IRC logs for #opendune on OFTC at 2011-03-29
            
00:48:44 *** glevans2 has joined #openDune
02:07:53 *** glx has quit IRC
13:02:02 *** glx has joined #openDune
13:02:02 *** ChanServ sets mode: +o glx
13:02:17 <glx> hello
13:05:01 <Xaroth|Work> o/
17:40:00 *** Alberth has joined #openDune
18:47:41 <Alberth> of course, the first field of csip32 is 'ip' :)
18:48:23 <TrueBrain> confusing as hell :D
18:58:43 <glx> that's why we have s.cs and s.ip ;)
19:03:26 <Alberth> yeah, but emu_get_memory16(emu_ss, emu_bp, 0x6) doesn't say what it is ;)
19:06:39 <TrueBrain> context :)
19:06:40 <glx> cs, ip, offset
19:07:01 <glx> ha yes :)
19:07:28 <TrueBrain> emu_[ad]x = emu_get_memory16();
19:07:34 <TrueBrain> emu_[da]x = emu_get_memory16()
19:07:39 <TrueBrain> either 32bit int, or csip :)
19:07:51 <glx> emu_ax | emu_dx == 0
19:08:37 <glx> usually when you see that it's or a csip32, or an uint32
19:09:02 <TrueBrain> I know we have a few items marked as csip32 which are uitn32, and visa versa :D
19:09:24 <glx> true, but that's fixed when we understand the code
19:09:38 <TrueBrain> yup
19:09:41 <TrueBrain> and indeed: when, not if :D
19:10:08 <Alberth> :)
19:11:24 <glx> when something is accessed with emu_lfp it's quite clear too :)
19:11:34 <Alberth> http://www.fpaste.org/eUf7/ is what I have currently
19:11:45 <Alberth> glx: that's a *p access, right?
19:11:54 <glx> yup
19:11:59 <glx> in 16bit world
19:12:07 <Alberth> how do I figure out what lines 29/30 is?
19:12:31 <glx> you ask me :)
19:12:59 <TrueBrain> (and seriously, you ask him :p)
19:13:08 <TrueBrain> not like: *shrug* you ask me, but like: sure, you ask me
19:13:09 <TrueBrain> :D
19:13:10 <Alberth> oh oracle glx, could thy please tell me the translation of those lines
19:13:31 <Alberth> ?
19:14:03 <Alberth> TrueBrain: yeah, that's how I interpreted it :)
19:15:35 <glx> http://devs.opendune.org/~glx/1DD7_010B.txt
19:15:41 <TrueBrain> he is a magic worker ;)
19:16:38 <TrueBrain> I assume you know how to read real assembly?
19:16:43 <TrueBrain> it is slightly different :p
19:17:12 <glx> hmm ipv6 access is gone again
19:17:21 <TrueBrain> gone where?
19:17:33 <glx> winscp fails
19:17:46 <Alberth> it is equally non-readable :)
19:18:04 <TrueBrain> glx: fixed
19:18:06 <TrueBrain> somehow it failed
19:18:10 <TrueBrain> IPv6 comes online a bit later
19:19:21 <glx> http://devs.opendune.org/~glx/decompile.sh.txt <-- that's what I use ;)
19:19:39 <glx> of course for overlays you first need to find the real CS
19:20:32 <glx> for that I open a txt from JIT and I search the bytes in hex editor
19:21:05 <glx> then I run the script with some guessing until I find the right one :)
19:21:26 <glx> and I write it on a paper for a next time
19:22:00 <glx> like B53B = 57BA
19:23:44 <Alberth> oh, objdump understand windows binaries, apparently
19:24:03 <Alberth> (mostly, probably :) )
19:24:30 <TrueBrain> yup
19:24:39 <TrueBrain> x86 assembly is still very much working :p
19:24:46 <TrueBrain> 8086 assembly, that is
19:24:48 <TrueBrain> :p
19:26:00 <glx> it's just you can't directly disassemble dune2.exe to get all the code ;)
19:26:19 <glx> but when you know what you want exactly it's ok
19:27:03 <glx> if there are calls in the unresolved code it starts to be harder to understand
19:27:29 <Alberth> http://www.fpaste.org/e6Bv/ I pasted some fragments in the original C code, it seems like I need to know about the 0x2452:0xb64 call
19:27:40 <glx> except for calls in segment "0000" which is 01F7 in our sources
19:29:20 <glx> hmm you have 0x2346 before this one
19:29:48 <glx> oh 0x2346 is the ReadWholeFIle
19:29:48 <Alberth> that's line 4
19:30:39 <Alberth> hmm, I can use your script of course, but can you relate that back to our code?
19:31:09 <glx> 2649:B64
19:31:35 <glx> always try to add 0x1F7 to CS first :)
19:32:30 <Alberth> :)
19:37:15 <glx> (in overlays it often doesn't work though)
19:46:48 <Alberth> the 'awk' part throws away the output in some way
19:49:22 <Alberth> DOH, missed the redirection :p
20:56:15 *** Alberth has left #openDune
23:52:10 *** glevans2 has quit IRC