[00:02:42] *** kb1ooo has quit IRC
[01:20:46] *** barra_ has quit IRC
[02:13:16] *** Walt_ has quit IRC
[02:58:02] <KittyCat> angasule, inf->ret is set in StarterFunc for when the user function returns. StopThread reads it after the thread is joined (StartFunc has exitted by then)
[03:00:21] <KittyCat> I'm also willing to bet the crash he's reporting is (one of many) pulseaudio issues..
[03:01:45] <KittyCat> the pastebin'd code works fine here..
[03:06:05] <angasule> yes, that was my guess as well
[03:06:15] <angasule> I'm gonna be heavily bruised tomorrow
[03:06:29] <angasule> btw, I realised what was the deal with inf->ret later
[03:07:04] <angasule> I hope my feet isn't bleeding, I hate taking off bloody socks, so sticky
[03:10:39] <KittyCat> what happened?
[03:12:24] <angasule> slippery floor + slippery slippers = half my body hit one wall, my slipping foot hit the opposite wall
[03:12:48] <angasule> I could say I was doing a real world test for ragdoll simulations
[03:13:08] <KittyCat> ouch :x
[03:13:36] <angasule> yeah, got lucky, didn't hit the head, and it's winter so if I get large bruises they won't be visible
[03:44:07] <angasule> I replied to the mail, and it's pizza time, bbl
[04:42:52] * KittyCat is back.
[04:45:23] *** angasule_ has joined #openal
[04:45:23] *** ChanServ sets mode: +v angasule_
[05:02:14] *** angasule has quit IRC
[05:29:08] *** tsaotsao has quit IRC
[07:00:14] *** Proteus_ has joined #openal
[07:00:14] *** ChanServ sets mode: +v Proteus_
[07:01:33] *** Proteus_ has quit IRC
[07:04:35] *** Proteus has joined #openal
[07:04:35] *** ChanServ sets mode: +v Proteus
[07:13:03] *** juanmabc has quit IRC
[07:14:54] *** Walt_ has joined #openal
[07:14:54] *** ChanServ sets mode: +v Walt_
[07:18:13] *** juanmabc has joined #openal
[07:18:13] *** irc.freenode.net sets mode: +v juanmabc
[07:52:38] *** Walt_ has quit IRC
[08:17:59] <KittyCat> hmm, should I change the default buffer size to 4096, instead of 8192?
[08:43:15] *** Walt_ has joined #openal
[08:43:15] *** ChanServ sets mode: +v Walt_
[09:19:36] *** per has joined #openal
[09:19:36] *** ChanServ sets mode: +v per
[09:19:52] <per> hello
[09:20:33] <KittyCat> hi
[09:22:41] <per> KittyCat: i have a small, openal test program that seems to fork-bomb my system if i run it under valgrind. it just opens and closes openal many times in a row. ever seen anything like this?
[09:22:58] <KittyCat> nope
[09:24:06] <per> the program is: http://rafb.net/p/XuTJ5q63.html
[09:24:25] <per> same result with both openal-si and openal-soft
[09:24:44] <per> it uses the alsa backend
[09:27:14] <KittyCat> hmm. have you tried with pure alsa instead of pulseaudio?
[09:27:40] <KittyCat> I saw your messages from before. I compiled and ran the program normally, and it worked fine..
[09:28:04] <per> i don't know how to skip pulseaudio
[09:28:30] <KittyCat> kill the pulseaudio daemon
[09:28:45] <KittyCat> not sure what it's called..
[09:31:04] <KittyCat> (also so you know, 0,0,-1 is a forward vector for OpenAL, not 0,0,1)
[09:34:51] <per> AL lib: alsa.c:344: Could not open playback device 'default': Connection refused
[09:34:59] <per> doesn't look like it can run without pulseaudio?
[09:35:46] <KittyCat> hmm. did it try to start again?
[09:35:49] <per> ah, it does run, just gives off a silly warning
[09:35:56] <per> no
[09:39:03] <per> my system does not seem to fork-bomb when pulseaudio is not running
[09:39:33] <per> although valgrind complains a lot about things happening inside alsa :/
[09:40:58] <KittyCat> figures. valgrind complains quite a bit about libdl and libGL for me
[09:41:23] <KittyCat> but if it doesn't fork bomb with alsa, then it seems like it's (yet another) pulseaudio problem
[09:42:47] <KittyCat> to be honest, I'm actually surprised openal soft's alsa backend is even working with pulseaudio. last I knew, it didn't.
[09:44:11] <per> worked fine here, i just uninstalled openal-si and installed openal-soft, and everything seemed to work, until i stress-tested it...
[09:45:43] <per> although i must admit i have no idea how these pieces actually work together
[09:47:42] <KittyCat> pulseaudio has an alsa plugin so that calls to alsa get forwarded to pulseaudio (which in turn get rerouted back to alsa..).
[09:57:20] <per> i'll try to complain to the pulseaudio people. thanks for your help.
[09:58:42] <KittyCat> I'd complain to your distro, too. too many people trying to jump ship from a perfectly fine audio api (alsa)
[10:03:39] *** Walt_ has quit IRC
[10:08:58] *** Proteus has quit IRC
[10:47:45] *** Walt_ has joined #openal
[10:47:46] *** ChanServ sets mode: +v Walt_
[11:01:23] *** Walt_ has quit IRC
[11:03:53] *** Walt_ has joined #openal
[11:03:53] *** ChanServ sets mode: +v Walt_
[11:21:43] <KittyCat> angasule_, btw, for an enhancement to the config utility, you can probably make the OSS device selectors use file selection (so you can browse for files, and make the box autocomplete filenames).
[15:16:42] *** barraAway has joined #openal
[15:16:42] *** ChanServ sets mode: +v barraAway
[15:17:00] *** barraAway is now known as barra_
[16:05:10] *** prophile has joined #openal
[16:05:10] *** ChanServ sets mode: +v prophile
[19:03:01] <angasule_> KittyCat: good idea :)
[19:04:08] *** barra_ is now known as barraAway
[19:04:22] <KittyCat> also wondering if maybe I should add an alsoft-specific extension to probe config settings (but not set)
[19:11:24] <angasule_> what do you mean? so that a game can set alsa, etc?
[19:12:10] <KittyCat> no, so things like your configurator can tell at run-time which backends are available, which are the default, and which are currently set
[19:12:47] <KittyCat> *which values are the default, etc..
[19:15:23] <angasule_> btw, regarding the killNow thread, you say "The worst
[19:15:23] <angasule_> that could happen is that it locks up/enters an infinite loop, in which case  a mutex won't do anything better. "
[19:15:51] <KittyCat> it = the thread
[19:15:52] <angasule_> you mean alsa would just output garbage if the thread is killed right after the killNow check?
[19:16:37] <angasule_> I don't think openal should ever segfault, being a library
[19:17:05] <KittyCat> what I mean is a bug in the code the thread runs would, at worst, cause the thread processing to lock up
[19:17:15] <KittyCat> what alsa outputs without getting new data is up to it
[19:17:50] <angasule_> ah, alright
[19:18:05] <KittyCat> and being that the thread would be locked up, whehter the loop checked a member variable or a mutex wouldn't matter, since it wouldn't get to checking it
[19:18:39] <angasule_> I think I had my last ibuprofen on sunday, and I'm low on coffee, I need my drugs :)
[19:18:58] <angasule_> well, the sequence I'm worried about is:
[19:19:01] <KittyCat> heh
[19:19:22] <angasule_> thread0: if !killNow <-- passes, still 0
[19:19:32] <angasule_> thread1: killNow = 1
[19:19:49] <angasule_> thread0: in the middle of the if !killNow block
[19:20:00] <angasule_> thread1: kills the thread
[19:20:04] <angasule_> wait
[19:20:09] <angasule_> it's a join? :?
[19:20:29] <KittyCat> yeah
[19:20:48] <KittyCat> you can't "kill" a thread like that (well, you can, but it's dangerous)
[19:21:12] <angasule_> ah, then my concern was rather silly
[19:21:15] <angasule_> good
[19:21:23] <KittyCat> mainly you'd either wait for the thread to stop, or set the thread to destroy itself on exit and ignore it from then on
[19:21:25] <angasule_> now brb, I'll dig around for some painkillers :P
[19:21:31] <angasule_> yeap
[19:21:37] <angasule_> don't know why I thought it was a kill
[19:21:45] <angasule_> other than the killNow, but I shouldn't be so literal :P
[19:27:35] <angasule_> is the Echo effect very expensive?
[19:27:37] *** angasule_ is now known as angasule
[19:28:49] <KittyCat> I guess there's something with it that's expensive
[19:29:41] <KittyCat> probably the damping and spread
[19:35:33] <KittyCat> can look at the effects extension guide (in the sdk) for what parameters individual effects need
[19:39:41] <angasule> yeap, I should finish reading it, but lately it's like I never have enough time, and when I do I try out ragdoll physics
[19:54:42] *** barraAway is now known as barra_
[19:59:13] * KittyCat is away: sleep
[20:00:28] <angasule> I'm guessing your question was because of OpenAL, so posting here too:
[20:00:29] <angasule> http://en.wikipedia.org/wiki/Quaternions_and_spatial_rotation
[20:00:40] <angasule> I used to do some work on it
[20:02:12] <KittyCat> http://connect.creativelabs.com/openal/Lists/OpenAL/Flat.aspx?RootFolder=%2fopenal%2fLists%2fOpenAL%2fCalculating%20listener%27s%20orientation%20from%20pitch%2c%20yaw%20and%20roll%20angles&FolderCTID=0x0120020001C624363C8DD3449DE26D1760F0364A&TopicsView=http%3A%2F%2Fconnect%2Ecreativelabs%2Ecom%2Fopenal%2FLists%2FOpenAL%2FAllItems%2Easpx
[20:02:16] <KittyCat> (lovely url)
[20:02:36] <Alam_Debian> what a mess!
[20:05:05] <angasule> KittyCat: you posted that?
[20:05:16] <KittyCat> not me
[20:05:22] <angasule> btw, does ANYONE use that PoS?
[20:06:06] <KittyCat> I try not to. but I still try to help people that end up using it
[20:06:36] <angasule> that message about the forum replacing it is silly
[20:08:11] * angasule checks the thread about mailing lists
[20:08:23] <angasule> sometimes these companies feel like they have a single employee working half time
[20:09:22] <KittyCat> heh
[20:10:10] <KittyCat> wonder if dan is the only one at creative that works with openal
[20:11:30] <angasule> no, seriously, a company the size of creative asking *us* about email list software? just do an aptitude search
[20:12:15] <angasule> isn't OpenAL the only way that ISVs can use their cards now? and they have a single guy?
[20:12:22] <angasule> I don't want to work in a big company :(
[20:12:49] <KittyCat> well, it's understandable to ask the list what known list software is good
[20:13:07] <KittyCat> not that we should be the only ones to ask, but among the ones to ask
[20:14:21] <angasule> well, it's more about the timing and such
[20:14:35] <angasule> a forum that requires posts to include HTML shouldn't have been an option
[20:14:46] <angasule> even if this was 1995
[20:15:55] <KittyCat> I don't really like the layout of the site at all
[20:16:06] <KittyCat> it's too cluttered and indistinctive
[20:16:25] <KittyCat> it's like, am I at creative's site, or openal's?
[20:18:52] <KittyCat> go to opengl.org, you're at opengl's site, not SGI's. go to this, you're at creative's site and oh yeah that openal thing.
[20:19:09] <angasule> well, creative behaves like they own openal
[20:19:23] <angasule> they never seemed too worried about creating an ARB
[20:19:47] <KittyCat> they might have to do actual work if people tried to move it forward :P
[20:20:21] <angasule> heh
[20:20:52] <angasule> what do you see as the biggest problems in the spec? there was a lot of talk about allowing doubles...
[20:23:16] <KittyCat> hmm. probably the lack of doubles. eg. alSourcedv
[20:23:37] <KittyCat> though actually the big problem is that one context is current for the whole process
[20:23:45] <KittyCat> though that can't be easilly changed without breaking things
[20:25:37] <angasule> well, the doubles could be added as an extension, I suppose
[20:25:55] <KittyCat> y eah
[20:27:09] <KittyCat> I think besides the context thing, the main problem is in alut. no support for streaming, or reading anything beyond basic wav/voc files
[20:27:18] <KittyCat> doesn't really make use of extensions
[20:28:50] * KittyCat really goes to bed now
[20:28:54] <angasule> well, I'm off for a few hours
[20:29:11] <per> g'nite then :)
[20:31:20] <angasule> ah, hejsann per
[20:31:48] <per> angasule: svensk?
[20:32:20] 
[20:32:26] <angasule> nei, argentinsk
[20:32:38] 
[20:32:39] <angasule> can't type worth a dime, though
[20:32:47] <angasule> ok, have fun :D
[21:54:24] *** Walt_ has quit IRC
[21:59:14] *** Walt_ has joined #openal
[21:59:15] *** ChanServ sets mode: +v Walt_
[22:09:19] *** Walt has joined #openal
[22:09:19] *** ChanServ sets mode: +v Walt
[22:09:24] *** Walt has quit IRC
[22:18:44] *** wild13 has joined #openal
[22:18:45] *** ChanServ sets mode: +v wild13
[22:58:59] *** per has quit IRC
[23:15:25] *** barra_ is now known as barra_fife





top