IRC logs for #opendune on OFTC at 2010-03-16
⏴ go to previous day
00:08:33 <DorpsGek> SVN: glx (r996) [JIT] -Add: mapped another function (voices + music/sound)
00:09:20 <DorpsGek> SVN: glx (r997) -Update (r996): update decompiled code
00:09:47 <glx> very instable (tends to hang)
00:14:03 <Xaroth|Work> glx just wants to claim r1000 tbh :P
00:15:43 <glx> instability is not surprising as we are not supposed to use mpu and sb at the same time
00:16:30 <glx> it's or sound blaster for sound/music/voices, or mpu for sound/music and no voices
07:42:21 *** ChanServ sets mode: +v Yexo_
09:32:08 <Xaroth|Work> sooooo cloose to r1000
10:31:02 *** Yexo__ has joined #openDune
10:50:57 <TrueBrain> welcome veK and arik181
12:56:33 <TrueBrain> what is the 'instability' factor you talked about?
13:00:49 <glx> btw voices "don't work" with mingw (timing issue)
13:01:59 <TrueBrain> any idea why exactly?
13:03:30 <glx> for some reason audio test takes a little more time on mingw
13:11:18 <glx> it's like on interrupt just force pic loop
13:17:49 <glx> oh and voices seem to play better with that indeed
13:26:29 <glx> seems to be more stable too
14:03:58 <glx> hmm voices don't work in VM today
14:19:26 *** Yexo__ is now known as Yexo
14:24:05 <glx> weird yesterday it worked
16:10:46 <TrueBrain> hmm .. I hear no voices :(
16:11:37 <TrueBrain> timing is much better now btw
16:11:46 <TrueBrain> the intro shows the DUNE II rotate much more clean
16:12:12 <TrueBrain> the intro voice is still scary as hell
16:13:19 <TrueBrain> shooting is much better :)
16:13:49 <TrueBrain> I would say almost perfect
16:14:03 <TrueBrain> segfault at end of intro
16:14:05 <TrueBrain> that is not so nice
16:14:37 <glx> happened to me but only in VM and not always
16:14:49 <TrueBrain> voices make the game more real :)
16:15:06 <TrueBrain> it does avoid the text messages too :)
16:15:41 <glx> yeah "construction complete" is better
16:16:03 <glx> or "enemy unit approaching from the south"
16:16:15 <TrueBrain> crash.bin at winning of map :)
16:16:44 <TrueBrain> or destroying a building .. let me check
16:16:54 <TrueBrain> yes, destroying a building
16:17:16 <TrueBrain> I like the Atreides voice :)
16:17:43 <TrueBrain> destroy your own building :p
16:18:40 <TrueBrain> running intro in gdb
16:18:48 <TrueBrain> while I am going to cook some dinner
16:19:19 <glx> emu_B483.c:94 last time I tried intro in gdb
16:22:17 <TrueBrain> 34 return __builtin___sprintf_chk (__s, __USE_FORTIFY_LEVEL - 1,
16:22:20 <TrueBrain> #0 0x0821abce in emu_Unknown_B483_0156 () at /usr/include/bits/stdio2.h:34
16:22:21 <TrueBrain> #1 0x0821ae7b in emu_Unknown_B483_0470 () at src/unknown/emu_B483.c:254
16:22:23 <TrueBrain> #2 0x0821b235 in emu_Unknown_B483_0363 () at src/unknown/emu_B483.c:206
16:22:42 <TrueBrain> sounds like something we can avoid?
16:23:12 <TrueBrain> every new sound a new thread is started, lol
16:24:14 <glx> theorically a new thread is started if sample rate changes
16:44:36 <TrueBrain> then it happens a lot :p
16:44:45 <TrueBrain> yup, segfault twice at the same place
16:45:27 <TrueBrain> on the sprintf indeed
16:45:36 <TrueBrain> not really suprising
16:45:40 <TrueBrain> how can it ever work, I wonder
16:47:21 <TrueBrain> too bad it takes for ever before it is triggered ...
16:48:05 <TrueBrain> glx: 'filename' is something like ?bla.voc, not?
16:48:10 <TrueBrain> where it should be prefixed with the house char?
16:48:17 <TrueBrain> as I really miss something like "%c%s" here
16:48:40 <TrueBrain> ah, so the mask is in the filename?
16:49:02 <TrueBrain> (I am btw amuzed that a glibc function can segfault :p)
16:50:09 <TrueBrain> hmm ... in this case, it is "?YOUR.VOC"
16:51:03 <glx> so it should just output "YOUR.VOC"
16:51:28 <glx> and don't care about prefixChar
16:52:42 <TrueBrain> but .. why is the ? ther ein the first place? :)
16:53:16 <TrueBrain> in fact, there are many which have a ? without a %c
16:58:25 <DorpsGek> SVN: truebrain (r998) -Fix: segfault at end of intro because of invalid houseID
16:58:38 <TrueBrain> ugly solution, but it works :)
16:58:47 <TrueBrain> best I could think off given the knowledge of the other functions :p
17:00:24 <glx> hehe I just translated dune2 source ;)
17:03:15 <TrueBrain> and there of course it is allowed :)
17:03:32 <TrueBrain> this is just bad luck it doesn't work ..
17:03:43 <TrueBrain> it ends up just outside reserved memory
17:03:56 <TrueBrain> many other values would have gone unnoticed
17:04:35 <glx> maybe I should try to use SDL_AudioCVT
17:06:12 <TrueBrain> and what does it do? And why would that be better?
17:06:39 <TrueBrain> it works now, not? And it works good, as far as I can tell :)
17:07:11 <glx> that way I could open audio with a fixed sample rate (like 22050) and convert all played voice files
17:07:24 <glx> without closing and reopening the device
17:12:58 <TrueBrain> I think we would do better if we uncover most of the code related to it, and convert VOC to some other format (one-time action)
18:52:26 <tneo> hello, what is "initializing the MT-32"?
18:52:42 <Xaroth|Work> that, is sound :)
18:52:42 <TrueBrain> sound driver 6 at work :)
18:53:18 <Xaroth|Work> pew pew lazor sound, and stuff
18:53:37 <TrueBrain> mostly the music stuff
18:54:12 <tneo> IC Xaroth got me thinking about effects ;-)
18:55:01 <tneo> the config sound setting doesn't seem to have effect on the volume level
18:56:58 <tneo> nope had to turn alsa timidity volume lower
18:57:22 <TrueBrain> volume is not controlled by OpenDUNE I believe :)
18:57:43 <tneo> then what does the setting do from 1 to 7?
18:58:11 <TrueBrain> 6 or 7 are only available
18:58:16 <TrueBrain> 7 is without the annoying MT32 initializing
18:58:48 <tneo> with 7 I see that message
18:59:06 <TrueBrain> voiceDrv = 1 works with latest trunk btw
19:00:04 <tneo> now sound is lost at all ...
22:21:45 <DorpsGek> SVN: glx (r999) -Add: completed C-ification of emu_B483_0363()
23:24:03 <TrueBrain> 94 voice clips? 94 messages on screen? Omg :p
23:26:48 <Xaroth|Work> and nice, r1000 en-route \o/
23:43:30 <DorpsGek> SVN: glx (r1000) [JIT] -Add: mapped more functions (voices and french)
23:43:56 <DorpsGek> SVN: glx (r1001) -Update (r1000): update decompiled code
continue to next day ⏵