IRC logs for #opendune on OFTC at 2011-07-29
⏴ go to previous day
08:18:56 *** Alberth has joined #openDune
08:18:57 *** ChanServ sets mode: +o Alberth
11:05:55 <TrueBrain> hhm, right, we are supposed to release 0.5 in a few days :D
11:06:02 <TrueBrain> do we have any open bugs ...
11:06:22 <TrueBrain> so that is positive
11:06:36 <TrueBrain> guess that we can just compile it on the platforms and release it :D
11:08:57 <Alberth> I remember having some problems with the new malloc overwriting parts of my screen, and then crashing the program while preparing for battle
11:09:19 <TrueBrain> there is an assert that will trigger than
11:09:41 <TrueBrain> and MrFlibble posted a few days ago that it happens in the original too, due to glitches voice files
11:09:41 <Alberth> hoe does an assert not cause a crash? :p
11:10:25 <TrueBrain> I tried to get more info on that from him, but he only posted the fixed voice files
11:10:28 <TrueBrain> I kinda want to know what is wrong
11:10:44 <TrueBrain> but so it has been in Dune2 (and with that in OpenDUNE) for a long time, so I am not too worried about that for a 0.5 release
11:11:51 <TrueBrain> so the malloc is working fine; it just mallocs too much memory over time
11:12:11 <TrueBrain> (and like with any system, that crashes after a while)
11:12:13 <Alberth> Ah, I missed that post
11:12:22 <TrueBrain> he posted it like 2 days ago
11:12:26 <TrueBrain> kinda explained a few things
11:12:43 <TrueBrain> and as far as I know the conversion is going slowly but steady for the rest ...
11:12:44 <Alberth> Yeah it does explain things :)
11:13:02 <TrueBrain> I only had issues compiling it on OSX for i386
11:13:21 <Alberth> I looked at the forums quite often, but mistakenly assumed I was logged in, so I could see new items.
11:13:22 <TrueBrain> but that might be due to my own, I have been fiddling with my OS
11:13:31 <TrueBrain> I make that mistake on forums often too :D
11:15:11 <Alberth> given its manufacturer, I would not trust OSX at all for working sanely ;)
11:15:44 <TrueBrain> it now complains it can't find stdarg.h :(
11:17:21 <TrueBrain> lets drop Mac OSX 10.4 as supported target I guess
11:18:22 <TrueBrain> still libSDL errors :(
11:23:00 <TrueBrain> ah, they switched the default to GCC 4.2
11:23:04 <TrueBrain> which is not support by 10.4u :D
11:27:47 <TrueBrain> works under OSX, including voices :D
11:30:01 <TrueBrain> also compiling ppc version
11:30:05 <TrueBrain> but that will fail so horrible
11:30:09 <TrueBrain> as we assume LE on so many places :D
11:30:57 <TrueBrain> Dune2 did, so little choice :)
11:31:07 <TrueBrain> we did add a few swap-byte routines, but I am pretty sure not all :)
11:31:22 * TrueBrain points to savegame handling
11:31:53 * Alberth points to all g_global constant arrays
11:34:41 <TrueBrain> they will hopefully fix theirself over time :D
11:36:37 * Alberth expects it will cost some human effort
11:36:55 <TrueBrain> so work faster! :P :P :P
11:37:31 <TrueBrain> meh; guess I have enough of erlang or a bit, and I can spend the day on OpenDUNE
11:44:08 <TrueBrain> can;t remember what I was doing ....
11:45:15 <Alberth> malloc & friends iirc
11:49:18 <TrueBrain> ah, I see I moved WSA already to normal malloc/free
11:49:28 <TrueBrain> thought I left off there
11:49:30 <TrueBrain> clearly I was wrong :P
11:52:52 <DorpsGek> SVN: truebrain (r2054) -Codechange: unify all 66EC usage
11:53:09 <TrueBrain> owh, lolz, thought it was a special array
11:53:13 <TrueBrain> then I checked the comments
11:56:06 <DorpsGek> SVN: truebrain (r2055) -Add: removed 66EC usage, as it is nothing but a simple array of offsets
11:56:21 <TrueBrain> wsa.c only has 2 emu_'s, left, both for screenID stuff :P
12:05:01 <DorpsGek> SVN: truebrain (r2056) -Codechange: useless amount of spaces :P
12:05:23 <TrueBrain> ugh, Dune2 plays around with memory too much .. hard to track
12:12:19 <TrueBrain> we have to push out a new release in a few days ;)
12:12:40 <TrueBrain> OSX compiles, I assume MSVC and linux do too .. so I don't feel like a feature freeze this time :P
12:14:34 <TrueBrain> after 0.5, I will be working on removing libemu a bit, like moving SDL to OpenDUNE etc etc
12:14:42 <TrueBrain> not sure how far it is possible with voices
12:15:26 <TrueBrain> I might just move the calls to OpenDUNE directly, and remove the emu_ part from it or something
12:15:33 <TrueBrain> (so instead of inb and outb, just calling the right function)
12:15:51 <TrueBrain> normally you can't just do that, but as we own the license to both projects, we can :D:D:D
12:16:46 <glx> I could try to replace some music driver functions too
12:16:52 <TrueBrain> okay, and we need to settle a naming convention, this is annoying me :p
12:17:02 <TrueBrain> local static variables, are now g_ and _
12:17:08 <TrueBrain> global variables are now g_ and _ :P
12:17:53 <TrueBrain> wiki mentions _ for local static
12:18:06 <TrueBrain> so g_ for global, _ for local, fine
12:18:10 <TrueBrain> s_ is maybe nicer thou
12:19:05 <TrueBrain> glx / Alberth: any objections if I make it s_ for static loca, and g_ for global?
12:19:53 <Alberth> basically s/\<_/s_/ ? fine
12:22:53 <Alberth> we currently have a s_local or so for the input segment data, I think
12:23:59 <DorpsGek> SVN: glx (r2057) -Add: removed useless emu_ functions
12:26:41 <TrueBrain> lol; ther ewas a nasty bug in iconsprite loading :D
12:26:46 <TrueBrain> but it fixes itself in what I am doing now :P
12:35:30 <glx> ok I'll try to replace f__AB01_08CE_005F_AC14() with midiOutShortMsg() (WIN32 only for now)
12:38:05 <TrueBrain> and I fucked up something :D:D
12:57:41 <TrueBrain> what do you guys think?
12:57:47 <TrueBrain> I am afraid it is a bit over-the-top :P
12:58:01 <TrueBrain> but it works, and it is much better
12:58:15 <TrueBrain> I just removed all references to screenID(5), and used malloc instead
12:58:53 <TrueBrain> and I shared the pointer between sprites.c and gfx.c, instead of moving them around between them
13:05:43 <DorpsGek> SVN: truebrain (r2058) -Add: don't 'optional' use ScreenID 5 for SpriteInfo, but malloc. Also cleanup the function a bit to make more sense, as it was messy :D
13:06:12 <Alberth> it seems shorter and simpler :)
13:06:39 <TrueBrain> the only thing I no longer verify is that the chunk length of the RTBL is equal to the amount of sprites in SpriteInfo
13:18:22 <TrueBrain> what do you think .. should I put effort in making sure all memroy is free'd on closing?
13:24:31 <Alberth> it would be better imho, as it makes tool output like valgrind more useful
13:24:47 <TrueBrain> so I guess I have to write a few Uninits
13:24:53 <TrueBrain> as Dune2 doesn't give a crap :P
13:25:14 <TrueBrain> it also uses pointers that are 'free'd on shutdown etc :P
13:25:27 <Alberth> they considered the program so brilliant they assumed it would never be stopped :p
13:26:54 <Alberth> much like some Sun machines did not have a power button :)
13:29:05 <TrueBrain> there are 2 orientation tables
13:29:16 <TrueBrain> how to name that ...
13:29:47 <TrueBrain> degress are from 0 .. 255 for all items
13:29:48 <Alberth> there is also a 0..255 table :)
13:29:55 <TrueBrain> then this is converted to either of the two
13:30:07 <TrueBrain> I guess one is for the turret, the other for the tank itself
13:30:20 <Alberth> 8directions and 16directions ?
13:30:54 <Alberth> or 'orientation8' and 'orientation16'
13:31:02 <Alberth> or merge the former into the latter :)
13:32:17 <Alberth> there are also many map offset tables, you'd think they would share such data, but apparently not
13:32:34 <TrueBrain> you can't merge them funny enough
13:32:38 <TrueBrain> they are initialized like this:
13:32:47 <TrueBrain> ((i + 16) / 32) & 0x7
13:32:55 <TrueBrain> ((i + 8) / 16) & 0xF
13:33:09 <TrueBrain> so at the 'halve' of the degree it already switches
13:33:30 <TrueBrain> so '7' for the first goes from 13.5 to 15.5 for the latter :P
13:33:56 <TrueBrain> well, of course it is possible to convert that
13:34:26 <Alberth> oh, it needs a >> 1 as well :)
13:34:49 <Alberth> but yeah, keep it 2 tables :)
13:35:03 <glx> hmm ok let's steal some files from libemu
13:37:48 <TrueBrain> keep that for after 0.5?
13:38:06 <TrueBrain> so we can also do that with SDL and shit :P
13:38:24 <Alberth> and ditch the emu repo :p
13:38:39 <TrueBrain> well, we ca't until all emu_ shit is gone :p
13:38:51 <TrueBrain> but I rather have 0.5 like it is now, and that from 0.5 we start working to remove libemu :)
13:39:02 <TrueBrain> if it helps, we can also just release 0.5 in a bit :P
13:39:21 <Alberth> sound like a good plan to wait until after 0.5
13:39:24 <TrueBrain> how to name this function ... Orientation_DirectionToOrientation8 ?
13:40:04 <Alberth> TrueBrain: sounds good to me
13:40:34 <TrueBrain> glx: let me finish up this patch, and lets just release 0.5 now :p
13:40:59 <glx> first step will be to make mpu optionnal in libemu :)
13:41:06 <TrueBrain> Alberth: so now we only have to start to differ between Direction, Orientation8 and Orientation16
13:41:12 <TrueBrain> lets call Direction Orientation256 :P
13:42:18 <Alberth> under the assumption that all 'Direction's are actually that ;)
13:42:41 <TrueBrain> ugh, I should hav eocmmited before renaming
13:44:59 <TrueBrain> I can only find things that either rotate in 4 (troops) or 8 (tanks)
13:45:03 <TrueBrain> don't see this 8 and 16 :P
13:46:31 <glx> hmm I don't get why there is emu_overlay in libemu
13:46:38 <DorpsGek> SVN: truebrain (r2059) -Add: used malloc() for iconMap, and use a static table for orientation conversion table
13:46:45 <TrueBrain> it is part of the overlay code :)
13:46:48 <glx> we set it to 1 after init but it's never used
13:47:09 <TrueBrain> to make sure I could still run apps that didn't use overlay :)
13:47:15 <glx> well at least not real use in libemu
13:47:28 <TrueBrain> anyway, shall we release?
13:47:33 <TrueBrain> so you can continue what you are doing :)
13:48:11 <TrueBrain> any patches in your queue that can be useful glx / Alberth?
13:48:25 <TrueBrain> k; gimme 5 to prepar
13:48:33 <TrueBrain> please compile on your local compiler to check for errors plz
13:49:32 <TrueBrain> gcc 4.2, 4.3 and 4.4 give no warnings (except for vsnprntf of course)
13:51:36 <glx> mingw and msvc debug are ok (checking release now)
13:51:39 <DorpsGek> SVN: truebrain (r2060) -Add: prepare for 0.5 release
13:52:46 <glx> Avertissement 3 warning C4700: variable locale 'backup' non initialisée utilisée d:\developpement\opendune\trunk2\src\gui\gui.c 2803
13:52:57 <Alberth> 4.6 gives a few extra 'variable XYZ set but not used' warnings
13:53:30 <TrueBrain> glx: that is indeed a weird one
13:53:30 <glx> oh the warning seems valid
13:53:50 <TrueBrain> seems the params are switched
13:53:56 <glx> but it warns only for release build
13:55:24 <DorpsGek> SVN: truebrain (r2061) -Fix (r1898): memcpy() is in Dune2 (src, dest), but in POSIX (dest, src). Tnx to glx for spotting this in the nick of time
13:55:54 <TrueBrain> clearly none of us have finish the game in a while :D :P
13:56:01 <DorpsGek> SVN: truebrain (r2062) -Release: 0.5
13:56:45 <Alberth> I have never finished the game for all houses, I think
13:57:32 <Alberth> without a savegame would be a huge achievement :p
13:57:53 <glx> I mean savegames ready to finish
13:58:05 <TrueBrain> glx: can you feed me with a .exe and SDL.dll plz?
13:58:34 <glx> IIRC SDL.dll is still the same
13:59:47 <TrueBrain> aint that always true? :D
14:00:14 <TrueBrain> hmm .. alsa is giving me a hard time again ..
14:05:10 <DorpsGek> SVN: truebrain (r2063) -Add: bump version for trunk
14:05:18 <TrueBrain> use 2061 or 2062 for release plz ;)
14:06:36 <TrueBrain> Xaroth|Work: feel like writing the post on forum and frontpage this time?
14:07:02 *** DorpsGek sets mode: +o TrueBrain
14:07:11 *** DorpsGek sets mode: -o TrueBrain
14:07:32 <TrueBrain> most amount of revisions between releases btw :P
14:07:36 <TrueBrain> concratz on that guys :D:D
14:07:37 <glx> hmm someone somehow failed to make the tag
14:07:48 <glx> Avertissement 3 warning C4700: variable locale 'backup' non initialisée utilisée d:\developpement\opendune\0.5\src\gui\gui.c 2803
14:08:54 <Alberth> commits are smaller, so you need more of them :p
14:09:34 <DorpsGek> SVN: truebrain (r2064) -Backport (r2061): release failed to copy from right revision
14:09:41 <Alberth> and you have to multiply by 2/3 for a fair comparison :)
14:09:49 <TrueBrain> I made my releases from trunk, so my binaries are fine :D
14:14:57 <TrueBrain> Xaroth|Work: nevermnd, did it already :P
14:17:10 <Alberth> TrueBrain: Worklist in /topic is also not so relevant any more, I think
14:23:32 <TrueBrain> planetmaker: if you have time, would you mind testing the MacOSX version?
14:24:36 <TrueBrain> on linux, I get an assert at src/driver.c :(
14:25:03 <TrueBrain> ah, missing data files of course
14:25:07 <TrueBrain> but it no longer gives this nice error :(
14:25:51 <TrueBrain> we really should fix that :)
14:26:03 <TrueBrain> amyway, glx, go ahead, copy/paste mpu* from libemu :P
14:26:34 <glx> before copy/paste I need a way to easily disable mpu in libemu
14:26:47 <TrueBrain> name it slightly different?
14:26:59 <glx> it's not a naming problem
14:27:10 <glx> you can init mpu only once
14:27:24 <TrueBrain> under linux it is easy, just don't define DMPU_ALSA
14:28:01 <glx> maybe a new libemu global
14:28:18 <TrueBrain> why not change the MSVC project locally to use mpu_none instead of mpu_win32
14:29:04 <glx> it's better to keep libemu opendune independant :)
14:29:22 <TrueBrain> owh, and I forgot ... the game still freezes for me when selecting a house in newgame
14:29:30 <TrueBrain> because it waits for voices to finish, which never started in the first place
14:36:20 <TrueBrain> glx: driver->dcontent, is the PACK of the struct still important? As I kinda want to make it a void* ...
14:37:09 <planetmaker> right... not all deps automatically build when I run make in tags/0.5: make: *** No rule to make target `libemu.so', needed by `opendune'. Stop.
14:37:28 <TrueBrain> that is because it is a separate project ;)
14:37:40 <planetmaker> then it should not be a dep
14:37:43 <TrueBrain> those should work out-of-the-box
14:38:11 <TrueBrain> planetmaker: I very much disagree :D For the same reason as libsdl is a dep :P
14:38:17 <TrueBrain> but I don't see us including libsdl in the source tree :D
14:39:57 <planetmaker> if you don't find a library, linking should fail. But not compilation...
14:40:22 <TrueBrain> who cares who warns about it :) You need it in any case :)
14:41:42 <planetmaker> let's say, I'm quickly bored, if error messages are pointing quite wrong directions ;-)
14:41:55 <Xaroth|Work> fun fact, the 'development team' list is alphabetical by both nickname and first name :P
14:41:59 <TrueBrain> glx: why not default it to 1?
14:42:18 <TrueBrain> planetmaker: that i why I asked f you could test the resulting binaries ;)
14:42:37 <TrueBrain> as dynamic linking in OSX fails anyway :P
14:42:49 <glx> both options are possible :)
14:43:07 <TrueBrain> glx: I would go with default of 1, because then nothing changes for any project :)
14:43:52 <Xaroth|Work> TrueBrain: shall i try linksubmitting to tweakers? :P
14:43:56 <Xaroth|Work> or do we not want to kill the site :P
14:44:02 <TrueBrain> wordpress would die!
14:44:06 <TrueBrain> make a better site plz :P
14:44:12 <planetmaker> TrueBrain: it's still stupid to not compile or link it, if I have the complete svn tree which includes libemu
14:45:35 <DorpsGek> SVN: glx (r2065) [LibEMU] -Add: make MPU externally disablable
14:45:52 <TrueBrain> planetmaker: it is also stupid to not read the instructions how to compile a piece of software ;)
14:46:36 <planetmaker> who reads instructions? Do you?
14:46:52 <Xaroth|Work> if at first I don't succeed, yes.
14:47:44 <TrueBrain> planetmaker: I don't know you being so bitter :p Wassup?
14:48:44 <DorpsGek> SVN: truebrain (r2066) -Add: make g_strings and g_stringsHint use malloc()
14:50:28 <Alberth> planetmaker: but libemu.so is a link dependency
14:51:01 <planetmaker> hm... I guess I got a classical hang at 100% cpu...
14:51:35 <planetmaker> and funnily ctrl+C is caught and not handled :S
14:52:06 <Alberth> you're probably stuck in the OS, where ^C has no effect
14:53:40 <Alberth> xterm passes it on to whatever you are running, the 'whatever you are running' is stuck in the kernel :)
14:53:42 <TrueBrain> but when/where? And witht he binary from binaries.opendune.org?
14:54:06 <planetmaker> yes, that binary. Loaded one of my savegames
14:54:10 <Alberth> to protect you from breaking the kernel
14:54:14 <TrueBrain> so it did start, you saw text etc etc?
14:54:39 <TrueBrain> first ^C is handled by libemu btw. The second one is passed through normal shissle
14:54:52 <planetmaker> and now a crash ingame...
14:54:53 <TrueBrain> did the map itself load?
14:55:03 <TrueBrain> yeah, but did you see it?
14:55:13 <planetmaker> yes. But never anything moved
14:55:29 <TrueBrain> and what kind of crash?
14:55:37 <TrueBrain> assert, libemu message, OS crash?
14:56:22 <TrueBrain> meh; at least it starts, I guess it means my compiling works :p I guess it crashes on voices ...
14:56:39 <planetmaker> Program Termination: jumped to 4352:0000, which is not decompiled.
14:56:39 <planetmaker> The jump was triggered at src/call.c:135
14:56:39 <planetmaker> The jump appears to originate from 0000:0000.
14:56:43 <TrueBrain> I am not too happy with asound ...
14:57:06 <TrueBrain> hmm, 16bit memory corruption ... that is weird ..
14:57:21 <TrueBrain> does the same happen if you start a new game?
14:58:33 <planetmaker> hm... how do I attach more than 1 file to a bug report?
14:58:41 <TrueBrain> no clue; believe you can't
14:59:01 <TrueBrain> I know that for a few revisions we made savegames which are invalid; it might happen that you catched thatone?
14:59:06 <Alberth> report the bug, add files after reporting
14:59:18 <planetmaker> that might be. Dunno :-)
15:02:45 <planetmaker> hm... any way to tell which savegame is which?
15:02:56 <TrueBrain> I Believe 00000 is always at the bottom
15:06:35 <TrueBrain> glx: driver is size-depending still, as it is in g_global :) It has a Tools_Malloc, but we will get there when we get there :)
15:10:13 <planetmaker> hm, there's no clear reproducable way to create the crashes it seems
15:10:42 <TrueBrain> do newgames cause it?
15:10:46 <TrueBrain> or only with that savegame?
15:11:16 <planetmaker> well. Even with that same savegame it doesn't always happen
15:11:32 <planetmaker> and... rockets fly 90° wrong at times
15:11:40 <planetmaker> with the exhaust 90° to the flight direction
15:12:22 <TrueBrain> hmm, that can be a valid bug with one of my latest changes ... I will have to test that ..
15:12:59 <planetmaker> but not all of course
15:14:35 <TrueBrain> the latter, I can reproduce :D
15:14:40 <planetmaker> seems to be the case for left-facing towers
15:15:40 <TrueBrain> I double checked that table .. hmm ... grrr
15:17:28 <DorpsGek> SVN: truebrain (r2067) -Add: free mouseSprite and mouseSpriteBuffer from the 16bit world. RUN FREE!
15:18:25 <Alberth> rather, get restricted in a 32 or 64 bit world :p
15:19:00 <planetmaker> feature request: don't interpret the scroll guesture as mouse click ;-)
15:19:04 <planetmaker> it's quite irritating
15:19:19 <planetmaker> when dealing with the file menu ;-)
15:19:30 <Alberth> planetmaker: yeah, very annoying
15:22:43 <planetmaker> But... now I don't manage to get it reproducable...
15:23:07 <planetmaker> hangs are "easier" but also not reproducable. Wait a bit. Wait long.... whatever
15:24:16 <TrueBrain> so the most logic reason would be voices
15:24:24 <TrueBrain> and the rockets are funny :p
15:24:28 <TrueBrain> everything to the left goes wrong :P
15:25:19 <planetmaker> just loaded the save and waited a bit... why ever
15:26:02 <TrueBrain> well, I hope glx can fix most of that for 0.6
15:26:12 <TrueBrain> for now it would be best to disable it I guess
15:26:28 <planetmaker> very much so. Playing "seriously" will seriously fail
15:26:41 <planetmaker> at least here :-(
15:27:10 <TrueBrain> trying to understand why rockets go wrong, but I can't find it :(
15:27:32 <TrueBrain> its because it is not my patch :D
15:27:46 <planetmaker> how do I disable sound... hm, I guess in options :-)
15:28:10 <TrueBrain> or in settings, maybe that is enoguh
15:28:30 <TrueBrain> r2050 broken ... backtracing ...
15:29:14 <planetmaker> nice. The option "sound = off" has no effect
15:29:34 <TrueBrain> no effect, in what way?
15:29:41 <planetmaker> I still go sound afterwards.
15:29:56 <planetmaker> Program Termination: jumped to FFEC:0000, which is not decompiled.
15:29:56 <planetmaker> The jump was triggered at src/call.c:135
15:29:56 <planetmaker> The jump appears to originate from 0000:0000.
15:30:08 <TrueBrain> null pointer writing data :(
15:30:12 <planetmaker> ^^ that subsequently
15:31:12 <planetmaker> probably still corrupted as game started with sound
15:31:16 <planetmaker> after restart it's silent
15:31:26 <TrueBrain> solves the crashes too?
15:32:29 <TrueBrain> r2010 fucks up rockets :P
15:32:38 <TrueBrain> Alberth's patch, any clue where to look for?
15:32:52 <TrueBrain> one of the tables flawed?
15:37:41 <TrueBrain> 0xFFFA, what is that?
15:38:02 <TrueBrain> two-complement, so 0xFFFF = 0, FE = -1, FD = -2, FC = -3, FB = -4, FA = -5, right?
15:38:58 <TrueBrain> (just wrote a small app :P)
15:41:56 <TrueBrain> the tables in r2010 seem fine
15:44:28 <planetmaker> without sound I didn't get another hang or crash so far
15:46:46 <TrueBrain> planetmaker: perfect :)
15:46:51 <planetmaker> hm. A converted tank would still follow the orders I gave it, even after it changed back its colour
15:47:12 <planetmaker> (i.e. run senselessly past my main attack forces)
15:55:04 <TrueBrain> nope, can't detect any problem with r2010 :( I must be overlooking something obvious :P
16:00:06 <TrueBrain> you got to love those bugs :D
16:06:32 <glx> and who destroyed the carryall ?
16:07:27 <TrueBrain> but now find the bug in r2010 :D
16:07:34 <TrueBrain> or plz confirm it is 2010 :P
16:14:58 <DorpsGek> SVN: truebrain (r2068) -Add: palettes can also live in a 32bit/64bit world
16:30:11 <DorpsGek> SVN: truebrain (r2069) -Revert (r2068): sadly, palettes have to remain in their jail of 16bit world for now
16:33:17 <TrueBrain> bit in a stale-mate with the rest of the mallocs :P
16:35:55 <Alberth> in 2010 table contents was printed and copy/pasted so unless I printed the wrong table, that should be ok
16:36:08 <TrueBrain> I double checked them, they seem fine
16:36:17 <Alberth> perhaps a signed/unsigned problem?
16:36:26 <TrueBrain> hmm, that is a good one
16:39:26 <glx> nice AB01 is full of never called functions :)
16:45:32 <DorpsGek> SVN: truebrain (r2070) -Add: reduce the amount of ScreenIDs used
16:47:56 <TrueBrain> Alberth: can't find an obvious problem with signed/unsigned
16:53:27 <TrueBrain> then again, I have no clue where it finds the right sprite :D
16:58:11 <glx> ok mpu withou libemu seems to work
16:59:50 <TrueBrain> diff is sadly enough partily (copies), so I can't test; but if it works, go for it :D
17:00:05 <TrueBrain> maybe add a comment
17:00:17 <TrueBrain> like: /* We don't need the MPU of libemu */
17:00:40 <glx> I'm removing more AB01 stuff now (never called functions)
17:01:03 <glx> I guess they appeared when testing different drivers ;)
17:02:47 <DorpsGek> SVN: truebrain (r2071) -Fix: global.h should only contain data required for the transition. Moved out enums that are required even after that time.
17:06:32 <DorpsGek> SVN: truebrain (r2072) -Fix: moved more stuff out of global.h that should not be there.
17:07:34 <TrueBrain> ugh, where to move lengthof to ...
17:10:25 <TrueBrain> fits there nicely I think, as some compilres do in fact have lengthof
17:10:32 <DorpsGek> SVN: glx (r2073) -Add: use MPU without libemu layer
17:11:20 <glx> if it works for me it should work for ALSA too
17:11:34 <TrueBrain> same errors as always :D
17:12:06 <DorpsGek> SVN: truebrain (r2074) -Fix (r2073): mute a GCC warning
17:13:42 <glx> hmm maybe hou need to add MPU_ALSA in configure
17:14:17 <TrueBrain> just realised that the 64bit binaries I released never had ALSA compiled in :D
17:14:56 <DorpsGek> SVN: truebrain (r2075) -Add (r2073): port the ALSA statement from libemu to OpenDUNE. Mind you: it does not autodetect yet. Run: make ALSA:=1 to compile with MPU support on Linux.
17:16:46 <DorpsGek> SVN: truebrain (r2076) -Fix (r2073): don't include files we don't need, and we are no longer [EMU]
17:18:44 <TrueBrain> the name given to ALSA as channel
17:18:58 <TrueBrain> I just wanted the emu_ name gone :P
17:19:43 <glx> emu_caption contains the window caption
17:20:39 <TrueBrain> ugh, you are right ...
17:21:01 <glx> and I guess you will have linking error
17:22:07 <DorpsGek> SVN: truebrain (r2077) -Fix (r2076): name the OpenDUNE MIDI channel as such
17:23:08 <DorpsGek> SVN: truebrain (r2078) -Add: cleanup includes a bit, don't include global.h when you don't have to
17:23:16 <TrueBrain> there, less includes == happy TrueBrain
17:27:36 <Alberth> orientation=12 <-- values_3304 is not that long :)
17:33:05 <glx> maybe 3324 is part of 3304 ;)
17:33:55 <Alberth> 14 is the highest number I got in my quick test
17:40:52 <Alberth> extending to [16][2] seems to fix the problem
17:44:29 <Alberth> go ahead then, I need to copy it hg -> svn :)
17:46:34 <DorpsGek> SVN: glx (r2079) -Fix (r1142, r2010): 3304 array was too short
17:47:11 <glx> hmm forgot to add the mantis tag
17:48:18 <Alberth> and fix the length :)
17:55:33 <DorpsGek> SVN: glx (r2080) -Fix (r2079): forgot to update the comment
17:55:59 <TrueBrain> how to call the black sprite
17:56:04 <TrueBrain> as in, we have bloomSpriteID
17:56:14 <TrueBrain> and this is the one that is totally black
17:56:51 <Alberth> yeah, that's the term I was looking for
17:56:59 <TrueBrain> is that valid english?
17:59:01 <Alberth> but veiledSpriteID seems like the correct technical term imho
17:59:49 <TrueBrain> veiledSpriteID it is
18:00:42 <DorpsGek> SVN: truebrain (r2081) -Add: named veiledSpriteID
18:00:50 <TrueBrain> now let me get them out of g_global :)
18:06:34 <TrueBrain> k, we need a wrapper IsPartiallyVeiled
18:06:42 <TrueBrain> but is that a Map_, Tile_ or Sprites_ ? :D
18:07:02 <TrueBrain> Sprites_IsUnveiled I guess
18:19:50 <TrueBrain> ugh, game hangs for me on radar on .. guess that is voice related again
18:20:29 <DorpsGek> SVN: truebrain (r2082) -Add: Sprite_IsUnveiled() to detect if a sprite is part of the veil-sprites, and moved all g_global spriteID variables out of there
18:20:45 <DorpsGek> SVN: truebrain (r2083) -Fix (r2082): minor typo
18:31:30 <DorpsGek> SVN: truebrain (r2084) -Add: move randomSeed to a static local variable with nice wrappers
18:32:53 <DorpsGek> SVN: truebrain (r2085) -Codechange: group Timer functions together
19:16:42 <DorpsGek> SVN: truebrain (r2086) -Add: moved all timer functions and variables to timer.[ch], and moved all the stuff for it out of g_global
19:30:48 <DorpsGek> SVN: truebrain (r2087) -Fix: obey coding style for static local variables (is now s_, consistant)
19:32:15 <DorpsGek> SVN: truebrain (r2088) -Fix (r2086): forgot to name the magic numbers with enums
19:34:38 <DorpsGek> SVN: truebrain (r2089) -Add: removed read-only variable 37AA.
19:37:25 <DorpsGek> SVN: truebrain (r2090) -Add: removed write-only varialbes 37A2 - 37A8 (some kind of script debugger it seems)
19:42:46 <DorpsGek> SVN: truebrain (r2091) -Add: 70A2 is g_paletteActive, updated when also sending to the screen
19:51:17 <DorpsGek> SVN: truebrain (r2092) -Add: made the unit-script-run code more readable by removing 2 global variables
19:53:48 <DorpsGek> SVN: truebrain (r2093) -Add: remove last g_global reference in team.c
19:56:37 <DorpsGek> SVN: truebrain (r2094) -Add: removed the last g_global reference from string.c
20:08:52 <DorpsGek> SVN: truebrain (r2095) -Add: moved the last emu_ piece in animation.c up one level, and removed the size restriction from Animation
20:09:00 <TrueBrain> tooo .. muucchhhh .. commmittsss ... :P
20:09:45 <TrueBrain> I found where most animation commands are stored
20:09:50 <TrueBrain> but I don't feel like making that nice now :P
20:10:58 <glx> you did your commit quota for 6 months ;)
20:11:42 <DorpsGek> SVN: truebrain (r2096) -Fix: no longer global.h is needed in editbox.c
20:11:46 <TrueBrain> awh, so I am no longer allowed to commit :(
20:12:48 <DorpsGek> SVN: truebrain (r2097) -Fix: emu_Global_GetIP() is no longer used
20:14:10 <DorpsGek> SVN: truebrain (r2098) -Fix: g_fontIntro is malloc()'d, so use free() on it
20:15:21 <glx> but emu_Global_GetCSIP() will be another major step
20:15:58 <DorpsGek> SVN: truebrain (r2099) -Add: g_paletteMapping[1]2 can switch to malloc/free
20:16:16 <TrueBrain> Interrupt_Video() also needs getCSIP()
20:16:22 <TrueBrain> guess it is time to start bypassing libemu
20:43:53 <Xaroth|Work> wasnt that one fixed in the latest release?
20:44:25 <glx> it will fixed in next release
20:44:38 <glx> (the version I have is fixed)
20:57:07 <TrueBrain> they finally fixed it?
20:58:19 <glx> it seems the dev has only IPv4 available (hard to test IPv6 then ;) )
20:58:33 <TrueBrain> I want to know the patch :P
21:02:29 <glx> the mails were sent the 26th
21:02:39 <glx> but I received them only today
21:08:35 <glx> it's quite impossible to see changes
21:27:31 <glx> anyway winscp now works correctly :)
21:27:46 <glx> btw putty has been updated recently too
21:43:54 <TrueBrain> so, back to OpenDUNE .. how to solve the Interrupt_Video easily
21:43:57 <TrueBrain> that is a very though one
21:44:01 <TrueBrain> as in libemu it is even ugly :P
21:46:02 <TrueBrain> hmm, I have another idea
21:46:16 <glx> replace all this with plain SDL
21:46:19 <TrueBrain> I can just temporary push it in a 16bit buffer for that function :)
21:46:29 <TrueBrain> but that is a long and hard road :P
21:47:06 <glx> of course will need a way to disable SDL_VIDEO in libemu ;)
21:47:21 <TrueBrain> problem is that you need to do everything at once
21:47:25 <TrueBrain> which is kinda a bitch to do :D
21:47:49 <TrueBrain> and disabling SDL in libemu is easy: never initialize video mode
21:48:09 <TrueBrain> SDL_Init is only called on change of gfx mode :)
21:49:24 <TrueBrain> but okay, lets try it then :)
21:56:01 <TrueBrain> we need a configure :D
21:58:50 <TrueBrain> okay, I have a black screen :P
22:00:24 <TrueBrain> how to do the event-loop
22:00:27 <TrueBrain> that is not so easy ...
22:04:40 <TrueBrain> I have image of some kind :D
22:09:57 <TrueBrain> now the on-the-fly palette thingy ...
22:11:19 <TrueBrain> not completely unexpected
22:11:23 <TrueBrain> infinite loops are easily possible
22:28:31 <DorpsGek> SVN: truebrain (r2100) -Add [ENHANCEMENT]: when fading in the screen on starting game, do it against black, so you see what is going on
22:28:39 <TrueBrain> finally :D:D:D:D Wanted to do that for so long now :P
22:29:19 <TrueBrain> it works, except for keys
22:30:45 <TrueBrain> it hooks itself on the timer
22:31:10 <TrueBrain> and the mouse actions are reinserted into libemu :D
22:43:12 <TrueBrain> keyboard kinda works, but it registers keystrokes twice :(
22:43:52 <TrueBrain> and it hangs often :(
22:45:36 <TrueBrain> this is weird, they are both registered as keydown
22:46:10 <TrueBrain> I did something horrible, and the compiler did not complain
22:46:17 <TrueBrain> switch (a) { int keyup = 1; case: blablabla
22:49:02 <TrueBrain> glx: can you try thatone in MSVC? :)
22:54:33 <TrueBrain> new patch, resolving hang when palette updates when the timer ticks :)
22:54:38 <TrueBrain> hihi, that was a nice deadlock ;)
23:02:31 <TrueBrain> another update, pompiedom :)
23:03:43 <glx> hmm mingw fails to compile :)
23:04:20 <TrueBrain> any more details, or do I need to guess? :D
23:04:42 <TrueBrain> don't forget to do: make WIN32:=1
23:05:28 <TrueBrain> same code as libemu
23:08:54 <glx> ha I remember I needed to specify SDL_INCLUDE for libemu too
23:09:11 <glx> (only needed when sdl.c needs to be recompiled)
23:09:12 <TrueBrain> like said, we need configure :p
23:09:18 <TrueBrain> might copy OpenTTD's soon .. I wrote most of it anyway :p
23:13:38 <glx> opendune hangs when I click
23:14:08 <glx> well the window hangs, but the intro continues to play
23:14:56 <glx> and keypresses have no effect
23:15:30 <TrueBrain> can you try MSVC first please?
23:15:41 <TrueBrain> mingw is always a bit weird with SDL if I am not mistaken
23:16:53 <glx> hmm previous commits added warnings
23:17:24 <TrueBrain> and yeah, please also make sure clean trunk works :D
23:18:14 <TrueBrain> hmm, no, mingw version should work, everything is correct as far as I can tell
23:18:22 <TrueBrain> do realise to update libemu ;)
23:18:27 <TrueBrain> 2 MPUs are not good :D
23:18:54 <TrueBrain> in the meantime I updated the patch a bit
23:19:02 <TrueBrain> might be worth updating, fixed some race-conditions
23:19:06 <glx> keyboard and mous fail in msvc too
23:19:25 <TrueBrain> define 'fail' please?
23:19:33 <TrueBrain> but it does compile btw? That already is awesome :D
23:20:17 <TrueBrain> oops, warning in patch; updated
23:22:51 <TrueBrain> weird, no reaction, as it really should work :P
23:23:14 <TrueBrain> clean trunk does work etc?
23:23:22 <TrueBrain> well, if screen updates, events are read
23:23:56 <glx> screen stops updating after a click
23:24:55 <TrueBrain> pfew, at least I didn't fuck up any other commit then :D:D
23:28:56 <DorpsGek> SVN: truebrain (r2101) [LibEMU] -Fix: don't think you do a right click on scrollwheel actions etc (silly really)
23:29:13 <TrueBrain> special for planetmaker ;)
23:34:32 <TrueBrain> (updated my patch too to reflect r2101 ;))
23:40:01 <TrueBrain> I have an idea why it might hang
23:40:28 <TrueBrain> can you try thatone?
23:42:40 <TrueBrain> can you add some printfs in video/sdl.c, see what is going on?
23:43:11 <glx> returning to mass effect now
23:54:47 <TrueBrain> if this works, I have to do work on mouse stuff, not too hard. Then I can move the pic in, with that the timer, and we should be able to remove libemu as soon as all the emu_ shit drops :P
continue to next day ⏵