IRC logs for #opendune on OFTC at 2009-10-26
⏴ go to previous day
08:57:59 *** boekabart has joined #openDune
09:03:31 <TrueBrain> I have applied a few of your patches, but not all
09:03:46 <TrueBrain> mostly because I could not test it :p (the applications I have never calls them :))
09:03:54 <TrueBrain> also, I think I fixed a few problems in your FCB handling :p
09:04:02 <TrueBrain> as now it shows FirePower at least :p
09:04:12 <TrueBrain> and your EGA driver is a mystery to me, as it doesn't work here :p
09:04:22 <boekabart> yeah, the screen didn't show - but that was also a EGA problem wasn't it?
09:04:52 <boekabart> Then I don't get how the ega video memory is mapped
09:05:07 <boekabart> i thought, every byte is 2 pixels, but it's not that easy apparently?
09:05:13 <TrueBrain> what I understood last time, is that every clocktick another colour is in the memory
09:05:37 <TrueBrain> when I try your patch, or a modified one of myself under that assumption, only the first 50 pixels (in height) have some kind of image
09:05:45 <TrueBrain> which might or might not hold the complete image ... I can't see/tell ;)
09:05:52 <boekabart> maybe it's paged or so
09:06:00 <boekabart> 4 pages make 4 bits per pixel
09:06:00 <TrueBrain> what I understood, it is
09:06:34 <TrueBrain> and Digger doesn't do that for me .. I get digger in CLI mode (no graphical mode), and I can't make anything out of it :p
09:06:53 <boekabart> it's not gfx because it doesn't use int10 to switch, but port 3d8
09:07:02 <boekabart> didn't I include that patch?
09:07:26 <TrueBrain> double checked, nope
09:07:40 <boekabart> it was the last thing I did, so that's possible
09:07:44 <TrueBrain> oh, and please pretty please, make your code a bit more like the rest? :$ :)
09:08:23 <TrueBrain> either way, in general, nice job :) Always good to have more more MORE! :)
09:08:49 <boekabart> my father-in-law is visiting for the next month so I might have more free time ...
09:09:26 <boekabart> well every voordeel has it's nadeel, of course, but still
09:10:42 <TrueBrain> ah, I remember about FCB: you had: fread(buf, 1, length, fp)
09:10:47 <TrueBrain> now 'res' is either 1 or 0
09:11:01 <TrueBrain> fread(buf, length, 1, fp) returns between 0 and length
09:11:17 <TrueBrain> so INT21:14 never returned SUCCESS, but always PARTLY :)
09:11:41 <TrueBrain> I noticed while typing over your patch, as I had the same problem with the other fread :p
09:22:06 <boekabart> so long ago i've used fread...
09:22:19 <boekabart> actually never understood I think :)
09:22:23 <TrueBrain> it is just a silly command, with his 2 params ..
09:22:31 <TrueBrain> but okay ... gimme your 3d8 patch! :p
09:22:53 <boekabart> i have to see if it's done
09:23:01 <boekabart> (although nothing ever is, right)
09:30:14 <Xaroth> TrueBrain: ignoring some of the annoying tile functions, I might have a patch ready for you to check tonight
09:34:15 <TrueBrain> video modes are so darn complex, it aint fun!
09:34:27 <TrueBrain> btw, FirePower fails to continue on keyboard handling .. dunno why :p
09:34:38 <TrueBrain> it registers the keypress, but ignores it somewhere
09:37:02 <TrueBrain> either the wrong scancode, or it is an application which really does assume he gets an extended code :p
09:37:19 <TrueBrain> (I never send extended codes :p It is kind of nasty ;) Also, keyup is sent directly after keydown :p)
09:37:21 <boekabart> anyway the kb handling seems like a big hack now
09:37:25 <TrueBrain> the shortcuts I took ....
09:37:45 <TrueBrain> in SDL you can fix keyboard pretty much .. but in ncurses, it is hard!
09:37:51 <boekabart> maybe one of us (*ugh* me probably) should invest a bit in writing it more like a real keyboard controller
09:37:52 <TrueBrain> but I guess you don't care about ncruses :p
09:38:05 <TrueBrain> way outside my experience box :)
09:38:12 <TrueBrain> boekabart: a CLI version of SDL, so to say :p
09:38:21 <TrueBrain> it handles CLI applications, easy drawing, easy colouring
09:38:25 <TrueBrain> for xterm linux applications
09:38:35 <TrueBrain> one suggested to me that I should do that in SDL, but that is also very hard :p
09:38:44 <boekabart> ah, the .Net 'Console' class :)
09:39:08 <TrueBrain> but I guess it is the only real suggestion, to make it in SDL .. removes the whole ncurses shit
09:39:20 <boekabart> but first, back to work - doing a side project which puts CSS, JavaScript on my resume :)
09:39:25 <TrueBrain> I use Snipes as test-app for CLI :p
09:39:59 <boekabart> Luckily, only 1 browser to support (custom browser on a tv settopbox)
09:40:41 <TrueBrain> I just finished making a website workable on all mainstream browsers .. IE6 was a bitch, as always
09:40:41 <boekabart> very: and the stb simulator has a nice debug console so I can put debuggin lines in my JS code..
09:41:01 <TrueBrain> and javascript is a very nice language, so you should do okay :)
09:41:04 <boekabart> plus, it compiles all JS code in advance so typo's are caught early
09:41:26 <TrueBrain> I need to start working on face-gesture-registration
09:41:32 <boekabart> I don't like the typelessness in it (same with py), i WANT to specify what a fn return and what it eats
09:41:50 <TrueBrain> I have the same problem .. but after a while you forget about that :p
09:41:56 <TrueBrain> then you get lazy :p
09:42:45 <boekabart> you don't happen to know a good (windows) editor for such a project? (notepad++ doesn't really have stuff like 'go to definition' and so)
09:43:12 <boekabart> (and don't say eclipse. too bloated)
09:43:51 <TrueBrain> I use netbeans lately
09:43:58 <TrueBrain> but it does its job, and it does it good
09:45:54 <boekabart> msvc ? for a JS project?
09:46:04 <boekabart> how do I set that up
09:46:05 <Xaroth> er, there might be a plugin for it
09:46:07 <TrueBrain> it is slow when loading in big projects, but in return it known a lot of languages, gives a lot of in-edit suggestions, etc etc
09:46:31 <TrueBrain> (netbeans tells me there will be a compile / runtime error before I even run it :p)
09:46:36 <boekabart> I like msvc, just didn't know it would help me with js
09:47:18 <boekabart> orcas is 2008 right
09:47:32 <Xaroth> part of asp.net tho :/
09:48:06 <TrueBrain> asp net ... who ever thought that would be a good idea?
09:48:10 <boekabart> actually you meant MSVS , not VC, of course :)
09:48:19 <TrueBrain> well, in all fairness, ASP.net just made one big flaw: on error resume next ;)
09:48:34 <Xaroth> VC is part of visual studio nowadays anyhow :P
09:48:36 <boekabart> TrueBrain sounds like basic!!
09:48:50 <boekabart> cool, auto try-catch-ignore around ever line of code :)
09:49:10 <TrueBrain> boekabart: you never read that before? In the early days (read: up to last year) you could crash any host running ASP.net by uploading a asp file with: on error resume next\ni = 1 / 0
09:49:43 <boekabart> I HAVE read (and written!) that before, in my quickbasic days... just never after I learned about exceptions
09:49:49 <boekabart> now I see what that actually does...
09:50:07 <TrueBrain> it is one of the worst thing you can do in a shared env .. allowing such statements
09:50:09 <Xaroth> once i realised how wrong i was taught, I dropped vb/qb/asp like a brick :P
09:50:12 <TrueBrain> that is just asking for shit
09:50:57 <TrueBrain> Xaroth: btw, I reconsidered, I really think we should not add any src/ things till the 1st of nov
09:51:00 <boekabart> netbeans Java SE is what I want?
09:51:07 <TrueBrain> boekabart: I guess :p
09:51:41 <TrueBrain> boekabart: it doesn't really matter which one you install, after install, it will automaticly add plugins it requires :)
09:52:21 <TrueBrain> Xaroth: anyway, that way, release 0.1 doesn't contain any possible mistake we make in things like tile functions or what ever :) Then after that we have some more time to test/stabilize such changes
09:52:35 <TrueBrain> (and that is the reason for the 2 week-before-release-no-new-shit stuff :p)
09:54:27 <Xaroth> I take it with the changes you made wrt the goto structuring there's no more stack issue for *dows?
09:54:50 <TrueBrain> btw, I will make a list of known problems in the next few days :)
09:55:23 <Xaroth> more/less compared to src/ ?
09:55:30 <boekabart> TrueBrain did you get this one: Fixed port 0x201 readbyte to return 0x00 (joystick port)
09:55:45 <TrueBrain> I have no patch related to ports
09:56:20 <Xaroth> the thing is, you want to not release with src/ to improve stability, which is counterproductive if the src/ additions actually make it more stable :P
09:56:36 <TrueBrain> Xaroth: then you missed something in the progress we made :)
09:56:44 <TrueBrain> the only required 'patch' in src is src/patched.[ch]
09:56:51 <Xaroth> that's what I asked :P
09:56:51 <TrueBrain> the rest is all optional, and doesn't improve the game as such
09:56:59 <Xaroth> how does stability compare with/without src/ :)
09:57:07 <TrueBrain> (only makes it worse inf act .. you can't load one of my savegames with it, because it cnotains a bug :p) (the savegame, that is)
09:57:30 <Xaroth> if it's negligable, I fully agree with your idea :)
09:57:31 <TrueBrain> [10:52] <TrueBrain> Xaroth: anyway, that way, release 0.1 doesn't contain any possible mistake we make in things like tile functions or what ever :) Then after that we have some more time to test/stabilize such changes <- clearly, without src/ it is more stable :p
09:57:47 <TrueBrain> in making things C, we can only make things less stable :)
09:59:13 <TrueBrain> take your tile stuff for example; if you wouldn't have tested it so well, it would have been a nasty thing to have for your first release :p
09:59:19 <boekabart> doesn't support the 'disable video' bit yet
10:00:20 <TrueBrain> hmm .. I fail to parse your comment :p
10:00:29 <TrueBrain> / bit 0 to 0 is: spk follows bit 1; bit 0 to 1 is: spk follows PIT channel 2 if bit 1 is 1 and unconnected if bit1 = 0
10:01:00 <boekabart> ok i've written better ones :)
10:01:16 <TrueBrain> and please please please, adapt to the coding style already there :)
10:01:21 <TrueBrain> I don't want to duplicate DOSBox :p
10:01:37 <boekabart> WAHT? no pc speaker support!?
10:02:06 <boekabart> i'd ignore the 61 stuff for now
10:02:29 <TrueBrain> I would love PC Speaker support :p
10:02:58 <boekabart> to drive the actual PC speaker, or to emulate it?
10:03:12 <TrueBrain> emulate, if possible ;)
10:03:20 <TrueBrain> (given I don't have a PC Speaker :p)
10:03:21 <boekabart> just write a linux + win32 driver that allows access to port 61 :)
10:09:00 <TrueBrain> wow, Digger works ...
10:09:09 <TrueBrain> well .. it has to initialize a lot of screens before it keeps one :p
10:09:26 <TrueBrain> and keyboard input fails
10:09:33 <TrueBrain> but ... it works :) I am impressed :)
10:09:43 <TrueBrain> again failure in keyboard driver :)
10:10:22 <boekabart> yea, can't get it to 'start'
10:10:43 <TrueBrain> uploaded new version
10:11:27 <TrueBrain> k ... soon I will split LibEMU to its own project, as this is getting annoying :p
10:11:33 <TrueBrain> (also to sync it back with OpenDUNE and stuff)
10:11:43 <TrueBrain> and it needs a good cleanup and splitting of some code, as it becomes messy!! :p
10:12:03 <boekabart> it calls 3d8 twice with 0x02, then twice with 0x0A : 0x2 == 0xA but with screen enabled. Maybe GFX_init should ignore re-inits with same value ...
10:12:10 <TrueBrain> but okay .. time time time time .. now first: university shit :) bbl later today
10:12:22 <TrueBrain> sounds like a plan :)
10:13:04 <boekabart> and add a '_gfx_blanked' boolean, and blit 'memset 0' instead of actual 0xa000 pixels :)
10:14:40 <boekabart> well 3d8 allows to setup a resulotion, but '0' the output
10:14:52 <TrueBrain> it already does that :p
10:15:02 <TrueBrain> mostly as there is nothing to show :p
10:15:10 <TrueBrain> oh well .. 15 minutes late, I think that is a nice time to go :)
10:15:36 <TrueBrain> nuh, me being annoyed with my lab-partner
10:16:15 <TrueBrain> either way, nice job on LibEMU so far, keep on going ;) :)
10:16:34 <TrueBrain> I really need my own domain or something for such projects .. I have nowhere to put them .. hmm ...
10:16:42 <TrueBrain> oh well, I will figure out something :)
10:16:45 <TrueBrain> now for real: bye :)
10:37:04 <TB2> point vor boekabart , he indeed is nowhere to be found :p
10:37:35 <boekabart> he left because you were late - just like last week
10:37:47 <boekabart> but you really do live close to the university
10:37:54 <boekabart> OR have a really fast car
10:38:05 <TB2> last week was insane .. I was here 12:30 .. he texted me 13:00 he would be here 14:00 .. I left 14:30, he texted me 15:00 that he meant 15:00 .. yeah .. sure ...
10:38:22 <boekabart> I recognize that :)
10:38:53 * boekabart has 2 x 7km to go by bike every day
10:39:15 <TB2> I believe my ride is like 3km or what ever
10:39:34 <TB2> bah, the normal room I use is booked
10:39:40 <TB2> and I have a webcam application to write
10:39:44 <TB2> not something you want to test in a public room
10:41:37 <TB2> I had a nice bikeride to think about a few things related to LibEMU
10:41:57 <TB2> I will split it up a lot, the ToC and LibEMU stuff, in a DOS section and others
10:42:05 <TB2> so later on some insane person can also add emulation code for any other type of OS :p
10:42:27 <TB2> then there is the problem that DOS runs on COFF binary format, and on 80386 instruction-set
10:42:40 <TB2> so that are 3 layers I need to add .. so we can support ELF too, and more modern instruction-sets
10:42:45 <TB2> (protected mode maybe some day :p)
10:43:48 <boekabart> ... let's write our own extender ....
10:44:17 <boekabart> WHY won't digger.vcproj link in libemu
10:44:27 <boekabart> I mean, it does, but finds no symbols
10:44:29 <TB2> copy over the OpenDUNE one :p
10:44:42 <boekabart> Error 1 error LNK2001: unresolved external symbol _emu_di cs__01E7.obj digger
10:45:09 <boekabart> is the HG libemu proejct different than the SVN one?
10:45:58 <TB2> they are (minus the latest update) the same
10:46:20 <TB2> but I know nothing about MSVC, so I can;t help you there :(
10:46:34 <TB2> I now need to think how I am going to follow a face and deduce gestures from it
10:46:44 <boekabart> no, with the vcproj from SVN it works
11:01:38 <TB2> sigh, I can't work here ... so much noise and people ...
11:01:41 <boekabart> TB2: digger just doesn't get to the routine where it reads keyboard input
11:02:16 <boekabart> it really does its own kb handling: reads 0x60 (scancode), writes bit 4 to 0x61 and resets bit 4 on 0x61
11:02:21 <TB2> boekabart: when you press a key you see INB:20 and 60/61/62 (cant remember which)
11:02:39 <TB2> dont I support that in LibEMU yet?
11:02:41 <boekabart> what i've just compiled, doesn't
11:02:43 <TB2> I thought I added code for that ...
11:03:00 <boekabart> but it doesn't call the routine where this is called
11:03:17 <TB2> when I ran it this morning, it showed me INB:20, and 60/61
11:03:30 <TB2> every key-press (in between the mess of INT08 prints)
11:03:40 <boekabart> it's reffed only in this line: " emu_hard_link(0x01E7, 0x0E48, &f__01E7_0E48_0022_EF68); "
11:03:56 <TB2> a hard link means that the libemu can call that function by pointing to the cs:ip
11:04:07 <TB2> for example, when an interrupt is taken over, this is used
11:04:21 <boekabart> i suppose it does take over the kb interrupt
11:04:31 <TB2> libemu calls int(9) or what ever, which points to 1E7:E48 .. this allows the right fucntion to be called :)
11:04:46 <TB2> (in a static version you of course need a static way in :p)
11:04:51 <boekabart> how do i find the int-override-programming code?
11:05:00 <TB2> the application does that
11:05:08 <TB2> it writes either in the Interrupt Vector Table (IVT)
11:05:13 <TB2> or it uses an INT21:?? for it
11:05:43 <TB2> I am pretty sure that if you set a breakpoint in that function, it is called every time you press a key
11:05:46 <boekabart> no int21 is every called in digger ...
11:05:47 <TB2> else MSVC fucked up something :p
11:05:56 <TB2> then it directly writes in the IVT :)
11:06:35 <TB2> Dune2 uses the INT21:SET VECTOR and writes in the IVT directly, depending on the interrupt he wants to take over
11:06:38 <TB2> a bit silly .. and annoying :)
11:06:42 <TB2> took me a while to find ;)
11:08:25 <boekabart> the whole SDL window isn't reacting at all
11:08:41 <boekabart> it's drawing but can't move it
11:08:54 <boekabart> can't bring it back up except by hiding all windows on top of it ...
11:09:03 *** mib_bjp8sa has joined #openDune
11:09:08 <boekabart> its windows loop seems to be either not running OR stuck
11:13:37 <boekabart> so if the winow can't get focus, no keys
11:26:54 <TrueBrain> got sick of the bad connection, and went home :p
11:27:17 <boekabart> TrueBrain SDL window doesn't really react here
11:27:23 <TrueBrain> boekabart: we had that problem before with MSVC .. try other combo of debug/release
11:27:27 <boekabart> and I think this is because the SDL_Init is done in the main thread
11:27:50 <boekabart> but (on win32) the update (SDL_Poll) is done in the PIC thread
11:27:52 <TrueBrain> SDL_Init is not done in the main thread, that took us a while to get correct ;)
11:28:10 <boekabart> digger still does it in main thread
11:29:59 <TrueBrain> SDL_Init is done in int10.c, inside the PIC thread
11:30:08 <TrueBrain> oh .. your latest patch of course does not do that
11:30:37 <boekabart> ok let me look at that
11:31:00 <TrueBrain> boekabart: gimme 1 sec to solve it :p
11:31:29 <boekabart> emu_get_memory8(BIOS_MEMORY_PAGE, 0, BIOS_VIDEO_MODE) = ...
11:34:05 <boekabart> hm ok: this is in a decompiled file:
11:34:19 <TrueBrain> and LOTS of them :)
11:34:38 <TrueBrain> no worries, any sane compiler removes them immediatly
11:34:42 <TrueBrain> and if not, no harm done
11:35:14 <TrueBrain> if I would get paid for the decompiler, that would have been fixed long ago :p
11:35:22 <TrueBrain> now I simply don't care enough to fix it :)
11:36:29 <boekabart> app doesn't ever receive the BREAK key code?
11:36:52 <boekabart> maybe that's the problem in case of digger..
11:37:05 <TrueBrain> why? Why would you give a break key code?
11:37:15 <TrueBrain> CTRL+BREAK? CTRL+C? Why would you ever give that to an app :p
11:37:33 <boekabart> scancode <0x80, scancode >0x80
11:37:46 <boekabart> that's what subsequent reads to 0x60 should do, basically
11:38:05 <TrueBrain> if you mean extended scancodes, as I said, I don't give them
11:38:10 <TrueBrain> so far no app cared enough to complain :p
11:38:33 <TrueBrain> then I don't follow what you try to say :)
11:38:38 <boekabart> keydown should put scancode on port 60 and fire int9, keyup should put scancode | 0x80, and fire int 9
11:39:21 <TrueBrain> just where ever there is a keydown, there is immediatly after it a keyup
11:40:12 <boekabart> so, why twice the if (emu_flags.inf) emu_hard_int(0x9);
11:40:53 <DorpsGek> SVN: truebrain (r421) -Fix: somehow HouseID 13 is loaded in campaign 9 savegames
11:40:54 <boekabart> in the end it will be called 4 times for each key down+up
11:42:28 <TrueBrain> I believe it was needed to make Dune2 work :p
11:44:30 <boekabart> if not, it exist immediately?
11:44:42 <boekabart> that's what happens here on keypress
11:44:47 <boekabart> if i comment one of them
11:45:04 <boekabart> I have this book somewhere... but can't find it
11:45:16 <DorpsGek> SVN: truebrain (r422) [LibEMU] -Fix: remove an ugly temporary hack
11:45:55 <DorpsGek> SVN: truebrain (r423) [JIT] -Add: mapped another 2 functions (removed hack in LibEMU requires it)
11:46:30 <DorpsGek> SVN: truebrain (r424) -Update (r423): update decompiled code
11:47:07 <TrueBrain> there you go boekabart
11:47:10 <TrueBrain> does Digger work now?
11:47:45 <TrueBrain> boekabart: my interest is not really in finishing libEMU .. I wanted it to be good enough to do a few apps, but I don't enjoy hardware enough to finish it :p
11:48:20 <boekabart> TrueBrain: svn doesn't have 3d8 yet?
11:48:30 <TrueBrain> the OpenDUNE one, no
11:48:36 <TrueBrain> there are 2 versions now :p
11:48:39 <TrueBrain> a bit annoying, I know :)
11:49:36 <boekabart> Now,. digger quits on keypress
11:49:48 <DorpsGek> SVN: truebrain (r425) [LibEMU] -Update: update LibEMU to latest version
11:49:53 <TrueBrain> there, now they are in sync
11:49:59 <boekabart> ah! need more recompilation
11:50:30 <TrueBrain> it just is too fast :p
11:50:50 <TrueBrain> it really works ... lol :)
11:50:53 <TrueBrain> totally unexpected :)
11:51:07 <boekabart> can you do the --static and send me the stuff, my VM isn't up atm
11:51:24 <boekabart> the decompiled and main.c, that is
11:51:25 <TrueBrain> firepower works too
11:51:30 <TrueBrain> forgetting the EGA output :p
11:51:53 * boekabart will figure it out if you can now help him get the stuff he can't do on windows
11:52:27 <boekabart> but I guess I really have to play the games on linux, right, in order to fully decompile them
11:52:58 <TrueBrain> I have been thinking it would be possible to add an emulator mode in it
11:53:08 <TrueBrain> but ... pff .. lot of work :p
11:53:20 <boekabart> so, what's the showstopper for toc on win32?
11:54:23 <boekabart> "tcc-0.9.25-win32-bin.zip Windows binary distribution " isn't what we're looking for?
11:54:45 <TrueBrain> but .. it doesn't work on windows
11:54:52 <TrueBrain> for the same reason it doesn't work on 64bit
11:55:08 <TrueBrain> (no real support :p)
11:55:46 <boekabart> digger now starts for 1 seconds, then similar error
11:56:20 <TrueBrain> you can send me the crash-log, or you can just wait for your VM to boot up again :p
11:56:31 <boekabart> i will do the latter, won't take up more of your time
11:56:41 <TrueBrain> that really is no problem
11:56:43 <boekabart> I can also download a VM again and set it up
11:56:55 <TrueBrain> you removed it?! :o
11:57:03 <boekabart> no, i installed it on my server
11:57:20 <boekabart> which one would you advise, last time i used fedora9
11:57:33 <TrueBrain> you need a GUI, so f9 is fine
11:57:58 <TrueBrain> okay, what SHOULD be done with LibEMU and ToC and stuff, is a complete rewrite
11:58:02 <TrueBrain> much more splitting of stuff
11:58:11 <TrueBrain> as in theory it is no problem to make a DOSBox like emulator
11:58:20 <TrueBrain> have a JIT like thing like we do now (MUCH faster)
11:58:22 <boekabart> yes, but that's there already
11:58:27 <TrueBrain> have a static compile stage
11:58:34 <TrueBrain> clear flow of AST and optimization
11:58:45 <TrueBrain> well, the difference is, that this emulator would also generate stats for static compilation :p
11:59:12 <TrueBrain> the reason I didn't picked the emulation approach, was because I was too afraid people would say: oh, but it is just DOSBox!
11:59:18 <TrueBrain> while it is much more :p
11:59:54 <TrueBrain> anyway, estimated time to do that: 120 hours :p
12:00:25 <TrueBrain> the app currently is 10k lines big :p
12:00:37 <TrueBrain> SLOC tells me it should have taken 8 months :p
12:00:58 <Xaroth> SLOC's time estimations are a bit off :P
12:01:17 <TrueBrain> it just assumes the code is more .. managable afterwards :p
12:03:27 <TrueBrain> too many fun projects
12:03:59 <boekabart> enjoy the amount of time you have now...
12:04:05 <boekabart> it will be less, probably, later
12:04:13 <TrueBrain> and much later it would get more again
12:04:19 <TrueBrain> but by then, this is of no relevance ;)
12:04:39 <TrueBrain> I should find a company which pays me for ToC and LibEMU :p
12:04:56 <TrueBrain> solves all the problems!
12:06:21 <DorpsGek> SVN: truebrain (r426) [LibEMU] -Add: allow via a dirty way to set caption name of window
12:06:39 <DorpsGek> SVN: truebrain (r427) -Add: set caption name of window
12:06:48 <TrueBrain> this is dirty, but it is something that needs to be done :p
12:08:18 <boekabart> TrueBrain: thinking of that... there HAVE to be companies still relying on old DOS software for some complex business process...
12:08:37 <boekabart> ...for which support has long since expired...
12:09:20 <boekabart> to be able to run that software on modern *and any* hardware, and even get the options for improvements...
12:10:29 <TrueBrain> Xaroth: I added known bugs to your post
12:10:56 <TrueBrain> boekabart: find them! I will be happy to make this application better if I would get paid for it :)
12:11:06 <TrueBrain> now I have to time-share with my daytime job :p
12:11:15 <TrueBrain> (next to my college, of course)
12:12:17 <boekabart> you have a daytime job?
12:12:23 <boekabart> haha warning C4099: 'ToC' : type name first seen using 'struct' now seen using 'class' d:\sandbox\toc\src\libtoc\node.hpp 210 toc
12:12:43 <TrueBrain> yup; I work for a IT company, doing all kinds of shit
12:16:00 <boekabart> what a detailed description!
12:19:07 <TrueBrain> lately I am deploying XenServers, building a few websites, advising on technical questions, running helpdesk, launching digital products, creating a full CMS system, creating deploy-software, ... (and that is what I did in the last 2 months :p)
12:19:14 <TrueBrain> so much more detail I can't give :p
12:19:50 <TrueBrain> boekabart: about your MSVC error: 'struct' and 'class' are the same, but 'struct' is only known in C, so I tend to use 'struct' in headers, and 'classes' in C++ files .. sadly, MSVC is the only compiler who bitches about that
12:22:08 <boekabart> well there's the obvious difference of the default accessibility - maybe that's why they added it, just to notify you of an inconsistency?
12:22:25 <TrueBrain> possible .. but how does a forward declare care about that?
12:22:49 <TrueBrain> that is something for the implementation to care about :)
12:26:26 <TrueBrain> I am very tempted to start a project with CouchDB .. just because I can :p
12:26:34 <TrueBrain> I want to know if it is really as fast as they claim :p
12:27:20 <boekabart> erlang is cool anyway
12:27:35 <TrueBrain> at least for once I see erlang used for what it is meant :)
12:28:10 <TrueBrain> but okay, I stil lhave WebTranslator to finish, which is a bit of a bitch, as it is very slow in data-access (large large large database)
12:28:41 <boekabart> maybe use couchdb for it :)
12:28:49 <TrueBrain> that is what I am wondering about
12:28:55 <TrueBrain> if it would improve the speed, or make it even worse
12:29:02 <TrueBrain> sadly enough, CouchDB only runs on Apache
12:41:12 <TrueBrain> okay, I am bored .. I should be doing something I guess
12:41:26 <TrueBrain> I can't do anything real on OpenDUNE, as I want a stable release on the 1st ..
12:41:47 <TrueBrain> I guess I can work a bit on WebTranslator again ..
12:51:00 <Xaroth> that you saved me the trouble of figuring out the known bugs ^^
12:51:52 <Xaroth> < TrueBrain> I can't do anything real on OpenDUNE, as I want a stable release on the 1st .. << that only means you can't commit stuff :P
12:52:27 <TrueBrain> Xaroth: I use commits to keep track of my development for a reason :)
12:52:33 <TrueBrain> without it you make mistakes which you can't revert
12:53:41 <TrueBrain> okay ... lets try this CouchDB :p
12:53:52 <TrueBrain> as it is almost exactly what WT needs :)
12:54:03 <Xaroth> fibble advertised for u :p
12:54:04 <TrueBrain> just I wonder about the speed ...
12:54:20 <Xaroth> that's why I call him fibble :P
12:54:36 <Xaroth> i'd call him fib if i was really lazy :P
12:55:15 <TrueBrain> GOOOOODD morning glx :)
12:56:33 * Xaroth can't wait til sunday :P
12:56:44 <TrueBrain> why are you so happy to get older?
12:57:06 <Xaroth> You're getting older every day
12:57:26 <Xaroth> but I do like the presents you get :)
12:58:16 <TrueBrain> how old are you? 15?
12:58:40 <TrueBrain> oh, was that too mean?
12:58:56 <glx> websvn is really not user friendly when you want to check latest diffs
12:59:01 <TrueBrain> I won't go any higher, so just say when we have to stop this game :p
12:59:06 <TrueBrain> glx: nope, not a bit
12:59:20 <Xaroth> TrueBrain: 23 for now.
12:59:27 <boekabart> the only thing good about getting older is that you make more money every year :)
12:59:48 <Xaroth> I actually went back 5% when I started here
13:00:04 <Xaroth> but that was for my trial period only
13:00:12 <boekabart> exception that proves the rule
13:00:16 <boekabart> plus, less time to spend it so more money per free hour!!
13:01:37 <boekabart> who do I have to f**k to get javascript support in netbeans!!
13:01:43 <boekabart> 'it's supported in base IDE'
13:01:49 <TrueBrain> boekabart: pick me pick me picke me
13:02:00 <TrueBrain> well, last time I opened a .js, it just worked :)
13:02:57 <TrueBrain> "Apache CouchDB has started, time to relax." <- I love it already
13:05:36 <glx> hmm maybe the same dirty way (r426) could be used for mouse disabling too
13:06:03 <TrueBrain> glx: it works fine now too, doesn't it? ;) :p
13:06:34 <TrueBrain> I really hope that we have seen the last of the crash logs ...
13:07:40 <TrueBrain> damn, the admin interface of CouchDB is pretty
13:21:56 <boekabart> TrueBrain: did you once suggest that toc could work for simple apps without jitting?
13:26:32 <TrueBrain> boekabart: no, I said the initial version did that :p
13:26:42 <TrueBrain> but 99% of the applications didn't fit in that simple category
13:30:57 <boekabart> I don't really get how JIT is getting called
13:31:26 <TrueBrain> LD_PRELOAD loads a library, where his functions overrule any functions any other library might define
13:31:38 <TrueBrain> libemu.so defined emu_call2, but because of the LD_PRELOAD, it is the one from libjit.so which is called
13:44:51 <Xaroth> TrueBrain: what's the issue with WT3 btw?
13:44:54 <TrueBrain> hmm .. I am trying to add 100k records to couchdb .. it takes a while :p
13:46:59 <glx> Xaroth: missing features, like website translations, different translations for branchs and trunk :)
13:48:22 <glx> and the search function is powerfull :)
13:52:27 <glx> a big improvment when you know WT2 search ;)
13:52:37 <TrueBrain> anything was an improvement :p
13:53:37 <TrueBrain> @calc 100000 * 10 / 1024 / 1024
13:53:37 <DorpsGek> TrueBrain: 0.953674316406
13:54:35 <TrueBrain> 10 minutes to write 100k records with a total of 0.6GB :p
13:54:43 <TrueBrain> CouchDB has a high overhead :p
13:56:06 <TrueBrain> I stored about 50 chars of data .. so about 50MiB worth of data .. 10:1 .. auch :p
14:06:29 <Xaroth> quite a bit of overhead
14:06:48 <TrueBrain> yeah, but it seems related more to Couchdbkit (python) then anything else
14:07:31 <TrueBrain> I request a first() object, but it fetches ALL objects to dismiss the rest :p
14:07:37 <TrueBrain> lovely library .. lets trash it
14:08:11 <TrueBrain> I found the NS one more useful :p
14:09:10 <TrueBrain> but it is kind of nice :)
14:19:23 <glx> hmm is it me or wt3 is broken ?
14:19:30 <TrueBrain> dunno, havent touched it in ages :p
14:19:41 <TrueBrain> hmm .. when I raw access CouchDB, 5000 records consume 20MB
14:32:05 <TrueBrain> back to being bored .. it is one of these days ..
14:35:02 <boekabart> ok then help me explain
14:35:21 <boekabart> how my first fopen() can fail with EMFILE (too many files open) as error
14:36:34 <TrueBrain> because you are already out of file descriptors on a higher level? :)
14:36:54 <TrueBrain> I think a simpler explenation is that you read the error code wrong :)
14:37:35 <Xaroth> TrueBrain: we can always use a good list of goals of what you want for opendune :P
14:37:44 <boekabart> then there must be another explanation that fopen ( ..., "wb" ) fails
14:38:01 <boekabart> folder exist, there is write access
14:38:05 <boekabart> filename is correct
14:39:37 <TrueBrain> strace is a typical *nix app which traces all syscalls of an application
14:39:41 <TrueBrain> showing you what goes wrong where and how
14:39:59 <TrueBrain> gives you a 'raw' overview of your app
14:40:04 <TrueBrain> but .. I doubt they have that for windows :p
14:40:13 <Xaroth> TrueBrain: I know something to cure your boredness, ask the fed2k folks what -they- want from opendune, and then spend a week explaining nyer how he fails? :P
14:40:57 <TrueBrain> I am packing my snowboard .. sounds more useful ;)
14:41:14 <Xaroth> wait, weren't you going snowboarding for another few weeks? o_O
14:41:32 <TrueBrain> I am going to snowboard tonight
14:41:37 <TrueBrain> then again at the first of January :)
14:41:49 <Xaroth> indoor i take it (today)
14:41:50 <TrueBrain> but that time, 10 whole days .. well .. 7 whole days of boarding :p
14:43:51 <TrueBrain> gha, everything fits in 1 bag
14:43:58 <TrueBrain> not that there is any room left, but it does fit :)
14:44:03 <Xaroth> what the hell @ the street names in zoetemeer
14:44:10 <Xaroth> Italielaan, frankrijklaan, europaweg
14:44:42 <TrueBrain> they are just insane
14:44:50 <TrueBrain> but okay ... Snowworld! :)
14:45:14 <TrueBrain> I have to be at NOI at 18:30 .. which means I have another 3 hours to kill
14:47:28 <glx> how does C conversion work ?
14:47:52 <glx> I mean converting some functions is easy, but how the decompiler handles it ?
14:49:20 <TrueBrain> glx: the @implements and @name takes care of it
14:49:41 <TrueBrain> you should simple copy and paste the function, then the decompiler will pick it up .. you can remove all CalledFrom, but none of the @ lines
14:49:59 <TrueBrain> btw, glx, for now we froze src/, so we don't make it less stable for the upcoming release
14:50:10 <TrueBrain> after that we have a few weeks to add as much C shit as we like ;)
14:50:18 <Xaroth> no worries, i'm not comitting stuff!
14:50:27 <glx> ok, let me try for shl and shr then
14:50:37 <glx> even if I not commit it :)
14:52:29 <TrueBrain> glx: and I mostly make a emu_ variant and a non-emu_ variant of a function, to make later transitions easier
14:52:34 <TrueBrain> but this is not always possible and/or doable :)
14:54:53 <Xaroth> or sane, in my case :P
14:55:09 <Xaroth> Tile_GetPosXY, hrr hrr
15:02:25 <boekabart> TrueBrain it seems I can have only 1 file open: because start.bin is never closed, that blocked the one file
15:02:32 <boekabart> what CRT setting might this be....
15:03:32 <TrueBrain> boekabart: sorry, I don't follow :p
15:03:59 <boekabart> wtf something is totally fucked up here: malloc returns null
15:04:02 <TrueBrain> boekabart: btw, if crash.bin fails to write, it is most likely due to stack overflow :)
15:04:25 <TrueBrain> (sorry, I didn't get from your previous questions it was a LIbEMU issue)
15:04:58 <boekabart> win32: I built a lib combining libemu, libtoc and libjit and libtcc
15:05:15 <boekabart> call.c in libjit was C++ by the way
15:05:27 <DorpsGek> SVN: truebrain (r428) -Change: more consistent naming (glx)
15:06:06 <TrueBrain> euh ... libjit is a C project
15:06:06 <boekabart> this basically seems to work fine, but strange things happen....
15:06:19 <TrueBrain> TCC is not stable for Windows :)
15:06:47 <TrueBrain> boekabart: I can read it as much as I want, it is C :)
15:06:50 <TrueBrain> not C89, what MSVC wants
15:06:53 <TrueBrain> but for sure it is C99
15:07:06 <boekabart> C99 allows variable init. later on in the function?
15:07:16 <TrueBrain> any version after C89 allows that
15:07:17 <boekabart> declaration, actually
15:07:21 <TrueBrain> it is the most stupid restriction ever
15:07:25 <TrueBrain> all and any sane compiler allows it
15:07:36 <boekabart> i know... strange that msvc still won't eat it?
15:07:39 <TrueBrain> just MSVC is the only one who disagress on it :(
15:07:44 <TrueBrain> VERY strange and VERY stupid
15:07:53 <glx> msvc is not C99 compatible
15:07:53 <TrueBrain> but it for sure doesn't make the file C++ :)
15:08:04 <boekabart> ok, i stand corrected
15:08:14 <boekabart> only 1 fopen seems to be allowed
15:08:16 <TrueBrain> either way, tcc on windows, expect random weirdness
15:08:23 <TrueBrain> LOL! That hopefully isn't true ;)
15:08:27 <TrueBrain> then Windows is broken :)
15:08:35 <boekabart> could this be tcc already
15:08:46 <TrueBrain> dunno, is emu_call2 called? :)
15:08:50 <boekabart> this is the frist call into jit_emu_call2
15:08:57 <TrueBrain> (the JIT overrides a few things, forcing itself to be called :))
15:10:20 <boekabart> toc_decompile was called
15:11:06 <boekabart> all well until jl.entries = (JumpFrom*)malloc(4 * sizeof(JumpFrom)); -> null!
15:11:29 <TrueBrain> then it is out of memory :p
15:11:49 <TrueBrain> for windows btw you might need to assign 'code' segement to the memory
15:13:53 <boekabart> ah... disable SELinux for windows
15:14:12 <TrueBrain> hmm .. that in fact should be sufficient on Windows yes :)
15:16:04 <boekabart> (on for essential services only)
15:17:12 <TrueBrain> but okay, for a moment I was confused Windows needed a code segment instead of a data segment to execute code
15:17:46 <TrueBrain> either way, stability of TCC on Windows is low, so it can be that .. or something else MSVC fucked up :p
15:17:52 <TrueBrain> maybe compile with gcc ;)
15:33:10 <TrueBrain> my roommate got disconnected from the internet for downloading a movie
15:33:15 <TrueBrain> downloading movies is not illegal in The Netherlands
15:33:18 <TrueBrain> so .. what the fuck?
15:37:31 <boekabart> how did they find out?
15:38:21 <TrueBrain> it is SurfNET we have here
15:38:27 <TrueBrain> so some organization reported the IP I guess
15:39:45 <TrueBrain> but as downloading is not illegal ... I wonder what the fuck it was about
15:39:51 <TrueBrain> she "downloaded a torrent"
15:40:02 <TrueBrain> well .. next time it happens, I will give them a call, and ask them what the fuck it is about
15:40:12 <TrueBrain> in the meantime, lets switch all my connections to SSL encrypted :p
15:44:27 <boekabart> well with torrents.. you also upload
15:45:18 <Xaroth> then they wouldn't download a torrent, they would upload a torrent
15:45:26 <Xaroth> the download bit is perfectly legal, as long as you don't upload
16:00:55 <TrueBrain> the guy next to it is the most fun :)
16:07:49 *** boekabart has left #openDune
18:11:56 <Xaroth> looks nice, can't really judge on if it works or not :P
18:18:06 <glx> well these functions are very easy
19:05:06 *** boekabart has joined #openDune
19:08:43 <boekabart> that's a weird reaction
19:08:51 <boekabart> if you're on dry land, just stay and watch
19:52:21 <Xaroth> I'd run over water if I had to..
21:42:21 <boekabart> TrueBrain: How do I make comments / annotations about decompiled code that won't get lost after re-running main or toc?
23:51:30 <glx> atreides were pacifist until A3 :)
23:52:00 <boekabart> and I found diggers' in-game inner loop
23:52:19 <boekabart> -> the place to put a SDL_Delay(60) which makes the game totally playable
23:53:48 <boekabart> it was designed for 4.77mhz
23:54:12 <boekabart> and didn't trigger on smth like Retrace or so - at least not in inner loop
continue to next day ⏵