[00:25:21] * KittyCat is back.
[01:23:59] <angasule> KittyCat: my butt-ugly config utility scared everyone away from the openal list :P
[01:33:34] <KittyCat> heh
[02:14:54] *** Walt_ has quit IRC
[02:17:36] *** Walt_ has joined #openal
[02:17:36] *** ChanServ sets mode: +v Walt_
[02:40:35] *** barra_ is now known as barraAway
[04:22:38] *** tsaotsao has quit IRC
[06:28:28] *** qknight has quit IRC
[06:28:42] *** qknight has joined #openal
[06:28:42] *** ChanServ sets mode: +v qknight
[09:18:47] *** barra_away has joined #openal
[09:18:47] *** ChanServ sets mode: +v barra_away
[09:35:50] *** barraAway has quit IRC
[13:57:06] <KittyCat> sweet. got OpenMW working with OpenAL
[14:03:16] *** Alam_Debian has quit IRC
[14:04:49] *** Alam_Debian has joined #openal
[14:04:50] *** ChanServ sets mode: +v Alam_Debian
[14:52:57] * KittyCat is away: sleep
[15:20:23] *** kb1ooo has joined #openal
[15:20:23] *** ChanServ sets mode: +v kb1ooo
[15:55:14] *** juanmabc has quit IRC
[16:05:48] *** juanmabc has joined #openal
[16:05:48] *** ChanServ sets mode: +v juanmabc
[16:55:29] *** tsaotsao has joined #openal
[16:55:29] *** ChanServ sets mode: +v tsaotsao
[17:47:53] *** juanmabc_ has joined #openal
[17:47:53] *** ChanServ sets mode: +v juanmabc_
[17:56:19] *** juanmabc has quit IRC
[17:56:50] *** juanmabc_ is now known as juanmabc
[18:02:58] *** tsaotsao has quit IRC
[18:14:27] *** prophile has joined #openal
[18:14:27] *** ChanServ sets mode: +v prophile
[18:46:30] *** per has joined #openal
[18:46:31] *** ChanServ sets mode: +v per
[18:46:42] <per> hello
[18:47:50] <per> anyone experienced initialization conflicts between opengl and openal? if i start and stop openal many times during a stress test after initializing the first opengl flip, sooner or later openal will crash during a thread operation, both with openal si and openal-soft
[18:48:21] <per> opengl+sdl, that is
[18:51:01] <juanmabc> opengl flip? can you show the code?
[18:53:13] <per> the code is a bit big, the test is internal to the warzone2100 trunk source code
[18:55:05] <per> but a perhaps related issue is that if i run this http://rafb.net/p/hZ2oQx93.html under valgrind, i seem to fork-bomb my system (works fine without valgrind), and alsa reports lots of errors from calls into alsa.
[18:57:31] <per> the bug backtraces look like this: http://rafb.net/p/mDy1pt27.html
[19:02:19] * angasule reads
[19:08:12] <angasule> per: did you test with the OSS backend?
[19:08:54] <angasule> per: you're on ubuntu, right?
[19:09:29] <per> angasule: i'm on fedora9
[19:09:46] <per> not sure if i have oss on fedora9?
[19:10:38] <angasule> fedora9 also uses pulseaudio?
[19:11:19] <angasule> btw, *the* guy to talk with is Chris Robinson aka KittyCat, but he's asleep right now (and probably for a few more hours)
[19:11:53] <per> yes, it uses pulseaudio
[19:12:23] <per> i'll see if i can catch him :)
[19:12:57] <angasule> ah, I guess a few more hours is a problem for you (just noticed the .no)
[19:13:28] <angasule> also, there is the openal email list, which has been quiet the last week but is helpful
[19:13:42] <per> it is just seven here yet
[19:14:14] <angasule> I'm looking at the code right now, anyway, maybe I'll find the problem
[19:15:03] <per> that would be awesome. can you reproduce the problem?
[19:16:15] <angasule> let me rty
[19:16:24] <angasule> not my day, as far as spelling is concerned, sorry
[19:18:53] <angasule> ah, don't have valgrind installed
[19:22:25] *** barra_away is now known as barra_
[19:36:20] <angasule> hmm
[19:36:30] <angasule> I've found something strange in the closing of threads
[19:42:20] <angasule> hmm, found another bit that doesn't make sense to me
[19:43:55] <angasule> in StopThread (pthread version), inf->ret isn't modified at all, what's the point of placing it in ret and returning its value?
[19:44:58] <angasule> in alsa_close_playback in alsa.c data->killNow = 1 seems to be completely pointless since that data can't be accessed by called functions and data is freed before returning
[19:46:36] <angasule> KittyCat: those two lines are for you, when you come back
[19:48:13] <per> i would be nice if the problem wasn't in our sound code, because it is such mess ;)
[19:48:23] <per> it would...
[19:48:53] <per> i am always nice ;)
[19:49:29] <angasule> heh
[19:49:38] <angasule> I have heard you reset connections :P
[19:50:11] <per> no, that is my longer brother, peer. i take no responsibility for his actions...
[19:51:01] <angasule> ah, the killNow is a signal for the thread
[19:51:07] <angasule> hehehehe
[19:51:09] <angasule> your evil twin? :)
[19:51:12] <per> yeah
[19:51:18] <per> killNow sends what kind of signal?
[19:51:51] <angasule> well, killNow is an int, used as a boolean to decide if alsa should process sound, it's internal to OpenAL-soft, don't worry :)
[19:53:34] * per pulls down the openal-soft code
[19:57:36] <angasule> it's always a good idea to review code, but I'm fearing this might be a pulseaudio issue
[19:59:10] <per> do you know if i can easily switch audio daemons under fedora9 to test this?
[19:59:30] <angasule> no, sorry, I've never used fedora 9
[19:59:38] <per> i'm completely noob about audio issues and programming
[20:04:05] <per> i can't see anything wrong with the openal-soft threading code
[20:08:36] <angasule> me neither, except I'd feel more comfortable with a 'volatile' for killNow
[20:10:10] *** juanmabc has quit IRC
[20:10:13] *** juanmabc_ has joined #openal
[20:10:13] *** ChanServ sets mode: +v juanmabc_
[20:15:45] <per> angasule: yes, that's true.
[20:16:00] <angasule> per: btw, why are you compiling warzone? you're packaging it?
[20:16:19] <per> angasule: i'm a developer
[20:16:34] <per> angasule: trying to figure out a bug that seems to be somewhere in the sound system
[20:17:43] *** juanmabc_ is now known as juanmabc
[20:17:56] <per> it is probably not related to this issue, but it is hard to debug some issues without valgrind working properly
[20:18:09] <angasule> but it's not originally your code?
[20:18:31] <per> no, it is the original GPLed code release ported almost directly to openal
[20:18:52] <per> by someone else, who is no longer active
[20:20:49] <per> the original was written for DirectX, and was quite a mess, too
[20:21:23] <per> sadly we don't have anyone who knows audio well enough to rewrite it
[20:26:44] <angasule> hmm, looks like a very cool game
[20:27:17] * angasule is installing it
[20:27:35] <barra_> hmm I seem to have overlooked the link :-(
[20:27:51] <angasule> what link?
[20:28:13] <per> http://wz2100.net/ ? :)
[20:28:32] <barra_> ohh yeah, they released the warzone code recently, nice
[20:28:42] <per> not just source code, game data, too
[20:29:02] <barra_> yep, nice website :-)
[20:29:18] <barra_> so I'll do some advertizing for FIFE as well :-) http://wiki.fifengine.de
[20:35:39] <angasule> brb
[20:35:56] *** angasule has quit IRC
[20:37:23] *** angasule has joined #openal
[20:37:24] *** ChanServ sets mode: +v angasule
[20:48:31] *** Walt_ has quit IRC
[20:58:12] *** angasule has quit IRC
[21:00:37] *** angasule has joined #openal
[21:00:37] *** ChanServ sets mode: +v angasule
[21:01:02] *** Walt_ has joined #openal
[21:01:03] *** ChanServ sets mode: +v Walt_
[21:08:11] *** tsaotsao has joined #openal
[21:08:11] *** ChanServ sets mode: +v tsaotsao
[21:41:44] *** tsaotsao has quit IRC
[22:00:43] <per> g'nite
[22:00:53] *** per has quit IRC
[22:36:52] *** tsaotsao has joined #openal
[22:36:53] *** ChanServ sets mode: +v tsaotsao
[23:19:46] *** prophile has quit IRC





top