[gst-devel] Daily IRC logs

wim.taymans at chello.be wim.taymans at chello.be
Fri May 18 06:28:58 CEST 2001

[07:25] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p7-max5.dun.ihug.co.nz]
[07:27] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) joined #gstreamer.
[07:36] Kuroyi_ (rick at ubr-35.82.223.wmelbourne.cfl.rr.com) joined #gstreamer.
[07:38] CHW (paul at temple-baptist.com) got netsplit.
[07:38] wtay-zZz (wim at cable-195-162-214-58.upc.chello.be) got netsplit.
[07:38] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) got netsplit.
[07:38] omega_away (omega at temple-baptist.com) got netsplit.
[07:38] Kuroyi (rick at ubr-35.82.223.wmelbourne.cfl.rr.com) got netsplit.
[07:40] wtay-zZz (wim at cable-195-162-214-58.upc.chello.be) returned to #gstreamer.
[07:40] omega_away (omega at temple-baptist.com) returned to #gstreamer.
[07:40] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) returned to #gstreamer.
[07:40] ChiefHighwater (paul at temple-baptist.com) joined #gstreamer.
[07:49] CHW (paul at temple-baptist.com) got lost in the net-split.
[07:49] Kuroyi (rick at ubr-35.82.223.wmelbourne.cfl.rr.com) got lost in the net-split.
[07:58] walken (foobar at c1583255-a.smateo1.sfba.home.com) joined #gstreamer.
[07:58] <walken> hi
[08:26] arik (arik at sdn-ar-001waseatP161.dialsprint.net) joined #gstreamer.
[08:26] <arik> lo
[08:27] <arik> bbiam
[08:27] arik (arik at sdn-ar-001waseatP161.dialsprint.net) left irc: leaving
[08:34] arik (arik at sdn-ar-001waseatP304.dialsprint.net) joined #gstreamer.
[08:34] <arik> hey
[08:36] <BBB> which is the direction in which functions of a plugin are executed?
[08:36] <BBB> open - init - caps?
[08:36] <BBB> (parse_caps)
[08:36] <BBB> or init - parse_caps - open?
[08:40] steveb (steveb at node1ee05.a2000.nl) joined #gstreamer.
[08:46] Nick change: taaz -> taazzzz
[08:50] omega_away (omega at temple-baptist.com) left irc: [x]chat
[08:54] ChiefHighwater (paul at temple-baptist.com) left irc: Ping timeout for ChiefHighwater[temple-baptist.com]
[09:43] arik (arik at sdn-ar-001waseatP304.dialsprint.net) left irc: Ping timeout for arik[sdn-ar-001waseatP304.dialsprint.net]
[09:48] arik (arik at sdn-ar-001waseatP304.dialsprint.net) joined #gstreamer.
[09:52] arik (arik at sdn-ar-001waseatP304.dialsprint.net) left irc: Ping timeout for arik[sdn-ar-001waseatP304.dialsprint.net]
[09:59] walken (foobar at c1583255-a.smateo1.sfba.home.com) left irc: zzzzZZzZzzzz
[10:02] steveb (steveb at node1ee05.a2000.nl) left irc: [x]chat
[10:19] arik (arik at sdn-ar-001waseatP304.dialsprint.net) joined #gstreamer.
[10:20] <arik> god damnit
[10:20] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p7-max5.dun.ihug.co.nz]
[10:24] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) joined #gstreamer.
[10:26] <BBB> ?!?!?!?!?
[10:26] <BBB> why isn't my libsdlaudiosink loaded?!?!?!?!?
[10:37] steveb (steveb at node1ee0c.a2000.nl) joined #gstreamer.
[10:38] arik (arik at sdn-ar-001waseatP304.dialsprint.net) left irc: Ping timeout for arik[sdn-ar-001waseatP304.dialsprint.net]
[10:42] arik (arik at sdn-ar-001waseatP252.dialsprint.net) joined #gstreamer.
[11:10] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p7-max5.dun.ihug.co.nz]
[11:21] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) joined #gstreamer.
[11:24] <arik> god damnit
[11:25] <arik> that sucked so hard
[11:25] <arik> how does gstreamer exose a bug in automake?
[11:25] <steveb> do you know about the patch to fix it?
[11:26] <arik> yep
[11:26] <arik> it still boggles my mind
[11:29] <steveb> yep
[11:30] <arik> don't think i applied the patch correctly though
[11:30] <arik> cause it is still sucking
[11:30] <steveb> mine is pretty stable - swap doesn't get used
[11:31] <arik> hmm
[11:32] <arik> oh yeah
[11:32] <arik> that helps _alot_
[11:32] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p7-max5.dun.ihug.co.nz]
[11:32] <arik> such a small patch too
[11:35] <steveb> someone made the comment that RMS makes tools that eat all your memory
[11:36] <arik> heh ;-)
[11:36] <arik> it's true too
[11:36] <arik> emacs gobs memory
[11:40] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[11:40] <steveb> yo
[11:40] <Uraeus> hi steveb
[11:41] <Uraeus> hi arik
[11:41] <steveb> what is the link to the roadmap?
[11:41] <Uraeus> http://www.linuxrising.com/files/gstreroadmap2.html
[11:42] <steveb> Could you add this to core...
[11:42] <arik> Uraeus: hey
[11:42] <Uraeus> arik: hope you didn't mind me putting your name in the summary :)
[11:43] <Uraeus> steveb: three dots?
[11:43] <arik> Uraeus: hehe, surprised me ;-)
[11:43] <arik> Uraeus: but no i didn't mind at all
[11:44] <steveb> Dynamic Parameters (red, yellow, green) development just started Steve Baker
[11:44] <Uraeus> steveb: ok
[11:44] <steveb> cheers
[11:47] <Uraeus> steveb: it is now added, thnx
[11:48] <steveb> ta
[11:48] <Uraeus> ok, think I have all known incoming plugins except maybe the DVD hardware one
[11:50] <arik> neeto
[11:50] <Uraeus> arik: I put you up on the mozilla plugin :)
[11:50] <steveb> i should probably claim mono2stereo
[11:51] <arik> Uraeus: oh god ;-)
[11:51] <arik> Uraeus: that means i have to do it
[11:52] <Uraeus> steveb: ok I add you to that
[11:52] <Uraeus> arik :) that was the idea
[11:52] <arik> *grin*
[11:53] <arik> i'm working on the media player am
[11:53] <arik> er atm
[11:53] <arik> i want to be the maintainer of it
[11:53] <Uraeus> arik: according to the roadmap you are :)
[11:53] <arik> Uraeus: oh
[11:53] <Uraeus> steveb: have I missed the mono2stereo plugin in the roadmap or is it the stereo plugin?
[11:53] <arik> neeto ;-)
[11:54] <arik> roadmaps don't lie ;-)
[11:55] Action: arik goes to read it
[11:57] <steveb> Uraeus: it ain't there yet :(
[11:57] <arik> Uraeus: looks good ;-)
[11:58] <Uraeus> steveb: what will the color codes be for that?
[12:00] <steveb> same as stereo2mono
[12:00] <Uraeus> ok
[12:01] <Uraeus> it is now in
[12:02] <arik> now that i've finally ditched incshed1 and moved back to head i can actually use gstmediaplay again ;-)
[12:02] <arik> it's been very much fun adding features without being able to test though
[12:02] <arik> i'll miss it
[12:02] <Uraeus> hehe
[12:02] <arik> *grin*
[12:02] Action: arik reminds himself to ask someone about cvs access at some point
[12:03] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) joined #gstreamer.
[12:03] <arik> ajmitch: wb
[12:03] <Uraeus> hi ajmitch
[12:04] <ajmitch> hi
[12:04] <Uraeus> arik: I think both wtay and omega can set that up for you
[12:04] <arik> Uraeus: *nod*
[12:04] <arik> just need to ask
[12:04] <arik> will do that soon
[12:07] <ajmitch> hmm, i need to figure out this caps nego stuff tomorrow ;)
[12:07] <arik> ;-)
[12:08] <arik> then you can explain it all to me
[12:08] <Uraeus> ajmitch: I thought I should volunteer you to make a GStreamer XScreensaver module :)
[12:08] <arik> hehe
[12:14] <steveb> Phase 1: underpants
[12:14] <arik> still trying to figure out the best way to do full screen play
[12:14] <steveb> http://www.eavel.com/gnome.mp3
[12:14] <Uraeus> arik: use the SDL output plugin?
[12:14] <steveb> (satire)
[12:14] <arik> Uraeus: that's not actually a bad idea
[12:15] <arik> atm i create a new window
[12:15] <arik> that's a popup, the size of the screen, and at 0,0
[12:15] <arik> and then play the movie in a xvideosink in that
[12:15] <arik> and seek to the current offset
[12:15] <arik> clumsy, and atm core dumping
[12:16] <Uraeus> arik: how do Galeon do fullscreen?
[12:17] <arik> Uraeus: prob the same way, should look at it though
[12:17] <arik> it's the way we did it in eog
[12:18] <steveb> how does xawtv do it?
[12:18] <arik> no idea
[12:18] <arik> ;-)
[12:18] <arik> i should do my researcg
[12:22] <Uraeus> Kuroyi_: are you around?
[12:24] <ajmitch> oops, /me should learn to check xchat more often ;)
[12:24] <ajmitch> too busy playing with the gimp to build a website....
[12:24] <arik> heh
[12:24] <Uraeus> ajmitch: since you delayed so long I took that as a yes
[12:24] <arik> hahaahahahahahahahaha
[12:24] <ajmitch> Uraeus: haha, haha, oh funny haha
[12:25] <Uraeus> there a saying which goes: those who keep silent concedes
[12:25] <Uraeus> or something like that, unsure about the translation :)
[12:25] <arik> hehe
[12:25] <ajmitch> hehe
[12:25] <ajmitch> an Xscreensaver module tho? ;)
[12:25] <arik> bwahahaha
[12:25] <Uraeus> yes, so you can have a videoplaylist as a screensaver
[12:25] <ajmitch> what would it do, play videos?
[12:25] <ajmitch> ugh
[12:26] <arik> that actually is kind of cool
[12:26] <ajmitch> i prefer the blank screensaver ;)
[12:26] <Uraeus> arik: maybe you could put such functionality onto the media-player
[12:27] <Uraeus> that would solve your fullscreen problem :)
[12:27] <arik> Uraeus: well, with full-screen play a playlist and loop you kinda have that
[12:27] <arik> ;-P
[12:27] <arik> Uraeus: i'll put that on my feature last, right after cup-holders ;-)
[12:28] <ajmitch> arik: oh, the cdrom eject feature? ;)
[12:28] <Uraeus> arik: you are making a cup-holder cool :)
[12:28] <Uraeus> arik: I hope you are adding functionality to choose output plugins, cause that is something I miss
[12:29] <arik> ajmitch: ;-)
[12:29] <arik> Uraeus: that's a good idea
[12:29] <arik> right now i need to figure out why gst_play_stop is causing a core dump ;-)
[12:29] <arik> ah screw it
[12:30] <arik> i'll rewrite full-screen again
[12:30] <arik> prob using the sdl plugin
[12:30] <arik> or something else
[12:30] <ajmitch> Uraeus: what do you think of a generic cd ripping app? i am thinking of extending my (idea & UI only) app from just using the osssrc plugin to using more than that
[12:30] <arik> neeto
[12:30] <arik> beos has that
[12:30] <ajmitch> Uraeus: so then i could use the same software for 'ripping' from tapes as for cds
[12:31] <arik> er nm
[12:31] <arik> kind of like grip but for any kind of media
[12:31] <arik> ?
[12:31] <ajmitch> yup
[12:31] <arik> neet
[12:31] <arik> damn full-screen!
[12:31] <ajmitch> i've got about 50 tapes (90min) that i want to copy onto my computer & compress & put on cd & redistribute
[12:32] <arik> *nods*
[12:32] <ajmitch> church talks ;)
[12:32] <Uraeus> ajmitch: sounds great
[12:32] <arik> heh
[12:32] <ajmitch> Uraeus: why do you think i was going to look at things like the vorbisenc plugin? ;)
[12:32] <Uraeus> ajmitch: I don't understand your motivation for wanting those church talks but ok :)
[12:32] <ajmitch> hehe
[12:32] <Uraeus> ajmitch: casuse you said so
[12:33] <Uraeus> ajmitch: when we discussed it last time you said you wanted to look at the vorbis stuff
[12:33] Action: ajmitch guesses Uraeus is not a deeply religious person ;)
[12:33] Nick change: arik -> arik|afk
[12:33] <ajmitch> yeah, i said i was gonna look at plugins that i would be using
[12:33] <Uraeus> dada
[12:34] Action: Uraeus notes that ajmitch guessed right, when younger he was even anti-religous :)
[12:35] <ajmitch> hehe
[12:37] <Uraeus> the only time these days I discuss religion is when drunk and we need a good topic to be loud about :)
[12:37] <ajmitch> lol
[12:37] <arik|afk> Uraeus: where are you from?
[12:37] <ajmitch> what do you think we spend the time doing in #freedevelopers? ;)
[12:37] <Uraeus> Oslo, Norwat
[12:37] <arik|afk> ah
[12:37] <arik|afk> neet
[12:37] <Uraeus> err Norway that is
[12:37] <arik|afk> hehe
[12:37] Action: ajmitch is just you average insane kiwi
[12:37] <arik|afk> hehehehe
[12:37] <arik|afk> arent they all insane?
[12:37] <arik|afk> ;-P
[12:37] <Uraeus> and arik is a California resident, now home visiting his parents :) 
[12:38] <Uraeus> according to Advogato that is
[12:38] <ajmitch> almost all, i think if you look hard enough there are some that are just plain nuts
[12:38] <ajmitch> hehe
[12:38] Action: ajmitch has to remember how to make a windows boot disk
[12:39] <Uraeus> ajmitch: think you can choose to do so from the explorer file menu
[12:39] <ajmitch> Uraeus: yeah, but this is on a win95 system ;)
[12:39] <ajmitch> Uraeus: parent's computer has stuffed up, grub just sits at a command prompt now
[12:39] <Uraeus> heh
[12:40] <ajmitch> so it's easier to do fdisk /mbr than to try & fix grub (at least until i get to the computer...)
[12:40] <arik|afk> Uraeus: correct-ish
[12:40] Action: ajmitch has his parents using kde 2 at times ;)
[12:41] <arik|afk> ewww
[12:41] <arik|afk> ;-P
[12:41] <Uraeus> kill -9 ajmitch
[12:41] <ajmitch> hey, it was either that or gnome 1.2 at the time...
[12:41] <ajmitch> they preferred kde
[12:41] <ajmitch> more windows-like for them ;)
[12:41] <Uraeus> hey cool, I now have even more credit on my card, more like that and I am ready to go bankrupt bigtime
[12:42] <Uraeus> ajmitch: give them nautilus they will probably like it
[12:42] <ajmitch> Uraeus: sure, on a p200 with 32MB ram....
[12:42] <Uraeus> ajmitch: ok, give em more ram first :)
[12:43] Nick change: arik|afk -> arik|food
[12:43] <arik|food> bbbiab
[12:43] <ajmitch> Uraeus: can't, it's a crappy fujitsu, and they can't afford it ;)
[12:43] <ajmitch> i prefer to use all the ram in this machine ;)
[12:43] <Uraeus> ajmitch: actually I not so sure it if Natilus can't run with all bells and whistles turned off
[12:44] <Uraeus> the AA stuff for instance takes a lot of juice if I remember correctly
[12:44] <arik|food> you can turn if off
[12:44] <ajmitch> i should try 1.0.3 before complaining, but 1.0 was slow on this computer (400Mhz, 384MB ram), with that stuff turned off
[12:44] <ajmitch> X, for some reason, slows down to a crawl on here at times
[12:44] <ajmitch> and uses 25% cpu time
[12:47] Action: ajmitch starts 1.0.3
[12:47] <ajmitch> no doubt it'll take over my desktop ;)
[12:47] <Uraeus> 1.0.3 made a huge difference on my system
[12:48] <ajmitch> damn, you don't realise how many windows are open until you go to minimise them all ;)
[12:49] <ajmitch> cool, 1.0.3 starts up in under 10 min ;)
[12:52] <ajmitch> wow, it shuts down cleanly for a change ;)
[12:54] <Uraeus> hehe
[12:55] <Uraeus> Well I need to get some food and get cleaned up, it is our national day today 
[12:55] <ajmitch> hehe, ok
[13:17] thomas (thomas at 212-100-172-175.adsl.easynet.be) left irc: Ping timeout for thomas[212-100-172-175.adsl.easynet.be]
[13:17] thomas (thomas at 212-100-172-175.adsl.easynet.be) joined #gstreamer.
[13:29] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p7-max5.dun.ihug.co.nz]
[13:31] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) joined #gstreamer.
[13:37] Nick change: ajmitch -> ajzzzz
[14:07] arik|food (arik at sdn-ar-001waseatP252.dialsprint.net) left irc: Ping timeout for arik|food[sdn-ar-001waseatP252.dialsprint.net]
[14:31] Nick change: Uraeus -> Ur_17may
[14:56] ajzzzz (ajmitch at p7-max5.dun.ihug.co.nz) left irc: Ping timeout for ajzzzz[p7-max5.dun.ihug.co.nz]
[15:15] ajzzzz (ajmitch at p49-max7.dun.ihug.co.nz) joined #gstreamer.
[15:16] dobey (dobey at dreadnought.ximian.com) joined #gstreamer.
[15:26] Nick change: matth_ -> matth-coffee
[15:27] matth-coffee (matth at qwest.dsplinux.net) left irc: [x]chat
[15:27] matth (matth at qwest.dsplinux.net) joined #gstreamer.
[15:27] Nick change: matth -> matth-coffee
[15:29] ajzzzz (ajmitch at p49-max7.dun.ihug.co.nz) left irc: Ping timeout for ajzzzz[p49-max7.dun.ihug.co.nz]
[15:30] big_T (thomas at fysgr763.sn.umu.se) joined #gstreamer.
[15:43] ajzzzz (ajmitch at p49-max7.dun.ihug.co.nz) joined #gstreamer.
[16:10] sienap (synap at ipc379c0d1.dial.wxs.nl) joined #gstreamer.
[16:21] <dobey> ack
[16:26] <sienap> FIN
[16:31] <dobey> i need to do some serious BitchX hacking
[16:44] <sienap> joker
[16:44] <sienap> i discovered something lately
[16:44] <sienap> bitchx sends os / client version in an udp packet to dimension6.bitchx.com
[16:44] <sienap> *undocumentated* yeah use the source luke..
[16:45] <sienap> we tried to get an /. message in
[16:45] <sienap> but negative propoganda :)
[16:45] <dobey> uh-huh
[16:47] <dobey> let's see
[16:47] <sienap> when MS does it is BLADIEBLA MS BLA BLA BLA
[16:48] <sienap> but when linux does it is kept silence
[16:48] <dobey> uhm
[16:48] <sienap> or linux software that is
[16:48] <sienap> i am pretty mad about this
[16:48] <sienap> however
[16:48] <dobey> os/client doesn't much matter
[16:48] <dobey> c:\windows\system.dat does
[16:48] <sienap> hehehe :)
[16:49] <sienap> aah shit happens 
[16:49] <sienap> :)
[16:49] <sienap> it just isn't nice to do
[16:49] <dobey> aim sends client/os to the server
[16:49] <dobey> so does icq, and every other messenger
[16:49] <steveb> the bitchx packet *is* documented
[16:49] <steveb> can't remember where - try the README
[16:50] <dobey> ooh look, you have an old version, would you like to upgrade/
[16:50] <sienap> steveb hmm
[16:50] <sienap> however
[16:51] <steveb> yeah, i know
[16:51] <dobey> Eterm is what's pissing me off
[16:52] <dobey> it doesn't like to redraw properly apparently
[16:53] <dobey> gah
[16:53] <sienap> dobey so what are you doing with gstreame r?
[16:53] <dobey> sienap: nothing atm
[16:54] <sienap> so any plans ?
[16:54] <dobey> i'm trying to get work done, so i can meet fucking deadlines
[16:55] <sienap> what is that work ?
[16:57] <dobey> mandrake and turbolinux are pieces of shit
[16:57] <sienap> ehehe
[16:57] <sienap> packager :)
[16:58] <dobey> schoolboy
[16:59] <dobey> bleh
[16:59] <dobey> at least gnome-terminal redraws properly
[16:59] <dobey> i should write a terminal
[16:59] <dobey> or hack gnome-terminal
[16:59] <dobey> or something
[17:02] <sienap> hehe
[17:02] <sienap> dobey yeah konsole from kde is wayyyyyyy better
[17:02] <dobey> shut up
[17:05] <sienap> he ?
[17:05] <sienap> it just is
[17:05] <sienap> open your eyes dude..
[17:05] <sienap> you should do that more often
[17:06] <dobey> no
[17:06] <dobey> i need to close my eyes, and go back to sleep
[17:07] <dobey> command.com is better too, maybe i should run that
[17:07] <sienap> also a great idea
[17:08] <dobey> me rewriting gnome-terminal wouldn't be a bad idea either
[17:09] <sienap> man dobey get a life
[17:09] <sienap> :)
[17:10] <dobey> don't smile at me
[17:10] <sienap> guess again
[17:10] <sienap> :)
[17:10] Action: dobey kicks sienap in the pants.
[17:12] <dobey> are you going to continue being the asshole you've been lately?
[17:12] <dobey> because i don't have time, or the patience to deal with assholes
[17:13] <sienap> dobey you really should look at your self..
[17:13] <sienap> and not blame other people for everything
[17:13] <sienap> however..
[17:13] <dobey> blame?
[17:13] <dobey> don't pull that shit with me
[17:15] <sienap>  /ignore dobey all
[17:15] <sienap> bladiebla
[17:15] <dobey> don't fake shit with me
[17:15] <dobey> sigh
[17:16] <dobey> all i've seen you do lately is bitch and complain
[17:16] <dobey> i'm not going to deal with it
[17:17] <BBB> uhm........
[17:17] <BBB> dobey
[17:17] <BBB> behave
[17:17] <BBB> please
[17:17] <dobey> i am behaving
[17:18] <BBB> uhm....
[17:18] <BBB> <dobey> don't fake shit with me
[17:18] <BBB> (?)
[17:18] <dobey> uh yeah
[17:18] <dobey> /ignore != /ignore
[17:19] <dobey> err
[17:19] <dobey> /say even
[17:19] <dobey> man
[17:19] <dobey> fuck this
[17:19] <dobey> i don't have time
[17:19] dobey (dobey at dreadnought.ximian.com) left irc: eh
[17:19] <BBB> sienap: is he always like this?
[17:43] big_T (thomas at fysgr763.sn.umu.se) left irc: BitchX: no additives or preservatives
[17:47] big_T (thomas at fysgr763.sn.umu.se) joined #gstreamer.
[17:56] <sienap> backkkkk
[17:56] <sienap> bbb yes :)
[17:56] <sienap> it is typical dobey
[17:56] <sienap> i used to talk with him a lot before but damn :) forget it
[18:09] <BBB> hmm.....
[18:09] <BBB> weird guy :)
[18:09] <BBB> of in het nederlands, die vent is vaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaag
[18:12] <sienap> lol
[18:12] <sienap> die vent heeft te lang geen pussy meer gevoelt
[18:12] <sienap> daar wordtie suicideable van ofzo
[18:12] <sienap> *G*
[18:12] <sienap> :)
[18:12] <sienap> hi big_t
[18:12] <BBB> haha
[18:13] Action: BBB denkt dat die jongen gewoon te gestresst is
[18:13] <BBB> hij moet eens lekker een maandje op vakantie gaan
[18:13] <BBB> ik las dat amerikanen de eerste paar jaar van hun werk *geen* vakantie mogen opnemen
[18:13] <BBB> best wel errug
[18:15] <sienap> mwha
[18:15] <sienap> hij is packager
[18:15] <sienap> bij ximian
[18:15] <sienap> daar kun je wel indeed gestresst van raken
[18:15] <sienap> hele dag rpmtjes maken >:)
[18:15] <BBB> dammit
[18:15] <BBB> dat meen je niet
[18:15] <BBB> wat een ONGELOFELIJK klotewerk lijkt me dat
[18:15] <sienap> jawel
[18:16] <BBB> dan ben ik liever vrijetijds programmeur zonder betaald te krijgen
[18:16] <sienap> issut ook
[18:16] <sienap> hij draaiter wel omheen altijd
[18:16] <sienap> | dobey was (dobey at dreadnought.ximian.com)
[18:16] <BBB> liever snel glukkig sterven dan langzaam ongelukkig dooslijten
[18:16] <sienap> wassie eerst ook
[18:16] <sienap> maarja geld is in america belangerijk dan hier
[18:16] <sienap> bbb agree with that :)
[18:16] Action: BBB had al gezien dattie van ximian was
[18:16] <BBB> maar.... wat doet zo iemand dan?
[18:16] <sienap> heheeh
[18:16] <sienap> mwha
[18:16] <sienap> hij blaat altijd wel van ook coding ook coding
[18:16] <BBB> spec-files schrijven enzo?
[18:16] <sienap> maar hij code maar amper of zelfs niet voor ximian
[18:16] Action: BBB vind spec files ondingen
[18:17] <sienap> weet ik veel :)
[18:17] <sienap> zorgen dat .rpm grote achterstand heeft >:)
[18:17] <BBB> een spec file is de configuratie file van een RPM pakketje :)
[18:17] <BBB> arme hij
[18:17] <BBB> RPMmetjes maken
[18:17] <BBB> de hele dag lang
[18:17] <BBB> daar is toch geen ene f*cking hol aan?
[18:17] Action: BBB zou liever vuilnisman zijn
[18:18] <BBB> ofzo
[18:18] <sienap> hehhe
[18:19] <sienap> vuilnis man he
[18:19] <sienap> dat dokt nog behoorlijk netjes ook
[18:19] <sienap> :)
[18:19] Action: BBB gaat helpdesk mannetje spelen
[18:19] <BBB> ik moet namelijk toch wat
[18:19] <BBB> dus ben ik volgend jaar helpdesker voor de universiteit
[18:19] <BBB> betaalt 15,-/uur
[18:19] <BBB> moet ik *wigh* windows help geven
[18:19] <BBB> maar ja
[18:20] <BBB> ik moet toch ergens beginne
[18:20] <sienap> lol
[18:20] <sienap> :)
[18:21] <BBB> achja....
[18:21] <BBB> zou jij zoiets ook niet leuk vinden?
[18:21] <BBB> we zoeken nog iemand voor de helpdesk ;)
[18:21] <sienap> lol
[18:22] <sienap> ik snap niks van wind0ze man
[18:22] <sienap> zo ffkes portis head op zetten
[18:22] <sienap> lekker deprimerent :)
[18:24] <BBB> lol
[18:24] <BBB> windoos hoef je niet te snappen
[18:24] <sienap> he
[18:24] <BBB> het is een doos
[18:24] <sienap> :)
[18:24] <BBB> met wind
[18:24] <BBB> windhoos
[18:24] <BBB> ohnee
[18:24] <BBB> wonddoos
[18:24] <BBB> dus windoos
[18:24] <sienap> ahja
[18:24] <BBB> en als ie het niet doet
[18:24] <BBB> dan moet je hem herstarten
[18:24] <BBB> met soms een harde schp erbij
[18:24] <BBB> dan gaat ie OF kapot, of er verandert niks, of hij doet et weer
[18:24] <sienap> en soms formateren ja
[18:25] <BBB> C) iedereen blij
[18:25] <sienap> *just done that*
[18:25] <BBB> B) nothing broken dus alles goed
[18:25] <BBB> A) windoos schuld, dus niet jouw schuld, klaar :)
[18:26] <sienap> ahja
[18:26] <sienap> maarja
[18:27] Action: BBB heeft net een examen gehad
[18:27] <BBB> nog twee nu
[18:27] <BBB> morgen biotechnologie en ICT-II
[18:27] <BBB> ICT wordt een ramp en biotech lukt wel
[18:28] <sienap> startrek
[18:28] <sienap> brb
[18:28] <BBB> doei :)
[18:29] <sienap> mwha
[18:29] <sienap> ik sla ut denk ik over
[18:29] <sienap> voor 1 keer
[18:30] <BBB> okee, nouja, moet je niet voor mij doen want ik ga zo biotech examen voor morgen leren
[18:33] <sienap> en jij dacht echt dat ik dat voor jou zou doen >:)
[18:33] <BBB> :P
[18:33] <sienap> >:)
[18:34] Action: BBB <-- beledigd
[18:35] <sienap> hahaha
[18:35] <sienap> geintje man :)
[18:36] <BBB> lol
[18:36] <BBB> :)
[18:37] <BBB> dat wist ik nou ook nog wel
[18:39] <sienap> zo zo
[18:41] <BBB> ik ga leren
[18:41] <BBB> seeya :)
[18:41] Action: BBB is away: leren
[18:41] Nick change: BBB -> BBB-not-here
[18:50] sienap (synap at ipc379c0d1.dial.wxs.nl) left #gstreamer.
[19:29] sienap (synap at ipc379c0d1.dial.wxs.nl) joined #gstreamer.
[19:29] <sienap> hi all
[19:29] justin_ (justin at h226-67.mahigan.albany.edu) left irc: new computer case
[19:49] <sienap>  /part
[19:49] sienap (synap at ipc379c0d1.dial.wxs.nl) left #gstreamer.
[20:05] dobey (dobey at dreadnought.ximian.com) joined #gstreamer.
[20:05] ChiefHighwater (paul at temple-baptist.com) joined #gstreamer.
[20:30] hadess (hadess at pc2-guil2-0-cust121.gui.cable.ntl.com) joined #gstreamer.
[20:31] <dobey> dude!
[20:31] <hadess> hey dobey
[20:31] <dobey> hadess: send me your .vimrc dude
[20:31] <hadess> it's on my website
[20:32] <hadess> files/config
[20:32] <dobey> oh
[20:32] <dobey> cool
[20:34] <hadess> Ur_17may: what's up ?
[20:35] Nick change: wtay-zZz -> wtay
[20:35] <wtay> hi
[20:36] <hadess> yo wtay
[20:36] <dobey> yo
[20:37] <wtay> 'sup ?
[20:37] <dobey> life sucks
[20:37] <wtay> yeah, what else?
[20:38] <dobey> moreso today
[20:38] <wtay> any particular reason?
[20:38] <dobey> turbolinux and mandrake
[20:38] <wtay> hmm?
[20:39] <wtay> ah right, you need to make packages for them :)
[20:39] <dobey> colorgcc is not very useable
[20:40] <wtay> what's that?
[20:40] <dobey> a shitty perl wrapper for gcc that spits errors in color
[20:40] <wtay> ugh
[20:42] <hadess> oh like gstreamer is a wrapper to aRTs ?
[20:42] Action: hadess runs
[20:42] <dobey> i thought arts was a wrapper for gstreamer
[20:51] <hadess> wtay: btw, feeling explaining caps nego ? i want to give a try to fixing v4lsrc
[20:52] <wtay> hadess: I'll try..
[20:52] <wtay> hadess: what do you know about caps in general?
[20:52] Action: wtay is taking a network trace 
[20:53] <hadess> probably enough, for video what's important is the colorspace
[20:53] <wtay> and the width/depth too of course
[20:53] <wtay> width/height, sorry
[20:53] <hadess> oh, yeah
[20:54] <wtay> also some colorspaces rewuire more info, like RGB (bits for red/blue etc..)
[20:54] <hadess> yeah
[20:55] <hadess> like "is it rgb 565 or rgb 888"
[20:55] <wtay> in gstreamer, caps are a list of key/value pairs
[20:55] <wtay> you have padtemplates and pad caps
[20:55] <hadess> ok
[20:55] <wtay> templates are attached to an element and describe the possible pads/caps this element will have
[20:56] <hadess> ok
[20:56] <wtay> so templates are defined for an element *before* it is actually run, like a template :)
[20:56] <hadess> like a "default"
[20:57] <wtay> not like the default, more like a description of what this element *can* take
[20:57] <wtay> not what it takes by default
[20:57] <hadess> oh, ok
[20:57] <dobey> hrmm
[20:57] <wtay> for example, the xvideosink has a template with a lot of possible colorspaces
[20:58] <wtay> we don't have a merit value for them yet, so there is not "default"
[20:58] <hadess> ok
[20:58] <wtay> at runtime, the element will know the exact type of the media stream
[20:59] <wtay> the first element will typycally know the type first
[20:59] <wtay> before sending out a buffer on one of its pads, the element has to refine (or set) the caps on the pad
[21:00] <wtay> this caps obviously has to be compatible with the template (ie more precise)
[21:00] <hadess> ok
[21:00] <wtay> for example:
[21:01] <wtay> disksrc pushes a buffer to mad
[21:01] <wtay> right before mad is going to push a buffer on its src pad, it'll know the exact type, like bitrate, channels etc..
[21:02] <wtay> so it sets the caps of its src pad to the exact caps
[21:02] <wtay> this will trigger capsnego with the element connected to the src pad (osssink for example)
[21:02] <wtay> now, what will happen then?
[21:03] <wtay> 1) the caps (set by mad) are compared to the template of the sinkpad of osssink
[21:03] <wtay> 2) if the caps match, two things can happen
[21:04] <wtay>   a) the sinkpad of osssink has a negotiate function, in which case it'll be called with the caps as an arg
[21:04] <wtay> osssink can then accept the caps by returning _AGREE 
[21:05] <hadess> right
[21:05] <wtay> osssink can choose to not accept this caps and send out an alternative
[21:05] <wtay> this procedure bounces back and forth between the pads until they agree
[21:06] <wtay> a pad can also choose to not agree and decide that it has run out of alternatives, this will result in a failed negotiation
[21:06] <wtay> the pads are then disconnected and a signal is fired
[21:06] <hadess> ok
[21:07] <hadess> and what is the state of the v4lsrc ?
[21:07] <wtay>   b) the sinkpad has no negotiate function, the caps are set on the src and sink pad (remember the caps matched the template) so all is fine
[21:07] <wtay> hadess: state?
[21:08] <hadess> how much of caps nego is implemented ? but finish your explanation first
[21:08] <wtay> hadess: first we'll have to decide how to handle caps for v4lsrc
[21:08] <wtay> 3) if the caps do not match the template and the srcpad has a negotiate function, it is called to get another caps
[21:09] <wtay> 4) if the caps do not match, and there is no negotiate function on the srcpad, negotiation fails
[21:09] <wtay> that's it, basically
[21:09] Action: dobey is trying to re-discover bitchx
[21:10] <wtay> hadess: this is one way to start the negotiation process
[21:11] <wtay> hadess: also not that I don't think the current implementation/design is optimal
[21:11] <wtay> s/not/note
[21:12] <wtay> hadess: now, looking at v4lsrc and trying to explain what it does:
[21:13] <wtay> v4lsrc, will never *decide* the caps (which is a first flaw)
[21:13] <wtay> in the _get function of v4lsrc, it'll check if the src pad has caps, if not it'll call gst_pad_renegotiate()
[21:13] <wtay> line 544 in gstv4lsrc.c
[21:14] <hadess> 544 is the last line ...
[21:14] Nick change: taazzzz -> taaz
[21:14] <wtay> oops, 296, sorry
[21:14] <hadess> ok
[21:15] <wtay> this wil start the negotiation process
[21:15] <wtay> since v4lsrc has no caps on the srcpad, the negotiate function will get the caps of the sinkpad of the element connected to v4lsrc
[21:16] <wtay> then the gst_v4lsrc_negotiate() function is called with these caps
[21:16] <wtay> line 176
[21:16] <hadess> ok
[21:16] Action: taaz suggests some brave soul put this discussion into the wiki
[21:16] <wtay> you'll see that you gat a pointer to a pointer to caps (so you can modify it)
[21:16] <wtay> taaz: it should be in the manual when I have time
[21:16] Action: hadess just saw taaz volunteer
[21:17] <ajzzzz> hehe
[21:17] <ajzzzz> it helps me (who has just got up...) ;)
[21:17] Nick change: ajzzzz -> aj_uni
[21:17] <wtay> hadess: line 183, v4lsrc does not agree with a NULL caps (it doesn't even propose an alternative, which is a second flaw)
[21:18] <wtay> so, it expects the peer element to provide a caps
[21:18] <wtay> if a caps is present, you'll enter line 190..
[21:18] <dobey> wee
[21:19] <wtay> hadess: then we go on converting the incomming caps to the v4l internal representation (VIDEO_XXX etc)
[21:19] <hadess> right
[21:20] <wtay> finally (if the caps can be converted) we call gst_v4lsrc_sync_parms()
[21:20] <wtay> this function will try to set the v4l device with the specified format.
[21:20] <hadess> GST_PAD_NEGOTIATE_AGREE <- tada
[21:20] <wtay> and all is well :)
[21:21] <wtay> if not... we set the incomming caps to NULL and return TRY
[21:21] <wtay> this will force the peer element to provide an alternative
[21:21] <hadess> oppsie, on the phone...
[21:22] <wtay> heh
[21:29] steveb (steveb at node1ee0c.a2000.nl) left irc: [x]chat
[21:36] <dobey> hrmm
[21:40] dobey (dobey at dreadnought.ximian.com) left irc: later
[21:50] steveb (steveb at node1ee0c.a2000.nl) joined #gstreamer.
[22:03] <steveb> the difference in latency between osssink and alsasink is huge
[22:03] <steveb> is there anything that can be done to make osssink a bit perkier? (internal buffer sizes?)
[22:04] <aj_uni> viagra? ;)
[22:05] Action: steveb forcefeeds viagra to aj
[22:05] <aj_uni> lol
[22:06] <steveb> there is about a 1 sec delay between doing it and hearing it
[22:06] <aj_uni> you thinking of lowering the buffer size? 
[22:06] <aj_uni> this computer is having major problems....
[22:06] <steveb> i don't know what internal buffers osssink has - it is already being fed buffers of 64 samples which should be v.fast
[22:07] <aj_uni> but does it dump buffers to the sound card when they are full?
[22:07] <steveb> dunno
[22:09] <aj_uni> source isn't telling me much...
[22:10] <steveb> it might be the oss layer of alsa causing it
[22:10] <aj_uni> hmm
[22:11] <aj_uni> i use alsa drivers with oss emulation...
[22:11] <aj_uni> on my wonderful SB16
[22:29] Nick change: aj_uni -> ajmitch
[22:30] dobey (dobey at dreadnought.ximian.com) joined #gstreamer.
[22:30] <dobey> Ur_17may: ?
[22:31] <dobey> blah
[22:31] <dobey> someone tell him to find me sometime
[22:31] dobey (dobey at dreadnought.ximian.com) left irc: eh
[22:32] karltk_ (karltk at tunnel-44-32.vpn.uib.no) joined #gstreamer.
[22:38] steveb (steveb at node1ee0c.a2000.nl) left irc: [x]chat
[22:49] omega_ (omega at omegacs.net) joined #gstreamer.
[22:50] <ajmitch> hi omega_
[22:50] <omega_> yo
[22:51] sienap (synap at ipc379c0d1.dial.wxs.nl) joined #gstreamer.
[23:00] <wtay> yo
[23:01] <omega_> I have a lot of questions about capsnego
[23:01] <wtay> right
[23:01] <wtay> go ahead
[23:01] <omega_> sec..
[23:02] <ajmitch> omega_: don't we all? ;)
[23:02] <omega_> yeah, but your questions come from the plugin side...
[23:03] <omega_> so, here's the scenario
[23:03] <omega_> mad ! autoplugger ! osssink (crippled to mono)
[23:03] <omega_> they connect up fine, then data comes into mad and it decodes the first buffer
[23:03] <wtay> ok
[23:03] <omega_> it sets caps on its src pad, which proxy to the osssink and *fail*
[23:03] <wtay> ok
[23:04] <omega_> the autoplugger catches the failed signal on autoplugcache:sink
[23:04] <omega_> nothing happens to autoplugcache:src, which is a potential bug
[23:04] <omega_> since there's no failed signal on osssink:sink either
[23:04] <omega_> anyway..
[23:04] <wtay> ok, I'm not sure this is implemented
[23:05] <omega_> in the handler, it disconnects the autoplugcache:src and osssink:sink
[23:05] <omega_> constructs an autobin to match caps, and puts it in place
[23:05] <omega_> now, lemme reproduce the major problem
[23:05] <wtay> good
[23:07] <omega_> useful tidbit: ridgerun is moving tomorrow, to offices that are much much closer to my hotel <g>
[23:08] <omega_> um, for some reason the autoplug is failing again ;-(
[23:08] <omega_> are you building that validator?
[23:08] <omega_> cause it needs to check properties sets
[23:09] <omega_> and there's got to be a better set of data to DEBUG for autoplug, it's very hard to see what's failing and why ;-(
[23:09] <wtay> nope, I've been quite busy for work today
[23:10] <wtay> s/today/this week/
[23:14] <omega_> ok, so first of all I need to get this working again....
[23:15] <wtay> what was wrong?
[23:15] <omega_> was or is?
[23:15] <wtay> was
[23:15] <omega_> um, lemme think
[23:15] <wtay> in general :)
[23:15] <omega_> basically, the autobin was being attached, and no caps nego was happening
[23:16] <wtay> ok
[23:16] <omega_> so my partial fix was to forcibly set the caps of the autoplugcache:sink to mad's caps before connecting
[23:16] <omega_> but that's just a hack
[23:16] <omega_> the base problem is that when you connect two pads, nego isn't happening as often as it should
[23:16] <wtay> I know why
[23:17] <omega_> specifically, negoproxy is confusin git
[23:17] <wtay> proxy is not very well implemented IMO
[23:17] <omega_> the check requires that one of the two pads have caps
[23:17] <wtay> anyway, what happens is this:
[23:17] <omega_> and when the sinkpad is a template, and the source is a proxy pad to something that *does* have caps, it fails
[23:18] <wtay> possibly, proxy uses the stack, whioch should be changed
[23:19] <wtay> when mad tried to set the caps, it failed
[23:19] <omega_> but there are advantages in doing it that way:
[23:19] <omega_> potentially, a nego function can decide [not] to proxy based on the caps
[23:20] <wtay> yes, that was the idea indeed
[23:20] <omega_> but I can't think of a specific example of that ever needing to be done, off-hand
[23:20] <wtay> colorspace does that
[23:20] <wtay> first it proxies, if that fails it sets up its internal conversion routines
[23:21] <omega_> ok
[23:21] <omega_> btw, if it succeeds, it should make use of the autoplugger's code to remove itself
[23:22] <wtay> that would be neat
[23:22] <wtay> so, when med set the caps and it failed, the caps were not set on the src pad
[23:23] <wtay> you then connected the bin and nothing happens because there are no caps on the pad
[23:23] <omega_> no, afaik the caps were set on the srcpad
[23:23] <wtay> mad should try to renegotiate if the caps failed
[23:23] <omega_> mad:src
[23:23] <omega_> but when I connect the autoplugcache:sink to the autobin, it doesn't nego
[23:23] <wtay> ok, the gst_pad_set_caps function was called, but that doesn't mean the caps are realy set
[23:24] <omega_> the upstream side of the cache is failing nego, and I'm putting the autobin on the downstream side
[23:24] <wtay> which is probably not the right way to handle it
[23:25] <wtay> you're connecting the bin to mad, right?
[23:25] <sienap> Wtay!
[23:25] <wtay> maybe mad should retry to set its caps if the nego has failed...
[23:25] <wtay> sienap: yo
[23:25] <omega_> no, connecting autobin to the *downstream* side of the cache
[23:26] <wtay> uh?
[23:27] <wtay> what did the pipeline look like then?
[23:27] <sienap> wtay done some cool stuff today ?
[23:27] <wtay> sienap: nope
[23:27] <omega_> the pipeline will end up looking like mad ! autoplugcache ! stereo2mono ! osssink
[23:28] <wtay> oh
[23:28] <omega_> the caps failure is between mad and the cache, I capture the cache:sink end
[23:28] <omega_> but osssink and cache:src never get that signal
[23:28] <wtay> and were is the autoplugger?
[23:28] <omega_> the autoplugger is everything between mad and osssink
[23:28] <sienap> wtay he damn
[23:28] <sienap> not much cool stuff  lately huh :)
[23:29] <wtay> sienap: I took 3 hours of tcpdump traces :-)
[23:35] <BBB-not-here> sienap: you might help me making sdlaudiosink
[23:35] <sienap> wtf
[23:35] <sienap> why that ? :)
[23:35] <BBB-not-here> I had some major problems there that I will work on next week
[23:36] <omega_> wtay: how about precomputing the compatibility graph?
[23:36] <sienap> BBB he
[23:36] <BBB-not-here> sienap: because I'm dumb and you've been around here longer than me ;)
[23:36] <sienap> BBB hmm i first recompiling the THANKS list of nautilus
[23:36] <BBB-not-here> sienap: and you don't do much else anyway :P
[23:36] <sienap> BBB he that doesn't say much :)
[23:36] <BBB-not-here> except compiling THANKS lists
[23:36] <BBB-not-here> :P
[23:36] <sienap> BBB for gstreamer :)
[23:36] <sienap> hehe *slap*
[23:37] <wtay> omega_: could be done, I actually did then when we still had the mime-types
[23:37] Action: BBB-not-here will try to work on lavsrc when he has some more gstrwamer experience
[23:37] <BBB-not-here> and I will, hopefully with help of Gernot and Andrew, be able to do just that :)
[23:37] <sienap> lavsrc ?
[23:37] <BBB-not-here> lavsrc will be the source-plugin for our MJPEG goody movies
[23:38] <BBB-not-here> our MJPEG goody movies are called 'lav' (linux audio/video)
[23:38] <BBB-not-here> so the source-plugin for that is lavsrc
[23:38] <BBB-not-here> :)
[23:38] <BBB-not-here> and later on probably also a lavsink to encode into MJPEG
[23:38] <sienap> ahja
[23:38] <BBB-not-here> but I will need to look into that
[23:38] <sienap> can't you better call it lavdec / lavenc ?
[23:39] <omega_> wtay: probably quite useful.  mulitple autoplugs mean multiple generations of the graph
[23:39] <omega_> sienap: good point <g>
[23:39] <sienap> hehe
[23:40] <wtay> right, that's why lavencode is called lavencode
[23:40] <omega_> wtay: then we can enable -inspect to print out compatible elements
[23:40] <sienap> ehm
[23:40] <wtay> omega_: yup
[23:41] <omega_> and save a lot of time <g>
[23:41] <BBB-not-here> so.....
[23:41] <BBB-not-here> lavenc/lavdec?
[23:41] <BBB-not-here> *eehw*
[23:41] <wtay> omega_: we can only do that for the padtemplates of course
[23:41] <BBB-not-here> that makes it look like a normal encoder/decoder
[23:41] <omega_> right
[23:41] <BBB-not-here> but we are special
[23:41] <BBB-not-here> we are cute, lovely
[23:41] <BBB-not-here> self-programmed, even
[23:41] <BBB-not-here> :)
[23:42] <omega_> wtay: with inspect, we instantiate the element, so we can compare padtemplates to the graph
[23:42] <omega_> hmmm, still requires matching
[23:42] <omega_> but better, since the graph is static until plugins are added live
[23:42] <wtay> but only for thos matching the template
[23:42] <omega_> right
[23:42] <wtay> hmm, not really actually
[23:43] <wtay> and instantiating isn't going to create the real pad caps anyway, that happens when media flows
[23:43] <omega_> depends, but I'm not worried about that
[23:43] <omega_> that's just -inspect, if it doesn't show everything, so be it
[23:44] <wtay> actually, creating the graph is pretty easy, it's O(n²) though
[23:44] <omega_> right, so do it once and update it as elementfactories are added
[23:44] <omega_> store it even
[23:44] <wtay> yes
[23:44] <omega_> but beware to store only nodes that make it into the registry otherwise
[23:45] <wtay> of course
[23:45] <omega_> anyway, what about this caps problem?
[23:45] <wtay> only element name, pad -> element1, element2, etc..
[23:45] <wtay> omega_: I have to see it
[23:45] <omega_> ?
[23:46] <omega_> the autoplug or caps problem?
[23:46] <wtay> caps prob..
[23:46] Nick change: ajmitch -> aj_uni
[23:46] <wtay> isn't it normal that osssink doesn't fire a signal when some other pad capsnego fails?
[23:47] <omega_> not when that pad *is* osssink's
[23:47] <wtay> right
[23:47] <omega_> in mad ! cache ! osssink, when mad sets caps
[23:47] <omega_> they fail
[23:47] <omega_> I get signals from mad:src and cache:sink
[23:47] <omega_> even though the nego was between mad:src and osssink:sink
[23:47] <wtay> ok, that's probably not implemented
[23:48] <wtay> that's a problem with the nego function
[23:48] thomas__ (thomas at urgent.rug.ac.be) joined #gstreamer.
[23:48] <omega_> right
[23:48] <thomas__> hi all
[23:48] <thomas__> anyone know how to get glib compiled on mac osx ?
[23:48] <wtay> I think proxying the nego should keep a stack 
[23:49] <wtay> hi thomas__
[23:49] <thomas__> It doesn't seem to know what thread implementation it has
[23:49] Action: omega_ has no idea ;-(
[23:49] <omega_> best guess would be to try #gtk+ or #gnome
[23:49] <wtay> thomas__: has anybody actually tried to get gtk working on macosx?
[23:49] <thomas__> wtay: yeah, me today ;)
[23:50] <thomas__> I have one on loan till monday
[23:50] <thomas__> and i really want to get aalib working ;)
[23:50] <thomas__> with gstreamer that is
[23:50] <thomas__> if they didn't cost so damn much, I'd buy them... I really like what they did, and I've never been an Apple fan
[23:51] <sienap> bbb but i'll look at the SDL vidsink if you want me too :)
[23:51] <wtay> I mean, getting gtk/gnome working on mac osx would be major /. news and I didn't see that one :)
[23:51] <sienap> it is the same as the aalib one ?
[23:52] <sienap> only then some more usefull pixels..
[23:52] <sienap> and colors
[23:52] <thomas__> wtay: are you "available" this weekend ?
[23:54] <thomas__>  
[23:54] <wtay> thomas__: depends :-)
[23:54] <wtay> thomas__: I'll be here most of the time
[23:55] <omega_> good grief.  pbd has just reinvented gstreamer
[23:55] <thomas__> wtay: ok, I'll get back then, I'd really like to clean up the audio plugins and do some new stuff with them
[23:55] <sienap> pbd ?
[23:55] <omega_> Paul Barton-Davis
[23:56] <sienap> who is here
[23:56] <sienap> where
[23:56] <sienap> url
[23:56] <sienap> tell me
[23:56] <sienap> explain
[23:56] <omega_> sienap: please do not flood the channel like that
[23:57] <wtay> omega_: did you add a capsnego_failed signal?
[23:57] <omega_> yes
[23:57] <wtay> ok
[23:57] <omega_> same targets as the caps_changed signal
[23:57] <omega_> and I moved caps_changed
[23:57] <wtay> in both the renegotiate and negotiate_proxy functions?
[23:57] <sienap> omega he grrr 
[23:57] <sienap> but please answer who he is.. 
[23:57] <omega_> checking
[23:58] <omega_> he's a developer working on audio stuff
[23:58] <wtay> sienap: LAD mailing list...
[23:58] <sienap> *now really confused*
[23:58] <sienap> and he now is starting an project like gstreamer ?
[23:58] <omega_> sienap: go search on the net for linux-audio-dev mailing list
[23:59] <sienap> he i am not that intrested :)
[00:00] --- Fri May 18 2001
[00:00] <omega_> wtay: updated gstautoplugger.c
[00:00] <wtay> ok
[00:01] <omega_> man.  he's got the gstreamer high-level concepts down :-)
[00:01] <omega_> but he's so NIH-happy that he just won't get it
[00:01] Action: thomas__ is away - Automatically set away. - messages will be saved.
[00:01] <sienap> damn
[00:01] <sienap> :)
[00:01] <omega_> he has the right quote at the bottom though:
[00:01] <omega_> "Those who do not understand aRtS are condemned to reinvent it" :)
[00:01] <sienap> shall i try to punch some sense in him.. :)
[00:02] <omega_> no
[00:02] <sienap> ":)"
[00:02] <wtay> let him suffer <g>
[00:02] <sienap> *G*
[00:02] <omega_> no, he's got to understand that he's reinventing what already exists
[00:02] <sienap> Childeren of Bodom is cute btw :)
[00:02] <sienap> omega tell him
[00:02] <sienap> :)
[00:03] <omega_> he seems to have taken some specific, highly optional parts of gstreamer (threads) and gone balistic
[00:03] <sienap> "Aah .. yeah yeah nice project nice framework.. only 2 years behind gstreamer so now come develop for us"
[00:03] <omega_> you have that message?
[00:03] <sienap> omega hmm then is being doing strange stuff.. why not just coding for gstreamer then
[00:04] <omega_> wtay: fwded in case
[00:05] <omega_> and he's moving away from DMA-based buffers where it's reasonable, towards having the engine responsible for all audio I/O, and copying from DMA to engine buffers
[00:05] <omega_> that's what bufferpools are for
[00:05] <omega_> I'm going to respond in detail to this message, which'll take me a while
[00:05] <omega_> but for the moment, I'd like to get this capsnego thing worked through, since that's what keeping incsched from going to HEAD
[00:06] <wtay> hmm
[00:06] <wtay> what should I test?
[00:06] <wtay> omega_: can you commit the pad changes too?
[00:06] <omega_> you should be able to replicate my test
[00:07] <omega_> but not sure that's needed, it's pretty clear what's happening (I think)
[00:07] thomas__ (thomas at urgent.rug.ac.be) left irc: I'm outta here!
[00:08] <sienap> so what was the site of hadess again ?
[00:08] <omega_> hadess.net
[00:09] <sienap> aah hadess.net
[00:09] <sienap> hmm
[00:09] <sienap> great minds / sools / bla
[00:09] <hadess> back
[00:09] <sienap> hej hadess
[00:09] <wtay> omega_: segfault?
[00:09] <omega_> ?
[00:09] <sienap> where were those itunes clone screenshots again
[00:09] <wtay> DEBUG( 4341: 5)gst_schedule_pullfunc_proxy:181: done switching
[00:09] <wtay> DEBUG( 4341: 5)gst_schedule_chain_wrapper:72: calling chain function of stereo2mono:sink
[00:09] <wtay> Segmentation fault
[00:09] <omega_> um
[00:09] <wtay> ./gstreamer-launch --gst-mask=-1 disksrc location=/opt/data/mp3/files.cgi ! mad ! autoplugger ! esdsink
[00:10] <omega_> esdsink won't work because it doesn't nego
[00:10] <omega_> but how'd you get stereo2mono to autoplug?
[00:10] <wtay> esdsink is hacked to mono
[00:10] <omega_> right, but it doesn't nego
[00:11] <wtay> you use  a hacked osssink then?
[00:11] <omega_> so you'll get autoplugger doing its thing, but you can't see what the problem is after that
[00:11] <omega_> yes
[00:12] <wtay> ok, ** CRITICAL **: file gstpad.c: line 1425 (gst_pad_push): assertion `peer != NULL' failed.
[00:12] <sienap> Ehm hadess ?
[00:12] <wtay> that's because capsnego failed, right?
[00:12] <omega_> yes
[00:12] <omega_> er
[00:12] <omega_> did the autoplug work for you?
[00:12] <wtay> yes
[00:13] <omega_> factories found in autoplugging:
[00:13] <omega_> "stereo2mono" ??
[00:13] <omega_> I don't get anything right now
[00:13] <wtay> hmm, ok, I don't get anything either
[00:14] <omega_> if I want to find the debug lines relevant, is stereo2mono going to be the first or second plugin on the line?
[00:14] <omega_> brb in 2min
[00:15] <wtay> the stereo2mono src padtemplate is not compatible with osssinks sink pad...
[00:15] ChiefHighwater (paul at temple-baptist.com) left irc: Ping timeout for ChiefHighwater[temple-baptist.com]
[00:16] <hadess> rats, 2nd usb channel is not working...
[00:16] <wtay> INFO ( 5257: 2)construct_path:515: factory: "stereo2mono"
[00:17] <omega_> DEBUG( 1634: 2)gst_autoplug_can_match:137: factory "stereo2mono" cannot connect with factory "osssink"
[00:18] <omega_> it's the stuff above that that tells why
[00:18] <omega_> ?
[00:18] <wtay> omega_: modify the s2m src pad to a max rate of 48000
[00:18] <omega_> DEBUG( 1634: 2)gst_props_check_compatibility:752: rate are not compatible: 
[00:18] <omega_> DEBUG( 1634: 2)gst_props_debug_entry:69: 8000-96000
[00:18] <omega_> DEBUG( 1634: 2)gst_props_debug_entry:69: 8000-48000
[00:18] <omega_> blegh, needs rethinking
[00:18] <wtay> yes, the incremental autoplugger should not try to use templates at all
[00:19] <omega_> it has to, but it should be more forgiving about template matches
[00:19] <omega_> ok, autoplugged
[00:19] <hadess> oh my good, my usb is badly broken, my camera is showing up 8 times in the devices...
[00:19] <omega_> oooops
[00:21] <wtay> omega_: more forgiving probably implemented with a weight..
[00:21] <omega_> so, when the autobin is connected, no caps nego happens
[00:21] <wtay> did capsnego happen before?
[00:21] <wtay> I mean, before connecting the bin
[00:22] <wtay> hmm, yes
[00:22] Nick change: big_T -> big_Tzzz
[00:24] <wtay> omega_: mad only tries to set the caps once, that's why it fails, it should retry until it succeeds
[00:25] <omega_> hrm, I guess that makes sense
[00:25] <wtay> maybe do it in a while loop and check for a stopping condition somehow
[00:26] <wtay> another option is to set the pad caps anyway
[00:26] <omega_> hmm
[00:26] <omega_> a loop can be really dangerous
[00:26] <wtay> yes
[00:27] <omega_> what should really happen is that the attaching of stereo2mono to the cache should cause nego back to mad
[00:27] <omega_> (O/T hint: if you ever need web-based email, don't use iname.com)
[00:28] <wtay> omega_: that'll need a complete redesign
[00:28] ChiefHighwater (paul at temple-baptist.com) joined #gstreamer.
[00:29] <wtay> hi
[00:29] <hadess> omega_: mail2web.com provides webmail for any pop3 mailbox
[00:30] <wtay> we could also retry the capsnego based on a return value of the capsnego_failed signal...
[00:30] <omega_> wtay: redesign?
[00:30] <omega_> hmmmm
[00:31] <wtay> currently mad takes the initiative to set caps, now you'll need something external to trigger it
[00:31] <omega_> is there any way we can 'always' do nego, and let the proxy code decide whether the attempt can be completed?
[00:31] <omega_> but pad_connect can also force renego
[00:31] <wtay> then nego must always use a pull based machanism
[00:32] <omega_> why?
[00:32] <omega_> nego is bidir, afaict
[00:32] <omega_> if you connect with src template caps and sink caps, it works just as well
[00:32] <wtay> you must be able to call an element method to start nego
[00:32] <omega_> right?
[00:33] <wtay> that's the nego function
[00:33] <wtay> mad doesn't have that
[00:33] Action: omega_ is lost
[00:33] <wtay> mad just calls, gst_pad_set_caps
[00:33] <omega_> right, that's fine
[00:33] <omega_> I'm talking about the pad_connect later on
[00:33] <omega_> linei 645 and on, gstpad.c
[00:34] <wtay> how are you going to tell mad to redo that?
[00:34] <omega_> mad doesn
[00:34] <omega_> doesn't have anything to do with that
[00:34] <omega_> if I connect two pads:
[00:34] <omega_> mad:src has caps already set
[00:35] <omega_> stereo2mono:sink has templatecaps
[00:35] <wtay> no it hasn't
[00:35] <omega_> yes it does
[00:35] <wtay> no, that's the problem
[00:35] <omega_> nego doesn't clear the caps on failure,and besides, even the signal handler hasn't finished yet, so it can't have cleared them yet anyway
[00:35] <wtay> it calls gst_pad_set_caps, but since nego fails, the caps aren't set on the pad
[00:36] <wtay> nego sets the caps *after* nego succeeds
[00:36] <omega_>   GST_PAD_CAPS(pad) = caps;
[00:36] <omega_>   return gst_pad_renegotiate (pad);
[00:36] <wtay> uhm
[00:37] <omega_> renego doesn't NULL the caps before it says 'failed'
[00:37] <omega_> afaict
[00:37] <wtay> you're right
[00:37] <omega_> 'NULL' doesn't apear in renego
[00:38] <wtay> that's weird
[00:39] <wtay> now I'm confused
[00:40] <omega_> regardless, there's still the issue that even if the caps are set, the pad_connect code won't even try renego, because it doesn't see caps set on the 'local' src, cache:src, since it's a proxy
[00:40] <wtay> true
[00:41] <wtay> but it should see them when connecting mad!src
[00:41] <omega_> I wonder if pad_connect should just always renego
[00:41] <omega_> and let renego figure out if it's possible
[00:41] ChiefHighwater (paul at temple-baptist.com) left irc: 
[00:41] <omega_> but mad!src was connected long ago
[00:42] <omega_> and is never disconnected until the autoplugger is finished, when the cache is yanked out
[00:42] <omega_> brb again
[00:44] hadess (hadess at pc2-guil2-0-cust121.gui.cable.ntl.com) left irc: mooooh!
[00:44] <omega_> back
[00:48] matth-coffee (matth at qwest.dsplinux.net) left irc: Ping timeout for matth-coffee[qwest.dsplinux.net]
[00:53] <wtay> actually, the cache pads do not have the mad src caps
[00:54] <omega_> right
[00:54] <omega_> I had to set them explicitely to force nego when the get connected
[00:54] <wtay> yes
[00:55] <omega_> autoplugger.c has pad_set_caps commented out in CVS
[00:55] <wtay> trying to figure out how to set them in negotiate
[00:56] <wtay> that's one of the reasons capsnego proxy should not be implemented with a call
[00:56] <omega_> can colorspace turn on the proxy bit, then catch signals sufficient to do its work by turning off the bit later?
[00:57] <omega_> say, if a proxied nego fails, the proxy gets a signal and a chance to patch it up before the real fail?
[00:57] <wtay> what do you mean?
[00:57] <omega_> where proxy is a flag on the pads or somesuch
[00:58] <omega_> so instead of calling pad->nego to find its way to the final pad, the nego function itself looks at the flag and determines the final pad to nego with, and just does it
[00:58] <wtay> that's a problem, the negotiation is never marked in the pad
[00:58] <omega_> ?
[00:58] <wtay> what you say here is what I was thinking about
[00:59] <wtay> the element does not call _proxy but sets some sort of flag
[00:59] Nick change: aj_uni -> ajmitch
[00:59] <wtay> capsnego should be redone
[01:00] Action: omega_ goes to CHW's house to get his DSL working
[01:00] <omega_> l8r
[01:00] omega_ (omega at omegacs.net) left irc: [x]chat
[01:00] <wtay> I'v got to sleep
[01:00] Nick change: wtay -> wtay-zZz
[01:00] <wtay-zZz> cya all
[02:02] sienap (synap at ipc379c0d1.dial.wxs.nl) left irc: sienap has no reason
[02:22] Nick change: Kuroyi_ -> Kuroyi
[02:27] karltk_ (karltk at tunnel-44-32.vpn.uib.no) left irc: restart X
[02:34] karltk_ (karltk at tunnel-44-32.vpn.uib.no) joined #gstreamer.
[02:56] matth-coffee (matth at qwest.dsplinux.net) joined #gstreamer.
[03:13] Nick change: Ur_17may -> Uraeus
[03:13] <Uraeus> hi ho
[03:17] <Uraeus> Kuroyi: are you here?
[03:20] Topic changed on #gstreamer by ChanServ!s at ChanServ: GStreamer: the ultimate multimedia framework
[03:20] Uraeus (cschalle at c224s9h5.upc.chello.no) left irc: syntax error - user imploded
[03:20] <Kuroyi> yeah
[03:20] <Kuroyi> gah
[04:05] BBB-not-here (BBB at ucu-105-116.ucu.uu.nl) left irc: Ping timeout for BBB-not-here[ucu-105-116.ucu.uu.nl]
[04:05] wtay-zZz (wim at cable-195-162-214-58.upc.chello.be) left irc: Ping timeout for wtay-zZz[cable-195-162-214-58.upc.chello.be]
[04:05] taaz (dlehn at left irc: Ping timeout for taaz[]
[04:06] taaz (dlehn at joined #gstreamer.
[04:06] BBB-not-here (BBB at ucu-105-116.ucu.uu.nl) joined #gstreamer.
[04:06] wtay-zZz (wim at cable-195-162-214-58.upc.chello.be) joined #gstreamer.
[05:22] ajmitch (ajmitch at p49-max7.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p49-max7.dun.ihug.co.nz]
[05:47] ajmitch (ajmitch at p37-max5.dun.ihug.co.nz) joined #gstreamer.
[05:59] Dr_Zaius (jeff at h00a0244bc361.ne.mediaone.net) joined #gstreamer.
[06:06] Dr_Zaius (jeff at h00a0244bc361.ne.mediaone.net) left irc: Read error to Dr_Zaius[h00a0244bc361.ne.mediaone.net]: EOF from client

More information about the gstreamer-devel mailing list