IRC logs for #opendune on OFTC at 2010-02-08
⏴ go to previous day
00:19:00 <TrueBrain> off too bed, classes tomorrow morning....
04:02:00 *** TinoDidriksen has joined #openDune
06:49:30 *** TinoDidriksen has joined #openDune
09:18:59 <TrueBrain> I hate it when customers call when my boss is away .. I really really can't help them :(
09:20:49 <Xaroth|Work> << fiddling with google docs to create a spreadsheet for production
09:21:06 <Xaroth|Work> we have a production sheet already, but it's cluttered as boot so I can't find anything :P
09:24:48 <TrueBrain> next question of customer: Outlook works at home, but not at office, what is wrong (via laptop)
09:24:56 <TrueBrain> how to explain the idea of closed firewall ...
09:31:16 <TrueBrain> and my boss made a TRUE mess out of the ticket system .. turns out that a few tickets never got a reply and the likes
09:34:22 <TrueBrain> tickets back from 4 months that he just never closed
09:34:29 <TrueBrain> I assume they can be by now .. don't care .. closed them
09:38:30 <Xaroth|Work> what do you use for ticketing system? RT?
09:38:38 <TrueBrain> some stupid OS system
09:42:57 <Xaroth|Work> this is starting to look quite funky
09:43:12 <TrueBrain> you do know you should work work, right?
09:43:36 <Xaroth|Work> but in between work I can fiddle with this
09:44:01 <TrueBrain> you have a lot of in-between time :p
09:44:13 <Xaroth|Work> lots of small things that require me to wait a while
09:44:34 <Xaroth|Work> you happen to have an account for google docs?
09:45:21 <TrueBrain> yes, but I dont want to talk about that ..
09:45:37 <TrueBrain> hmm .. odd .. I finished a skill which should reduce CPU power with 5% for launchers .. but there is nothing more free :(
09:46:55 <Xaroth|Work> did you try unfitting a launcher, then refitting it?
09:47:56 <TrueBrain> that indeed I Was trying, and then it works
09:47:58 <TrueBrain> that I call insanity
09:48:03 <TrueBrain> still I miss a bit of CPU for my salvager :p
09:48:35 <TrueBrain> you really have to strip your ship from time to time for your new skills to take effect ... WEIRD
09:52:13 <Xaroth|Work> you can also try ejecting from ship, then re-entering
09:52:21 <Xaroth|Work> else somebody might yoink your ship :P
09:59:56 <TrueBrain> one fun thing: had to change a MX record .. but the SPF was: +mx
10:00:07 <TrueBrain> so now mail from the old MX is no longer valid ;)
11:55:01 *** TinoDidriksen has joined #openDune
12:05:07 <TrueBrain> I can never tell if Dumpert stuff is real, or fake
12:05:31 <Xaroth|Work> yeh, but still, cool concept
12:05:42 <TrueBrain> Container DC concept is old
12:05:49 <Xaroth|Work> ok, technically, the bistot BPO can yield 2b/year
12:05:51 <TrueBrain> SUN had it a few years ago
12:49:44 <glx> I found the cause for the wrong emu_cs :)
12:50:14 <TrueBrain> it is always in the details
12:51:00 <TrueBrain> I decided not to C-ify the File_Open code, but to write my own
12:51:38 <TrueBrain> so first I do C-ify them, so I have a good overview
12:51:43 <TrueBrain> but in non-working condition :)
13:05:19 <glx> btw it's almost playable now, just some random unresolved jumps
13:05:54 <TrueBrain> I am not going to bother implementing a function that is never used ...
13:06:07 <TrueBrain> (Dune2 allows you to load a file in memory on open, and work from that .. )
13:06:13 <TrueBrain> maybe an XMS feature :p
13:06:33 <TrueBrain> well .. that is not even such a bad guess
13:06:50 <TrueBrain> then I have to look up fileInfo is created, to find that out
13:06:51 <glx> voices are disabled before drivers init in opendune :)
13:07:36 <glx> because it checks for XMS I guess
13:07:37 <TrueBrain> well, without EMS/XMS, it doesn't have enough memory to keep any of those files in its memory
13:08:37 <TrueBrain> seems to be a pointer indicating such memory
13:08:44 <TrueBrain> (being 0x0 for OpenDUNE)
13:09:36 <TrueBrain> return from DOS_Multiplex 0x4300 request
13:10:40 <TrueBrain> so yes, I am right :)
13:10:57 <TrueBrain> (and you too, most likely :))
13:11:46 <glx> lol it seems to check for enabled XMS in cfg (and disable voices in this case) after some other checks which can have already disabled voices
13:13:14 <glx> btw I don't understand how if (emu_cs == 0x3480) { overlay(0x3480, 1); } can work
13:13:35 <TrueBrain> so .... we need to make a choice here .. do we keep reading stuff from disk, or do we load a PAK file in memory upon first opening?
13:13:51 <TrueBrain> don't try to understand the overlay, just know it works :)
13:14:00 <glx> for B480 emu_cs is indeed 3FF1, and I don't get how it can be 0x3480 after a call
13:14:29 <glx> some kind of stack magic if an overlay is unloaded?
13:14:47 <TrueBrain> when you are in an overlay, and call another overlay, the overlay manager can unload the current overlay
13:14:55 <TrueBrain> it then changes the return cs to 3480 for example
13:15:02 <TrueBrain> where normally the bootstrap for overlay loading is
13:15:15 <TrueBrain> note that B480 == 3480 & 0x8000
13:15:52 <glx> the question was just about how emu_cs could become 0x3480, now it's clear
13:16:10 <TrueBrain> I was just explaining :)
13:16:19 <TrueBrain> making sure you realise B480 is a fake value, it doesn't realy exist ;)
13:16:42 <glx> yes I know the & 0x8000 thing
13:17:00 <glx> but the 44AF -> AB00 is weird
13:20:59 <TrueBrain> so what do you guys think? Load all in memory, or not?
13:21:40 <glx> seeking may be easier then :)
13:21:41 <TrueBrain> we talk about increasing the memory footprint with about 2 MiB or something
13:21:57 <TrueBrain> well, it is just that now files are opened ALL THE TIME
13:22:36 <Xaroth|Work> 5MiB footprint should be enough for all the files?
13:22:39 <Xaroth|Work> if so, fine with me
13:23:02 <glx> Xaroth|Work: just add size of all .PAK :)
13:23:18 <Xaroth|Work> glx: at work so can't check how big it is :), but if you say so, fine with me
13:23:29 * Xaroth|Work has 8gb ram, so really doesn't care about 5mb :P
13:23:31 <TrueBrain> I just said: 2 MiB :p
13:23:43 <Xaroth|Work> ah, wait, 2mib, even better
13:23:52 <Xaroth|Work> dunno where I got 5 from o_O
13:23:57 <Xaroth|Work> too much spreadsheets I think
13:24:19 <Xaroth|Work> for modern pcs that's still nothing
13:24:26 <TrueBrain> so slowly I do understand why that function is so complex :p
13:24:41 <glx> if all .PAK are open at the same time (and we don't open all languages at the same time ;)
13:27:13 <TrueBrain> 2649 is the extended memory handler, I guess
13:27:58 <TrueBrain> okay, I start to appreciate the beauty of this all
13:28:49 <glx> 3480:0152 does some XMS check I think
13:29:19 <TrueBrain> f__2BF2_0000_0007_42B5 runs the check for 6E3E
13:29:49 <TrueBrain> lol, f__2BF2_008E_0013_1D73 is weird :)
13:29:52 <TrueBrain> I think it is broken :p
13:33:00 <TrueBrain> btw, with my latest SaveLoad stuff, it no longer is safely possible to resume memory dumps
13:33:09 <TrueBrain> (as the local file pointers are not stored :p)
13:33:24 <TrueBrain> and with this work I am planning, in general it becomes a no-go
13:34:01 <TrueBrain> either way, glx, I see no way to fake XMS support so voices can be loaded, so you can at least see that code
13:34:23 <TrueBrain> I guess that needs to be done the old-fashioned way .... using DOSBox and its DEBUG command :)
13:34:47 <glx> we can look at that later ;)
13:35:05 <glx> having music/sound will be already nice
13:35:23 <TrueBrain> hmm ... this is a bit problemetic ..
13:35:29 <TrueBrain> I would like to use malloc, to allocate memory for files
13:35:36 <TrueBrain> but ... how to reach that from the 16bit world
13:36:37 <Xaroth|Work> wrap all the file read/write/seek operations?
13:36:39 <TrueBrain> owh, all access goes via wrappers
13:36:51 <glx> fake address and a wrapper, but won't help for direct memory accesses
13:37:17 <TrueBrain> I just need to validate all 2E9C acess
13:38:08 <glx> but I think if dune2 needs data from a file it usually map it somewhere in its memory (like 44AF)
13:38:33 <glx> I remember seeing strings data there with no sound/music drivers
13:39:27 <TrueBrain> ha, all access is already restricted to 1FB5
13:40:02 <TrueBrain> so I can freely do what ever I want, as long as I feed the right data on request :)
13:47:40 <TrueBrain> yeah, fuck the current code, I am going to write my own free implementation :)
13:48:53 <glx> hmm ok A1 and H1 are playable, let's see what I can get from all this decompiled stuff
13:52:24 <TrueBrain> I wonder .. if either flag 0x4 or 0x2 is set, it loads a file in memory
13:52:28 <TrueBrain> fine .. but how can they ever be set?
13:52:33 <TrueBrain> I have no code which suggests it can do that
13:52:56 <glx> static value in FileInfo?
13:53:09 <TrueBrain> yes .. static the key word here
13:53:35 <TrueBrain> so something has to flag any of those files as either 0x2 or 0x4?
13:53:40 <TrueBrain> maybe some never-decompiled subroutine
13:56:08 <TrueBrain> ah, yes, that is true
13:56:14 <TrueBrain> k, fine, we don't need it
13:56:17 <TrueBrain> I CAN DO IT ALL MYSELF
13:56:27 <TrueBrain> (it is nice to know this is an isolated library with no external references :)
14:12:07 <TrueBrain> 676 indexed files .. wow .. that is a lot :p
14:12:26 <glx> nice emu_Drivers_Sound_Init and emu_Drivers_Music_Init looks very similar (the only difference seems to be the struct used to store some infos)
14:13:20 <TrueBrain> do I need to make the malloc stuff optional?
14:13:22 <TrueBrain> runtime or compiletime?
14:13:51 <TrueBrain> possible, of course
14:13:54 <TrueBrain> but is it needed :)
14:15:16 <TrueBrain> funny enough ... changing scenarions runtime without XMS does work, with XMS doesn't work ;)
14:15:36 <TrueBrain> but in the first case it most likely results in disaster
14:16:57 <glx> hmm maybe I should commit the decompiled stuff (as it's disabled by default)
14:17:13 <TrueBrain> if you don't damage anything, it is fine by me
14:17:23 <TrueBrain> we can always revert :p
14:17:38 <glx> well it just needs the right config file to be executed
14:18:34 <TrueBrain> which people might have
14:18:41 <TrueBrain> so maybe add temporary code in dune.cfg which disables it?
14:19:22 <TrueBrain> maybe the safest :)
14:23:29 <TrueBrain> [LibEMU] in front, but yes
14:23:44 <DorpsGek> SVN: glx (r874) [LibEMU] -Add: fake MPU device
14:24:39 <TrueBrain> so many options and posibilities
14:24:48 <TrueBrain> for example, you can only have the same file open just one time
14:24:55 <TrueBrain> if you open it twice, things break
14:29:38 <TrueBrain> glx: I wonder .. runtime switching between malloc mode and file access ... where to switch? :)
14:29:57 <TrueBrain> maybe 'abuse' the XMS setting in the dune.cfg?
14:30:30 <glx> editing dune.cfg is not easy :)
14:31:32 <TrueBrain> the install.exe does that :p
14:31:38 <TrueBrain> but okay .. other suggestions?
14:41:33 <glx> . /* We only support one driver type. */
14:41:33 <glx> if (config->soundDrv != 7) config->soundDrv = 0;
14:41:33 <glx> if (config->musicDrv != 7) config->musicDrv = 0;
14:41:41 <glx> should be enough for now ;)
14:41:57 <TrueBrain> if not seven, than zero?
14:42:25 <glx> so we allow 0 (no music/sound) or the decompiled one
14:42:49 <TrueBrain> but does the decompiled one really work?
14:43:04 <TrueBrain> and then I would redirect all > 0 to 7
14:43:12 <TrueBrain> as if they want music, give them a music driver that works :)
14:43:16 <glx> there are still unresolved jumps
14:43:23 <TrueBrain> then put them all to 0 for now :)
14:44:53 <glx> hmm yes I still can "patch" this part when I need to test
14:46:44 <glx> ok all to 0, so the change should not break anything
14:49:42 <glx> [JIT] -Add: started to map one music/sound driver ?
14:50:59 <glx> oops forgot to run the fix_spaces script
14:51:46 <DorpsGek> SVN: glx (r875) [JIT] -Add: started to map one music/sound driver
14:54:40 <DorpsGek> SVN: glx (r876) -Update (r875): update decompiled code (and disable all audio drivers for now)
14:55:51 <TrueBrain> hmm .. for 'exists' function it can also change the drive it seems
14:55:57 <TrueBrain> maybe you could store savegames on a:?
14:56:39 <glx> hmm I forgot to set props for cs__AB00.c
14:58:56 <DorpsGek> SVN: glx (r877) -Fix (r876): forgot to set props for added file
14:59:27 <glx> let's configure autoprops in the VM
15:01:23 <TrueBrain> how to name a function that, in case the file was loaded in the memory, unloads it?
15:01:27 <TrueBrain> for example, highscore
15:01:29 <TrueBrain> normally is in memory
15:01:34 <TrueBrain> but when you click: clear
15:01:42 <TrueBrain> it tells the File stuff to unload that file
15:02:45 <TrueBrain> owh, it even deletes the file? hmm ..
15:07:25 <Xaroth|Work> why does it even load it in memory... i mean.. at most you'll be using it once every hour :P
15:07:51 <glx> you need 1 hour to finish first mission ?
15:08:06 <TrueBrain> Xaroth|Work: it is not about that .. either it loads all files in memory, and use them, or uses the disk all the time
15:08:14 <TrueBrain> more the question is, why it allows such files to be inside a PAK file
15:08:17 <TrueBrain> _that_ is the insane part :p
15:08:21 *** ChanServ sets mode: +v Yexo
15:08:49 <glx> I guess the lib used is some kind of all or nothing
15:09:22 <TrueBrain> hmm .. it seems nothing is loaded in the memory for Dune2 ... closer reading of a few bits give me the strong impression it was disabled
15:09:28 <TrueBrain> I remember some fuzz about this on the forums
15:20:35 <TrueBrain> haha, yes, I love my memory sometimes :)
15:20:45 <TrueBrain> In the EU version (the one we use), they disabled the load to XMS/EMS flag
15:21:12 <TrueBrain> it is the only version where it isn't used
15:28:13 <TrueBrain> f__1FB5_18D3_000E_76AC initializes the file stuff
15:31:20 <TrueBrain> k .. named all functions, and managed to get it split off
15:35:16 <TrueBrain> starting now fails .. strange :p
15:37:41 <TrueBrain> it even opens DUNE2.EXE via this function
15:44:10 <TrueBrain> Q: should I open all PAK files and index all files in them, or use the static array like we have now?
15:46:34 <Xaroth|Work> static array? as in, you have pre-set information what pak file has what content?
15:46:51 <TrueBrain> that is what is going on currently, yes
15:47:09 <TrueBrain> well, there still will a small list of the PAK files to scan
15:47:22 <TrueBrain> but that has to do with the fact that dir-listing is not in any sane cross-platform way :p
15:47:40 <Xaroth|Work> *.pak would have done wonders :P
15:49:55 <TrueBrain> hmm .. no, it should either be: scan all *.pak
15:50:00 <TrueBrain> or use this complete long static list
15:50:05 <TrueBrain> I am not going to make 'something that is in between' :p
15:50:14 <TrueBrain> so first the latter, later the first
15:51:10 <Xaroth|Work> scan all *.pak :)
15:52:36 <TrueBrain> feel free to write that code a bit later this week :p
15:54:45 <Xaroth|Work> I wouldn't even know how! :P
16:00:18 <TrueBrain> hmm .. what to use to open a PAK file
16:00:24 <TrueBrain> first allows paks in paks
16:00:29 <TrueBrain> second is most likely more sane
16:09:17 <Xaroth|Work> who's going to put paks into paks anyhow?
16:12:03 <TrueBrain> it never validates if a file is really in a PAK :o
16:12:08 <TrueBrain> it just assumes it should be there, so it is there
16:21:05 <TrueBrain> Unable to load font new8p.fnt
16:24:05 <Xaroth|Work> I think something messed up there :P
16:24:20 <TrueBrain> I think it does what IE xpected it to do :)
16:35:14 <TrueBrain> bah, I can't get it to work correctly .. hmm ..
16:47:43 <TrueBrain> File_Seek(0, 12, 0)
16:47:45 <TrueBrain> File_Seek(0, 12, 1)
16:47:46 <TrueBrain> File_Seek(0, 306, 1)
16:47:48 <TrueBrain> File_Seek(0, 5482, 1)
16:47:49 <TrueBrain> File_Seek(0, 8708, 1)
16:47:51 <TrueBrain> File_Seek(0, 185074440, 1)
16:47:52 <TrueBrain> somehow I think it seeks wrong
16:48:46 <Xaroth|Work> so it first tries, then does?
16:49:00 <TrueBrain> SEEK_SET and SEEK_CUR
16:49:19 <Xaroth|Work> so it goes to pos 12, then to +12
16:53:23 <Xaroth|Work> I thought I was late at work
16:53:34 <Xaroth|Work> then I realised I desynced this machine's clock :P
16:55:01 <TrueBrain> well ... almost then :p
16:55:08 <Xaroth|Work> time for me to go home
16:55:12 <TrueBrain> there is some corruption
16:55:16 <TrueBrain> I am going to do some rockclimbing
16:55:25 <Xaroth|Work> cya ingame tonight, probably :P
20:16:55 <Xaroth> 6 Weeks, 5 Days, 9 Hours, 46 Minutes << production time of a titan...
20:17:42 <Xaroth> @calc (((((6 * 7) + 5) * 24) + 9) * 60) + 46
20:18:20 <Xaroth> @calc (((((8 * 7) + 3) * 24) + 6) * 60) + 14
20:18:35 <DorpsGek> Xaroth: 0.79998593761
23:22:57 <TrueBrain> GAH! Everything works. ... except the ingame pallete
23:23:37 <SmatZ> what's your problem with File_Seek() ?
23:23:46 <TrueBrain> I fixed it, I thought
23:23:48 <TrueBrain> still something is wrong ..
23:24:44 <TrueBrain> my validations are too strict, so it seems
23:24:50 <TrueBrain> you can seek past the file
23:25:33 <SmatZ> does it work? sell it...
23:28:32 <TrueBrain> there ... now it can seek past the file, but the next read returns a 0 :)
23:49:46 <TrueBrain> glx: can you please test it on Windows?
23:50:26 <TrueBrain> and, if you like, validate my code?
23:50:31 <TrueBrain> it is a complete rewrite, where I left out tons of things
23:50:33 <glx> Xaroth: I still miss a "download patch" button ;)
23:50:36 <TrueBrain> for now it doesn;t load files into the memory
23:52:14 <SmatZ> TrueBrain: how can I download that patch?
23:52:21 <TrueBrain> SmatZ: replace 'svn' with 'devs'
23:53:01 <glx> hmm doesn't apply cleanly
23:53:08 <TrueBrain> ah, didn't update yet
23:53:59 <SmatZ> "not cleanly" = "with quite many failures" ;)
23:57:35 <TrueBrain> SmatZ: SVN still doesn't like removing the same file twice
23:57:39 <TrueBrain> not something I can fix :)
23:57:48 <TrueBrain> maybe you should learn to revert correctly :p :p :p
23:57:55 <glx> just remove 1fb5 manually :)
23:58:13 <TrueBrain> SVN really sucks in adding and removing of files
23:58:28 <glx> Avertissement,...1,...warning C4244: 'return' : conversion de 'uint16' en 'uint8', perte possible de données,...d:\developpement\opendune2\src\file.c,...115
23:58:28 <glx> Avertissement,...2,...warning C4244: 'return' : conversion de 'uint16' en 'uint8', perte possible de données,...d:\developpement\opendune2\src\file.c,...203
23:59:07 <SmatZ> TrueBrain: ok, though there is still that warning with file.c
continue to next day ⏵