IRC logs for #opendune on OFTC at 2011-06-17
⏴ go to previous day
07:37:06 *** Tomaz^W has joined #openDune
08:44:44 <Tomaz^W> Is this project still active?
08:46:21 *** ChanServ sets mode: +o Xaroth|Work
08:47:03 *** Alberth has joined #openDune
08:50:55 <Tomaz^W> If one is interested in discussing decompileg/disassembling with TrueBrain, would this IRC channel be a good way to get in contact with him?
08:51:08 * Xaroth|Work points at TrueBrain
08:51:24 <Xaroth|Work> he might not respond, but he's online 23/7 :)
08:52:37 <Tomaz^W> Great, yeah I was mainly interested in knnowing if he just idled here or actually check his IRC from time to time :)
08:53:27 <Tomaz^W> I'm working on 2 similar projects which is why I'm interested in his knowledge about disassembling
08:53:44 <Tomaz^W> The 2 projects being freeSynd ( might have heard of it? ) and OpenThemePark
08:55:15 <Xaroth|Work> then I hope you got the legal jibber jabber out of the way :P
08:55:36 <Tomaz^W> Yeah, absolutely... :P
08:55:51 <Xaroth|Work> that shit can get real messy
08:56:19 <Tomaz^W> YEah I have contacted Harvey Elliot ( CEO at EA LIght Bright ) which currently own all old bullfrog Trademarks
08:56:53 <Xaroth|Work> Harvey doesn't read his mail :P
08:57:12 <Tomaz^W> You know that from personal experience?
08:57:38 <Xaroth|Work> contacting EA, in general, to ask about licensing stuff like that, usually ends up in /dev/null
08:57:48 <Xaroth|Work> took us a lot of trouble to even get a single response
08:58:00 <Tomaz^W> I have backup plans tho
08:58:25 <Tomaz^W> I know a few people at EA DiCE and one of them would take it up with some marketing guy
08:58:41 <Tomaz^W> So I'm doing what I can
08:58:56 <Tomaz^W> Slowly crawling towards the top of the EA mountain
08:59:39 <Tomaz^W> Right, Dune2 is also owned by EA
08:59:50 <Tomaz^W> You guys sorted legal issues with Dune2?
09:00:14 <Tomaz^W> If so then who did you finally get in contact with in the end?
09:04:33 <Tomaz^W> just a thought, ever considered using ScummVM instead of SDL?
09:04:54 <Tomaz^W> You'd get a much wider cross-plattform support that way
09:05:05 <Tomaz^W> that's what I'm doing with OpenThemePark
09:18:03 <Xaroth|Work> er, I tend to stay away from the code decisions :)
09:18:10 <Xaroth|Work> hence why I'm lead manager :P
09:18:53 <Xaroth|Work> and officially the Dune brand is owned by multiple parties
09:19:07 <Xaroth|Work> and none of em wants to claim full ownership of it
09:28:17 <Alberth> why use ScummVM instead of eg Java?
09:28:31 <Xaroth> because java is shit, duh
09:29:45 <Alberth> Xaroth I know :p but if portability is your goal, java seems to be another option. I am interested in pros and cons of both solutions
09:47:02 *** Alberth has joined #openDune
09:47:02 *** Tomaz^W has joined #openDune
09:47:02 *** TigerP has joined #openDune
09:47:02 *** TrueBrain has joined #openDune
09:47:02 *** planetmaker has joined #openDune
09:47:02 *** TinoDidriksen has joined #openDune
09:47:02 *** Xaroth|Work has joined #openDune
09:47:02 *** charon.oftc.net sets mode: +vo planetmaker Xaroth|Work
09:47:02 *** ChanServ sets mode: +v DorpsGek
09:47:02 *** ChanServ sets mode: +v Xaroth
09:52:34 <Tomaz^W> as Xaroth said, JAva s*cks, and I knew some of the ScummVM devs so the choice wasn't so hard
09:55:28 <Tomaz^W> Xaroth: so what you're saying is, you have not solved the legal issues?
10:00:20 <Xaroth|Work> Tomaz^W: we did, only not through EA alone
10:00:33 <Xaroth|Work> there were multiple parties that we had to contact
10:01:39 <planetmaker> Xaroth: solved or tried to solve w/o anyone claiming ownership, thus nothing really resolved?
10:01:42 <Tomaz^W> May I ask who at EA you contacted?
10:01:51 <Xaroth|Work> Tomaz^W: dunno I didn't do the EA bit
10:02:22 <Xaroth|Work> planetmaker: well some of them said 'we don't care, do wtf you want'
10:02:40 <Xaroth|Work> Tomaz^W: the other lead of this team :)
10:02:52 <Tomaz^W> Which would be? TrueBrain?
10:02:57 <planetmaker> the brain of the project :-P
10:03:35 <planetmaker> btw, you're missing an op here, Xaroth;-)
10:03:41 <Tomaz^W> Se, as I said, I'm slowing crawling my way towards the EA mountain ;)
10:04:37 <Xaroth|Work> 12:04 -ChanServ(services@services.oftc.net)- User Xaroth|Work is already an operator on channel #openDune.
10:05:05 <Xaroth|Work> I suspect your client playing tricks on you, planetmaker :)
10:05:51 * planetmaker points at the person listed first in alphabetical order ;-)
10:06:24 <Tomaz^W> DorpsGek is oped too ;)
10:07:25 <Tomaz^W> So, Xaroth|Work, I assume you, just like me, have a computer related job, since your on IRC from work
10:07:26 <planetmaker> but 'D' is hardly the first letter of the alphabet ;-)
10:10:18 <Xaroth|Work> planetmaker: that person failed to identify to nickserv prior to joining the channel ;)
10:10:28 <Xaroth|Work> and yes, Tomaz^W, i'm a sysadmin
10:11:13 <Tomaz^W> neat, if i knew what a sysadmin does ;)
10:15:49 <Alberth> keeping the servers and the network up&running
10:16:38 <Alberth> upgrading software at machines, fight with printers to accept print jobs, that kind of stuff :p
10:27:24 <Tomaz> we call that network tecnichian where i work
10:42:03 <TrueBrain> appadmin - the one managing the applications
10:42:08 <TrueBrain> netadmin - the one managing the network
10:42:13 <TrueBrain> sysadmin - the one doing everything
10:45:17 <Tomaz> the guy I was looking for
10:47:35 <Alberth> I have to go now, not sure whether I am around again later today.
10:48:41 <TrueBrain> Tomaz: yeah, read your PM; bit busy atm :P
10:49:17 <Tomaz> will do once im back at work, lunch break :)
11:19:28 *** Tomaz^W has joined #openDune
11:25:14 <TrueBrain> no Tomaz, I read your PM; bit busy atm :P
11:34:36 <Tomaz^W> sorry, thought you meant I should read it and you had replied :P
12:13:33 <TrueBrain> the decompiler used for this project, I wrote myself
12:13:43 <TrueBrain> it is somewhat released
12:13:47 <TrueBrain> but I don't give any support on it
12:13:55 <TrueBrain> some people have tried to apply it to other projects, and failed :P
12:14:15 <TrueBrain> ToC started general, but focused more and more on Dune2
12:14:29 <TrueBrain> mostly what is missing is libemu (DOS layer) support
12:14:32 <TrueBrain> and some other stuff
12:14:51 <TrueBrain> feel free to try it
12:23:08 <TrueBrain> basically, we found out that the concept of ToC was wrong; it now depends on a JIT which compiles the code for real
12:23:12 <TrueBrain> it would have been enough to emulate it
12:23:20 <TrueBrain> but as it already works, I can't be arshed to change it
12:23:52 <Tomaz^W> ill fire up the laptop then which has ubuntu on it
12:24:46 <Tomaz^W> so no interest in turning it into a more general 16bit DOS disassembler?/decompiler?
12:25:02 <TrueBrain> I started it for this project
12:25:08 <TrueBrain> and I rewrote it 3 times for that
12:25:13 <TrueBrain> don't want to do a 4th atm :D
12:25:16 <TrueBrain> too many other kewl stuff to work on
12:25:20 <Tomaz^W> ok, shame, lots of projects out there that would benefit greatly from it
12:25:32 <Tomaz^W> like the 2 i work on, freeSynd and openThemePark
12:25:43 <TrueBrain> also the illegal part of the internet would, which makes me a bit reluctant tbh :)
12:26:15 <Tomaz^W> I'm not sure what illegal stuff you can do with something that decompiles 16bit DOS applications, tbh
12:26:38 <TrueBrain> in my countries reverse engineering is illegal if it is with the goal of publication
12:27:33 <Tomaz^W> We'll i guess I have my weekend setup for me, trying to compile and figure out how to use this :/
12:48:21 <Tomaz^W> TrueBrain: no chance of having you just try to dump the exe for themepark? just to see if it works or not?
12:48:35 <Tomaz^W> it should be a regular MZ exe
12:51:42 <Tomaz^W> Yeah I didn't mean this particular second ;)
12:52:09 <TrueBrain> yeah; bit busy for me means a few days :D
12:52:24 <TrueBrain> but if you pack it up including all datafiles and in a way it runs in DOSBox, I am sure one of us can try it soonish
12:52:43 <Tomaz^W> oh absolutely, thats easy because that's what I have it as ;)
13:48:59 *** fjb is now known as Guest4971
13:53:03 *** Alberth has joined #openDune
13:53:03 *** ChanServ sets mode: +o Alberth
14:19:39 <TrueBrain> Tomaz^W: if you can link us such package, I might be able to slip it true when I have 10 minutes or whatever :)
14:24:02 <TrueBrain> my typing goes horrible lately
14:33:42 <Tomaz^W> TrueBrain: WIll do once I get home, work day ends in 30 minutes
14:34:25 <Tomaz^W> Brainfucked entire week, got back home from LA on tuesday, total time from leaving hotel to back home was 30 hours
14:34:54 <Tomaz^W> about 165 on plane, 12h on airport and finally a 3h drive
14:43:34 <Xaroth|Work> TrueBrain: you're clearly not spending enough time with Petern.
14:45:43 <TrueBrain> Xaroth|Work: with reasons
14:59:39 <TrueBrain> lol, we only have a few emu_ files left, didn't know that :D
15:00:20 <TrueBrain> in fact ... only drivers left
15:01:01 <TrueBrain> that is just awesome :D
16:19:34 <DorpsGek> SVN: truebrain (r1792) -Add: make run-gdb
17:25:58 <TrueBrain> Tomaz: I might have time a bit later tonight :)
17:53:04 <TrueBrain> this is like LONG ago that I tried ToC .. lol
17:53:46 <TrueBrain> sorry Tomaz, it won't work :)
17:53:50 <TrueBrain> this application is not 16bit :)
17:54:10 <TrueBrain> (Well, the instruction-set is, but the memory isn't)
17:54:34 <TrueBrain> ToC can only decode the so-called Real Mode
17:54:44 <TrueBrain> this application is a Protected Mode application, which is vastly more difficult
17:54:57 <TrueBrain> many more instruction etc (the 80386 instruction set)
17:55:07 <TrueBrain> of which many are extended instructions (32bit)
17:56:27 <TrueBrain> it has WATCOM C support (the compiler for this)
18:17:47 <TrueBrain> glx: I wrote an easy and to-the-point implementation for Malloc / Free, and it works .. kinda :p At a certain point it throws an INB 5620 ... :(
18:20:00 <TrueBrain> comes from DSP_WriteAndWait
18:20:24 <TrueBrain> might be the slightly wrong binding of memory I guess ...
18:24:24 <TrueBrain> I kinda don't want to convert their code, although I know how they do memory manegement ..
18:24:28 <TrueBrain> guess I should not be so lazy ...
18:35:31 <Tomaz> I just checked the header that said MZ and figured it was a proper 16bit app
18:35:59 <TrueBrain> its a proper Protected Mode 16bit application with extended instruction set :D
18:36:46 <TrueBrain> dos4gw helps a lot with higher memory and stuff, if I am not mistaken
18:36:52 <TrueBrain> but it always runs in protected mode
18:37:07 <TrueBrain> I guess it is the beginning of modern OSes :P
18:37:15 <Tomaz> if the game/app has dos4gw then it is a protected app
18:37:28 <Tomaz> Syndicate is the same then
18:37:32 <TrueBrain> and it would be possible to extend ToC to run those apps
18:37:41 <TrueBrain> it would just have a lot of garbage from dos4gw and other stuff
18:37:45 <Tomaz> but its not in your interest
18:37:46 <TrueBrain> and requires a lot of love :D
18:37:57 <TrueBrain> not really; not now anyway
18:38:25 <Tomaz> I have huge portions of the game coded already
18:38:40 <Tomaz> it would just feel better to use the "correct" values for everything rather than guess
18:39:26 <TrueBrain> IDA is an application that can handle almost all executables
18:39:31 <Tomaz> didn't understand much of it tho
18:39:34 <TrueBrain> should give you correct values etc :)
18:39:56 <Tomaz> I'm an engine coder by profession, not reverse enginerrer
18:40:27 <Xaroth|Work> well it's just the reverse of what you're doing :P
18:40:33 <TrueBrain> assembly knowledge is all it takes ;) :D
18:41:03 <Tomaz> so much simpler to code DirectX11 in C++ !
18:41:14 <TrueBrain> ugh, windows ... bah
18:41:28 <Tomaz> well makign a game engine using OpenGL is also simpler
18:43:47 <TrueBrain> glx: I have this idea the audio drivers also do malloc
18:43:50 <TrueBrain> but I can't find the entry for that
18:44:32 <TrueBrain> they don't call 01F7 .. still ... I see minor offset differences
18:48:25 <glx> they don't malloc, they use buffers allocated by the game for them
18:55:17 <glx> MPU_GetDataSize() tells how much to allocate
19:15:02 <TrueBrain> something is really odd in memory allocation ...
19:35:45 *** Alberth has joined #openDune
19:35:45 *** ChanServ sets mode: +o Alberth
19:39:46 <TrueBrain> lol; the memory manager wastes on average 10 bytes per allocation :D
19:40:01 <TrueBrain> the header for a block is 4 bytes
19:40:05 <TrueBrain> and they want to be on a page
19:40:12 <TrueBrain> so instead they put those 4 bytes on a page
19:40:21 <TrueBrain> they just add 15 + 4 to the size, and then put that on a page
19:40:33 <TrueBrain> in result, often blocks are 1 page bigger than needed
19:41:10 <TrueBrain> for example, if the size you alloc is 10 byes, is allocates 2 pages (of 16 bytes)
19:41:15 <TrueBrain> where 1 page would have been plenty :D
19:43:42 <Alberth> that is to be expected when you add 15, right?
19:43:59 <TrueBrain> yes; it should have been + 11
19:44:18 <TrueBrain> hmm .. either one :P
19:46:14 * Alberth finds it very confusing :p
19:46:28 <TrueBrain> was wondering why I had slightly different values ;)
19:49:22 <TrueBrain> Alberth: owh, I think you miss the piece where I tell you it always rounds up :D
19:50:37 <Alberth> I assumed as much, as until now I have never managed to squash 17 bytes into a page of 16 bytes :D
19:50:56 <TrueBrain> a 10 byte block with 4 bytes of header information ends up on 2 pages
19:51:05 <TrueBrain> that is wasting 18 bytes, not 2 :)
19:51:25 <TrueBrain> glx: no, it is a 4 bit align
19:51:34 <TrueBrain> how ... well .. by wrong math :D
19:51:42 <TrueBrain> or maybe twice the safety
19:51:53 <TrueBrain> dunno .. I just spend 10 minutes figuring out THAT it happened :P
19:52:09 <Alberth> oh, the round the header to 16, and the data to 16 ?
19:52:23 <TrueBrain> no, round the data to 16, add the header, round to 16
19:52:39 <TrueBrain> hmm ... either I have random memory corruption, or something is wrong ... hmm ..
19:53:03 <TrueBrain> I made a linked-list of free pages, and it is damaged ...
19:53:20 <glx> wrong write somewhere is possible
19:53:29 <glx> especially in dune2 code :)
20:10:04 <TrueBrain> circular references in my free linked list
20:10:09 <DorpsGek> SVN: glx (r1793) -Fix: more emu_ax removal
20:23:28 <TrueBrain> I think I lost my thinking or something :D I can't even make simple linked lists ... *sigh*
20:23:35 <TrueBrain> okay ... lets rethink this ... I want an easier solution :D
20:28:51 <Alberth> a bitmap of free pages :p
20:34:14 <TrueBrain> k .. redid my initial implementation
20:34:18 <TrueBrain> this time knowing what is expected
20:35:47 <TrueBrain> src/gui/widget.c:151 assert :(
20:37:45 <TrueBrain> and it is really my patch
20:44:25 <TrueBrain> I start to wonder if it is luck that Trunk works tbh :D
20:44:29 <TrueBrain> keep hitting weird errors
20:45:53 <glx> you were drunk when you did it ?
20:46:06 <TrueBrain> owh ... I think it now works ...
20:46:25 <TrueBrain> you see subtitles at end screen?
20:46:52 <glx> voices are only for intro :)
20:47:27 <TrueBrain> run the intro, play the game, do strange things
20:53:19 <Alberth> intro does not do much with malloc, does it?
20:53:45 <TrueBrain> hmm .. btw, selecting a house no longer works for me (trunk)
20:53:52 <TrueBrain> when I click Harkonnen, the game seemly hangs
20:54:10 <TrueBrain> trace shows Voice_IsPlaying
20:55:07 <glx> I need to check something I noticed before
20:55:25 <glx> you here sounds from other side of the map
20:56:01 <TrueBrain> yeah, just double checked; selecting a house fails here (on voices)
20:56:28 <glx> intro, load, end game works
20:56:31 <TrueBrain> guess DSP_GetStatus never returns the right status
20:56:45 <TrueBrain> glx: also tried mentat and stuff?
20:57:21 <TrueBrain> wow, voices hang the game a lot
20:57:27 <TrueBrain> placing windtrap also freezes up the game
20:57:28 <glx> selecting house works for me
20:57:41 <TrueBrain> remember that my sound-card doesn't really work
20:58:30 <Alberth> house selection seems to work
20:58:35 <TrueBrain> again in IsPlaying and GetStatus
20:58:45 <TrueBrain> src/driver.c:675 and decompiled/cs__AB00.c:2180
20:58:50 <glx> Alberth: we force voice driver
20:59:00 <TrueBrain> but okay; any problems with my patch?
21:01:27 <TrueBrain> src/house.c:482 is an entrance for example
21:01:37 <TrueBrain> but okay, I hit ctrl+c glx
21:01:44 <TrueBrain> so it stops somewhere in that function at 2180
21:02:40 <glx> but yes if voice is not played, emu_get_memory16(emu_cs, 0x00, 0x14C) is not "updated"
21:03:30 <TrueBrain> so I guess we do need a detection for voices :D
21:03:43 <TrueBrain> Alberth: you found any problems? Or didn't you test? (both okay of course :D)
21:04:05 <Alberth> I was busy testing :)
21:04:18 <Alberth> but no problems found
21:05:12 <DorpsGek> SVN: truebrain (r1794) -Add: implemented my own simplified version of Malloc and Free in 16bit world. It is slow. It is bad. It works.
21:06:31 <Alberth> I was commenting that adding a uint16 * at the top of the while would result in cleaner code
21:07:10 <TrueBrain> hmm .. that would work now yes, but then I need to do << 3
21:07:14 <TrueBrain> which might be confusing .. hmm ..
21:07:20 <TrueBrain> you mean making buf uint16?
21:07:59 <DorpsGek> SVN: truebrain (r1795) -Fix (r1794): cleanup the emu_ tree a bit
21:08:43 <TrueBrain> using a struct would be best
21:08:50 <TrueBrain> but I am kinda lazy ... this will be gone sooner or later anyway :D
21:10:54 <Alberth> but that holds for all code, so that's no excuse imho :p
21:13:09 <TrueBrain> lol; don't hope we remove all code :D
21:13:19 <TrueBrain> but tbh, I was hoping this would remove most of 01F7 ..
21:18:43 <DorpsGek> SVN: truebrain (r1796) -Fix: remove a few emu_?? variable references
21:18:51 <TrueBrain> that fixes up most emu_ax, emu_si, ...
21:18:54 <TrueBrain> replaced it with locsi :D
21:19:16 <Alberth> glx: seems a nice step, except 'language' is not in that patch, it seems
21:20:31 <glx> was always a copy of config.language
21:21:18 <Alberth> ah, then it is a good idea by definition :p
21:24:23 <TrueBrain> (it works because there are not TSR running :P)
21:29:24 <DorpsGek> SVN: glx (r1797) -Fix: g_global->language is always g_globlal->config.language, so use that instead g_globlal->language
21:35:53 <TrueBrain> typical dune: we start off with a out-of-bound check, and then we just ignore it and read some random piece of memory :D
21:43:01 <glx> that's when there's a check
21:49:19 <DorpsGek> SVN: truebrain (r1798) -Add: initialize the video driver directly; don't go via 10 layers of useless stuff
21:51:13 <DorpsGek> SVN: truebrain (r1799) -Fix: removing 1 of the 2 patched emulator functions
21:52:21 <TrueBrain> okay ... next step ... manually setting the interrupts
21:53:16 <TrueBrain> as I can now skip the whole 01F7, but the mouse acts weird :D
21:56:29 <TrueBrain> hmm .. I forgot how interrupts are linked in .. hmm ..
21:58:08 <glx> the game uses 2 timer interrupts
21:58:16 <TrueBrain> where does it write doen Input_Keyboard_EventHandler for example
21:58:23 <TrueBrain> it is mapped at 29E8:0D47
21:58:27 <TrueBrain> but when does it install that ...
21:59:22 <glx> emu_get_memory16(emu_cs, 0x00, 0x258) = emu_bx;
21:59:22 <glx> emu_get_memory16(emu_cs, 0x00, 0x25A) = emu_es;
21:59:23 <glx> emu_pushf(); emu_flags.inf = 0; emu_push(emu_cs); emu_cs = 0x0070; emu_push(0x09F8); Interrupt_DOS();
21:59:35 <TrueBrain> I was looking for 0D47 :D
22:00:47 <glx> I did that first (then I used my brain ;) )
22:00:49 <TrueBrain> k, timer interrupts seem to go correct
22:01:14 <TrueBrain> WSA animation works
22:01:59 <TrueBrain> well, tbh, everything works fine
22:02:03 <TrueBrain> only the mouse feels weird
22:02:10 <glx> animations are managed via Game_Timer_Interrupt IIRC
22:02:19 <TrueBrain> acts up when leaving the screen only
22:02:58 <TrueBrain> can you see what windows does?
22:04:35 <TrueBrain> dunno exactly why the mouse acts up .. guess something is not send or something
22:04:39 <glx> windows warns at compile ;)
22:04:58 <TrueBrain> yeah, so does linux
22:05:51 <TrueBrain> owh, I think it is missing a / 2
22:05:55 <TrueBrain> just a guess, but I tihnk that is it
22:06:46 <TrueBrain> lets convert Mouse_Init then :D
22:13:22 <Alberth> iirc, if not double-width: x /= 2 :)
22:14:40 <glx> no x is always *2 for DOS
22:17:21 <DorpsGek> SVN: truebrain (r1800) -Add: C-ified Mouse_Init()
22:21:45 <DorpsGek> SVN: truebrain (r1801) -Add: no longer initialize the game via 01F7, but go directly to Main(). This is huge progress for this project
22:23:35 <Alberth> we are finished now :)
22:30:23 <glx> I think a lot of 01F7 can be removed
22:31:21 <DorpsGek> SVN: truebrain (r1802) -Add: remove code that installs the overlay interrupt. It is long gone.
22:35:25 <DorpsGek> SVN: truebrain (r1803) -Add: removed last patch function. Also: goodbye 01F7. You have always been here. I won't miss you.
22:38:07 <glx> let me remove another file
22:38:08 <TrueBrain> Functions done: 1157 / 1229 (94%)
22:38:09 <TrueBrain> Functions named: 1178 / 1229 (95%)
22:38:51 <DorpsGek> SVN: truebrain (r1804) -Add: C-ified Mouse_CallbackClear()
22:39:52 <glx> it made me think 2BEE was uneeded
22:42:28 <DorpsGek> SVN: truebrain (r1805) -Fix: we don't need a vsync check. We don't care.
22:42:29 <glx> even if 0x3DA is totally faked :)
22:42:33 <TrueBrain> in a sense he was right btw :P
22:43:57 <DorpsGek> SVN: truebrain (r1806) -Fix: more cleaning up
22:44:32 <DorpsGek> SVN: truebrain (r1807) -Fix: more useless externs
22:44:58 <TrueBrain> VERY little emu_ left
22:45:11 <TrueBrain> not sure what to do with WaitForNextVsync
22:45:14 <TrueBrain> I think a sleep() is enough
22:47:15 <TrueBrain> I have no clue when that is called :D
22:47:29 <glx> I know that part of code :)
22:47:37 <TrueBrain> so yeah, replace it wiith a sleep(0), run it; if it works, remove 2BEE :D
22:48:10 <TrueBrain> now the hard part starts ... removing libemu :D
22:51:24 <glx> seems to work correctly without sleep
23:01:11 <DorpsGek> SVN: glx (r1808) -Fix: no need to wait for next vsync. Bye 2BEE.
continue to next day ⏵