IRC logs for #opendune on OFTC at 2011-08-05
⏴ go to previous day
00:01:09 <glx> there's nothing uninitialised there
00:01:53 <TrueBrain> well, [16] would be :p
00:02:26 <TrueBrain> I can access 0 .. 10 withtout problem
00:02:47 <TrueBrain> at 11 is starts to complain
00:02:50 <TrueBrain> that is just weird and wrong
00:03:40 <TrueBrain> there is one fair point in this all
00:03:49 <TrueBrain> this is not initialized
00:03:52 <TrueBrain> and is copied to the unit
00:05:14 <TrueBrain> but it should all be set .. so what goes wrong there ...
00:06:15 <glx> indeed it should have read from the save
00:07:18 <DorpsGek> SVN: truebrain (r2348) -Fix: if SaveLoad_Save() or SaveLoad_Load() fail in recursion, fail completely
00:08:23 <TrueBrain> doesn't solve the issue btw :p
00:08:27 <TrueBrain> but that was a bug :p
00:09:40 <TrueBrain> how can only 10 of the 15 elements be read ..
00:09:42 <TrueBrain> that makes no sense
00:09:50 <TrueBrain> and it is index 3, so that should be early on
00:10:28 <glx> maybe *u = ul doesn't work as we expect
00:11:04 <TrueBrain> nope, doesn't matter
00:11:09 <TrueBrain> so *u = ul does exactly what we expect
00:11:54 <TrueBrain> I am now reading [12] when loading
00:11:56 <TrueBrain> nothing weird happens
00:13:10 <TrueBrain> with some WEIRD value
00:13:49 <TrueBrain> so how can it be uninitialized
00:15:03 <TrueBrain> it feels like hunting ghosts
00:17:22 <TrueBrain> but when in Unit_Load I print stack[12]
00:17:27 <TrueBrain> then later on in Script_Run
00:17:48 <TrueBrain> well, I am not going to worry too much about it, as it is this broken unit anyway ...
00:18:01 <TrueBrain> we add the stack check, and I will play a few games with valgrind on, on a new game
00:18:47 <TrueBrain> I just hope something like this tackles the bug from pm
00:18:53 <TrueBrain> as I really would like to have a stable release now :p
00:19:00 <TrueBrain> bu tfor now, night!
00:19:05 <glx> at least with stack check we can be sure it's not our fault
00:47:48 <DorpsGek> SVN: glx (r2349) -Fix (r2290, #81): silenced warning broke walls
01:06:46 <glx> hmm no after some tests it's unsigned, but the compiler is silly :)
08:54:11 *** Alberth has joined #openDune
08:54:12 *** ChanServ sets mode: +o Alberth
09:50:07 <TrueBrain> Alberth: have you seen Scale3x in action? :D:D
09:53:20 <Alberth> not yet, I am fighting with be & fedora bug thingies
09:53:56 <Alberth> 'be init' fails on a hg repo. fix: drop hg support. this should fix your issue....
09:54:26 <Alberth> report for F15, get an update link for F16 :p
10:32:22 <Alberth> oh, 640x400 is smaller than I thought :p
10:32:36 <Alberth> 3x looks weird, but 2 is nice
10:34:34 <Alberth> even at 2x they are a bit wobbly
10:35:27 <Alberth> enhancement: also when a harvester is driving to the factory by itself, let it return to the spice field after unloading?
10:35:38 <TrueBrain> it does ... but nto always it seems
10:35:42 <TrueBrain> will have to look into it yes :)
10:35:44 <Alberth> mainly a problem with the lower levels
10:35:45 <TrueBrain> annoys the hell out of me :p
10:36:06 <Alberth> it probably looks around from the factory
10:37:05 <TrueBrain> but it knows its return position
10:37:06 <TrueBrain> so I don't really get it
10:38:05 <TrueBrain> might be a script error
10:38:11 <TrueBrain> we are having some issues with scripts anyway
10:41:00 <Alberth> oh, path finding is working :p
11:15:00 <Alberth> oops, harkonnen were not very happy with my visit :)
11:22:19 <TrueBrain> in a list, I have push and pop
11:22:30 <TrueBrain> what is the common name to replace the top?
11:24:28 <Alberth> there is no common name afaik
11:24:43 <TrueBrain> I can also just pop/push I guess
11:24:47 <TrueBrain> better for the general flow
11:25:31 <Alberth> at least better for general understanding :)
11:27:03 <Alberth> forth only has stack data manipulations
11:43:22 <TrueBrain> comments before I do all files?
11:52:05 <Alberth> so it even has stack-based computations. this is all new code to me :)
11:52:14 <Alberth> yeah, it looks like a good move to me
11:52:17 <TrueBrain> it has a scripting engine :p
11:52:59 <TrueBrain> Dune2 has so many highly advanced shit
11:52:59 <TrueBrain> that got lost for a few years before it got back as 'new'
11:52:59 <TrueBrain> for years it was said that a script per unit was too expensive
11:53:00 <TrueBrain> here we have a game, 20 years old, which has it ;)
11:53:11 <Alberth> I heard about that, but never looked at the code
11:53:36 <Alberth> it depends on what you do with the script
11:54:22 <TrueBrain> see if valgrind is happy now
11:54:57 <TrueBrain> almost all savegames have a damaged unit in them :p
11:55:03 <TrueBrain> I think it has to do with old corruption
11:55:18 <TrueBrain> but we will see during a normal game, if it happens
11:55:29 <Alberth> Corsix-TH uses lua for scripting, which works in its own, but their problem is that the interactions are very complex, many many corner cases
11:55:51 <TrueBrain> you have to stay primitive
11:56:14 <TrueBrain> don't do complex things in scripts
11:56:27 <TrueBrain> there are stuff like:
11:56:36 <TrueBrain> which is just a waste
11:57:09 <DorpsGek> SVN: truebrain (r2350) -Add: protect stack access by wrappers, which validate for out of bound.
11:57:12 <Alberth> one of them was a parameter probably
11:57:20 <TrueBrain> guess I now have to play a game to see if it happens in normal games ..
11:58:41 <TrueBrain> I really need to know if it happened because of older corruption
11:58:48 <TrueBrain> or if the scripts are flawed :)
12:00:56 <TrueBrain> I wonder if fog uncover should be radius
12:01:56 <TrueBrain> owh, btw, pathfinder ignores what you can see :p
12:09:28 <TrueBrain> starports are bugged
12:09:38 <TrueBrain> it just put down my stuff AT THE ENEMY BASE
12:09:56 <Xaroth|Work> instant-attack! :P
12:10:14 <TrueBrain> what the fuck went wrong there
12:20:49 <TrueBrain> trying again, and it does work correctly
12:21:06 <glx> heavy corruption somewhere ?
12:21:14 <TrueBrain> I started a new game .. :(
12:21:43 <TrueBrain> as in, I modified the game for starport units to cost 10 credits
12:22:01 <TrueBrain> so I am very sure it was clean memory etc
12:22:50 <glx> grr hang in _File_Open again
12:24:26 <TrueBrain> attach debugger, and please tell me what both threads are doing :)
12:24:26 <TrueBrain> btw, glx, the STACK_PUSH etc patch I committed just before you joined (lol) does resolve all valgrind warnings
12:24:26 <TrueBrain> including the weird on
12:25:00 <glx> there are more than 2 threads ;)
12:25:10 <TrueBrain> those you want to worry about
12:26:04 <glx> main is loading screen.cps
12:26:32 <glx> timer is in mpu_interrupt
12:27:00 <TrueBrain> mpu_send does a free or malloc?
12:27:23 <glx> there's another thread doing a free
12:27:40 <TrueBrain> which thread? And from where?
12:28:15 <TrueBrain> SDL audio thread, I wanna bet ..
12:40:30 <TrueBrain> found a valgrind error
12:40:39 <TrueBrain> invalid read of size 1
12:40:49 <TrueBrain> guess packed exceeded his max
12:42:16 <TrueBrain> I am tempted to make it an assert
12:42:24 <TrueBrain> so next time it gives a backtrace
12:42:32 <TrueBrain> I have the backtrace of course
12:43:26 <TrueBrain> hmm, yes, that is weird
12:43:38 <TrueBrain> GUI_Widget_Cancel_Click sets g_structureActivePosition to 0xFFFF
12:43:55 <TrueBrain> ChangeSelectionType on its turn can set Map_SetSelection to that value :p
12:44:34 <TrueBrain> build something, place it, cancel it
12:47:45 <TrueBrain> and I really don't want to bisect today :P
12:48:25 <TrueBrain> today I am going to try to compile this with tcc
12:49:17 <TrueBrain> sadly, tcc hasnt had an update for 2 years
12:49:19 <TrueBrain> that makes me sad :(
12:49:56 <TrueBrain> holy crap, he is fast
12:50:19 <TrueBrain> it doesn't know byte swap
12:53:52 <TrueBrain> holy crap, tcc works
12:54:14 <TrueBrain> well, not completely
12:54:31 <TrueBrain> 1.2 seconds to compile
12:55:54 <TrueBrain> static uint32 __builtin_bswap32(uint32 x)
12:55:58 <TrueBrain> r |= (x & 0xFF000000) >> 24;
12:56:00 <TrueBrain> r |= (x & 0x00FF0000) >> 8;
12:56:01 <TrueBrain> r |= (x & 0x0000FF00) << 8;
12:56:03 <TrueBrain> r |= (x & 0x000000FF) << 24;
12:56:07 <TrueBrain> am I doing it wrong?
12:56:41 <Xaroth|Work> even i get that :P
12:56:48 <Xaroth|Work> so it must be wrong..
13:01:31 <TrueBrain> scripts seem to fail
13:01:34 <TrueBrain> but intro is running nicely
13:04:01 <TrueBrain> hall of fame works, mentat works
13:04:01 <TrueBrain> only script fails, and it doesn't see savegames
13:07:20 <TrueBrain> ah, TCC doesn't like multichar shit
13:11:10 <TrueBrain> I did not expect thatone to be the cause
13:14:14 <TrueBrain> check that function ;)
13:14:21 <TrueBrain> you now get a multiplication over multiplication
13:15:30 <TrueBrain> yeah, but the functionality is gone too :D
13:15:55 <TrueBrain> I also don't see how the two are related tbh
13:16:23 <TrueBrain> and I wonder if the crash is related to the valgrind error
13:17:14 <glx> hmm svn revert, recompile, no crash
13:17:54 <TrueBrain> it doesn't always crash I guess
13:19:05 <TrueBrain> it just randomly overflows
13:19:05 <TrueBrain> with gdb attached, it never crashes here for example
13:19:05 <TrueBrain> valgrind always complains :)
13:19:06 <TrueBrain> I guess a != 0xFFFF is missing
13:22:05 <TrueBrain> hmm .. what to do about TCC I wonder ..
13:25:42 <TrueBrain> it indeed is code-wise not supported
13:25:42 <TrueBrain> '' is always 1 byte, or 2 in case of wchar
13:27:38 <TrueBrain> can we use "NAME", as string?
13:27:42 <TrueBrain> bit complex I guess
14:04:13 <TrueBrain> I expected an if (packed == 0xFFFF) return tbh :p
14:04:19 <TrueBrain> but I didn't check what the function does really
14:06:08 <TrueBrain> right .. I want a script that converts (blabla)'NAME'(blabla) to (blabla)0x43<whatever>(blabla)
14:09:13 <DorpsGek> SVN: glx (r2351) -Fix (#82): invalid read
14:10:56 <TrueBrain> how do you convert NAME to hex in bash ...
14:15:08 <TrueBrain> echo NAME | od -t x4 | head -n 1 | cut -d\ -f2
14:20:27 <TrueBrain> I like mine more, sorry :)
14:23:05 <TrueBrain> ugh, my script acts weird ... it resolves /
14:24:16 <TrueBrain> for i in `cat $1`; do echo $i; done
14:24:19 <TrueBrain> and it resolves / ....
14:26:05 <TrueBrain> why the ... would it resolve?
14:26:18 <TrueBrain> I am not the best in bash, but I thought I could walk through a file if I wanted too ...
14:29:14 <TrueBrain> what the fuck .... any of you any idea why it happens?
14:55:58 <TrueBrain> whiiee, tnx to blathijs, OpenDUNE now runs :D
14:58:28 <DorpsGek> SVN: truebrain (r2352) -Add: TinyC support (doesn't know byteswapping)
15:00:04 <DorpsGek> SVN: truebrain (r2353) -Add: a Makefile for TCC, as it wants stuff slightly different (big thanks to blathijs for supplying the needed perl line)
15:00:29 <TrueBrain> it runs very smooth
15:01:59 <TrueBrain> it is smaller than gcc compiler binary, after optimization and stripping :P
15:02:37 <TrueBrain> means we are fully ISO C btw :p
15:04:37 <Alberth> {'N', 'A', 'M', 'E'} ?
15:06:54 <TrueBrain> might work too .. but this works too :)
15:07:03 <TrueBrain> kinda didn't want to make our current code uglier for a silly compiler
15:14:10 <TrueBrain> I now want a BE machine
15:14:17 <TrueBrain> not much should be needed to get that to work
15:33:22 * Alberth thinks that BE != my be this morning :)
16:45:07 <TrueBrain> k, I am sick of this server, it keeps timing out .. time for a reconnect
16:45:13 *** TrueBrain has joined #openDune
16:45:45 <TrueBrain> lets see if thisone is more stable
18:27:27 <TrueBrain> I haven't seen the AI rebuild anything
18:27:31 <TrueBrain> might be that I am just doing H4 atm
18:58:23 <Alberth> I always go for the construction yard first :)
18:59:07 <Alberth> I *think* I have seen it, but that was an awful long time ago, and might also just be my imagination, or /me being confused with dune 2k
19:47:22 <Alberth> would a different 'unveiled' colour for the radar not be nice? it is very hard to see discovered rock
19:54:09 <Alberth> my game very often says 'atreides harvester deployed', is that normal?
20:12:31 <glx> if you play atreides it might
20:25:25 <TrueBrain> every time it deploys
20:25:31 <TrueBrain> kind of nice, as that is the time to check it :D
20:26:18 <TrueBrain> so far OpenDUNE works very stable
20:26:23 <TrueBrain> no scripting errors
20:26:25 <TrueBrain> but I am just at H5
20:28:41 <Alberth> it also says it when in the field, it seems
20:30:04 <TrueBrain> with carry alls that seems the way, yes, but here I only have seen it while just getting out of the refinery
20:33:51 <Alberth> perhaps when it needs to drive to another spice tile (ie with some sand in-between)?
20:34:07 <TrueBrain> haven't seen that yet
20:34:20 <TrueBrain> but with 4 harvesters, I kinda lose track anyway
20:34:39 <TrueBrain> and remember, those messages are queued behind "acknowledged" of giving commands and clicking units
20:36:39 <TrueBrain> trapping sandworms is fun
20:37:59 <TrueBrain> I hate harkonnen for not having trikes
20:38:03 <TrueBrain> so slow to explore the map
20:40:13 <TrueBrain> yeah, got 2 harvesters for 270
20:44:19 <TrueBrain> a sandworm only comes for units you see
21:00:54 <TrueBrain> could you always order trikes via starport as harkonnen?
21:13:40 <TrueBrain> I get the enemy aproaching signal on a harvester, lolz
21:22:11 <Alberth> yes, as far as I remember you could get trikes via starport
21:24:13 <TrueBrain> I only use the starport to get stuff below costprice
21:24:16 <TrueBrain> its funny like that :)
21:25:51 <Alberth> that also helps in not depleting it completely :)
21:46:53 <TrueBrain> starport requests are bugged somehow
21:46:57 <TrueBrain> I just got tile corruption ...
21:59:38 *** Jaunt_srsly has joined #openDune
22:00:15 <Jaunt_srsly> I downloaded your latest opendune .5
22:00:50 <Jaunt_srsly> I read about it on blue's news website
22:01:31 <TrueBrain> on who's news website?
22:02:54 <TrueBrain> whiiee, we are on a news website
22:03:34 <Jaunt_srsly> I just have a few questions about actually getting to play the openDune2
22:03:59 <Jaunt_srsly> It seems I need the original Dune2 game, is that right?
22:05:59 <Jaunt_srsly> can i do anything cool with the opendune 0.5 without the original Dune2 game?
22:06:29 <TrueBrain> or you must be an artist and you can draw .. 500 images or so? :D
22:07:59 <TrueBrain> (please say yes :D)
22:08:51 <Jaunt_srsly> oh, sorry i was looking for dune2. no im not an artist
22:09:38 <TrueBrain> then you really need the original dune2 (more specific 1.07eu version) before you can play OpenDUNE
22:23:10 <Jaunt_srsly> 1.07 is an unofficial patch?
22:24:29 <Jaunt_srsly> nvm i see that 1.07 is a real patch
22:27:26 *** TinoDidriksen has joined #openDune
22:29:20 <Jaunt_srsly> i can only find the 1.07us version, Dune II: The Building of a Dynasty. I'm looking for "Battle for Arrakis"
22:34:57 *** ChanServ sets mode: +o TrueBrain
22:48:05 <Jaunt_srsly> well thanks anyway. I read that this is just one usage of your nifty DOS decompiler
continue to next day ⏵