IRC logs for #opendune on OFTC at 2009-10-14
⏴ go to previous day
02:13:32 <DorpsGek> SVN: glx (r282) -Add: figured out a few variables
02:15:25 <glx> dune.cfg contains "encrypted" data with a checksum
04:39:41 <proyvind> (if anyone would feel like doing a windows build, I'd be very greatful! :)
04:40:39 <proyvind> oh, I see you already talked about it earlier
04:41:36 <proyvind> if you ever find something useful in the code you'd like to make use of, but experience the license being too restrictive, lemme know and I can prolly' grant you an exception ;)
04:49:12 <proyvind> or dual-license it, or whatever.. :o)
05:13:47 <proyvind> other funny details like that is the looping of animation frames etc..
05:18:42 <proyvind> or how midi instruments are used for some sound effects, ie. like when the harkonnen troopers are firing their guns in the intro :o)
07:20:10 <Xaroth> proyvind: yeh, you did nice work on it :)
07:28:50 <proyvind> Xaroth: you wouldn't happen to be on windows and have the chance to try building it for me? :o)
07:34:13 <Xaroth> yes, and no unfortunately :(
07:38:15 <proyvind> you're excused then :p
07:47:42 <DorpsGek> SVN: truebrain (r283) -Add: figured out a few more function names (glx)
07:47:53 <Xaroth> the beast has awaken! :P
08:06:43 <TrueBrain> proyvind: linux and OSX only here ;) Am I excused too? :) :)
09:15:56 <Xaroth> esx can be such a twat from time to time
09:55:43 <TrueBrain> oaky ... now I have a laptop, I need to find a simple way to synchronize data back and forth
10:50:10 <TrueBrain> I dunno what to think of Nyerguds .. he always knows better .. he always has a comment that tops the last post .. dunno
11:00:27 <blathijs> TrueBrain: unison seems to be a fairly ok syncing tool
11:01:04 <blathijs> TrueBrain: Haven't tried it in practice yet, but it is rsync-like, but with two-way support (it keeps some metadata about the last sync)
11:01:08 <blathijs> It is commandline, though
11:01:24 <TrueBrain> I love commandline :)
11:39:49 <Xaroth> TrueBrain: I think he's too used to his position as a person with answers
11:44:11 <TrueBrain> good morning to you glx :)
11:54:50 <TrueBrain> so, what part of Dune2 will we uncover today?
11:57:29 <TrueBrain> oh, btw, glx, I had to move your Read_Wrapper
11:57:38 <TrueBrain> it turns out it was only called from one function, which did only a goto to that place
11:57:48 <TrueBrain> so the decompiled code was weird ;)
11:58:03 <glx> yes goto end, goto begin :)
11:58:25 <TrueBrain> that is now the result ;) Which is slightly better than a function which only calls one other function ;)
11:58:53 <glx> Program Termination: jumped to 24E6:0172, which is not decompiled.
11:58:53 <glx> The jump was triggered at decompiled/cs__24E6.c:133
11:58:53 <glx> The jump appears to originate from 24E6:0169.
11:59:02 <TrueBrain> unmodified version?
11:59:12 <glx> for invalid checksum in dune.cfg
12:00:42 <glx> maybe you also need the modified dune.cfg
12:01:04 <TrueBrain> that alone should be enough
12:01:19 <TrueBrain> when I start the crashlog, it ends up in nowhere land .. most likely because of invalid file descriptor ;)
12:02:27 <glx> hmm shouldn't matter as the file is already closed
12:03:26 <DorpsGek> SVN: truebrain (r284) [JIT] -Add: mapped another 2 functions (dune.cfg crc failure)
12:03:54 <glx> ha 2 functions, I guessed only one
12:04:12 <TrueBrain> one is a goto over the other :p
12:04:16 <TrueBrain> and the only code in between is an xor :p
12:04:29 <glx> at least the missing part in emu_Config_Decode
12:04:37 <DorpsGek> SVN: truebrain (r285) -Update (r284): update decompiled code
12:06:04 <glx> I expected that indeed :)
12:06:56 <TrueBrain> I need to find a way to name overlay functions
12:07:11 <TrueBrain> ovl__34B8(2); <- thisone handles the whole intro + menu (to which you can never return btw)
12:07:37 <TrueBrain> and it links to case 2: f__B4B8_16D3_0022_3C6A(); return;
12:07:43 <TrueBrain> I can name the latter .. but not the first ..
12:07:56 <TrueBrain> as it needs to load/unload overlays correctly ..
12:10:37 <glx> now I'm trying to find why it doesn't print an error for invalid dune.cfg :)
12:11:03 <glx> because this is the "setup must be run first" string
12:11:43 <glx> btw I like msvc for debugging :)
12:13:31 <glx> I can follow emu_memory + ((emu_ss) << 4) + (((emu_sp) + (0)) & 0xFFFF) in real time
12:13:40 <glx> just with a memory window
12:15:10 <TrueBrain> well, I didn't mean the naming as such, I meant that I can't name it ...
12:15:33 <TrueBrain> does DOSBox do print the error?
12:15:44 <TrueBrain> my emulator layer does not support all types of printing btw :p
12:16:23 <TrueBrain> hmm .. maybe use enums for overlay numbers?
12:16:26 <TrueBrain> dunno ... tricky ..
12:17:00 <glx> in dosbox it displays the string
12:17:18 <TrueBrain> so it is a bug in the emu layer most likely :)
12:17:41 * glx follows step by step in debugger :)
12:17:58 <TrueBrain> in non-int-debug it never displays
12:19:08 <glx> that can be the cause yes :)
12:20:09 <TrueBrain> pff .. main-loop functions are HUGE
12:21:02 <glx> hmm something like line 722 should work
12:21:30 <TrueBrain> yup, but there were a few problems with it .. can't remember :)
12:21:36 <TrueBrain> oh, yes, in non-GUI mode that fails ;)
12:22:23 <TrueBrain> hmm .. or maybe I should fix overlays different .. like: ovl__34B8(load); f__B4B8_16D3_0022_3C6A()
12:23:31 <glx> I don't fully understand overlays yet :)
12:24:27 <TrueBrain> just a thin wrapper in between
12:24:39 <TrueBrain> but as long as overlays use data from their 'cs' (static data), it is a pain :)
12:25:45 <glx> ,...if (emu_cs == 0x3480) { ovl__3480(0xFF); } <-- like this ;)
12:25:49 <TrueBrain> yeah, I will work on that next week or something, to make that a bit easier ..
12:26:11 <TrueBrain> yeah .. problem of leaving the current cs, is that it can happen it unloads that cs :) So on return, we need to make sure it is reloaded when needed
12:27:03 <TrueBrain> hmm .. the check is not at all places :)
12:29:21 <TrueBrain> so many functions .... printfs, where are you? :)
12:40:20 <DorpsGek> SVN: truebrain (r286) -Add: identified the functions which take care of the intro and the intro+menu
12:48:24 <TrueBrain> k, gone for the day :) Have fun all ;)
13:54:44 <glx> oh of course it doesn't print anything, it uses int21:40 with file 1 which is not allowed by libemu :)
14:11:36 <Xaroth> now I know why I hate msexchange.
17:52:59 <DorpsGek> SVN: glx (r287) -Add: figured out a few variables
20:00:34 <Xaroth> hah, monthy python on telly :P
21:17:53 <TrueBrain> - if (emu_bx < 5 || emu_bx >= 20) {
21:17:54 <TrueBrain> + if ((emu_bx < 5 && _int21_filemap[emu_bx] == NULL) || emu_bx >= 20) {
21:18:08 <TrueBrain> glx: how about: if (emu_bx >= 20 || _int21_filemap[emu_bx] == NULL)?
21:19:45 <glx> it's a better check indeed
21:24:48 <TrueBrain> so change it and commit it, I say ;)
21:24:53 <TrueBrain> (you have also permissions in that dir)
21:32:32 <DorpsGek> SVN: glx (r288) [LibEMU] -Fix (r73): protection against invalid file handle access was to strict
21:33:01 <TrueBrain> too strict .. well .. before your commit, it was even non-existing for < 5 ;)
21:33:24 <glx> it just considered <5 as invalid
21:33:32 <TrueBrain> it never assigned a value to it too ;)
21:36:02 <glx> I discovered 2 nice math functions ;)
21:36:41 <TrueBrain> I will process both your other requests tomorrow :)
21:36:51 <TrueBrain> if I can find the time inb etween ... I have a big project ahead :(
22:03:34 <proyvind> anyone willing to do a win32 build of libeastwood now perhaps? :o)
22:03:49 <TrueBrain> if you give glx a nice cookie he might want to give it a shot? :)
22:04:38 <proyvind> poly wants a cracker?
22:04:49 <proyvind> glx: poly wants a cracker?
22:18:33 <proyvind> no cracker for polly then!
22:38:48 <DorpsGek> SVN: truebrain (r289) -Add: resolved 4 more functionnames (glx)
22:39:54 <TrueBrain> glx: your cfg gives me NO new functions
22:40:17 <glx> it should give more data for a function, not a new function
22:40:33 <TrueBrain> it also doesn't write anything to the output btw
22:41:12 <glx> Program Termination: jumped to 24E6:0092, which is not decompiled.
22:41:13 <glx> The jump was triggered at decompiled/cs__24E6.c:65
22:41:13 <glx> The jump appears to originate from 24E6:00B9.
22:41:25 <glx> that's what I get with the cfg
22:41:42 <TrueBrain> with that dune.cfg, I have a clean run (it abort, but okay, that was to be expected :p
22:42:45 <TrueBrain> owh .. hmm .. it seems it loaded libemu.diff as dune.cfg :p
22:44:03 <DorpsGek> SVN: truebrain (r290) [JIT] -Add: mapped another function (dune.cfg crc failure)
22:44:34 <DorpsGek> SVN: truebrain (r291) -Update (r290): update decompiled code
22:44:47 <TrueBrain> weird tomorrow, I know :)
22:45:16 <glx> it was not a crc failure in this case ;)
22:45:40 <TrueBrain> I just copied/pasted it :p
22:45:48 <TrueBrain> either way, now I am going to bed :)
22:46:01 <glx> anyway the result is config is broken :)
22:57:29 <glx> proyvind: what do you want exactly ?
23:02:56 <proyvind> glx: I'd like someone to try do a win32 build of my library and it's python bindings
23:03:40 <proyvind> I *think* the source should mostly build fine, there might be some minor issues with cmake though..
23:03:46 <proyvind> you might as well go with trunk :)
23:04:40 <glx> hmm cmake never worked very well for me
23:05:04 <glx> but last project I tried to build using it was rigsofrods
23:11:36 <proyvind> (I've never cared about windows portability before or have any experience with development on windows whatsoever, so let's just hope it's not too awful.. :p)
23:11:37 <glx> http says relocated to https and https fails
23:11:45 <proyvind> bzr branch lp:libeastwood
23:14:34 <proyvind> 'bzr branch lp:libeastwood' won't work?
23:14:46 <Xaroth> proyvind: he's in windows? :P
23:14:58 <proyvind> bzr is available for windows as welll :)
23:15:12 <proyvind> but I can generate a tarball
23:15:24 <proyvind> (cpack blows btw. :()
23:19:53 <glx> of course it's not in path
23:26:20 <glx> looks like I never reinstalled cmake ;)
23:47:45 <glx> of course I don't have sdl_mixer
23:50:27 <proyvind> you can build without SDL
23:51:35 <proyvind> oh, I haven't fixed so the check doesn't run when SDL build has been disabled?
23:51:55 <proyvind> hm, I seem to have ensured that..
23:52:10 <proyvind> set LIBEASTWOOD_BUILD_SDL to 0
23:52:22 <proyvind> cmake -DLIBEASTWOOD_BUILD_SDL=0
23:52:23 <glx> too late I compiled and installed sdl_music :)
23:52:35 <proyvind> sdl_music != sdl_mixer..?
23:53:40 <glx> nice it found sdl_mixer.dll but not sdl include
continue to next day ⏵