IRC logs for #opendune on OFTC at 2011-04-09
⏴ go to previous day
07:25:47 *** Alberth has joined #openDune
12:02:21 <TrueBrain> ugh, wish I didn't had to battle matlab :(
12:08:43 <Alberth> I programmed java/eclipse/EMF a whole week, only to discover friday evening while walking home that the problem is much simpler than I thought
12:10:53 <TrueBrain> that always sucks ...
12:11:33 <TrueBrain> I now have the issue that all my algorithms find a worse solution then shourtest tour, in a TSP problem :p
12:13:27 <TrueBrain> then agian, I am running this on a machine ... they won't like it if they find out I do :)
12:13:30 <TrueBrain> but I have nothing else atm :p
12:13:45 <Alberth> hmm, is matlab of any use for programming such a problem?
12:14:04 <TrueBrain> to program it rapidly and see graphical feedback, yes
12:14:11 <TrueBrain> is for a class about natural algorithms
12:14:35 <Alberth> for a sysop you seem to lack useful machines, lately :)
12:16:03 <Alberth> isn't it a property of natural algorithms that they don't always find the really shortest path?
12:16:28 <TrueBrain> but they should give better results then a simple nearest neighbour algorithm :D
12:16:35 <TrueBrain> owh, sorry, I mistyped :p
12:16:40 <TrueBrain> shourtest = nearest neighbour
12:26:22 <Alberth> no hurries, I am currently enjoying f__B4DA_0000_002C_B3C2() (some security function)
12:28:13 <glx> tools_free_wrapper_remove seems wrong,
12:28:44 <glx> mainly wrong emu_cs when calling emu_Tools_Free
12:30:32 <Alberth> it should just be a rename of the function, so we can leave it alone until the end?
12:30:40 <Alberth> or perhaps not even touch it at all
12:32:54 <glx> they are in different segment, so a rename may not be enough
12:34:49 <Alberth> I was thinking something like emu_Tools_Free_Wrapper_B520() as new name, so it is clear it is just a wrapper
12:35:04 <Alberth> but maybe that is TMWFTLB
12:35:14 <glx> ha yes you can name it emu_Tools_Free_Wrapper()
12:36:05 <glx> if you name it add it to function_names.txt
12:39:02 <TrueBrain> or also replace the things like: emu_push(emu_cs); emu_push(0x0231); emu_cs = 0x23E1; emu_Tools_Free();
12:39:11 <TrueBrain> emu_push(emu_cs); emu_push(0x0BD4); emu_cs = 0x3520; overlay(0x3520, 0); f__B520_0223_000E_38A3();
12:39:22 <TrueBrain> emu_push(emu_cs); emu_push(0x0BD4); emu_cs = 0x23E1; emu_Tools_Free();
12:39:37 <TrueBrain> but yeah ... maybe just naming is enough :)
13:16:54 <TrueBrain> seems one algorithm now does what it should do ...
14:11:11 <glx> +B520:0223:000E:38A3 emu_Tools_Free_Wrapper # (csip) Forwards the call to emu_Tools_Free() <-- I'd say "stupid wrapper" ;)
14:13:51 <Alberth> that's a quality judgement :p
14:18:28 <Alberth> emu_get_memory8(0x2C6C, emu_di * 8, 0x0); emu_get_memory16(0x2DCE, i * 4, 0x47A); <-- any knowledge what these base addresses are?
14:19:17 <Alberth> 0x2C6C is an array of an array of 8 bytes
14:23:18 <glx> g_sprites = (csip32 *)&emu_get_memory8(0x2DCE, 0x0, 0x440);
14:23:59 <DorpsGek> glx: An error has occurred and has been logged. Please contact this bot's administrator for more information.
15:21:46 <Alberth> same patch, second one ignores white space changes so it may be easier to follow
15:21:46 <Alberth> 2 questions: emu_al = g_houseInfo[unknown].index; /* XXX index is a 16bit value! */ <-- looks wrong due to size, but I see no other translation
15:21:46 <Alberth> there is a subroutine call using data from a fixed address. Can I remove everything but the one call?
15:21:46 <Alberth> Despite the still large number of emu_* calls, I don't really see many more possibilities otherwise, except the finishing touches: a rename, a move to a different file, and a short test.
15:21:47 <Alberth> but first time for dinner, and some other stuff I need to do.
16:20:53 <glx> + emu_ip = emu_get_memory16(0x353F, 0x00, 0x6630); /* XXX 0x04a5 */
16:20:53 <glx> - emu_cs = emu_get_memory16(emu_es, 0x00, 0x6632);
16:20:53 <glx> + emu_cs = emu_get_memory16(0x353F, 0x00, 0x6632); /* XXX 0x20af */
16:20:53 <glx> it's g_global->variable_6630 and it's always 0x22A604A5
17:52:45 <Alberth> fixed the houseinfo puzzle, I had the wrong structure. removed the switch.
19:12:56 <Alberth> hmm, voices block the whole game :(
19:13:46 <Alberth> 'sound effects' may be a better word
19:23:07 <glx> voices use a lot of cpu yes
19:23:20 <glx> it's because emulator layer
19:24:28 <Alberth> hmm, my harvester is '100% full and awaiting pickup', but it is A2, ie no carry-alls
19:47:45 <glx> ha do 22A6:0FAE and 22A6:0FD7 are mouse callbacks
19:49:00 <glx> they move write only stuff
19:59:59 <Alberth> or mouse speed perhaps
20:00:37 <glx> one copies 22A6:0026-22A6:002C to 22A6:0031-22A6:0037
20:01:08 <glx> the other does the reverse
20:01:41 <glx> and set 31-37 to 0,40,0,200
20:02:37 <Alberth> could be a mouse box, except for the second 0
20:03:14 <glx> anyway these variables are never read outside these 2 functions
20:04:49 <Alberth> probably because we all use the same mouse interface
20:11:32 <glx> hmm maybe you wrote that somewhere, what is base for B536 ?
20:41:54 <Alberth> ha ha: Ram Free Error! Press any key to exit to DOS. :)
20:42:29 <glx> but usually means error in conversion somewhere :)
20:42:44 <Alberth> just after the 'good job done' talk
20:44:20 <Alberth> that would be emu_Unknown_259E_00B1.patch
20:46:49 <Alberth> or the tool-rename patch, but that is less likely :)
23:54:46 <DorpsGek> SVN: glx (r1469) -Add: C-ified Driver_Voice_0248()
continue to next day ⏵