IRC logs for #opendune on OFTC at 2011-06-02
⏴ go to previous day
00:30:48 *** Osai^2 has joined #openDune
02:25:52 <glx> ok it's not a timing issue, it's a luck issue :)
02:26:49 <glx> it always do g_houseInfo[0xFFFE].prefixChar, but sometime that segfault :)
03:02:30 <DorpsGek> SVN: glx (r1732) -Fix: replaced Drivers_CallFunction() when possible. Removed some global variables. Fixed an out of bound array access.
07:09:12 *** Alberth has joined #openDune
09:58:12 *** fjb is now known as Guest3141
10:34:15 <Alberth> lol, swapping source and destination is not working :p
10:42:27 *** HannesHH has joined #openDune
10:42:37 <HannesHH> i am on archlinux. using nvidia and ossv4
10:44:03 <HannesHH> oh, ouch. "Copy the original Dune2 1.07eu data files (including dune2.exe) to data/."
10:47:48 <HannesHH> hm, "ERROR: unable to open file 'WESTWOOD.PAL'." no matter why i try
10:49:18 <Alberth> % grep -l WESTWOOD data/*
10:49:58 <Alberth> my guess it is in the intro.pak file
10:50:29 <Alberth> are you sure you have the exactly right version?
10:50:30 <HannesHH> silly me, i forgot to lowercase the old dos filenames
10:51:11 <HannesHH> intro showed, i got to the "team" selection screen, chose atreides, saw the lady and then i got noise from my speaker and:
10:51:13 <HannesHH> [PCM] [022C] Not Yet Implemented D1
10:51:13 <HannesHH> Program Termination: jumped to 3FF1:20EC, which is not decompiled.
10:51:13 <HannesHH> The jump was triggered at decompiled/cs__B4DA.c:1914
10:51:13 <HannesHH> The jump appears to originate from B4DA:0C5C.
10:51:13 <HannesHH> Please send the file 'memory/crash.bin' to upstream developers.
10:51:15 <HannesHH> Creating crash-dump ...
10:51:17 <HannesHH> Crash-dump in 'memory/crash.bin'
10:52:31 <HannesHH> known_bugs.txt and changelog.txt where not in the archive i downloaded
10:53:08 <Alberth> they don't exist afaik
10:53:27 <HannesHH> the readme.txt references them ;)
10:54:05 <Alberth> I don't know how to help you with the crash dump
10:54:26 <Alberth> the file doesn't exist any more in current trunk
10:55:02 <Alberth> the only thing I can think of is that your binary may not be the right version
10:55:22 <HannesHH> are there checksums available somewhere?
10:56:05 <Alberth> md5sum data/dune2.exe
10:56:06 <Alberth> 8b650651e6c6d5bfd8199da68fccddfa data/dune2.exe
10:56:21 <HannesHH> ah, i got 9e3dfe97aa04341d599b79f2848903cb dune2.exe
10:59:00 <HannesHH> damn, none of my copies has that sum
11:05:23 <HannesHH> sound is garbled though (probably wrong rate)
11:06:35 <Alberth> it runs out of sync here, but is understandable
11:23:11 <DorpsGek> SVN: alberth (r1733) -Add: C-ified CopyToScreen() and XorToScreen(). Bye 2AE1.
11:25:37 <planetmaker> ho ho. :-) A new commiter here ;-)
11:26:15 <Alberth> for some reason devs tends to give such access to me after some time :)
11:29:36 <planetmaker> pure lazyness ;-)
11:29:37 <TrueBrain> nice work Alberth :D
11:29:43 <TrueBrain> and yes, what planetmaker says :D
11:29:52 <planetmaker> why should they commit your patches when you can do yourself just as well ;-)
11:30:05 <planetmaker> I just wondered that it took that long here, quite honestly ;-)
11:30:25 <Alberth> do you know how long it took with openttd? :)
11:30:45 <planetmaker> approx. yes. But then people didn't know you. All these people knew you already
11:31:10 <planetmaker> which kinda makes a big difference. Or would to me
11:33:58 <planetmaker> as there are two parts - as I see it - in granting commit acces: quality work and trust
11:34:36 <TrueBrain> in this case it was a third reason ;)
11:34:47 <Alberth> I don't mind much about commit access; I care more for writing nice patches.
11:35:38 <Alberth> the fact that my patches survive a review without change indicates the quality already
11:35:55 <TrueBrain> it took me (and glx) weeks before we knew how to C-ify this shit; so keeping taps on someone till then is the best thing to do. Now we are done, and I no longer care :D
11:37:13 <planetmaker> Alberth: yes. exactly. People would know 'yes' as an answer quite from the start here ;-)
11:37:57 <TrueBrain> I remember ludde in the beginning .. wanted to keep so close watch on what changed in OpenTTD, he only wanted to commit
11:38:00 <TrueBrain> that was SO SLOW GOING
11:38:24 <planetmaker> that's the trust part
11:38:50 <planetmaker> one has to be reasonably sure that one shares about the same vision and that major changes or additions are discussed and agreed
11:39:16 <planetmaker> well, there it was a lack of trust
11:39:37 <TrueBrain> it also has to do that you want to know what happens
11:39:53 <TrueBrain> not in relation to trust, but because you want to know :)
11:41:03 <Alberth> I want to know what you do?
11:41:46 <planetmaker> well, sure. But that's about communication, isn't it?
11:42:08 <planetmaker> Like offering patches for review before commit
11:43:51 <Alberth> TrueBrain: so, any decompiled/* interesting enough to convert any more? iirc you said no
11:44:33 <TrueBrain> Input_ and Mouse_ if you like
11:44:45 <TrueBrain> or at least we need to figure out all functions and parameters that are called from src/
11:44:51 <TrueBrain> so we have a chance of converting that correctly :)
11:44:58 <TrueBrain> was hoping to avoid it, but that is unlikely
11:45:16 <Alberth> we could try a bunch of monkeys :)
13:19:13 <Alberth> oh, g_global has a different segment
13:20:05 <Alberth> I was wondering about "emu_ax = 0x353F;" versus "29e8:0d8b 2d89b: b8 48 33 mov ax,0x3348"
13:20:52 <TrueBrain> its still 353F I hope :D but he means assembly vs memory I guess
13:21:04 <TrueBrain> Alberth: dune2 is relocated in the memory :)
13:21:37 <glx> don't forget to add 0x1F7 to cs in the asm ;)
13:24:12 <Alberth> apparently, I shouldn't :)
13:25:34 <glx> I don't know how the relocation knows what to change and what to not, but it knows
13:25:54 <TrueBrain> relocation table :)
13:26:11 <TrueBrain> still used nowedays
13:29:56 <glx> anyway the segfault was not directly caused by 1724 ;)
13:30:29 <glx> 1724 just made happen more often because voice driver was always enabled :)
13:39:12 <glx> but it was somehow random
13:40:38 <glx> grr we have projects/generate for a good reason Alberth ;)
13:41:30 <DorpsGek> SVN: glx (r1734) -Fix (r1733): MSVC project file again
13:45:54 <Alberth> it's not even executable :p
17:20:31 <Alberth> hmm, my strategy for the decompiled input code fails, I need a new one
17:22:39 <TrueBrain> everything not decompiled is useless? :D
17:22:49 <TrueBrain> not a good strategy maybe, but it might work :D
17:23:40 <Alberth> rather converting all as a queue of patches is too much work
19:38:02 *** HannesHH has left #openDune
20:35:07 <DorpsGek> SVN: truebrain (r1735) -Add: removed two now unused emu_ functions
21:32:01 <TrueBrain> weird pointer returns :(
21:32:42 <TrueBrain> memorySize is kinda invalid :D
21:38:01 <DorpsGek> SVN: truebrain (r1736) -Add: first C-ified a function, discovered it was no longer used, removed it, cascade remove. See ya 23E1!
21:47:10 <TrueBrain> okay ... reduced the 01F7 calls a bit more now ... drivers.c has one, Interrupt_Vector_[SG]et, and Malloc/Free
21:48:15 <TrueBrain> the rest is in 263B and 2649 .. almost there :D
21:53:59 <DorpsGek> SVN: truebrain (r1737) -Fix (r1736): for some reason not everything was committed
21:59:03 <DorpsGek> SVN: truebrain (r1738) -Fix: in no version of usage, buffer is NULL for Driver_Voice_LoadFile. This might be reintroduced later, but this simplifies highmem calls for now
22:00:40 <glx> hey I'm doing some cleanup there too
22:01:09 <TrueBrain> there is now only 1 alloc for highmem
22:01:12 <TrueBrain> going to replace it with a malloc
22:01:42 <glx> highmem stuff is used by voices
22:01:47 <TrueBrain> yup; and only there
22:02:22 <TrueBrain> how badly are you touching this stuff?
22:02:40 <TrueBrain> I want to move 1 function from emu_B483.c to sound.c, and replace 3E54 by a static pointer array
22:03:00 <glx> hmm let me test my stuff and commit it first :)
22:03:29 <TrueBrain> from what I can see, it can be fully replaced by real memory
22:03:34 <TrueBrain> and only here and there is a Tools_Memmove
22:04:12 <glx> there are 3 Tools_Memmove
22:04:20 <TrueBrain> all 3 will be gone after I did this :)
22:04:24 <TrueBrain> but I will wait for your commit then :)
22:04:29 <TrueBrain> after that, I can remove all highmem stuff :D
22:06:09 <glx> checking compilation with gcc, and I'll can commit
22:06:40 <DorpsGek> SVN: glx (r1739) -Fix: removed useless functions
22:06:59 <glx> (like always return NULL and things like that ;) )
22:08:36 <TrueBrain> do you know how to name the B483 functions?
22:08:43 <TrueBrain> else I will now rename them to Sound_Unknown<ip>
22:10:03 <glx> Unknown_B483_0156() is a kind of Voice_Play(), but we already have one
22:10:35 <TrueBrain> going to name it Sound_Unknown<IP> then :D
22:12:53 <DorpsGek> SVN: truebrain (r1740) -Add: B483 functions are related to sound, so renamed to Sound_Unknown<IP> and moved
22:13:14 <TrueBrain> right, then I can remove highmem :D
22:13:16 <TrueBrain> first, some food :D
22:16:18 <glx> hmm Unknown_B4B8_110D() defines the recolor map or something like that
22:24:28 <DorpsGek> SVN: glx (r1741) -Fix: renamed Unknown_B4B8_110D() to GUI_Palette_CreateRemap() and moved it
22:48:59 <glx> btw I don't understand why they don't preload ?XXX files
22:49:42 <glx> in the list of .voc files
22:50:08 <TrueBrain> the code can also preload PAK files in highmem; it doesn't do this either
22:50:40 <glx> all .voc are preloaded in Voice_LoadVoices() except ? ones
22:51:18 <TrueBrain> owh, shit, they have a function that returns the size of a voice file ...
22:51:26 <TrueBrain> ah; maybe less frequent onces?
22:51:57 <glx> maybe, but that's only 8 files out of 131
23:02:30 <TrueBrain> lol ... one weird one ... getting the size of highmem, then allocate it in lower mem ...
23:13:34 <TrueBrain> can you test this patch please?
23:13:37 <TrueBrain> see if voices still work etc
23:18:39 <DorpsGek> SVN: glx (r1742) -Fix: replace a csip32 arg with uint8*
23:22:27 <DorpsGek> SVN: truebrain (r1743) -Add: remove highmem, replace it with malloc/free
23:27:19 <DorpsGek> SVN: truebrain (r1744) -Add: remove highmem and xms driver code
23:27:20 <TrueBrain> test again please :D
23:31:44 <TrueBrain> that was a huge step :D
23:31:51 <TrueBrain> removed a complete driver :D :P
23:34:23 <glx> yeah I can replace g_global->variable_998A
23:35:20 <TrueBrain> so much stuff we can replace :)
23:35:31 <TrueBrain> how is voice conversion going?
23:36:51 <TrueBrain> well, you named a lot, which is very nice :D
23:37:29 <TrueBrain> src/drives.c 01F7 call, and Tools_Malloc and Tools_Free
23:48:30 <TrueBrain> 01F7_27FD does getenv I think
23:50:19 <DorpsGek> SVN: truebrain (r1745) -Fix: don't call a function that returns fake data anyway; just inject the fake data
23:51:06 <DorpsGek> SVN: truebrain (r1746) -Fix (r1745): forgot to remove an extern
23:51:17 <TrueBrain> so ... that is one more less :)
23:51:23 <TrueBrain> forgot how we insterted the BLASTER information
23:51:25 <TrueBrain> was something tricky
23:51:34 <TrueBrain> I believe I embedded it in the memory
23:53:37 <TrueBrain> what string is at 353F:7B54?
23:56:29 <DorpsGek> SVN: truebrain (r1747) -Add: named a function
continue to next day ⏵