: sorry for me being stupid...
[00:12] <ChiefHighwater> omega-europe has digital camera and video
[00:12] <zaheer> we have a superb whiteboard at uni, it has a print button that prints the contents of the whiteboard onto faxroll style paper
[00:13] <taaz> you know... the only portable digital device i own is a $5 crappy watch ;)
[00:13] <zaheer> and your brain i hope :)
[00:14] <zaheer> its digital :)
[00:14] <zaheer> as far as we know anyway...
[00:14] <zaheer> wtay: should i therefore comment out the set_caps lines?
[00:14] <zaheer> in _init that is
[00:15] <wtay> zaheer: yes
[00:15] <wtay> I spotted a capsnego error... not sure if it will matter in this case...
[00:16] <taaz> any chance that test cases are being written to check for all these errors so they don't appear again? hint hint
[00:17] <wtay> taaz: hehe, yeah I will... I will...
[00:18] <zaheer> i got a segfault doing the unref i spoke about earlier :)
[00:18] <wtay> zaheer: ok, I got to a segfault now and I know why...
[00:18] <wtay> hmm
[00:19] <zaheer> interesting...i got a segfault in line 257
[00:20] <wtay> zaheer: me too
[00:20] <zaheer> (gdb) print mono_data
[00:20] <zaheer> $1 = (gint16 **) 0x80e18a0
[00:20] <zaheer> (gdb) print *mono_data
[00:20] <zaheer> $2 = (gint16 *) 0x0
[00:20] Action: zaheer thinks....
[00:21] <wtay> capsnego works 100% now I think, osssink is set to mono
[00:21] <zaheer> mono_data = (gint16*)GST_BUFFER_DATA(outbuf);
[00:21] <zaheer> aaaaah
[00:21] Action: zaheer spotted the prob :P
[00:22] <zaheer> stupid mistake
[00:22] <zaheer> doh!
[00:22] <wtay> yeah alloc first then access the mem :-)
[00:22] Action: wtay hands zaheer a BPB
[00:22] <taaz> well... if you were to go by extreme programming guidelines you would write the test that fails cause of the bug, add it to test suite, then fix bug until test suite runs again
[00:22] <wtay> ermm sorry you did that
[00:23] <zaheer> noo
[00:23] <zaheer> its not the alloc
[00:23] Action: wtay takes the BPB from zaheer
[00:23] <zaheer> its im passing the poitner instead of the address to the pointer
[00:24] <wtay> oh, hence the compiler warnings...
[00:24] <zaheer> yah which i skipped over stupidly :)
[00:25] <zaheer> however new problem :)
[00:25] <wtay> Floating point exception
[00:25] <zaheer> this makes no sense....
[00:25] <zaheer> #0  0x404977f6 in gst_stereo2mono_fast_16bit_chain (data=0x80f3160, 
[00:25] <zaheer>     mono_data=0xbffff8b8, numbytes=4608) at stereo2mono.c:257
[00:25] <zaheer> 257	     (*mono_data)[i]=(data[i]+data[i+1])/2;
[00:25] <zaheer> (gdb) print data
[00:25] <zaheer> $5 = (gint16 *) 0x0
[00:25] <zaheer> look at data in #0
[00:26] <zaheer> then look at the result of my print data
[00:26] <wtay> heh
[00:26] <wtay> printf is your friend...
[00:26] <zaheer> dodgy or what :P
[00:27] <zaheer> yes i will use printf :)
[00:27] <wtay> you also need to unref buf when you are done with it in the chain function...
[00:28] <zaheer> and its at i=1612 where it segfaults....
[00:28] <zaheer> not i=0 :)
[00:29] <wtay> It doesnt crash here on that statement...
[00:29] <zaheer> where it should do if data points to 0x0
[00:29] <zaheer> maybe somin else is overwriting the pointer then...
[00:30] <zaheer> s2m 16bit: data=0x80e0698 numbytes=4608
[00:30] <zaheer> s2m 16bit: data=0x80e3008 numbytes=4608
[00:30] <zaheer> s2m 16bit: data=0x80e6008 numbytes=4608
[00:30] <zaheer> s2m 16bit: data=0x80e9008 numbytes=4608
[00:30] <zaheer> s2m 16bit: data=0x80ec008 numbytes=4608
[00:30] <zaheer> s2m 16bit: data=0x80ef008 numbytes=4608
[00:30] <zaheer> s2m 16bit: data=0x80f3160 numbytes=4608
[00:30] <zaheer> Segmentation fault (core dumped)
[00:30] Action: zaheer thinks its to do with mono_data
[00:31] <zaheer> thats only thing im writing to...
[00:31] <wtay> hmm no idea..
[00:31] <wtay> doesn't crash here...
[00:31] <zaheer> how does one allocate a buffer?
[00:31] <wtay> g_malloc
[00:31] <zaheer> i just did:
[00:31] <zaheer> GST_BUFFER_DATA(outbuf) = g_new(gint16,GST_BUFFER_SIZE(buf)/4);
[00:31] <zaheer>   GST_BUFFER_SIZE(outbuf) = GST_BUFFER_SIZE(buf)/2;
[00:31] <zaheer>   mono_data = (gint16*)GST_BUFFER_DATA(outbuf);
[00:32] <zaheer> ouch
[00:32] <zaheer> divided by 4 :P
[00:32] <zaheer> should be by 2 :)
[00:32] <zaheer> no
[00:32] <zaheer> it is by 4
[00:33] <zaheer> coz its gint16's not gint8's
[00:33] <wtay> I was looking at that too... try with /2
[00:33] <zaheer> GST_BUFFER_SIZE returns bytes...
[00:33] <wtay> yes
[00:33] <zaheer> so g_new is allocating space for numbytes/4 gint16's
[00:33] <zaheer> which is correct
[00:34] <vektor> all these g's make me nervous. :)
[00:34] <wtay> so it seems if should alloc sizeof(guint16)*size..
[00:34] <wtay> zaheer: just divide by 2 to make sure the problem is not there...
[00:35] <zaheer> ok will do :)
[00:35] <zaheer> vektor: welcome to glib :)
[00:35] <zaheer> it works now
[00:35] <zaheer> so strange :P
[00:35] <wtay> zaheer: look at line 255 :)
[00:36] <wtay> by how many do you increment i?
[00:36] <zaheer> yes
[00:36] <zaheer> stupid :P
[00:36] <wtay> isn't that an array index to mono_data too?
[00:36] <wtay> :)
[00:36] <zaheer> yah ive just realised
[00:36] <zaheer> what a mockery ive made of myself :P
[00:37] <wtay> for once it's not me who's having to wear the BPB :)
[00:37] <zaheer> hehe
[00:38] <zaheer> thats why i got the problem
[00:38] <zaheer> not the alloc :P
[00:38] <wtay> ok works perfect now here
[00:38] <zaheer> i added a j
[00:38] <wtay> me too
[00:38] <zaheer> counter
[00:38] <zaheer> can you hear it mono? :)
[00:38] <wtay> yes!
[00:39] <zaheer> cool :)
[00:39] <zaheer> does it sound ok? :)
[00:39] <wtay> perfect
[00:39] Action: zaheer doesnt have speakers to test here :)
[00:39] <wtay> oops
[00:40] <zaheer> i was gonna test on my laptop once i had the prob fixed :)
[00:41] <zaheer> i have to unref buf right?
[00:41] Action: wtay is listening to a stereo mp3 converted to mono for maximun audio experience.. :)
[00:41] <wtay> zaheer: yes
[00:41] <wtay> ok something else...
[00:42] <wtay> before the caps are sent to the nego function, it is checked against the padtemplate so the (gst_caps_get_int(*caps,"channels")!=1 check is not needed
[00:43] <zaheer> ok
[00:43] <zaheer> so the padtemplate is always checked first, right?
[00:43] <wtay> yes
[00:43] <zaheer> before even bothering to nego
[00:43] <wtay> else the _get functions on the caps might fail
[00:43] <zaheer> makes life easier :)
[00:44] <taaz> does this beast just handle stereo?  or is it a generic N-to-1 component?
[00:44] <wtay> capsnego is hard enough allready
[00:44] <zaheer> taaz: only stereo at mo...
[00:44] <wtay> taaz: just stereo
[00:44] <zaheer> to do many channels is trivial addition...
[00:44] <taaz> that doesnt sound too challenging ;)
[00:44] <zaheer> but i want stereo2mono working first
[00:44] <wtay> very trivial to add indeed
[00:44] <zaheer> taaz: its the first capsnego plugin i tried
[00:45] <zaheer> theres about 6 real lines of code, the rest is to make it a plugin or to handle caps :P
[00:45] <zaheer> what is nontrivial however is to do arbitrary widths...
[00:46] <zaheer> i have fast functions for 16bit and 8bit widths
[00:46] <zaheer> i need a slow generic one for others
[00:47] Action: zaheer thinks he should check signedness and unsignedness
[00:47] Action: zaheer thinks he should also check endianness
[00:47] <wtay> zaheer: think so?
[00:48] <zaheer> yah i wanna be able to handle non standard widths...
[00:48] <wtay> the capsnego function could even be made very short in this case
[00:48] <wtay> just adjust the channels property and return the result of _proxy
[00:48] ChiefHighwater (paul at temple-baptist.com) left irc: 
[00:49] <wtay> store the last encountered channels for future use...
[00:49] <wtay> the idea is that reffing/unreffing the caps is not needed, it'll use copy_on_write and stuff
[00:50] <zaheer> so i dont need to change the caps passed in?
[00:50] <zaheer> for _TRY
[00:51] <wtay> lemme create me ideal solution...
[00:51] <wtay> s/me/my
[00:52] <zaheer> ok :)
[00:52] <wtay>   if (*caps==NULL) 
[00:52] <wtay>     return GST_PAD_NEGOTIATE_FAIL;
[00:52] <wtay>   gst_caps_set(*caps,"channels",GST_PROPS_INT(1));
[00:52] <wtay>   return gst_pad_negotiate_proxy(pad,stereo->srcpad,&tempcaps,counter);
[00:54] <zaheer> you mean i dont need to copy the caps like i have done?
[00:54] <wtay> return gst_pad_negotiate_proxy(pad,stereo->srcpad,caps,counter);
[00:54] <wtay> that's the correct line
[00:55] <wtay> zaheer: no need to, just modify it and let the peer elements negotiate the adjusted caps
[00:55] <zaheer> ok cool
[00:55] <wtay> and it work!
[00:55] <zaheer> i didnt wana take the risk of affecting it :)
[00:55] <wtay> s/work/works
[00:55] <zaheer> ok im changing my copy :)
[00:55] <wtay> zaheer: the api should copy_on_write it automagically
[00:56] <zaheer> aaah ok cool
[00:56] <zaheer> should = does now or will do soon? :)
[00:56] <wtay> reffing/unreffing is only needed if you want to keep a pointer to it yourself
[00:56] <wtay> does
[00:56] <wtay> I tried it
[00:56] <zaheer> perfect :)
[00:56] <zaheer> thats mmakes the nego functions a cpla lines long :)
[00:57] <wtay> 4 :-)
[00:57] <zaheer> yah
[00:57] <zaheer> 4 as opposed to my 197 :)
[00:57] <wtay> lol
[00:59] <wtay> you could even do with just one function and switch based on the input pad...
[00:59] <zaheer> yah true :P
[00:59] <wtay> if (pad == s2m->srcpad) etc..
[00:59] <zaheer> ill leave it as 2 for the moment :P
[01:00] <wtay> sure, for future enhancements... <g>
[01:00] Action: zaheer is cleaning up the stereo2mono code
[01:00] <zaheer> i wish i knew it was this easy :)
[01:00] <wtay> actually you don't need capsnego...
[01:01] <zaheer> i just have pad templates? :)
[01:01] <zaheer> how do you do an undo in vi?
[01:01] <wtay> and set the caps of the pads, the peer elements will try to adjust 
[01:01] <zaheer> even with incomplete caps?
[01:01] <wtay> ah no... sorry
[01:02] <wtay> you're right
[01:02] <zaheer> you know the undo keystroke in vi? :)
[01:02] <wtay> you need to pass the unmodified ones too
[01:02] <wtay> you
[01:02] <wtay>   you
[01:02] <wtay> damn
[01:02] <zaheer> "u" ?
[01:02] <wtay> yes
[01:02] <zaheer> thx :)
[01:02] <wtay> autocompletion misery
[01:03] <zaheer> hehe
[01:04] <wtay> you can even remove the stereo2mono_sink_caps and src caps...
[01:04] <zaheer> i have done :)
[01:05] matth (matth at qwest.dsplinux.net) left irc: Ping timeout for matth[qwest.dsplinux.net]
[01:05] <wtay> hehe this is a neat plugin :-)
[01:06] <zaheer> its a good one to learn from :)
[01:06] <wtay> yes
[01:06] <zaheer> thats why i did it first
[01:06] <wtay>  width=gst_caps_get_int(GST_PAD_CAPS(pad),"width");
[01:06] <wtay> hehe, cool
[01:06] <wtay> never done before
[01:07] <zaheer> i get a core....
[01:07] <wtay> :(
[01:07] <zaheer> #0  0x403003ad in gst_props_check_compatibility (fromprops=0x80bcb70, 
[01:07] <zaheer>     toprops=0x8076aec) at gstprops.c:796
[01:07] <zaheer> 796	    entry1 = (GstPropsEntry *)sourcelist->data;
[01:07] <zaheer> (gdb) bt
[01:07] <zaheer> #0  0x403003ad in gst_props_check_compatibility (fromprops=0x80bcb70, 
[01:07] <zaheer>     toprops=0x8076aec) at gstprops.c:796
[01:07] <zaheer> #1  0x402f1c39 in gst_caps_check_compatibility_func (fromcaps=0xbffff8b8, 
[01:07] <zaheer>     tocaps=0x80ae638) at gstcaps.c:513
[01:07] <zaheer> #2  0x402f1e9e in gst_caps_check_compatibility (fromcaps=0xbffff8b8, 
[01:07] <zaheer>     tocaps=0x80ae638) at gstcaps.c:560
[01:07] <zaheer> #3  0x402fabad in gst_pad_renegotiate_func (pad=0x80bcb70, peerpad=0x80b1668, 
[01:07] <zaheer>     newcaps=0xbffff86c, counter=0xbffff840) at gstpad.c:960
[01:07] <zaheer> #4  0x402fb8b6 in gst_pad_negotiate_proxy (srcpad=0x80bce30, 
[01:07] <zaheer>     destpad=0x80b13b8, caps=0xbffff86c, counter=0) at gstpad.c:1101
[01:08] <zaheer> #5  0x40497641 in stereo2mono_negotiate_sink (pad=0x80bce30, caps=0xbffff8b8, 
[01:08] <zaheer>     counter=1) at stereo2mono.c:97
[01:08] <zaheer> sorry for the flood
[01:08] <wtay> better add the copy_on_write statement too then...
[01:08] <zaheer> do i need to cvs update?
[01:08] <wtay>   *caps = gst_copy_on_write (*caps);
[01:08] <wtay>   gst_caps_set(*caps,"channels",GST_PROPS_INT(1));
[01:08] <zaheer> ok :)
[01:08] <wtay> when was you last update?
[01:08] <zaheer> earlier today
[01:08] <wtay> should be ok
[01:09] <zaheer> it might not be...
[01:09] <wtay> looks like my pc is very stable on bogus mem access...
[01:09] <zaheer> that was my laptop :)
[01:09] <zaheer> yah..
[01:09] <zaheer> what kernel you running?
[01:09] <wtay> a typo:  *caps = gst_caps_copy_on_write (*caps);
[01:10] <wtay> 2.4.1
[01:10] <zaheer> wow lots of changes :)
[01:10] <zaheer> im running 2.4.2 on this one
[01:10] Action: zaheer goes and recompiles...
[01:10] <wtay> I need to work on my uptime :)
[01:11] <wtay> 3days :(
[01:11] <zaheer> hehe
[01:11] <zaheer> [zaheer at odin gst]$ uptime
[01:11] <zaheer>  12:20am  up 23 days,  5:31, 17 users,  load average: 0.82, 0.26, 0.08
[01:11] <zaheer> [zaheer at odin gst]$ uname -a
[01:11] <zaheer> Linux odin.cs.ucl.ac.uk 2.4.2 #1 Mon Feb 26 17:47:06 GMT 2001 i686 unknown
[01:12] <wtay> I had to reboot after apt-get updated libc...
[01:12] <zaheer> i think i had last updated a few days ago...u think ill still need the gst_caps_copy_on_write ?
[01:13] <wtay> not sure, automatic copy on write is not done yet so it depends on where/how the caps was generated
[01:13] <zaheer> ok ill add it in, to be on safe side
[01:13] <wtay> yes
[01:14] <wtay> we can remove it later on when a testsuite has been build
[01:15] <zaheer> i do it before the set, write?
[01:15] <wtay> before the set, yes
[01:15] <wtay>   *caps = gst_caps_copy_on_write (*caps);
[01:15] <wtay>   gst_caps_set(*caps,"channels",GST_PROPS_INT(1));
[01:15] <zaheer> yah in both nego funcs...
[01:16] <wtay> yup
[01:16] <zaheer> really this is like the gstQueue in terms of nego apart from it has one constraint on each side
[01:17] <wtay> yes, very similar indeed
[01:17] <zaheer> the ulaw converter is kinda similar
[01:18] <zaheer> it just requires 16bit int on sink
[01:18] <zaheer> and has 8bit int mulaw on src pad
[01:19] <wtay> other properties are irrelevant too?
[01:19] <zaheer> yep
[01:19] <zaheer> as long as its 16bit int PCM
[01:19] <wtay> audio is never going to be a problem I feel...
[01:19] <zaheer> thats good to hear :)
[01:19] <wtay> dunno what will happen with floats and stuff though
[01:20] <zaheer> just need a converter
[01:20] <zaheer> for both ways
[01:20] <wtay> yeah
[01:21] <wtay> something else..
[01:21] <zaheer> take this scenario
[01:21] <zaheer> if the app knows, that what is coming from rtpsrc is mu law x sample rate y numchannels
[01:21] <zaheer> can it set the caps of the rtpsrc src pad?
[01:22] <wtay> yes, although I was thinking of a property on the element
[01:23] <wtay> setting the caps like that would bypass any element logic
[01:23] <wtay> setting it on the pad I mean
[01:23] <zaheer> yah, let it set a property on the element
[01:23] <zaheer> thats better
[01:24] <zaheer> then the element can set its src caps accordingly
[01:24] <wtay> I think so, it can also check to see if the caps make sense etc...
[01:24] <zaheer> yah
[01:25] <zaheer> that is cool
[01:26] <wtay> I should also think about adding a hook in the negotiation process so the app can interfere in some way...
[01:26] <zaheer> therefore gstreamer will soon be in use in a live environment converting voicemail msgs into mp3 format :)
[01:26] <dobey> so, how am i supposed to debug gstreamer, or are you doing that now?
[01:27] <wtay> dobey: I'm wondering why the --gst-info-maks=-1 parameter doesn't get parsed...
[01:27] <zaheer> if you spot bugs notify the list :)
[01:28] <wtay> dobey: I propose to add a simple g_print in gst.c on every argument it parses...
[01:28] <dobey> wtay: it printed info
[01:28] <wtay> dobey: in gstmediaplay it didn't, you told me
[01:28] Action: zaheer thinks he had this problem a cpla weeks ago...
[01:29] <dobey> wtay: what? it printed it to console, well, it did once
[01:29] <wtay> dobey: the first INFO lines are always printed, the log you sent me didn't have INFO lines after the first ones
[01:30] <wtay> so I figured the -info-mask didn't work too well..
[01:30] <dobey> wtay: the last one i did (test.log)
[01:31] <wtay> dobey: that was the log of gstreamer-register, no?
[01:32] <dobey> no
[01:32] <dobey> not afaict anyway
[01:32] <wtay> oh
[01:33] <wtay> there is no sign of any pipeline setup, just plugin loading AFAICS
[01:34] <dobey> oh
[01:34] <zaheer> wtay:
[01:34] <zaheer> ** ERROR **: compiled without Xvideo extention support
[01:34] <zaheer> aborting...
[01:34] <zaheer> on helloworld.c ?
[01:35] Action: zaheer gets a bit worried...
[01:35] <wtay> zaheer: hum
[01:36] <wtay> zaheer: oh yeah.. I don't check for Xv in the new videosink.. sec...
[01:36] Action: zaheer wonders how it came to affect an audio app :)
[01:37] <wtay> zaheer: it's because all plugins are loaded (gstreamer-register was not run)
[01:37] <zaheer> aaah
[01:37] <wtay> anyway running gstreamer-register will cause the same probs
[01:37] <wtay> so it need urgent fixing
[01:38] <zaheer> ok :)
[01:39] Action: zaheer starts setting up his laptop to put the new stereo2mono plugin in
[01:45] <wtay> zaheer: fix in CVS
[01:46] Action: zaheer is updating :)
[01:46] <dobey> bbl
[01:46] dobey (dobey at dreadnought.ximian.com) left #gstreamer.
[01:46] Action: zaheer is recompiling
[01:46] <zaheer> gstximage.c:78: macro `GST_PROPS_FOURCC' used with just one arg
[01:46] <zaheer> make[3]: *** [gstximage.lo] Error 1
[01:47] <wtay> it should give a warning now
[01:47] <wtay> zaheer: yup, I changed the FOURCC API a bit
[01:47] <zaheer> thats gstximage.c ....
[01:48] <wtay> yes
[01:48] <zaheer> have i not update'd some file?
[01:48] <zaheer> to get that error?
[01:48] <wtay> humm.. maybe I didn't commit that yet...
[01:48] <wtay> ok my fault.. hold on
[01:49] Action: zaheer smiles
[01:49] <wtay> in CVS
[01:50] <zaheer> wow quite a few files there..
[01:50] <wtay> and there goes our helloworld.c example too...
[01:50] <wtay> 9
[01:50] <wtay> the plugins using a FOURCC in their caps...
[01:51] <zaheer> hehehe, ppl will have mp3's in mono if they use helloworld :)
[01:52] <wtay> will correct that with a next update...
[01:52] <zaheer> i wanna start working on glib2/gobject stuff for gstreamer on monday
[01:52] <wtay> oh cool
[01:52] <zaheer> maybe have somin working for GUADEC in glib2
[01:53] <wtay> are you going?
[01:53] <zaheer> 80% chance
[01:53] <zaheer> that I am going
[01:53] <zaheer> i have registered to go, my flights are provisionally booked
[01:53] <wtay> cool
[01:53] <zaheer> i need to confirm a cpla things here, then i can make that a 100%
[01:54] <zaheer> and book the accomodation
[01:54] <zaheer> lots of compilation happening :)
[01:55] <zaheer> maybe gstreamer will become part of the standard gnome distribution come gnome 2.0 :)
[01:55] <wtay> yeah I heard :)
[01:55] <wtay> TOP Hotel Hebron, Helgolandsgade 4, DK-1653 Copenhagen V
[01:55] <zaheer> Hotel Hebron?
[01:55] <wtay> in case you are wondering where Erik and I reside
[01:55] <wtay> yes
[01:55] <zaheer> is that where youre satying? :)
[01:55] <wtay> yes
[01:55] Action: zaheer will look that up
[01:56] <wtay>  www.hebron.dk
[01:56] Action: zaheer looks
[01:57] <zaheer> you staying there friday night, saturday night and sunday night?
[01:58] <wtay> 4 nights
[01:58] <wtay> 5-9
[01:58] <wtay> I'll arrive on thursday
[01:59] <zaheer> ok
[01:59] <zaheer> you lleaving on monday?
[01:59] Action: zaheer wonders why the compile is taking so long...must be a major header file that was updated :)
[02:00] <wtay> tuesday or wednesday
[02:00] <wtay> zaheer: yup gstcaps.h :)
[02:00] <zaheer> hehe
[02:00] <wtay> most likely tuesday...
[02:00] <zaheer> unfirtunately i wont have the luxury of an extended stay :)
[02:00] <zaheer> il prob be there saturday - monday
[02:00] zaheer (zaheer at tnt-16-171.easynet.co.uk) left irc: [x]chat
[02:01] zaheer (zaheer at tnt-16-171.easynet.co.uk) joined #gstreamer.
[02:01] <zaheer> hmmm
[02:01] <wtay> whup
[02:01] <zaheer> i musta pressed the x :)
[02:01] <zaheer> caps.c:28: macro `GST_PROPS_FOURCC' used with too many (4) args
[02:01] <zaheer> caps.c:50: macro `GST_PROPS_FOURCC' used with too many (4) args
[02:01] <zaheer> make[3]: *** [caps.o] Error 1
[02:02] <wtay> where?
[02:02] <zaheer> tests
[02:02] <zaheer> :)
[02:02] <wtay> sec... I must have missed that one...
[02:02] <zaheer> GST_PROPS_FOURCC ('Y','V','1','2'),
[02:03] <zaheer> GST_PROPS_FOURCC ('Y','U','Y','V')
[02:04] <zaheer> are a lot of the gstreamer "gang" going?
[02:04] <zaheer> i know of hadess, you and omega
[02:05] <wtay> Uraes
[02:05] <zaheer> aah yes Uraeus :)
[02:07] <wtay> in cvs
[02:07] <wtay> doing a complete rebuild to make sure...
[02:09] <wtay> ok compiles fine now
[02:09] <zaheer> compiling....
[02:09] <zaheer> eeew
[02:09] <zaheer> helloworld.c :)
[02:10] <zaheer> it didnt merge properly with my changes :)
[02:10] <wtay> oooh, severe conflics probably :-)
[02:10] Action: zaheer rm -f'ed his helloworld.c :)
[02:12] <zaheer> would autoplug use stereo2mono automatically if required, now?
[02:12] <wtay> nope
[02:13] <wtay> it only operates on padtemplates so...
[02:13] <wtay> and the audiosink can handle stereo
[02:13] <wtay> ermm
[02:13] <zaheer> but if a sink requires mono....
[02:13] <wtay> ok, then it will be used I guess
[02:14] <wtay> but you'll have to make sure it also supports mono input
[02:14] <zaheer> yes
[02:14] <wtay> which is a nop the of course
[02:14] Action: zaheer is getting a core...
[02:15] Action: zaheer wonders why
[02:15] <wtay> where?
[02:15] Action: zaheer is in gsb...
[02:15] <zaheer> gdb that is
[02:15] <zaheer> still reading sumbols
[02:15] <zaheer> same place by looks
[02:15] <wtay> unref?
[02:15] <zaheer> #0  0x403004cd in gst_props_check_compatibility (fromprops=0x80b1338, 
[02:15] <zaheer>     toprops=0x8076ae4) at gstprops.c:796
[02:15] <zaheer> 796	    entry1 = (GstPropsEntry *)sourcelist->data;
[02:15] <zaheer> (gdb) bt
[02:16] <zaheer> #0  0x403004cd in gst_props_check_compatibility (fromprops=0x80b1338, 
[02:16] <zaheer>     toprops=0x8076ae4) at gstprops.c:796
[02:16] <zaheer> #1  0x402f1c39 in gst_caps_check_compatibility_func (fromcaps=0xbffff8a8, 
[02:16] <zaheer>     tocaps=0x80ae630) at gstcaps.c:517
[02:16] <zaheer> #2  0x402f1e9e in gst_caps_check_compatibility (fromcaps=0xbffff8a8, 
[02:16] <zaheer>     tocaps=0x80ae630) at gstcaps.c:564
[02:16] <zaheer> #3  0x402facd1 in gst_pad_renegotiate_func (pad=0x80b1338, peerpad=0x80bce60, 
[02:16] <zaheer>     newcaps=0xbffff85c, counter=0xbffff82c) at gstpad.c:972
[02:16] <zaheer> #4  0x402fb9d0 in gst_pad_negotiate_proxy (srcpad=0x80b15f8, 
[02:16] <zaheer>     destpad=0x80bcbb0, caps=0xbffff85c, counter=0) at gstpad.c:1112
[02:16] <zaheer> #5  0x404976a2 in stereo2mono_negotiate_sink (pad=0x80b15f8, caps=0xbffff8a8, 
[02:16] <zaheer>     counter=1) at stereo2mono.c:99
[02:16] Action: zaheer has the copy_on_write stuff in...
[02:16] <wtay> well...
[02:17] <wtay> make that a _copy then...
[02:17] <zaheer> i should make a copy..
[02:18] Action: wtay should sort out the refcounting RSN
[02:18] <wtay> I suggest making a copy...
[02:18] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[02:18] <Uraeus> hi
[02:18] <zaheer> yo uraeus
[02:18] <wtay> hi Uraeus
[02:19] <Uraeus> I am a bit depressed today ;(
[02:19] <zaheer> whys that? :(
[02:19] <wtay> ? :(
[02:19] <Uraeus> the girl I ahave been pursuing the last weeks left the part with another guy
[02:19] <wtay> ouch :(
[02:19] <zaheer> sorry man....
[02:21] <Uraeus> yup, story of my life 
[02:22] <Uraeus> actually the story of my life has usually been more hopeless, since the last two girls I have been dating actually lived with someone else
[02:22] Action: Uraeus must get his act togheter
[02:22] <zaheer> wtay, i think i spotted the bug
[02:22] <wtay> ah
[02:22] <zaheer> my fault as usual
[02:23] <Uraeus> wtay: did you mail the overflow guy?
[02:23] <wtay> yup
[02:23] <wtay> he joined here last night
[02:23] <zaheer> he came and spoke to omega i believe
[02:23] <Uraeus> cool
[02:23] <wtay> he very enthousiastic AFAICT
[02:24] <zaheer> yah its working!!! :)
[02:24] <wtay> overflow is a very valuable addition to gstreamer
[02:24] <zaheer> nothing to do with the caps stuff :P
[02:24] <wtay> and so is gstreamer to overflow
[02:24] <wtay> zaheer: cool
[02:25] Action: zaheer passed a *** instead of a **
[02:25] <zaheer> ok should i commit my stereo2mono stuff?
[02:26] <wtay> sure
[02:27] Action: wtay is tired...
[02:30] Action: zaheer is getting a stupid error from cvs server
[02:30] <zaheer> cvs server: cannot add file on non-branch tag HEAD
[02:31] Last message repeated 2 time(s).
[02:31] <greg_> greg wants no more SIGSEGV tonight - going to bed now
[02:31] Action: zaheer has to drive to airport in 3 hrs :P
[02:32] <zaheer> i wonder why cvs is being bitchy to me tonight :)
[02:33] greg_ (greg at home.sente.pl) left irc: see ya tommorow
[02:33] <wtay> zaheer: doesn't work?
[02:34] <zaheer> it wont let me add file
[02:34] <wtay> --> Using per-directory sticky tag `HEAD'
[02:34] <zaheer> however its fine adding directory
[02:34] <wtay> that was the extra log message when you added that dir...
[02:35] <wtay> did you checkout with -r HEAD or something?
[02:35] <zaheer> yah i did
[02:35] <zaheer> i think
[02:35] <wtay> maybe that's the problem then...
[02:35] <zaheer> ok
[02:35] <wtay> dunno..
[02:35] <zaheer> ill copy it to one i co'ed with no tag
[02:35] <wtay> yup
[02:36] <Uraeus> well, time for sleep, see ya
[02:37] Uraeus (cschalle at c224s9h5.upc.chello.no) left #gstreamer.
[02:45] <zaheer> well its committed :)
[02:45] Action: zaheer is gonna test a rebuild of newly committed tree be4 going to sleep
[02:46] <wtay> zaheer: no configure.in changes?
[02:46] <zaheer> i did
[02:46] <taaz> hmm.. if you make a N-to-1 version maybe change name to multi2mono
[02:46] <wtay> or multi2multi :)
[02:46] <zaheer> configure.in and plugins/filters/Makefile.am
[02:47] <zaheer> multi2multi :)
[02:47] <zaheer> channel-reducer :)
[02:47] <wtay> ueber mixer
[02:48] <wtay> zaheer: shall I add something to configure.in or do you have something to commit?
[02:48] <zaheer> i committed something i believe
[02:48] <zaheer> aaargh
[02:48] <zaheer> it failed :)
[02:48] <wtay> eh?
[02:48] <zaheer> ill do it in a sec :)
[02:49] <zaheer> committed :)
[02:49] jmv (jmv at joined #gstreamer.
[02:50] <wtay> yo
[02:50] <zaheer> configure.in version 1.124 :)
[02:50] <zaheer> thats a sign of a healthy project :)
[02:51] <wtay> heh
[02:52] <wtay> or a sign that the developers know nothing about autoconf and hence have to try 5 times before it works...
[02:52] <zaheer> at least it works :)
[02:53] <zaheer> i have seen projects with the most archaic config scripts :)
[02:53] <wtay> INFO:gst_plugin_load_all:235: loaded 57 plugins with 75 elements and 10 types
[02:53] <wtay> 75 plugins!!
[02:54] <zaheer> its cool eh :)
[02:54] Action: zaheer will have another cpl in by end of weekend :)
[02:54] <wtay> what bothers me is that the s2m plugin doesn't show the right padtemplates in -inspect...
[02:54] <zaheer> what does it show?
[02:55] <wtay> nothing
[02:55] <wtay> I know what the prob is though...
[02:55] <zaheer> is that coz the padtemplates are set in _init not statically?
[02:55] <wtay> you should add the padtemplate to the factory...
[02:56] <wtay> instead of creating them in _init..
[02:57] <wtay> do you care if I correct that?
[02:57] Action: taaz debates working on porting oms to gstreamer or python bindings... hmm.
[02:57] <zaheer> gst_elementfactory_add_padtemplate
[02:57] <zaheer> ?
[02:57] <wtay> yup
[02:57] <zaheer> go for it :)
[02:57] <wtay> ok
[02:57] <wtay> taaz: python bindings!
[02:58] <taaz> ok
[02:58] <zaheer> using gstreamer from python would be cool :)
[03:02] <wtay> there
[03:05] <zaheer> ok cool :)
[03:05] <wtay> I'm now keeping my hands of your puppy :)
[03:05] <zaheer> hehe
[03:06] <zaheer> is there a standard tag to specify whether its pcm or mulaw or alaw?
[03:06] <zaheer> tag = property name :)
[03:09] <wtay> hmm yeah althoug I don't really agree with it..
[03:09] <wtay> it called "law"
[03:09] <zaheer> how do you think it should be done?
[03:09] <wtay>  law  - The law used to describe the data. This is an integer for which there are three valid values: 0 for linear, 1 for mu law, 2 for A law.
[03:10] <wtay> does signed endianness, depth, width etc make sense for those types?
[03:10] <zaheer> so thats what id be needing for my mu law converter
[03:10] <zaheer> yes they do make sense
[03:11] <wtay> ok then I agree :)
[03:11] <zaheer> oh hold
[03:11] <zaheer> mu law and A law are always 8 bit
[03:11] <wtay> hmm
[03:12] <wtay> hehe, I changed it already :)
[03:12] <wtay> "format",       GST_PROPS_STRING ("int"),
[03:12] <wtay> format is now a subtype
[03:12] <wtay> the int case has law/endianness/width/depth etc..
[03:13] <zaheer> yah
[03:13] <wtay> you could create a "mulaw" type and only use end/sign
[03:13] <zaheer> i dont know if sign is appropriate either :)
[03:13] <wtay> then just drop it :)
[03:13] <zaheer> but endianness is required
[03:13] <wtay> ok
[03:13] matth (matth at qwest.dsplinux.net) joined #gstreamer.
[03:13] <wtay> hi
[03:14] <wtay> there still is a "law" property for the "int" format though...
[03:14] <zaheer> you think a mulaw format would be better than a property of format int
[03:15] <wtay> both are still possible (look at gstosssink.c)
[03:16] <wtay> it doesn't matter really, law is then a subsubtype of "int"
[03:16] <zaheer> yah
[03:16] <wtay> and you can drop the width/depth props
[03:17] <wtay> zaheer: are you sure mulaw is only 8bit?
[03:17] <zaheer> so format="int" law=1 
[03:17] <zaheer> wtay: 95%
[03:18] <zaheer> "In a typical mu-law system,
[03:18] <zaheer> linear samples of 14 to 16 bits are companded to 8 bits.
[03:18] <zaheer> "
[03:19] <zaheer> i have yet to see mu law in anything but 8 bits
[03:19] <wtay> format=int, law=1, endianness=G_BYTE_ORDER, signed=1, width=8, depth=8, rate...
[03:19] <wtay> zaheer: I think mulaw 16 bits is also possible then..
[03:20] <zaheer> wtay: i have yet to see a case...
[03:21] <zaheer> 16bit pcm (or 14bits or whatever) gets converted to 8bit mulaw
[03:21] <zaheer> theiretically maybe its possible to have 16bit mu law
[03:21] <zaheer> but i havent seen it "allowed" in any of the algorithms
[03:22] <wtay> ok
[03:22] <zaheer> Typically, mu-law compressed speech is carried in 8-bit samples. It is a companding technqiue. That means that carries more information about the smaller signals than about larger signals.
[03:23] <zaheer> that implies that it is possible to have in other sizes....
[03:23] <wtay> theoretically yes...
[03:24] <wtay> isn't it just a way to encode the delta's?
[03:24] <zaheer> yah
[03:24] <zaheer> btw
[03:24] Action: taaz appologizes... i'm going to work on the oms port now <g>
[03:24] <zaheer> spectrum.c
[03:24] wtay (wim at cable-195-162-214-198.upc.chello.be) left irc: Read error to wtay[cable-195-162-214-198.upc.chello.be]: EOF from client
[03:24] <zaheer> doesnt have the right include paths
[03:25] wtay (wim at cable-195-162-214-198.upc.chello.be) joined #gstreamer.
[03:25] <wtay> hmm, xchat has this weird tendency to close itself upon certain keypresses
[03:25] <zaheer> spectrum.c
[03:25] <zaheer> doesnt have the right include paths
[03:25] <zaheer> for gnome include files
[03:25] <wtay> oh
[03:26] <zaheer> in tests/
[03:26] <zaheer> i mean test/
[03:26] <wtay> hmm
[03:27] <zaheer> gnomesupport.h not found..
[03:27] <zaheer> its in /usr/libs/gnome-libs/include
[03:27] <zaheer> which is -I'd when doing gnome-config --cflags gnome
[03:27] <wtay> strange
[03:28] <zaheer> however it compiled on odin..
[03:28] <zaheer> just on my laptop and my home machine it doesnt
[03:28] <wtay> only with spectrum?
[03:28] <zaheer> yah
[03:29] <taaz> question about ac3dec
[03:29] <taaz> let's say you wanted to do >2 output channels
[03:30] <wtay> yes
[03:30] <zaheer> im gonna hace to sleep in a min...gotta leave house in less than 2 hrs for airport
[03:30] <wtay> you need a new plugin for that
[03:30] <taaz> how would you go about that?  i'm trying to convince someone in #livid to do it in gstreamer instead of with a oms plugin ;)
[03:30] <wtay> just like zaheers new stereo2mono plugin
[03:31] <wtay> brb
[03:32] Nick change: matth -> matth-gone
[03:34] <wtay> back
[03:35] <wtay> taaz: ac3dec would have several output pads or one pad with multichannel audio
[03:36] <wtay> taaz: better one pad with multichannel data
[03:36] <wtay> taaz: that data goes straight to the audiosink or passes a mixer plugin
[03:37] <wtay> when the audiosink is not capable of handling >2 channels that is
[03:37] <zaheer> wtay: spectrum when linking doesnt get linked with gnome libraries
[03:38] <wtay> zaheer: strange, there are other gnome dependent test apps in that dir using the same Makefile.am rules
[03:39] <zaheer> hmmm
[03:39] <zaheer> its the only one that doesnt compile..
[03:39] <zaheer> or link properly
[03:39] Doomer (Doomfreak at 208-58-112-121.c3-0.rdl-ubr1.trpr-rdl.pa.cable.rcn.com) joined #gstreamer.
[03:40] Doomer (Doomfreak at 208-58-112-121.c3-0.rdl-ubr1.trpr-rdl.pa.cable.rcn.com) left #gstreamer.
[03:40] <wtay> zaheer: very weird
[03:41] <zaheer> whats weird is it compiled on odin
[03:41] <zaheer> i lie
[03:42] <wtay> ?
[03:42] <zaheer> it hadnt got round to compiling it on odin :)
[03:42] <wtay> oh
[03:42] <zaheer> it ran out of diskspace :)
[03:43] <wtay> doh
[03:43] <zaheer> so i have rerun make on odin
[03:44] <zaheer> spectrum compiled fine on odin
[03:44] <zaheer> very strange
[03:45] <zaheer> the only difference between odin and here (desktop and laptop)
[03:45] <zaheer> is gnome 1.4 on odin
[03:45] <zaheer> gnome 1.2 on both desktop and laptop
[03:45] <wtay> hmm
[03:47] <zaheer> that should make no difference
[03:47] PartyZan (root at ip84-52.dialup.wplus.net) joined #gstreamer.
[03:47] <zaheer> anyways
[03:47] <zaheer> i must get sleep
[03:47] <wtay> me too
[03:47] <zaheer> before i become a zombie on the road :P
[03:48] <wtay> cya guys ..
[03:48] <zaheer> 1 1/2 hrs before i drive off
[03:48] <zaheer> thx wtay
[03:48] <PartyZan> Hi
[03:48] <zaheer> cya sunday..
[03:48] <wtay> hello
[03:48] <zaheer> bye all
[03:48] <wtay> yup
[03:48] <wtay> zzzzz
[03:48] zaheer (zaheer at tnt-16-171.easynet.co.uk) left irc: sleep
[03:48] <wtay> need to sleep too (3:50am)... l8r
[03:49] Nick change: wtay -> wtay-sleeping
[03:50] <PartyZan> Could I find all libraries needed by gstreamer on the ftp?
[03:57] <taaz> not on the gstreamer site
[04:00] <PartyZan> And where?
[04:01] <taaz> where ever they are from
[04:01] <PartyZan> OK
[04:26] matth_ (matth at qwest.dsplinux.net) joined #gstreamer.
[04:29] matth-gone (matth at qwest.dsplinux.net) got netsplit.
[04:40] matth-gone (matth at qwest.dsplinux.net) got lost in the net-split.
[05:26] PartyZan (root at ip84-52.dialup.wplus.net) left irc: Ping timeout for PartyZan[ip84-52.dialup.wplus.net]
[05:28] PartyZan (root at ip84-52.dialup.wplus.net) joined #gstreamer.
[06:16] chillywilly (baumannd at d165.as6.nwbl0.wi.voyager.net) joined #gstreamer.

More information about the gstreamer-devel mailing list