IRC logs for #opendune on OFTC at 2010-03-15
⏴ go to previous day
00:13:27 <Xaroth|Work> sound noise, or just.. noise?
00:14:08 <glx> seems to come from provided data :)
00:14:20 <TrueBrain> refresh even higher? :)
00:14:21 <Xaroth|Work> even more interesting, is the noise subject to stuff happening, like your mouse moving?
00:14:26 <Xaroth|Work> I had that once with a broken driver
00:14:31 <Xaroth|Work> move mouse => hear noise
00:14:38 <TrueBrain> nothing to do with driver
00:14:47 <TrueBrain> all to do with how things are interconnected, mostly the ground
00:14:49 <Xaroth|Work> no, this one -was- due to driver :P
00:14:54 <glx> anyway I can't really test on windows because I usually trigger an unresolved jump ;)
00:15:19 <TrueBrain> if (len <= (_dma.count + 1)) { <- useless () :p
00:16:12 <TrueBrain> nice progress, seen the smallness of the patch
00:17:41 <TrueBrain> which sound driver?
00:18:35 <glx> the most basic sound blaster :)
00:19:03 <TrueBrain> D1 not yet implemented :o :p
00:19:32 <TrueBrain> it makes more then noise
00:19:38 <TrueBrain> but it repeat itself over and over
00:20:00 <glx> well there are 2 ticks at start
00:20:20 <TrueBrain> but later in the intro
00:20:30 <glx> and I hear something at the intro start yes
00:20:32 <TrueBrain> lol, JIT of course find a whole new set of functions in AB00 :p
00:20:47 <TrueBrain> I should have disabled music :p
00:21:16 <TrueBrain> do you have a dune.cfg with voice on, and rest off?
00:22:36 <TrueBrain> emu_get_memory8(_dma.page << 12, _dma.offset, 0)
00:23:45 <glx> yes page is the high nibble
00:24:18 <TrueBrain> wow, it can damage your ears :p
00:24:31 <glx> full address is page << 16 + offset
00:24:56 <TrueBrain> 0XFFFF <- hihi, nice typo :P
00:25:17 <TrueBrain> it doesn't hurt anything
00:25:19 <TrueBrain> just reads funny :)
00:27:31 <TrueBrain> that gives much less new JIT functions, tnx :)
00:28:51 <TrueBrain> you clearly hear the begin of the track, or something that looks VERY much like that
00:28:56 <TrueBrain> it just keeps running the same fragment :)
00:29:49 <glx> oh of course, maybe I should increase offset ;)
00:29:55 <TrueBrain> something I just did ;)
00:30:20 <TrueBrain> OMG! What an annoying voice
00:30:33 <TrueBrain> and what a poor annoying quality!
00:30:49 <TrueBrain> I am getting all ... scared here
00:31:00 <TrueBrain> the sync is not yet 100% correct
00:31:17 <TrueBrain> I expected another voice for the empirer
00:31:38 <TrueBrain> amazing job glx, really :)
00:31:48 <TrueBrain> the shooting scene is amazing
00:31:55 <TrueBrain> video syncs with audio and visa versa, but .. broken :)
00:33:21 <TrueBrain> so 0.3 will have voices too :)
00:33:55 <TrueBrain> glx: reverting the int10 changes make it better
00:34:13 <TrueBrain> the voices are so fucking primitive
00:34:19 <TrueBrain> it is just amuzing :)
00:34:59 <TrueBrain> glx: maybe we should make the int10 thingy run on the real clock
00:35:06 <Xaroth|Work> I am -so- going to try to get this machine to compile opendune again.. tomorrow
00:35:19 <Xaroth|Work> it's half past one, so i'm going to sleep first.
00:35:26 <glx> TrueBrain: I let you do that ;)
00:35:44 <TrueBrain> glx: if you have the frequency for me it should run on :)
00:36:16 <glx> well it's based on screen vertical sync
00:36:47 <glx> so something like 50Hz or 60Hz I guess
00:36:55 <TrueBrain> now for me too, good night
00:36:57 <TrueBrain> glx: amazing job :)
00:37:08 <TrueBrain> finish it up, check your patch a bit for coding style (mostly spaces), and it is good to go :)
00:37:20 <glx> jit in a vm is not that good
00:37:21 <TrueBrain> owh, and you need a few JIT runs :p
00:37:38 <TrueBrain> I can give you my current JIT patch
00:37:42 <TrueBrain> but it contains too much I noticed
00:37:46 <TrueBrain> at startup it immediatly found things
00:39:18 <TrueBrain> make sure to filter it ;)
00:40:02 <TrueBrain> toc segfaults on it
00:40:31 <TrueBrain> owh, dah ... I had voice.patch in my root
00:41:55 <TrueBrain> well, you know better than me what is important, so I leave that to you :)
00:41:58 <TrueBrain> for now, good night :)
00:42:02 <TrueBrain> voices in OpenDUNE .. lol :)
00:49:40 <glx> arg you left some references of your C-ified functions ;)
00:59:51 <DorpsGek> SVN: glx (r987) -Fix (r986): forgot to remove a reference
13:22:44 <TrueBrain> I found that those unknowns you did, of which I said they were dequeues, are the text you see on the screen, like "Construction complete"
13:22:47 <TrueBrain> but the array is HUGE :p
13:22:56 <TrueBrain> found at least 0x41
13:23:03 <glx> I noticed there's a bug in at least one of the converted B483 ;)
13:24:26 <glx> with emu_B483 stuff, intro says only one string (and it's used for all other messages too)
13:24:47 <glx> but if I reenable decompiled stuff the speak is correct
13:25:07 <glx> so there's an error somewhere ;)
13:25:13 <TrueBrain> the voice? Works fine here, as far as I can tell :p
13:25:26 <TrueBrain> but okay :) The last function there queues messages
13:25:31 <TrueBrain> ah, yes, of course, I only tried JIT :)
13:26:18 <glx> 22A6:1200 syncs with video too
13:28:11 <TrueBrain> but either way, the 312 array is I believe [N][6] big?, the [6] is a bit odd, in that way that [1] .. [3] seem never used
13:28:30 <glx> 218 is probably wrong too :)
13:28:33 <TrueBrain> but okay, it is some static data in the binary, where [4] is the stringID and [5] the voice clip
13:28:39 <glx> as 222 seems to be another array
13:30:02 <glx> 312 has 0xE bytes elements
14:02:54 <glx> oh I think the error is in B483_0470 :)
15:05:04 <DorpsGek> SVN: glx (r988) -Fix (r983): two little mistakes
15:27:01 <TrueBrain> ah, [i], so they are all used after all :)
15:46:22 <glx> and probably voice index ;)
16:51:15 <glx> TrueBrain: I found the toc bug I had
17:05:07 <TrueBrain> it treats ALL tiles
17:13:39 <glx> I think it should treat only .c and .h, or at least not *~
17:15:10 <TrueBrain> how is that a problem?
17:15:18 <TrueBrain> and yes, it should, but it doesn't :p
17:15:20 <TrueBrain> it should only use .txt files
17:15:32 <TrueBrain> it should only use .c and .h
17:16:07 <glx> it's a problem because *~ files are usually "backup" files with outdated info
17:16:28 <glx> like if I remove a reference to a function and save it will still see it
17:16:39 <glx> and of course it will readd it
17:16:48 <TrueBrain> I disabled any and all backup files :)
17:19:15 <TrueBrain> if (len <= (_dma.count + 1)) {
17:19:47 <TrueBrain> _dma.offset = high ? _dma.offset | (value << 8) : value; <- for clearity, please put () around middle value
17:20:11 <TrueBrain> case 0x003: *dest = emu_io_read_003(); return; /* TODO -- DMA * <- not so TODO anymore
17:20:35 <TrueBrain> line 58:61 have useless space to align :p
17:20:52 <TrueBrain> rest looks absolutely fine, and I say: commit ;)
17:21:52 <glx> maybe buffer size will need tweaking later
17:35:01 <DorpsGek> SVN: glx (r989) [LibEMU] -Add: SoundBlaster support
17:39:56 <DorpsGek> SVN: glx (r990) -Add: completed C-ification of emu_Unknown_B483_0156()
18:54:17 <DorpsGek> SVN: glx (r991) [JIT] -Add: mapped more voices related functions
18:55:18 <DorpsGek> SVN: glx (r992) -Update (r991): update decompiled code
18:57:43 <TrueBrain> vvooooiiicccceeeesssss :)
18:59:48 <glx> but in the VM I sometimes get segfault or X error after "only one house will prevail"
20:37:45 <glx> oh BTW I kept the unneeded 01F7 change (the stuff after the return) ;)
21:24:24 *** Yexo is now known as Guest1076
22:10:37 <DorpsGek> SVN: glx (r993) [LibEMU] -Add: SoundBlaster DSP command 0x80
22:12:27 <DorpsGek> SVN: glx (r994) [JIT] -Add: mapped more voices related functions
22:12:57 <DorpsGek> SVN: glx (r995) -Update (r994): update decompiled code
22:44:11 <glx> voices support is not optimal but the game is playable
22:44:57 * TrueBrain is a happy bunny :)
continue to next day ⏵