IRC Logs

*******************************************************************
[03:45] <LeRoutier> grrr, out of band data is such a mess in tv broadcasting (tons of competing standards)
[03:46] <LeRoutier> by "out of band", i mean things other than audio & video (channel/program info)
[03:48] <Rounin> Mhm
[03:50] <LeRoutier> say : PDC, VPS, XDS, Teletex, ...
[03:50] <Rounin> CC ?
[03:50] <Rounin> Although that might be only for VHS
[03:51] <LeRoutier> i don't know CC, what does it do ?
[03:51] <Rounin> Closed Captioning... Provides text for the deaf
[03:51] <Rounin> Or something
[03:52] <LeRoutier> CC exists too for broadcast, but only in NTSC zones
[03:52] olimar ([email protected]) joined #gstreamer.
[03:52] <olimar> Hi all
[03:52] <olimar> compiled totem against gstreamer but totem starts on the command line and that was it
[03:52] <olimar> !!!
[03:53] <olimar> [root@localhost totem-0.99.10]# totem
[03:53] <BBB> ?
[03:53] <olimar>  
[03:53] <olimar> [root@localhost totem-0.99.10]#
[03:53] <BBB> you're not making any sense to me
[03:53] <olimar> I type totem and nothing happens!
[03:53] Nick change: harshyGone -> harshyWork
[03:53] <wheels> BBB: have another drink
[03:53] <BBB> right, that'll help
[03:53] Action: ds-work needs a drink
[03:53] <olimar> I'm sure drinks help ;) but what about my totem!
[03:53] Action: BBB needs sleep
[03:54] <BBB> olimar: I don't know, did you compile it in a special way?
[03:54] <BBB> did you run gst-register-0.8?
[03:54] <olimar> I did
[03:54] <BBB> do othe gst-based apps (e.g. rhythmbox) work?
[03:54] <olimar> well ok forget about totem
[03:54] <olimar> here is what gst-player tells me
[03:55] <olimar> command line:
[03:55] <olimar> ** (gst-player:16969): WARNING **: libgstplay: failed initializing pipeline, error: Could not link video output thread (cs and balance)
[03:55] <olimar> and a dialog with:
[03:55] <olimar> could not link video output thread (cs and balance)
[03:55] <olimar> and then a blank window appears
[03:55] <BBB> it means something's wrong with your installation, I guess...
[03:56] <BBB> did you run gst-register-0.8 both as user and as root?
[03:56] <olimar> hmm
[03:56] <olimar> Ah
[03:56] <olimar> same problem
[03:57] <BBB> hm...
[03:57] <BBB> do you have gst-ffmpeg?
[03:57] <BBB> or only gstreamer core and gst-plugins?
[03:58] <ds-work> dude, navseek is cool
[03:58] <olimar> only gstreamer core and gst-plugins I suppose as I've downloaded them from a fedora mirror
[03:58] <BBB> ah, ok, so RPMs?
[03:58] <BBB> which version?
[03:58] <olimar> yep
[03:58] <olimar> 0.8.0
[03:58] <BBB> in that case I don't know... ask thomasvs (he'll be up in a few hours) or leave a bug report on bugzilla...
[03:59] <BBB> as far as my knowledge goes, the fedora RPMs should "just work"
[03:59] <BBB> I don't know how/when they could fail
[03:59] <olimar> ic
[03:59] <ds-work> few == 6
[03:59] <BBB> hm, yeah
[04:00] <olimar> but as I have you here anyway BBB is there a possibility to play movies with gst without a gui?
[04:00] <olimar> directly form the CL?
[04:01] <BBB> gst-launch-ext
[04:01] <BBB> or gst-launch-ext-0.8
[04:01] <BBB> I guess it was versioned
[04:07] <ds-work> there's also examples/gstplay/player
[04:07] <ds-work> I use that a lot
[04:07] <LeRoutier> ds-work, is that installed ? (he used RPMs)
[04:07] <ds-work> no
[04:08] <LeRoutier> good night ppl, i need some sleep
[04:09] <LeRoutier> BBB: go to bed too (as long as it isn't mine)
[04:09] <LeRoutier> ++
[04:09] LeRoutier ([email protected]) left irc: "Leaving"
[04:10] <ds-work> is Company gone for a while?
[04:11] <BBB> I guess he's asleep?
[04:11] <ds-work> I haven't seen him today or yesterday
[04:12] Action: BBB doesn't know
[04:12] <wheels> Yeah, he hasn't been around for a couple-o-days...
[04:13] <wheels> And this would be early for him to be sleeping anyway...
[04:13] bilboed ([email protected]) left irc: Read error: 60 (Operation timed out)
[04:15] Action: BBB doesn't know
[04:15] <BBB> he'll be back, I guess he's working on his scheduler or so
[04:15] Action: BBB sleep, bbl :)
[04:15] BBB ([email protected]) left irc: "Client exiting"
[04:19] olimar ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:37] jimmy_dean ([email protected]) left irc: "Carrier Signal Lost"
[04:39] jimmy_dean ([email protected]) joined #gstreamer.
[05:27] Rounin ([email protected]) left irc: "Client exiting"
[05:56] <thomasvs> night peeps
[06:02] <ds> eh?
[06:03] <ds> isn't it a bit early/late?
[06:08] <taaz> hackers don't believe in time
[06:09] <thomasvs> ds: not with barcelona's nightlife it is :)
[06:10] <ds> heh
[06:10] <ds> well, if it's any consolation, I've been up until 4 recently
[06:10] <taaz> you all ever looked at TRON?
[06:10] <thomasvs> taaz: yeah, quite some time ago
[06:10] <thomasvs> tron is a poor ripoff of the strokes' video 12:51
[06:11] <taaz> uh... what?
[06:11] <taaz> http://www.tron.org/
[06:11] <thomasvs> you haven't seen that video ?
[06:11] <thomasvs> oh, different tron
[06:11] <taaz> yeah ;)
[06:11] <thomasvs> well, i'm quite sure I don't speak that language
[06:12] <taaz> there are english specs
[06:12] <taaz> i was just noticing their task state machine model
[06:13] <taaz> made me think of gst states
[06:13] <taaz> http://www.na.assoc.tron.org/documents.html
[06:13] <taaz> READY, RUNNING, WAITING, WAITING-SUSPENDED, SUSPENDED, DORMANT, and NON-EXISTENT
[06:14] <taaz> mutron-400e spec section 3.2 if anyone wants to see a state transition diagram
[06:15] <taaz> mentions a blocked state too
[06:15] <taaz> maybe this has nothing to do with gst, i dunno ;)  it just seemed like something to mention here
[06:16] Action: ds would like to let everyone know that he hates Solaris
[06:28] Nick change: spyder_away -> spyder482
[06:36] eliman ([email protected]) left irc: "Leaving"
[06:47] sub_pop ([email protected]) left irc: "Client exiting"
[07:15] sub_pop ([email protected]) joined #gstreamer.
[08:51] <taaz> ds: ping?
[08:55] ChrisHJW ([email protected]) joined #gstreamer.
[08:57] <ds> moo
[08:58] <taaz> ds: are you good with C macro magic? ;)
[08:58] <ds> yes
[08:58] <taaz> is there someway to use the preprocessor to conditionally pass a parameter to a function?
[08:59] Action: spyder482 wonders why cows are the coice animals for programmers to imitate...
[08:59] <taaz> i have a specific use case
[08:59] <ds> spyder482: apt-get moo
[08:59] <spyder482> lmao
[08:59] <ds> #ifdef HAVE_2_PARAMS
[08:59] <ds> #define function(a,b) function(a,b)
[08:59] <ds> #else
[09:00] <ds> #define function(a,b) function(a)
[09:00] <ds> #endif
[09:00] <taaz> i want to pass a sim environment param to functions in sim mode but optimize that param out in runtime mode
[09:00] <ds> huh?
[09:01] <taaz> it's embedded coding stuff.  i'd like reentrant functions that gen some state var in a simulator but on a microcontroller use static state and avoid extra function parameter
[09:02] <taaz> foo(state, a) in sim mode, but just foo(a) in runtime mode
[09:02] <taaz> maybe i'm not being clear ;)
[09:03] <ds> I'd do what I mentioned above
[09:04] <ds> probably calling the real function function_real() or something
[09:05] <ds> and then #ifdefs in the function definition
[09:05] <taaz> yeah... i started with something similar.  it's just verbose
[09:06] <ds> yeah, it's going to be
[09:11] Action: taaz finds it amazing when he solders up a uC and various parts on a board and it works first try
[09:13] abo ([email protected]) joined #gstreamer.
[09:21] spyder482 ([email protected]) left irc: "Leaving"
[09:24] Nick change: abo -> abo_
[09:41] kwm ([email protected]) joined #gstreamer.
[09:55] ChrHJW ([email protected]) joined #gstreamer.
[09:57] sxpert_work ([email protected]) left irc: "Leaving"
[09:59] markey ([email protected]) joined #gstreamer.
[10:03] ChrisHJW ([email protected]) left irc: Read error: 60 (Operation timed out)
[10:08] ChriHJW_log ([email protected]) joined #gstreamer.
[10:13] ChrisHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:14] sxpert_work ([email protected]) joined #gstreamer.
[10:14] bilboed ([email protected]) joined #gstreamer.
[10:16] teuf ([email protected]) joined #gstreamer.
[10:38] ploum ([email protected]) joined #gstreamer.
[10:47] wingo ([email protected]) joined #gstreamer.
[11:33] bilboed_school ([email protected]) joined #gstreamer.
[11:34] sublett ([email protected]) joined #gstreamer.
[11:37] bilboed_school ([email protected]) left irc: Remote closed the connection
[11:45] sxpert ([email protected]) got netsplit.
[11:46] Toastie ([email protected]) joined #gstreamer.
[11:47] Toastie ([email protected]) left irc: Client Quit
[11:47] sxpert ([email protected]) returned to #gstreamer.
[11:50] pb_ (~pb@2002:5160:4550:1:240:95ff:fe30:aa84) joined #gstreamer.
[11:52] ploum ([email protected]) left irc: "Leaving"
[11:55] olimar ([email protected]) joined #gstreamer.
[11:55] <olimar> Hi all
[11:55] <wingo> hi
[11:55] <olimar> I compiled gst-player against gstreamer 0.8.0 rpm packs but I get this error:
[11:55] <olimar> hi wingo btw :)
[11:56] <olimar> ** (gst-player:4351): WARNING **: libgstplay: failed initializing pipeline, error: Could not link video output thread (cs and balance)
[11:56] <teuf> I got this kind of error when I didn't have gst-ffmpeg installed iirc
[11:56] <olimar> Aha
[11:57] <olimar> I got yesterday also a hint about gst-ffmpeg, but teuf I am using fedora rpms the gst-plugins are all bundeld in one rpm
[11:57] <teuf> olimar: gst-ffmpeg is now a separate cvs module and is no longer part of gst-plugins
[11:57] <olimar> I can't install any other gst plugin in this setup (well at least I think so)
[11:57] <olimar> Aha
[11:57] <olimar> cool
[11:58] <olimar> ok I'll try it right now
[11:58] <wingo> hm, is gtk 2.4 in debian experimental?
[11:58] <teuf> wingo: debian experimental has packages for most of gnome 2.6
[11:59] <teuf> so gtk 2.4 must be in there
[12:00] <wingo> i must be a moron, then...
[12:00] <teuf> you can't find it ? 
[12:01] <teuf> which arch are you on ? x86/ppc or something else ?
[12:01] <wingo> x86. i have experimental in my sources.list, but it's just saying 2.2.4 is available.
[12:01] <wingo> anyway, i'll continue poking it.
[12:02] <teuf> wingo: you need to use -t experimental if you want apt-get to look in experimental
[12:02] <wingo> hm -- this is dselect
[12:02] Action: wingo checks what apt-get has to say
[12:03] olimar_ ([email protected]) joined #gstreamer.
[12:03] <olimar_> got kicked out :/
[12:03] <olimar_> teuf, do I have to install gst-ffmpeg from a tarball or is there any RH9/fedora rpm?
[12:03] <teuf> olimar_: no idea
[12:04] <olimar_> and hwat have you done?
[12:04] <olimar_> what
[12:04] <teuf> I'm on debian, and I compile gstreamer from source anyway :)
[12:04] <olimar_> oh man these debian folks .....
[12:04] <olimar_> I'll switch to Debian no way ;)
[12:08] olimar ([email protected]) left irc: Read error: 60 (Operation timed out)
[12:09] Shoragan ([email protected]) joined #gstreamer.
[12:09] iain ([email protected]) joined #gstreamer.
[12:13] <wingo> hey iain
[12:13] <iain> hihihi
[12:13] <olimar_> I found some ffmpeg rpms do you think this is ok? or do I need something called gst-ffmpeg?
[12:13] <iain> wingo: wassup in africa today?
[12:13] bilboed_school ([email protected]) joined #gstreamer.
[12:13] <olimar_> no way :/
[12:13] <olimar_> I installed them but it doesn't work
[12:15] <teuf> olimar_: it seems to be available from http://gstreamer.freedesktop.org/download/fedora.html
[12:15] <teuf> (http://gstreamer.freedesktop.org/pkg/fedora/1/i386/RPMS.gst/)
[12:16] <wingo> iain: nothin, yo. long weekend :)
[12:16] <olimar_> cool thanx for the tip
[12:17] <olimar_> woohoo
[12:17] <olimar_> it finally works!!!
[12:19] <olimar_> hmm but why do I get now this error? (excuse me pls)
[12:19] <olimar_> Alsa device "default" is already in use by another program.
[12:19] <olimar_> but no app is using Alsa!
[12:20] <wingo> dunno, yo
[12:20] <iain> wingo: lucky you...I've been doing insane amounts of time at work, and its really starting to get me down
[12:25] <wingo> that sucks, yo.
[12:25] <wingo> anything happening with marlin?
[12:26] <olimar_> can someone tell me when we'll finally have any mpeg working under gstreamer? (btw I am myself a translator so I need what hard work is...)
[12:26] <olimar_> know
[12:26] Nick change: ChrHJW -> ChristianHJW
[12:26] <olimar_> well I don't need it ;)
[12:28] <olimar_> cz to be honest every time I wait for the latest gstreamer spend hours over hours to make it work and at the end niente nados... whereas with mplayer after one compile everything wokrs, I know the goal of gstreamer but man, when well it finally work?
[12:30] <iain> wingo: company fixed some bugs in it, but I've been having life "issues" over the past few months, so I have no desire to do anything at all
[12:30] <iain> olimar_: it all works for me
[12:31] <olimar_> can I ask which distro?
[12:31] <iain> it used to be redhat 8, but it doesn't have much original stuff left
[12:31] <olimar_> well same thing for me with RH9
[12:31] <olimar_> so what's the problem ???
[12:32] <wingo> iain: well i'm still set to steal a lot of your code :P
[12:33] <iain> wingo: feel free. I hope these issues will sort themselves out...but I don't see that happening soon
[12:33] <olimar_> reinstalled the rpms and now I am using the ones from the gstreamer site
[12:33] <iain> I don't use rpms
[12:33] <iain> I broke RPM in fact
[12:34] <iain> and whenever I run it, I need to reboot my computer, because it locks up the hard drive
[12:34] <olimar_> well that's messy a distro without apt or rpm ;)
[12:35] <olimar_> I could never live without rpm it takes too much time to compile every pack
[12:36] <olimar_> could you make totem run with gstreamer?
[12:36] <iain> yup
[12:36] <olimar_> I compiled it rigt now but it hangs on the command line, at least gst-player starts
[12:36] <olimar_> uh
[12:37] <olimar_> then you have a devil hat not really a redhat distro ;)
[12:38] olimar_ ([email protected]) left irc: Remote closed the connection
[12:47] Nick change: markey -> markey|afk
[12:49] steveb_away ([email protected]) left irc: Remote closed the connection
[12:52] LeRoutier ([email protected]) joined #gstreamer.
[12:52] <LeRoutier> hello ppl
[12:54] <LeRoutier> iain, did you make a conditional patch for this one http://bugzilla.gnome.org/show_bug.cgi?id=137488 (depending on gtk version used) ?
[12:57] <wingo> hi LeRoutier
[12:57] <LeRoutier> hi wingo
[12:58] herzi ([email protected]) left irc: Remote closed the connection
[13:03] sublett ([email protected]) left irc: "I quit, therefore I was"
[13:03] bilboed_school ([email protected]) left irc: Remote closed the connection
[13:03] bilboed_school ([email protected]) joined #gstreamer.
[13:03] steve_b ([email protected]) joined #gstreamer.
[13:03] Nick change: steve_b -> steveb_away
[13:09] <iain> LeRoutier: no, I dont like doing weird autoconf magic
[13:10] <LeRoutier> iain, i'm doing it now
[13:10] <LeRoutier> i'll attach the modified patch once tested
[13:15] Shoragan ([email protected]) left irc: "Leaving"
[13:40] sublett ([email protected]) joined #gstreamer.
[13:47] <LeRoutier> iain, was long but i have it working
[13:47] <LeRoutier> (i hate this autotools stuff too)
[13:49] <lilo_booter> i don't think i've ever met anyone who likes that stuff :-) - guess there must be someone, somewhere...
[13:50] <teuf> LeRoutier: hmm, I just realized that you could just have ripped the rhythmbox stuff, it already does that
[13:51] <LeRoutier> teuf, no problem, at least, i see how it works in doing it myself 
[13:51] tomtom20154 ([email protected]) joined #gstreamer.
[13:52] <LeRoutier> lut tomtom20154
[13:52] <tomtom20154> slt
[13:52] <tomtom20154> peut etre tu pourrais m'aider ?
[13:53] <LeRoutier> dis toujours
[13:53] <tomtom20154> je veux utiliser l'encodeur aac
[13:53] <tomtom20154> par contre j'ai pas la la lib faac
[13:53] <LeRoutier> donc tu ne peux pas. c pas une dépendance pour rien
[13:54] <tomtom20154> sur www.audiocoding.com/, leur lib se compile pas
[13:54] <LeRoutier> quelle distrib ?
[13:54] <tomtom20154> et dans les paquets debian alors que je tourne sous testing, ya pas faac
[13:54] <tomtom20154> par contre ya faad2
[13:54] <LeRoutier> faad c que le decodeur
[13:54] <tomtom20154> 2.4.23
[13:54] <tomtom20154> je pensais bien
[13:55] <tomtom20154> mais il aurait pu mettre l'encodeur avec
[13:55] <teuf> y a peut etre des pb de brevets avec l'encodeur
[13:55] <LeRoutier> quel version de faac tu as essayé ? 1.23.5 ?
[13:56] <tomtom20154> oui
[13:56] <tomtom20154> General FAAC compiling instructions
[13:56] <tomtom20154> 1. Make sure you have autoconf, automake and libtool installed.
[13:56] <tomtom20154> 2. cd to FAAC source dir
[13:56] <tomtom20154> 3. Run:
[13:56] <tomtom20154>         ./bootstrap
[13:56] <tomtom20154>         ./configure
[13:56] <tomtom20154>         make
[13:56] <tomtom20154>         make install
[13:56] <tomtom20154> j'ai ce qui est demandé
[13:56] <tomtom20154> et ./configure connait pas
[13:56] <teuf> thomasvs: http://apt-get.org/search.php?query=faac&submit=&arch%5B%5D=i386&arch%5B%5D=all ils marcheront peut etre sur une testing les paquets, sinon tu peux tenter un apt-get source -n b dessus
[13:56] <teuf> merde
[13:56] <teuf> tomtom20154 plutot
[13:56] <tomtom20154> ci
[13:57] <LeRoutier> tomtom20154, le configure est sensé être généré par le bootstrap donc si il ne le génère pas, c qu'il y a une couille
[13:58] <tomtom20154> ./bootstrap
[13:58] <tomtom20154> bash: ./bootstrap: Permission non accordée
[13:58] <teuf> chmod 755 ./bootstrap peut etre
[14:00] <LeRoutier> arf
[14:00] <LeRoutier> vivi, ça serait mieux
[14:00] <tomtom20154> pauv pomme que je suis
[14:00] <tomtom20154> j'avais pas regardé les droits
[14:01] Company ([email protected]) joined #gstreamer.
[14:01] <LeRoutier> hello Company
[14:01] Action: teuf kicks Company's ass for the sexyteuf link ;)
[14:01] <LeRoutier> iain, modified patch uploaded to bugzilla
[14:02] <LeRoutier> tested to work
[14:02] <tomtom20154> la ca marche impec, le plugin faacenc est reconnu avec gst-register-0.8
[14:02] <tomtom20154> merci bcp
[14:02] <LeRoutier> de rien, et la prochaine fois, quand tu vois "permission denied", vérifie les droits
[14:02] <tomtom20154> ct pas malin de ma part
[14:02] <tomtom20154> c noté
[14:04] <teuf> LeRoutier: what happens when gtk 2.4 can't be found ?
[14:04] <LeRoutier> it takes the old code path, using selector
[14:04] <teuf> did you test it ?
[14:05] <LeRoutier> iain did the main patch to switch from selector to chooser. i modified it to compile one or the other depending on gtk 2.4 presence
[14:05] <LeRoutier> yep
[14:05] <teuf> adding a -DHAVE_GTK24=.... will unconditionnaly enable the filechooser code path, won't it ?
[14:05] <LeRoutier> i suppose it won't if HAVE_GTK24 is not defined
[14:05] tomtom20154 ([email protected]) left #gstreamer.
[14:05] tomtom20154 ([email protected]) joined #gstreamer.
[14:09] Nick change: harshyWork -> harshyZ
[14:27] Nick change: markey|afk -> markey
[14:31] Toastie ([email protected]) joined #gstreamer.
[14:32] Toastie ([email protected]) left #gstreamer.
[14:36] Nick change: LeRoutier -> LeRaway
[14:43] <Company> good morning angels
[14:44] <Company> teuf: it was the first hit for "teuf" on google ;)
[14:44] <teuf> Company: and you are actively contributing at keeping it to the top ;)
[14:44] <teuf> here I get http://www.teufeurs.com/ as the first result
[14:44] <teuf> sexyteuf is second
[14:44] <wingo> howdy Company
[14:45] <Company> teuf: so my contributions didn't work enough ;)
[14:45] <Company> teuf: you should get a blog that i could link to instead :p
[14:45] <Company> hi wingo
[14:46] <wingo> nice blog on companies, btw :)
[14:46] <Company> wingo: i wasn't even sure i should write that
[14:46] <teuf> your blog was nice
[14:46] tomtom20154 ([email protected]) left #gstreamer.
[14:47] <teuf> Company: did you read  http://www.kdedevelopers.org/node/view/408 otoh ?
[14:47] <teuf> it was funny to read that after your blog
[14:47] <teuf> « KDE tends to look around and pick from what works if there isnt something already home grown. even when there is, cooperation is often a key point in decision making. GNOME tends to not care and just come up with their thing, whether or not its extensible beyond their own platform and whether or not it sucks »
[14:50] <Company> teuf: everything he says is essentially correct
[14:50] <Company> teuf: it's just worded the KDE way ;)
[14:51] <Company> teuf: at least if you apply the "gnome tends to not care and do their own thing" to gnome core technologies like the file chooser
[14:52] tomtom20154 ([email protected]) joined #gstreamer.
[14:52] tomtom20154 ([email protected]) left #gstreamer.
[14:52] Action: teuf rereads the blog entry and wonders what is so wrong with the vfs layer that he gets mentioned there
[14:53] tomtom20154 ([email protected]) joined #gstreamer.
[14:53] tomtom20154 ([email protected]) left #gstreamer.
[14:53] <teuf> Company: the way it is written, I read it "kde follows as much as possible freedesktop standards while gnome don't care", which is totally flawed ;)
[14:55] <Company> teuf: i read that as "kde picks an app that works good enough when there is one while gnome has to cook their own to satisfy hig, consistency, etc"
[14:55] TomTom635 ([email protected]) joined #gstreamer.
[14:55] <Company> teuf: which makes a lot of sense to me ;)
[14:56] <wingo> TomTom635: you really shouldn't use irc as root :)
[14:56] <TomTom635> yes
[14:57] Action: Company hides
[14:57] <Company> oh, i'm not root anymore, go me!
[14:57] <teuf> yeah, depends on the apps. imo this is more or less true for gnome-terminal and metacity, but I don't find that many examples
[14:58] TomTom635 ([email protected]) left #gstreamer.
[14:58] <wingo> hehe
[14:59] smoke ([email protected]) joined #gstreamer.
[15:00] <Company> teuf: KDE has the problem of too often selecting something that's not of good enough quality and especially maintainership quality
[15:00] <Company> teuf: arts is dead (though esd is, too), and they're having 15 different media players that all suck to half of the users ;)
[15:01] <teuf> don't say esd is dead, I know esd maintainer!!! (ok, he'd like to get rid of it but...)
[15:01] <teuf> :)
[15:02] <LeRaway> bbl
[15:02] TomTom631 ([email protected]) joined #gstreamer.
[15:02] LeRaway ([email protected]) left irc: "Leaving"
[15:03] <Company> we all know fcrozat and how much he likes esd
[15:05] <TomTom631> does mp4muxer or ffmux_mp4 exist ?
[15:08] <Company> TomTom631: ask BBB (whern he's around)
[15:09] <TomTom631> ok
[15:09] <teuf> Company: did you look at rhythmbox tagging code at all ?
[15:09] <Company> teuf: no, not yet - there was still enough gst code to fix ;)
[15:10] <teuf> I need to try that code some day, but haven't had time yet
[15:10] <teuf> what's up with gstreamer and win32? Is it working at all ?
[15:11] <Company> teuf: haven't heard from the matroska guys anymore
[15:11] <Company> teuf: but ds was working on trying to compile gst under mingw
[15:12] <teuf> k
[15:12] Action: Company found a bug in rb
[15:12] <teuf> I'll be stuck with windows computers for a week, maybe I'll give it a try
[15:12] <Company> if you ctrl-c it during first import dialog, you don't get that dialog again
[15:12] <teuf> Company: hehe, I have exactly the opposite bug :)
[15:13] <teuf> Company: I reenabled this dialog through gconf and can't get rid of it 
[15:13] <Company> d'oh ;)
[15:14] <teuf> you can also try to enable the "quality" column from the prefs if you want an easy to reproduce bug ;)
[15:16] <Company> i just want to find memleaks for now
[15:16] <Company> that was quite successful so far
[15:17] <teuf> yeah :)
[15:17] <teuf> ds complained it still leaked 50MB to import a 5000 songs library 
[15:17] <teuf> that probably would have leaked several gigabytes before all the leak fixes
[15:18] <Company> yeah
[15:19] <Company> if i open rb with all songs already importet, it takes 50mb
[15:19] <Company> if i import all my songs, it takes 90MB now
[15:19] <Company> if i imported them before, it took 500MB
[15:19] Nick change: wingo -> wingo-away
[15:28] <ChristianHJW> <teuf> what's up with gstreamer and win32? Is it working at all ?
[15:28] <ChristianHJW> <Company> teuf: haven't heard from the matroska guys anymore
[15:28] <ChristianHJW> <Company> teuf: but ds was working on trying to compile gst under mingw
[15:28] <ChristianHJW> <---- jcsston isnt working on it anymore, probably because he prefers the native MSVC port instaed of using mingw
[15:28] <ChristianHJW> i ask him on next occasion
[15:28] <teuf> there is a native msvc gstreamer port ?
[15:29] <ChristianHJW> it seems using mingw is not really a nice solution
[15:29] <ChristianHJW> teuf : yes
[15:31] <teuf> any idea what's wrong with mingw ?
[15:32] TomTom631 ([email protected]) left #gstreamer.
[15:35] <teuf> I don't know if that was already mentioned on the chan, but there's an article about gstreamer (using gstreamer in apps I think) in a french paper magazine
[15:36] Action: Company wonders why people from those magazines don't show up in the channels of the projects to ask about/announce it
[15:36] <teuf> yeah, or at least on the mailing lists
[15:37] <teuf> I dunno who wrote the article, I guess I could buy the magazine to get his email address ;)
[15:38] <teuf> Description of the article for those who don't fear french « 56 - Programmation multimédia avec GStreamer
[15:38] <teuf> GStreamer est une API libre permettant de développer facilement des applications multimédias complexes. Celle-ci reprend beaucoup de mécanismes de Gnome et Gtk et ne devrait donc pas dépayser les développeurs habitués à ces environnements.  »
[15:38] <Company> "ne devrait donc pas dépayser" ?
[15:39] <teuf> ne devrait pas dépayser les développeurs habitués a ces environnements => people used to gnome/gtk shouldn't have difficulties getting used to it
[15:40] <Company> ah
[15:47] jdahlin ([email protected]) joined #gstreamer.
[15:54] <ChristianHJW> <teuf> any idea what's wrong with mingw ?  <--- dont ask me, i am not a programmer
[15:54] <ChristianHJW> i would use it, thats for sure
[15:55] <ChristianHJW> i wouldnt care about what compiler was used to make it work
[15:56] wheels ([email protected]) left irc: Read error: 113 (No route to host)
[16:10] <Company> hey, if someone makes gst work with MSVC i'm all for it
[16:11] <Company> as long as it's incorporated into the official build and doesn't break stuff that works otherwise, of course
[16:15] wheels ([email protected]) joined #gstreamer.
[16:21] ploum ([email protected]) joined #gstreamer.
[16:22] <jdahlin> Company: shouldn't be *too* difficult, most of the portability is already in glib
[16:24] <Company> jdahlin: there's quite a bit of issues, like MSVC not doing autotooling and not knowing what vararg macros are
[16:30] <jdahlin> Company: how is that worked around in gtk+?
[16:30] <jdahlin> and autotooling can be replaced by MSVC projects I guess. Someone just needs to create it
[16:32] thomasvs ([email protected]) left irc: "Client exiting"
[16:33] <Company> jdahlin: glib and gtk don't use vararg macros very much, and when they use it they use inline functions in the MSVC case i think
[16:34] <Company> jdahlin: it's mostly a "just needs to create it" thing, and that's a lot of boring work (as is keeping up to date)
[16:34] thomasvs ([email protected]) joined #gstreamer.
[16:35] <jdahlin> Company: true. but some point someone is going to do it
[16:37] <Company> yeah
[16:40] Nick change: wingo-away -> wingo
[16:41] thomasvs ([email protected]) left irc: "Client exiting"
[16:48] trow ([email protected]) joined #gstreamer.
[16:48] thomasvs ([email protected]) joined #gstreamer.
[17:01] Action: wingo wants to hear more about scheduling changes from Company
[17:01] jdahlin ([email protected]) left irc: "Leaving"
[17:02] foser ([email protected]) joined #gstreamer.
[17:03] <teuf> hi foser
[17:03] <teuf> does the rhythmbox ebuild have a reminder that bugs should be reported to bugs.gentoo.org and not directly to bugzilla.gnome.org ?
[17:03] <foser> oi
[17:03] <foser> nope
[17:03] <foser> would be something if we had to move that into every ebuild
[17:04] <Company> wingo: the idea is pretty simple
[17:04] <teuf> well, it would be nice to at least have that in the masked ebuilds
[17:04] <teuf> the broken 0.6.8 and the broken unofficial 0.7.2 ebuild generated a bunch of duplicates in bugzilla.gnome.org
[17:04] <Company> wingo: you have element->iterate (element, reason)
[17:04] <foser> teuf: hmmm ? hows that.. 
[17:05] <foser> teuf: wo don't supply 0.7.2, so don't look at us.. it's bmg fools messing things up again (yeah they're the ones giving the most trouble upstream :/)
[17:05] <Company> wingo: in that function you set up your element for the next iteration - especially what kind of iteration is possible
[17:05] <teuf> foser: I got 2 reports for the 0.7.2 ebuild, but I assume it was based on your masked ebuild, so the warning would have gone through if there had been one in yours ;)
[17:05] <teuf> that's why I mention it
[17:05] <teuf> but there was about 10 reports for the 0.6.8 ebuild (maybe a bit less)
[17:06] <Company> wingo: possible reasons for an iteration range from "data available on sink" or "data requested on src" or "fd available for writing" to "requested time reached"
[17:06] <foser> teuf: stupid users.. i dunno what to do about it.. we try to emphasize that they should always report to our bugzilla, but nobody reads a damn waord
[17:06] <wingo> Company: can you still have chain functions?
[17:07] <teuf> foser: yeah, that's what I was told
[17:07] <wingo> or is it all s/loop/process/ functions?
[17:07] <teuf> foser: I especially liked that bug http://bugzilla.gnome.org/show_bug.cgi?id=139269
[17:07] <teuf> « should I forward this on the
[17:07] <teuf> gentoo people if it is a distro specific issue. »
[17:07] <Company> wingo: you'd replace the chain function with iterate (REASON_PAD_HAS_DATA, data) or something
[17:07] <foser> teuf: you shouldnt be sorry to sound harsh, they wont listen anyway :)
[17:08] <foser> teuf: ill bring it up again on our dev mailinglist once again and try to get a note in GWN this or next week ok ?
[17:09] <Company> wingo: and instead of using gst_pad_push you'd flag the sourcepad as having data available and return
[17:09] <teuf> foser: sure. A warning in rhythmbox masked ebuilds would be enough, no need to make too much fuss about it ;)
[17:09] <foser> teuf: general note..
[17:09] <wingo> interesting thoughts. could you mail the list so we can chew on them a bit? :-)
[17:10] <foser> teuf: i can tell you a note won't work.. ppl think our ~arch is stable or something, it is meant to test stuff and fix those little issues (like the one you posted there is imo)
[17:11] <Company> wingo: dunno, i don'T think it's ready for consumption yet
[17:11] <teuf> foser: I'll guess I flame more people next time I get bug reports because of a broken masked ebuild then
[17:11] <wingo> i do want to sit down and think about it at some point though. anyway, when you've sorted it out.
[17:11] <teuf> I don't care if you masked ebuilds are broken as long as I don't get more work because of it ;)
[17:12] <foser> teuf: well for most ppl they work, because it wont affect an update, only a clean install
[17:12] <foser> teuf: so thats why we didn't come across it ourselves
[17:13] <teuf> the gconf key causing the crash wasn't in 0.6.1, was it ?
[17:13] <teuf> I don't remember at which point it was added
[17:14] <foser> teuf: mmm 0.6.1 has it.. it wasnt in our later ~arch ebuilds.. 
[17:17] <foser> teuf: i'm gonna add some stuff to our eclass to detect such cases earlier i think..
[17:17] <teuf> "such cases" ? non installation of gconf schema files ?
[17:18] <foser> teuf: yep..
[17:19] <foser> teuf: just checked.... 0.6.1 always installed schemas fine
[17:19] <teuf> wouldn't it be possible to automatically detect the app installs a schema and to automatically do the right thing ?
[17:20] <foser> teuf: hmm i'm not sure what you mean ?
[17:23] <teuf> foser: I have absolutely no idea how one makes a gentoo ebuild, but can't you detect that the app will need a schema somehow, and in such cases, have a tool which adds export GCONF_DISABLE_MAKEFILE_SCHEMA_INSTALL=1 before running make install in the ebuild and which runs gconftool to install the schema in the postinstall hook instead ?
[17:24] <Company> wingo: the discussion mostly happens informally here in the chat when someone had an idea about how something could be done  and another one wants to listen to it
[17:24] <Company> wingo: so just hang out in here and listen ;)
[17:24] <foser> teuf: thats what we do already.. and it worked in 0.6.1 obviously, but somewhere between there and 0.6.10 it stops working
[17:24] Action: teuf posts that on the mailing list so that Company gets flamed ;)
[17:24] <foser> teuf: in 0.7 it works again
[17:24] <Company> wingo: some day it'll end up in a branch - that's most likely the point when it'll go "public"
[17:25] <Company> teuf: i don't want that on the lists, that'd end up in too much useless posts ;)
[17:25] <teuf> foser: 0.6.10 should work
[17:25] <Company> teuf: most of the time me answering "no clue, haven't thought about that yet"
[17:26] <teuf> Company: people can raise potential problems you hadn't thought about though
[17:26] <teuf> foser: I think 0.6.10 and 0.7.2 schema installation are identical
[17:27] <teuf> schema installation was broken in 0.6.7 or 0.6.8 though
[17:27] <foser> teuf: hmm ok.. then that was the cause..
[17:28] <Company> teuf: yes, but it doesn't scale like this
[17:29] <Company> teuf: until you have a good idea of what you want you should get lots of people looking at it but only those that are really interested in it
[17:29] <Company> teuf: then when you have a ghood idea, you want those
[17:29] warner ([email protected]) joined #gstreamer.
[17:29] Nick change: warner -> mat1t
[17:34] <Company> woah, we're at 48 people on a holiday
[17:38] <teuf> it's not a holiday here :)
[17:39] <Company> that's even better
[17:39] ploum ([email protected]) left irc: "Leaving"
[17:40] <teuf> there are cheaters though, 2 bilboed, 2 ChriHJW_log, 2 ds, ...  :)
[17:43] bilboed_school ([email protected]) left irc: "Leaving"
[17:43] <Company> hm
[17:43] mat1t ([email protected]) left irc: Remote closed the connection
[17:44] <Company> should i attach a 88k textfile or should i zip it when mailing the list?
[17:44] <teuf> put a link to an online version :)
[17:44] <teuf> otherwise zip it
[17:44] Action: Company has no webspace
[17:45] <teuf> oh ? how is that possible ? :)
[17:45] <Company> noone felt the need to provide me with some and I didn't feel the need to get some
[17:46] <teuf> lazy guy :)
[17:47] <teuf> I got some free webspace from free.fr for my "personal web pages" which I use as a file repo
[17:47] spyder482 ([email protected]) joined #gstreamer.
[18:00] <teuf> hmm, I should try to run rhythmbox in valgrind too
[18:00] <Company> it's too slow for playback here
[18:01] <teuf> yeah, I only plan to run it, add a few songs and quit it
[18:01] <Company> but it's nice to have it running in the background if you do other stuff
[18:01] <Company> "some songs" didn't cut it for me
[18:02] <teuf> there weren't any leaks with only a few songs ?
[18:03] <Company> i got it working without leaks for a few songs
[18:03] <teuf> ah ok
[18:04] <teuf> with recent gstreamer cvs, all I need to do is to run valgrind rhythmbox with the appropraite parameters ?
[18:06] <Company> teuf: yeah
[18:07] <Company> teuf: the comand i used is in the mail ;)
[18:07] Nick change: markey -> markey|RUN
[18:09] <Company> teuf: rhythmbox/widgets/rb-tree-view.c:306 - shouldn't that be "g_array_free (expand_col_widths, TRUE);" =
[18:09] <Company> ?
[18:10] jdahlin ([email protected]) joined #gstreamer.
[18:15] <teuf> Company: looks like it
[18:20] Nick change: spyder482 -> spyder_away
[18:26] teuf ([email protected]) left irc: "Client exiting"
[18:27] <Company> GstData needs a "last pad that looked at it" debugging field so you have a chance of finding where it leaked
[18:36] <Company> anybody in here that knows MPEG?
[19:05] <taaz> i've heard of mpeg
[19:17] yippi ([email protected]) joined #gstreamer.
[19:19] bilboed_school ([email protected]) joined #gstreamer.
[19:20] <bilboed_school> hi all, I need an emergency answer :( Anybody knows when DirectShow was created ?
[19:21] <bilboed_school> I'd say 98, but I'm not sure
[19:28] Nick change: markey|RUN -> markey
[19:42] teuf ([email protected]) joined #gstreamer.
[19:54] <ds> Company!
[19:54] <Company> ds!
[19:54] <ds> you've been away just when I've needed to ask questions
[19:54] <ds> anyway, now I must go to work
[19:55] <Company> i'll be away again tonight
[19:55] <Company> easter weekend, girlfrienbd at home ;)
[19:56] jdahlin ([email protected]) left irc: "sloff"
[20:06] apoc ([email protected]) joined #gstreamer.
[20:14] Action: Company wants gst_exit
[20:34] Action: Company wants BBB now
[20:34] <ds-work> omega and wtay added a patch for it, but they called it before checking it in
[20:37] <Company> it shouldn't be too hard to have it run gst_exit
[20:37] <Company> it's just a lot of boiring work
[20:37] <ds-work> it would be a good job for that unknown newcomer
[20:37] <Company> especially cleaning up the plugins
[20:37] <ds-work> you know, the ones that never want to do boring work...
[20:38] <Company> yeah
[20:38] <Company> one day i'll just add it and then i'm gonna kick people with gst-register valgrinding
[20:38] <Company> after that it's pretty much done
[20:42] <Company> ds-work: could you look at gst-plugins/gst/typefind/gsttypefindfunctions.c:675 and tell me if the data[4] reading is correct?
[20:43] <Company> ds-work: or if that should be data[3]
[20:44] <ds-work> I'd have to look at the spec
[20:44] <ds-work> but I think it's correct, from what I remember
[20:45] <Company> ok, i'll fix it by making sure 5 bytes are available
[20:53] <lilo_booter> hmm - do you guys have any kind of luma/organic wipe effect for video?
[20:53] <ds-work> lilo_booter: see smpte
[20:54] <lilo_booter> yeah - i know the standards - just curious if you have an implementation of it
[20:54] <ds-work> heh, there's an element called smpte
[20:54] <lilo_booter> ah :-)
[20:55] <lilo_booter> to save me hunting around - are the lumas provided as fixed images or generated on the fly?
[20:55] <ds-work> generated
[20:56] <lilo_booter> 16 bit? full complement?
[20:56] <ds-work> I think the processing is only 8 bit, don't remember
[20:57] <dilinger> wow
[20:57] <lilo_booter> ah - k - will check further then (i have a partial set of generated 16bit ones - getting tedious though :-))
[20:57] <dilinger> is there a libtunepimp plugin?
[20:57] <ds-work> dilinger: no
[20:58] <Company> ds-work: can i commit the no-memchunks-for-objects patch?
[20:58] <ds-work> Company: yes, please
[21:01] <lilo_booter> ds-work: thanks - found it - thanks
[21:10] Zeenix ([email protected]) joined #gstreamer.
[21:12] <Zeenix> hi
[21:18] <Company> ok, i found the last big rb "leak"
[21:21] <ds-work> is it possible to deregister a GType?
[21:22] wingo ([email protected]) left irc: Read error: 113 (No route to host)
[21:23] wingo ([email protected]) joined #gstreamer.
[21:25] <Company> ds-work: it is if you use gTypePlugin
[21:25] <ds-work> which is cleverly undocumented :)
[21:25] <Company> so you have shouldn't use g_type_register_static
[21:26] <ds-work> Company: ok, so my question
[21:27] <ds-work> i'm working on the autoplugger
[21:27] <ds-work> after plugging the first element (typefind), I need to iterate a few times for typefind to get buffers
[21:27] <ds-work> however, my pipeline is
[21:27] <ds-work> filesrc ! plugidentity ! typefind   plugidentity ! osssink
[21:28] <ds-work> with typefind not connected to plugidentity
[21:28] <ds-work> (actually, there might be yet another plugidentity there somewhere)
[21:28] <ds-work> this pipeline doesn't schedule correctly
[21:28] <ds-work> in fact, it appears to deadlock
[21:29] <Company> ds-work: if plugidentity just call gst_element_interrupt it should work
[21:29] <Company> ds-work: that's what spider does
[21:29] ploum ([email protected]) joined #gstreamer.
[21:29] <ds-work> ok, I see
[21:32] <ds-work> btw, back on the module loading topic -- we need to change how plugin loading works in order to get it working on windows
[21:32] <ds-work> and from what I've been reading, it's the right direction anyway
[21:32] <Company> why?
[21:32] <Company> and to what?
[21:32] <ds-work> we should link plugins against all dependent libraries (including libgstreamer and other plugins)
[21:33] <ds-work> and load the plugins with RTLD_LOCAL
[21:33] <ds-work> the problem on windows is that you can't link a DLL with unresolved symbols
[21:34] <Company> isn't linking plugins unportable?
[21:35] <ds-work> not if we do it very carefully :)
[21:35] <ds-work> the way we do it now, yes, it's unportable
[21:36] <Company> i'll leave that to you to figure out, i'll just ask dumb questions later ;)
[21:36] <ds-work> http://people.redhat.com/drepper/dsohowto.pdf
[21:36] <ds-work> if you are curious
[21:37] <Company> ah yeah, i wanted to read that, but didn'T to yet
[21:39] <ds-work> anyway, I haven't worked out all the details, but I don't think it can reasonably done without breaking ABI
[21:39] <ds-work> unless we want to have two classes of plugins
[21:39] <ds-work> also, it would make sense to start using GTypePlugin
[21:41] <Company> agreed on both points
[21:54] arana ([email protected]) joined #gstreamer.
[21:59] jdahlin ([email protected]) joined #gstreamer.
[22:01] teuf ([email protected]) left irc: "Client exiting"
[22:02] teuf ([email protected]) joined #gstreamer.
[22:03] <jdahlin> Company: why should plugins use malloc instead of memchunk?
[22:04] <Company> jdahlin: 1) valgrind 2) it doesn't buy anything if it's not something that's allocated/freed constantly
[22:06] <jdahlin> Company: is it not possible to have a "emulation layer" for valgrind, so memchunks are malloced?
[22:09] <Company> jdahlin: you could definitely add valgrindability to memchunks, but that would still not make it needed in this case
[22:09] wingo ([email protected]) left irc: Read error: 60 (Operation timed out)
[22:10] spyder_away ([email protected]) left irc: Read error: 113 (No route to host)
[22:10] abo_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[22:11] <Zeenix> ds-work: you write/edit docbook docs in ordinary editors (vi/emacs) ?
[22:14] hallibaby ([email protected]) joined #gstreamer.
[22:16] <ds-work> yes
[22:22] <jdahlin> Company: ok
[22:22] <jdahlin> tonight I'm going to a Jazz club called Jamboree
[22:24] wingo ([email protected]) joined #gstreamer.
[22:41] Nick change: jdahlin -> jdahlin_jamboree
[22:46] spyder__ ([email protected]) joined #gstreamer.
[22:46] arana ([email protected]) left irc: Read error: 60 (Operation timed out)
[22:47] walters ([email protected]) joined #gstreamer.
[22:55] Action: wingo is back
[22:55] <wingo> and hacking all night :-)
[22:59] <wingo> hi walters and ds-work
[23:01] <ds-work> yo
[23:02] <walters> hey
[23:08] Action: ds-work makes sure everyone knows that he hates Solaris
[23:09] Action: teuf makes sure everyone knows he hates gentoo
[23:11] markey ([email protected]) left irc: "g'nite, I'm gonna dream in infrared"
[23:12] Action: wheels high fives teuf
[23:12] <wheels> ...and I guess ds-work while I'm at  it.
[23:13] <wheels> teuf: But look at it this way -- previously people argued over whether they hated SuSE or Redhat or Caldera or whatever.  Now we can all agree on hating Gentoo.  It's really a community building experience.
[23:13] sub_pop ([email protected]) left irc: "Client exiting"
[23:13] <teuf> hehe
[23:14] <ds-work> curiously, Gentoo absorbed all those Debian users that Debian developers used to get annoyed at
[23:14] <teuf> of course, debian suX0rZ it's not l33t enuf
[23:15] sub_pop ([email protected]) joined #gstreamer.
[23:15] <ds-work> bingo
[23:15] <ds-work> "Debian sucks because it doesn't have a libc optimized for i686"
[23:15] <teuf> wheels: if you were a real unix hacker, you'd know you have to use gentoo
[23:16] <wheels> ds-work: Ok, well, Debian does suck for that.  :-)
[23:16] Action: wheels had to learn how to build a deb for a i686 optimized glibc a while back...
[23:16] <ds-work> wheels: actually, Debian has that, but it makes 0 difference in performance (except causing more bugs)
[23:17] <ds-work> over the last few years, I've gotten *very* comfortable with not having to search for/compile/install software
[23:17] <ds-work> Solaris is making me compile things like GNU m4
[23:18] <wheels> ds-work: In our case (work) it was for some stack bounds checking that's only done in the 686 build...
[23:18] <ds-work> wheels: libc6-i686.  You know you want it
[23:20] <wheels> ds-work: Couldn't find it  at the time (Hmm, I guess about 4 months ago) and at this point it's a long passed issue...
[23:20] <ds-work> yeah, it's new
[23:21] Action: wingo has pity on ds-work for having to compile the whole stack on solaris
[23:22] <Company> ds-work should install JDS
[23:22] <Company> that pretty much requires having GNU m4
[23:22] <teuf> it could be worse, he could have to compile the whole stack on aix
[23:23] <Company> glib needs valgrind support for memchunks
[23:28] <wingo> so in all this talk recently i've been thinking about realtime operation of gst.
[23:28] <wingo> generally the problem boils down to rt-safe dynamic allocation.
[23:29] <wingo> memchunks solve half the problem -- for data objects of known size.
[23:29] <wingo> but perhaps we need to think about providing an rt-safe allocation library for allocations of unspecified (but limited) size.
[23:32] <Company> it sounds easier to me to find a realtime libc
[23:34] kwm ([email protected]) left irc: "shutdown -p now"
[23:36] <ds-work> memchunks don't solve the problem
[23:37] <ds-work> we have a pretty serious problem with plugging while playing in rt, since negotiation uses caps
[23:38] <ds-work> we use malloc _heavily_
[23:39] Action: Zeenix remembers his program once got atleast three times slower on shifting to memchunks from malloc
[23:39] <ds-work> my recommendation would be to use a malloc implementation that is mostly immune to latency issues
[23:39] <Company> realtime is not about fast
[23:40] <Company> it's about not slow
[23:41] CompIBook ([email protected]) joined #gstreamer.
[23:50] Action: ds-work should move swfdec CVS to fdo
[23:52] <CompIBook> fdo: the new sourceforge for gnome or kde endorsed projects
[23:53] <ds-work> fdo is a sourceforge refugee camp
[23:54] guaiacum ([email protected]) joined #gstreamer.
[23:55] <Zeenix> ds-work: do you know why the <email> element is'nt allowed under the <author> element in an <artheader> by the docbook?
[23:56] <ds-work> Zeenix: isn't it something like <affiliation><address>[email protected]...?
[23:56] <Zeenix> ds-work: let me see the ref.
[23:57] <ds-work> that's off the top of my head
[23:58] Action: CompIBook wonders if big projects can reasonably do a 6month release cycle
[23:58] <ds-work> like GNOME?
[23:58] <CompIBook> like gst or gtk
[23:58] <ds-work> I think so
[23:58] <CompIBook> gnome is lots of small projects
[23:58] <ds-work> we just need some experience at it
[23:58] <Zeenix> ds-work: the description for the <affiliation> says "The institutional affiliation of an author, contributor, or other individual."
[23:59] <ds-work> Zeenix: <email> is a subtag of <affiliation>, iirc
[23:59] <CompIBook> ds-work: the problem is if you want to do big changes
[00:00] --- Sat Apr 10 2004
[00:00] <CompIBook> ds-work: like gst switching scheduling or gtk embracing the rendering model of render/cairo
[00:01] <ds-work> CompIBook: as long as 25 other things aren't tried to be stuffed into the same cycle, I think it's ok
[00:01] <CompIBook> ds-work: i don't think it works in 6 months
[00:01] <ds-work> we tried to stuff too many things into 0.8, and moreover, we were attempting to be both a feature-based and time-based schedule
[00:01] yippi ([email protected]) left irc: "Client exiting"
[00:01] <CompIBook> ds-work: you need at least 2 months for stabilizing
[00:01] LeRoutier ([email protected]) joined #gstreamer.
[00:02] <CompIBook> ds-work: that leaves 4 months at most for implementing
[00:02] <LeRoutier> night
[00:02] <CompIBook> ds-work: and if you do something that requires touching lots of code in a non perl-doable way
[00:02] <Zeenix> ds-work: oh ok, got it now
[00:03] <ds-work> you can have much of the basic implementation done before branching
[00:03] <ds-work> so in effect, yeah, it would be longer than 6 months
[00:03] <CompIBook> that would require core devs working on a non-HEAD branch
[00:03] <ds-work> yup
[00:04] <ds-work> at least for a while
[00:04] <CompIBook> yeah
[00:04] <CompIBook> which would make the realase pretty useless
[00:04] <LeRoutier> did i ear 'perl' ?
[00:04] <ds-work> realistically, it would only be one or two developers
[00:05] <CompIBook> in our case that"d be pretty much all core devs
[00:05] <ds-work> yes, but that wouldn't affect gst-plugin developers
[00:06] <ds-work> and during the real cycle, the patch gets applied, and everyone works on cleaning up all the parts that are broken
[00:07] <CompIBook> but you can't do it in 6 months from start to finish
[00:09] <CompIBook> i guess this is why gtk doesn't follow the 6 months cycle
[00:11] <CompIBook> it's probably impossible to switch bigger parts of a project hierarchy from one thing to another
[00:12] <CompIBook> i don't think all of gnome uses action based menus for example
[00:13] <CompIBook> i wonder how the bonobo->dbus transition will work
[00:14] <teuf> ssshh don't say that, dbus is not a replacement for bonobo, nothing is decided yet, blah blah blah
[00:17] <CompIBook> yeah, we'll probably end up with two systems
[00:17] <CompIBook> and do a "gradual" transition
[00:17] <CompIBook> or translated: there'll be a time where nothing works but after that everyone will use dbus
[00:18] walters ([email protected]) left irc: Client Quit
[00:21] Nick change: harshyZ -> harshy
[00:21] guaiacum ([email protected]) left irc: "Leaving"
[00:23] bilboed ([email protected]) left irc: "Client exiting"
[00:25] <CompIBook> the probelm they'll have is trying to replace corba in 6 months
[00:25] <CompIBook> gnome will not do any big changes in 6 months anyway
[00:27] <dilinger> i'm still kind of annoyed i wasted time on corba back in like '99
[00:27] <CompIBook> you can't just switch a technology in 6 months
[00:27] <CompIBook> well, i dunno how good or bad corba is
[00:27] <CompIBook> i just know that the interface sucks
[00:27] <CompIBook> or else people wouldn't remove bonobo but embrace it
[00:27] <dilinger> the concept of corba is cool, it's just far too complex
[00:28] <CompIBook> i mean hey, they embrace gst
[00:28] <LeRoutier>  
[00:29] <LeRoutier> got : gst-launch-0.8 v4l2src device=/dev/video0 ! spider ! ximagesink  : crashing, anyone interrested i should i bugzilla ?
[00:29] <LeRoutier> i should i => or should i
[00:29] <ds-work> that reminds me -- we should push off some of our crap to glib or another library
[00:31] <CompIBook> what's "some of our crap apart from debugging?
[00:34] <LeRoutier> opened new bug with back-trace : http://bugzilla.gnome.org/show_bug.cgi?id=139617
[00:37] Nick change: spyder__ -> spyder482
[00:37] <ds-work> GstStructure
[00:37] <CompIBook> ah right, and all the gstvalue stuff
[00:37] <ds-work> GstAtomic
[00:38] <ds-work> gstmemchunk
[00:39] <ds-work> gsttrashstack
[00:40] <ds-work> wow, we have a lot of crud
[00:41] <CompIBook> gstmemchunk and gsttrashstack should die
[00:41] <CompIBook> GstAtomic is already dead
[00:41] NFusi0n ([email protected]) joined #gstreamer.
[00:42] <NFusi0n> Hey, do you think this is portable to gstreamer? http://xmms-projectm.sourceforge.net/
[00:42] <ds-work> wow, I want to start 0.9 just to remove junk from the core
[00:43] <CompIBook> heh
[00:43] <LeRoutier> NFusi0n, so, it's a VuMeter no ?
[00:44] <CompIBook> NFusi0n: not nicely
[00:44] <CompIBook> NFusi0n: you'd have to make it an output plugin because it requires hardware access
[00:44] <CompIBook> NFusi0n: and gst has issues with opengl because of threading
[00:47] <CompIBook> LeRoutier: try updating core to see if your segfault has magically disappeared ;)
[00:47] <LeRoutier> ok, i saw your commit, recompiling now
[00:49] <LeRoutier> i retry it as soon as it is recompiled
[00:49] <CompIBook> k
[00:54] <teuf> foser: do you disable bug-buddy in gentoo ?
[00:54] <foser> teuf: nope
[00:54] <teuf> so this comment http://bugzilla.gnome.org/show_bug.cgi?id=139269#c5 is valid :)
[00:55] <teuf> foser: any idea if it would be doable to disable it/redirect it to bugs.gentoo.org ?
[00:56] <LeRoutier> Company, thanks, your fix worked perfectly
[00:56] <foser> teuf: valid.. 
[00:56] <foser> teuf: disable it.. piece of cake.. it's hooked somewhere in libgnome isn't it (?)
[00:56] <LeRoutier> and as v4l2 is really better than v4l, i don't have video size nego problems i had with the older
[00:56] <foser> teuf: redirect.. it's not the bugs that are useful to us i guess...
[00:57] <LeRoutier> foser : doesn't it use extended info in .desktop files ?
[00:57] <foser> teuf: dunno.. most distros have it enabled (?)
[00:57] <teuf> I mean, if people get a nice dialog asking them to report a bug, they can't be blamed for doing so, so the comment has a point :)
[00:57] <teuf> foser: mandrake disable it
[00:57] ploum ([email protected]) left irc: "Leaving"
[00:57] <teuf> I wouldn't be surprised if redhat/fedora did something to it too
[00:57] <foser> teuf: mmm..
[00:57] <LeRoutier> X-GNOME-Bugzilla-* in .desktop files i think
[00:58] <foser> dunno.. its 2 sided.. either you want a lot of bugs and screen the correct ones or you want little bugs and hope they are good
[00:58] <foser> if we get them, chances are lots of them just aren't gentoo bugs
[00:59] <CompIBook> someone patch bug-buddy to download a list of known crashers and only show up if it's unknown :)
[00:59] <LeRoutier> most bug-buddy based gdb dump are useless as they come from stripped binaries
[00:59] <foser> LeRoutier: uh yeah.. you stil have a handler somewhere
[00:59] <CompIBook> in that case kick the packager for stripping binaries
[00:59] <foser> they sure are.. even if it's unstripped it is usually useless
[00:59] <teuf> yeah, there's a sigsegv handler in libgnomeui or something like that
[01:00] <foser> but backtraces aren't always needed to give a bug away
[01:00] <foser> you should question the validity of bug-buddy if you think its a lot of bogus stuff..
[01:01] <foser> it now pops up a question doesn't it ... like 'do you want to report this bug' or something
[01:02] Zeenix ([email protected]) left irc: "The whole problem with the world is that fools and fanatics are always so certain of themselves, but wiser people so full of d"
[01:02] <teuf> bugbuddy backtraces can give you an idea about how many people get a given crash, then you can try to get backtraces with symbols for crashes which happen often and are repoduceable
[01:02] <CompIBook> the problem is that bugbuddy doesn't scale
[01:02] <CompIBook> so people will need to invent something sooner or later
[01:03] <LeRoutier> there was a "unique stack trace dupe finder" that was looking for dupes in gnome some monthes ago. don't know if it is still in activity
[01:03] <foser> we could probably hack something on the gentoo level to only suggest to report if binaries aren't stripped and built with debug info..
[01:05] <LeRoutier> and that could propose to rebuild the app with no optimisations and without striping
[01:06] hallibaby ([email protected]) left irc: "Bye bye"
[01:06] <foser> nah.. a link at best.. but ppl don't care to go trough much trouble... and a backtrace from just one app with debug info still isn't that useful i think
[01:07] teuf ([email protected]) left irc: "Client exiting"
[01:10] <CompIBook> depends on what the user does
[01:11] <CompIBook> there's a lot of reports for gstreamer that i can fix by looking at the bt
[01:11] <CompIBook> like the thing from LeRoutier i just fixed
[01:11] Nick change: wingo -> wingoz
[01:15] <foser> but was that with or without debug info ?
[01:15] <LeRoutier> foser : full debug
[01:16] <LeRoutier> i compile full gstreamer related packages with -O0 -g + nostrip
[01:16] <iain> yo
[01:16] <LeRoutier> hi iain
[01:16] spyder482 ([email protected]) left irc: "Leaving"
[01:17] <LeRoutier> (all but gst-ffmpeg because any lib/program using ffmpeg can't compile well with anything else than -O1 on my system)
[01:17] <foser> LeRoutier: well.. thats my point..
[01:18] <LeRoutier> or to rephrase what some gcc people wrote on a mailing list : ffmpeg code is pure shit
[01:18] <foser> well the lower -Os usually don't make that much difference in my opinion.. it's more the wierd stuff
[01:19] <CompIBook> -O2 is fine
[01:20] <CompIBook> since up to -O2 gcc doesn't do fancy stuff like reordering etc
[01:20] <foser> that -O's of course.. not Os ;)
[01:22] <LeRoutier> Company, i did a nasty  pipeline (one more) and got a : Internal GStreamer error: pad problem.  File a bug
[01:23] <LeRoutier> is that one correct ? v4l2src device=/dev/video0 name=v ! ffmpegcolorspace ! ximagesink ! .v alsasink device=hw:0
[01:24] <CompIBook> i don't think v4l2src does audio output
[01:24] smoke ([email protected]) got netsplit.
[01:25] <CompIBook> and anyway, it would have been  ... ximagesink  v. ! alsasink
[01:25] <CompIBook> there should be a warning that it can'd link ximagesink and alsasink
[01:26] Action: CompIBook pokes wheels with a mail on the list on a pkgconfig stick
[01:26] <LeRoutier> hum, from where does sound from TV card is supposed to come so ?
[01:27] <LeRoutier> oh, an alsasrc device, i see
[01:27] <CompIBook> be happy that yours is provided as an alsa device already
[01:27] <CompIBook> i'm using alsa but need to use oss for my TV card ;)
[01:29] <LeRoutier> got a bt878, alsa does not support it from old times
[01:29] <LeRoutier> it was new in 1.0.0 or around
[01:30] smoke ([email protected]) got lost in the net-split.
[01:34] <LeRoutier> to be precise : 2003-12-05  	1.0.0rc2 release | ENJOY! | bt87x driver added
[01:35] <CompIBook> heh
[01:35] <CompIBook> i need an saa7134 driver though
[01:38] CompIBook ([email protected]) left irc: "Client exiting"
[01:43] ploum ([email protected]) joined #gstreamer.
[01:44] <LeRoutier> hello ploum
[01:45] <ploum> hello
[02:01] <LeRoutier> anyone know a VBI decoding lib that is quite complete ?
[02:02] <LeRoutier> say it decodes XDS, VPS, PDC and the like
[02:02] Nick change: jdahlin_jamboree -> jdahlin
[02:05] <Company> gnight
[02:05] <LeRoutier> bye Company
[02:05] Company ([email protected]) left irc: Remote closed the connection
[02:08] BBB ([email protected]) joined #gstreamer.
[02:09] ploum ([email protected]) left irc: "Leaving"
[02:09] <LeRoutier> hi ronald, still ill ?
[02:10] <BBB> yes
[02:10] <BBB> went to work today
[02:10] <BBB> shouldn't have done that :(
[02:10] Action: BBB dropped half-dead halfway the working day
[02:10] <BBB> bah
[02:13] <LeRoutier> Company just fixed a basic sched bug
[02:13] <BBB> where is he?
[02:13] <BBB> :)
[02:14] Action: BBB wonders how 'allow-overwrite' is clearer than 'erase-ask'
[02:14] <LeRoutier> he just went away a minute before you joined
[02:14] iain ([email protected]) left irc: "Novocaine for the soul"
[02:15] <BBB> ah
[02:16] ChristianHJW ([email protected]) left irc: Read error: 110 (Connection timed out)
[02:42] <LeRoutier> BBB: know anyone working on vbidec those days ? (or is it an old unmaintained piece of code)
[02:44] <BBB> it's unmaintained
[02:46] <LeRoutier> ok, looking inside it, i see it only decodes XDS & CC, which are north american standards only
[02:48] <LeRoutier> i'll try to see if there's a way to re-use zvbi lib to do all the work as it has far more features
[02:53] sublett ([email protected]) left irc: "I quit, therefore I was"


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click