#exult@irc.freenode.net logs for 6 Apr 2002 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage

[00:55:54] <-- wjp has left IRC ("Zzzz....")
[01:47:13] --> Kirben has joined #exult
[01:47:13] --- ChanServ gives channel operator status to Kirben
[02:27:02] --> Darke has joined #exult
[02:27:02] --- ChanServ gives channel operator status to Darke
[02:27:43] * Darke bows. Hi.
[03:12:38] --- Darke is now known as Darke|afk
[03:50:33] --- Darke|afk is now known as Darke
[06:33:35] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[07:01:22] --> Kirben has joined #exult
[07:01:22] --- ChanServ gives channel operator status to Kirben
[07:15:13] --> Kirben2 has joined #exult
[07:15:13] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[07:15:58] --- Darke is now known as Darke|afk
[07:37:23] --- Darke|afk is now known as Darke
[09:21:22] --- Kirben2 is now known as Kirben
[09:57:00] --> wjp has joined #exult
[09:57:00] --- ChanServ gives channel operator status to wjp
[09:57:03] <wjp> hi
[09:57:06] <Darke> Hi.
[09:58:56] <Darke> Incidentally, is there a 'Pentagram Team' entity defined? I just noticed in the code I commited I used that for the (C), and thought it might be a good idea to check if it exists. <grin>
[09:59:15] <wjp> no, there isn't really
[10:02:45] * Darke wonders if we should make an AUTHORS file like exult and start dropping names in there, or wait until we have a bit more 'stuff' completed.
[10:03:19] <wjp> fine by me
[10:19:19] <Darke> Ok, done. I've only included the names of others who have got copyrights to the source files in the pentagram/ tree. If I've missed anyone, they can add themselves. <grin>
[10:22:03] * Darke watches his 'cvs commit's Just Not Work, and wonders if sf.net is having problems again...
[10:25:13] <Darke> The entire front page of sourceforge.net: "An error occured in the logger."
[10:26:06] * Darke woo!s. It's working again. But the forum appears to have canabalised itself again.
[10:28:12] <Darke> Hmm... authenticating by their webpage is giving errors. Looks like an authserv is down somewhere.
[10:50:45] --> matto has joined #exult
[10:54:22] --> Fingolfin has joined #exult
[10:54:49] <Fingolfin> hi
[10:54:58] --- ChanServ gives channel operator status to Fingolfin
[10:55:05] <wjp> hi
[10:56:29] <Darke> Hi.
[10:58:47] <matto> hi!
[11:04:24] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[11:22:41] --> Kirben has joined #exult
[11:22:42] --- ChanServ gives channel operator status to Kirben
[11:24:45] <Fingolfin> SF is foobared
[11:25:09] <Fingolfin> but at least the web is back up now
[11:27:22] --> bj0ern has joined #exult
[11:27:32] <bj0ern> hi
[11:28:04] <Fingolfin> hi!
[11:29:59] <Darke> Hi.
[11:30:33] <matto> hey bj0ern
[11:31:41] * Darke yays and can actually commit not. <grin>
[11:31:48] <Darke> s/not/now/
[11:35:48] <Fingolfin> SF is back up, wooohooo
[11:35:53] <Fingolfin> yeah
[11:36:02] <Fingolfin> coax just told me that they got the stuff working again
[11:38:56] <Kirben> Darke: do you really need to use readline for ucfiddle ? I noticed scummvm has similar menu functions but readline is optional
[11:39:52] <matto> I see that Kirben is lurking in #scummvm :)
[11:40:27] <Darke> Kirben: Not really. I just want a way of getting a user's input, without having to go to all the effort of writing a basic 'readline' library myself, and I knew that existed, so I used it.
[11:41:32] <Darke> As I've said previously, I can use getline() just to grab the raw text. But that doesn't support backspace/delete handling, which is the primary thing I would want.
[11:43:42] <Fingolfin> welll
[11:43:53] <Fingolfin> why don't you use readline if it is available, and fall back to getline otherwise?
[11:43:56] * Darke could, alternately, use raw curses, which is what readline also does. But that still requires yet another library.
[11:44:05] <Fingolfin> sure it is not as comfortable, but better than not being able to use this at all
[11:44:14] <Fingolfin> curses is not there on all systems either
[11:45:57] <Kirben> I can't read code well but scummvm seems to do that in just a few extra lines.
[11:46:32] <Darke> Fingolfin: <nod> That's what I was intending upon doing if I had no other option, I just haven't coded it yet. There's a HAVE_READLINE_READLINE_H define if it finds the header, so I can probably use that for the handling of it.
[11:46:47] <Fingolfin> exactly, just do it (TM)
[11:47:13] <Kirben> http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/*checkout*/scummvm/scummvm/debug.cpp?rev=HEAD&content-type=text/plain
[11:48:35] <Darke> Kirben: Thanks. I was just looking for that.
[11:48:59] <bj0ern> i never tried that. is a missing header file a warning or an error for the compiler?
[11:50:11] <Fingolfin> ohhhhhh!!! scummvm now support loom!!!
[11:50:31] <Fingolfin> damn, only the CD version, sigh
[11:50:33] <bj0ern> sweet
[11:50:36] <bj0ern> oh
[11:50:49] <Fingolfin> didn't even know there was a 256 color version of loom :-) he
[11:50:56] <bj0ern> loom was cool, though i never played it very far
[11:52:47] <matto> looooom
[11:54:07] <matto> hehe.. I just found an old usenet post of mine from 1994!
[11:54:32] <matto> who in the world saves these old messages??
[11:55:40] <Fingolfin> hehe
[11:55:42] <Darke> Kirben: It doesn't look like they're doing anything fancy, just grabbing text. It appears to work though. <grin>
[11:55:49] <Fingolfin> google, for example, bought a huge newsnet archive :-)
[11:56:28] <Fingolfin> hm, my Loom copy and that "backup" copy here *cough* are slightly different... mine is newer, too, from 31. Jan 1991 vs 1. Jul 1990 =)
[11:56:48] <Fingolfin> hm, wonder what version the Loom in my games box is
[11:57:34] * Fingolfin has a box with Indy III, Loom, MI 1, Zak + Maniac Mansion
[11:58:18] <Kirben> the 256 colors versions of loom and zak256 are from japanese fm-towns system.
[11:58:46] <Kirben> Seems really odd the LucasArts never released them elsewhere.
[12:00:24] <Fingolfin> woa
[12:00:31] <Fingolfin> that would explain maybe, yeah
[12:01:05] <Fingolfin> yet another version on my CD, from 13. Jun 1990 ?=
[12:07:40] <Darke> Ok. It works. The question now is, how do I get configure.in to automatically add the three libraries it needs if it finds the headers? Any idea if there is there a $have_readline_h variable available after I've done the AC_CHECK_HEADERS()? I'm not very fluent with configure.
[12:12:28] <Fingolfin> what do you mean, it has to add libraries to find a header ?!? hu?
[12:14:32] <bj0ern> no, it has to add the libs to -L if it finds the headers too :)
[12:14:49] <bj0ern> or was it -l? ;)
[12:15:21] <Fingolfin> well
[12:15:25] <Darke> I need to manually add libraries to the LIBS variable, if it finds the header.
[12:15:40] <Fingolfin> I will look into it, right now I fix up your Makefile to automatically regenerated configure / config.status if needed
[12:15:47] <Darke> But I don't know how in configure.in to tell if it finds the header. And the docs are, afaict, worse then useless.
[12:16:36] * Darke nods. No problem. He's working on it anyway.
[12:19:21] <Fingolfin> the docs areN#t that bad, go and take a look at the autobook :-)
[12:24:45] <Darke> Cool. I've got it working, although it checks for readline.h twice, once to put the #define into the config.h, the second time to add the libraries to LIBS. But considering it checks for flex twice, I'm not complaining. <grin>
[12:24:53] <Darke> (docs) Umm... no thanks. <grin>
[12:27:53] * Darke is mainly griping, because he can't find the 'Autoconf for Dummies' section, where it gives you half a dozen examples of 'I want to do this. Then do this'. Instead it seems it sprinkles all these little examples throughout the entire document. Good once you know the program well, but not so good if you are new to it, and want to use it to accomplish something. <grin>
[12:40:57] <Fingolfin> hrm
[12:41:23] <Fingolfin> weird
[12:41:59] <Fingolfin> pentagram's configure.in seems to think that I have no gawk, no gcc (my gcc is called cc, but still a GNU-C compiler), no readline, and sizeof(int) == 0 -> this is not exactly correct
[12:42:44] <Darke> <very-big-grin> You don't say?
[12:48:34] <Fingolfin> wo
[12:48:40] <Fingolfin> and I doN't have stdlib.h, btw =) lol
[12:48:54] --> Colourless has joined #Exult
[12:48:54] --- ChanServ gives channel operator status to Colourless
[12:49:03] <Colourless> hi
[12:49:04] <Fingolfin> hi ryan
[12:49:16] <Darke> Have you tried doing what 'configure' does manually to find gcc? It's rather insane and convoluted, but it seems to work for me. <grin>
[12:49:22] <Darke> Hi Colourless.
[12:51:06] <Darke> Fingolfin: It, at the very least, might give you a hint as to what is wrong.
[12:51:28] <Fingolfin> I am looking at the config.log, yeah, and I see waht goes wrong
[12:51:35] <Fingolfin> I just have no clue why, so I need to look a bit more :-)
[12:51:38] <matto> to the pit with autoconf, automake, and the rest of those wretched tools :)
[12:51:46] <Fingolfin> I like them a lot, actually
[12:51:57] <Fingolfin> well, maybe not "Like"
[12:52:06] <Fingolfin> but they have been extremly useful to me in the past
[12:53:12] <matto> I'm sure once you get passed the learning curve they can prove useful
[12:53:23] <matto> I never made it passed.. hehe
[12:54:59] * Darke likes autoconf/automake, but they can be _extremely_ frustrating to work with at times.
[12:56:15] <matto> I finally did figure out how to add dependancies to my Makefiles though... hehe
[12:57:49] <Fingolfin> well it's way easier to gtok than the source of apt, I think
[12:58:33] <Fingolfin> boy, how can one name classes so badly, I wonder... at least they should have put in a few comments that explain why some stuff is as it is, if they can't structure/name it in a more logical fashion, grumble
[13:02:00] * Darke often worries why people can't follow that logic. Admittedly though it's not as if he isn't prone to name things relatively generically, umm... like, 'ucfiddle' perhaps? <grin>
[13:03:23] <Colourless> well people would just ask. "WTF is that?" and then wonder what it does and after figuring it out that it doesn't do anything seemingly that useful, at least not at the moment :-)
[13:04:04] <Darke> Incidentally, I'm throwing together a basic 'Hacking' file together, of 'filename' and a couple of lines description of what it does, at least for the files I know about. Where should I drop it? In the pentagram root directory like I've seen other projects do it, or into docs/?
[13:04:27] <Colourless> docs... all text files goes in docs
[13:04:34] * Darke nods. Makes sense.
[13:05:00] <Colourless> things that could be thought of being somehow related to pentagram's program design should go into docs/design
[13:06:19] <Darke> Hmm... I'm documenting the documents directory. That's scary in a meta way.
[13:07:09] <Fingolfin> I mean, look at this:
[13:07:10] <Fingolfin> enum DepStateFlags {DepNowPolicy = (1 << 0), DepNowMin = (1 << 1),
[13:07:10] <Fingolfin> DepInstPolicy = (1 << 2), DepInstMin = (1 << 3),
[13:07:10] <Fingolfin> DepCandPolicy = (1 << 4), DepCandMin = (1 << 5)};
[13:07:20] <Fingolfin> there is a comment with that, actually...
[13:07:25] <Fingolfin> // These flags are used in StateCache::DepState
[13:07:28] <Fingolfin> very helpful :-)
[13:07:52] * Colourless doesn't have a clue
[13:07:56] * Darke 's first thought is a wonder as to what a 'Cand' is...
[13:08:09] <Fingolfin> this is on package versions and their install state...
[13:08:41] <Fingolfin> there is exactly one place where those are changed (and one other place where they are checked), it seems. Commen excerpot:
[13:08:43] <Fingolfin> /* This determines the combined dependency representation of a package
[13:08:43] <Fingolfin> for its two states now and install. This is done by using the pre-generated
[13:08:43] <Fingolfin> dependency information. */
[13:08:53] <Fingolfin> argh
[13:08:56] <Fingolfin> !
[13:09:14] <Darke> Nice! Even the comments are encrypted. <grin>
[13:09:46] <Fingolfin> so somhow, pakcage can have a now, an install, and a candidate state, or so
[13:09:56] <Fingolfin> and there is a "policy" and a "min" flag for each such state....
[13:10:12] <Fingolfin> maybe I can track by RTFS what the states represent, sigh
[13:11:26] <matto> bye!
[13:11:28] <-- matto has left IRC ("Play Dragon's Lair in linux - http://www.daphne-emu.com - Developers welcome :)")
[13:13:50] <Fingolfin> my configure bug was caused by autoconf 2.52, btw
[13:13:55] <Fingolfin> with 2.13 it seems to work fine again
[13:15:58] * Darke upgraded to 2.52, then downgraded back to 2.13 because exult suddently stopped configuring correctly, and he couldn't figure out what the problem was.
[13:18:38] <Fingolfin> =)
[13:18:47] <Fingolfin> we should go over our configure.in one day and fix that
[13:23:59] <Darke> Probably a good idea. <grin> I wasn't even sure if it was my problem or not since, IIRC, I upgraded and downgraded automake at the same time. It could have just been the interactions between the two.
[13:24:21] <Fingolfin> I think there are really problems
[13:24:25] <Fingolfin> at leas in pentagram there are :-)
[13:24:40] * Darke noticed. <grin>
[13:26:11] <Fingolfin> Darke: BTW, the AC_CHECK_HEADERS can be at the top level of the configure.in (and should be). it just will check for their existance, and you can do that safely on any system. Anyway,I'll fix that up, I am working on the confiugre.in anyway
[13:26:17] * Darke _really_ liked the fact it determined `sizeof(int)==0`, it's trying to tell you something, he's just not sure what. <grin>
[13:26:34] <Darke> Thanks.
[13:32:29] <Darke> Colourless: You'll need to update u8sfxfmt.txt now you've appearntly figured out how to decompress the sound, or get rid of it altogether. <grin>
[13:33:03] <Colourless> hmm, true
[13:33:47] <Colourless> a quote from that doc
[13:33:53] <Colourless> "I tried to do some reverse engineering but because of U8's weird memory management I didn't get anywhere. Sorry, but unless someone has a brilliant idea"
[13:34:02] <Colourless> :-)
[13:34:12] <Colourless> i actually did both :-)
[13:34:45] * Darke snickers.
[13:35:10] <Colourless> i did reverse engineer u8.exe ... but not in the way that generally you would... and the brilliant idea was to use u8.exe to decompress the sounds... not attempt to recreate the algorithm ;-)
[13:40:51] * Darke thinks it's certainly an unconventional idea. <grin>
[13:41:16] * Darke thinks he's sick of documenting already, and he's only just finished documenting the docs.
[13:42:05] <Colourless> hey wjp: small changes around the place has sped up the decompression a bit. i'm not down to 381 ms from about 450 ms for the 4 second sound
[13:42:12] <Colourless> s/not/now/
[13:46:57] <Colourless> hmm it appears that the C++ implementations of the opcode may in some cases be faster than the more correct ASM implementations ;-)
[13:47:26] <Colourless> i'm down to about 360 ms with some opcodes now in C++
[13:47:38] <wjp> Colourless: you know, that brilliant idea really was mine ;-P
[13:47:52] * Darke mumbles something about 'optimising compilers'.
[13:48:01] <Colourless> well the intepreter was my idea :-)
[13:48:06] <wjp> no, Fingolfins :-)
[13:48:32] <wjp> logs of 21 March, 16:58:45, 17:01:23
[13:48:47] <Colourless> hmm, i could swear i started on it before anyone said anything
[13:49:25] <wjp> that's possible
[13:49:44] <Colourless> hmm, maybe not it seems :-)
[13:50:07] <wjp> of course, both of these comments were said jokingly
[13:50:12] <Colourless> i think i started doing everything about 2 or 3 days later
[13:50:26] * wjp never thought it would be actually doable
[13:51:33] <Darke> "Colourless: Making the impossible possible."
[13:51:46] <Colourless> well,
[13:51:47] <Colourless> yes
[13:51:57] <Colourless> :-)
[13:52:23] * Darke grins.
[13:52:42] <Colourless> up until us, no one had even manage to partly decompress the u8 sounds :-)
[13:57:02] <Colourless> changing Or, And and Xor to C++ has slowed things down a bit
[14:06:11] <Colourless> ok, when using no asm the decompression takes 390 ms which is better than I expected
[14:06:38] <Colourless> the faster combination of some asm and some c++ code take between 360 and 370 ms
[14:12:51] <Colourless> now for the 'fun' part... attempting to make is platform portable and to run on Big Endian systems... when i don't have one to test on :-)
[14:13:11] <Colourless> but i think i know how to do it
[14:14:15] <Darke> Nice! Good luck! <grin>
[14:14:37] <Colourless> all registers will be in the format of the host platform. the stack and obviously memory and the executing code will be in little endian format. byte swaps would need to performed on the go for any memory accesses, pushes and pops
[14:16:12] <Colourless> making those changes will be fairly easy, but i just can't test if they work :-)
[14:19:19] <Colourless> first though, see if it compiles with gcc
[14:32:49] * Fingolfin is back
[14:32:57] <Fingolfin> yo folks
[14:33:01] <Colourless> hi
[14:33:03] <Colourless> wb
[14:34:43] * Fingolfin is still working on revisiing pentagram's configure.in
[14:34:51] <Fingolfin> for some reasons, it's not using my compile rule anymore, arg
[14:35:54] <Colourless> ow, that's a lot of errors :-)
[14:38:16] * Fingolfin notes an old mail on the pentagram build system that he never sent a reply to, with some bug mentioned in it, which actually explain the problems he sees now
[14:38:44] <Colourless> all errors gone
[14:39:10] <Colourless> as expected gcc was a little slower than msvc
[14:39:31] <Colourless> about 440 ms
[14:56:18] <Fingolfin> hm, I guess /dev/null only exists on Unix like systems, right :-)
[14:56:42] <Colourless> windows has null, which is the same
[14:56:55] <Fingolfin> ahhh GOOD!
[14:57:13] <Fingolfin> you see, right now, I get messages like these:
[14:57:13] <Fingolfin> common.mk:14: no file name for `-include'
[14:57:25] <Fingolfin> I can fix those by adding an empty file to the end of the include list, so that it is never empty
[14:57:28] <Fingolfin> cool
[14:57:37] <Colourless> actually is nul not null
[14:58:06] <Fingolfin> I would like to get my build system in a shape that allows it to be used on Cygwin, BTW, if somebody could help me on that, I'd be grateful... i.e. I would make changes, but I'd need somebody to test if it works :-)
[14:58:42] <Colourless> kirben would probably be best for that sort of thing
[14:59:00] * wjp has to go
[14:59:01] <wjp> bye
[14:59:02] <Colourless> you know, he does use cygwin and all... i don't
[14:59:04] <Colourless> cya
[14:59:06] <-- wjp has left IRC ("[x]chat")
[14:59:25] <Fingolfin> Colourless: sure. I'll page him, once I have my current mods in CVS
[15:01:26] <Fingolfin> Colourless: so I could say something like: "echo foo > nul" ?
[15:01:32] <Colourless> yeah
[15:01:42] <Fingolfin> and also "type nul" (type == cat, IIRC)
[15:01:43] <Colourless> that wont do anything
[15:01:48] <Fingolfin> good
[15:01:58] <Colourless> type nul will also do nothing
[15:02:09] <Fingolfin> fine
[15:02:11] <Fingolfin> what I need
[15:40:16] <Kirben> no mingw can't use cygwin's buld system and mingw doesn't have one
[15:41:04] <Kirben> there is mingw through cygwin compiler switch but thats usually too limited.
[15:47:59] <Fingolfin> Kirben: I am talking about modifyin the main build system to work on cygwin as well
[15:48:05] <Fingolfin> this would reduce the maintance work a little bit
[15:48:26] <Kirben> Just depends on what that requires
[15:48:27] <Colourless> and mingw as well i take it?
[15:49:11] <Fingolfin> Colourless: if possible, I can't judge it
[15:49:21] <Fingolfin> Kirben: well, I'll look at the Makefile.cygwin now
[15:49:44] <Kirben> cygwin is used as nothing more than shell for file commands
[15:49:48] <Fingolfin> and then I'll try to adjust configure.in / Makefile.in to do stuff right on cygwin; and then one of you must test it =)
[15:50:17] <Kirben> it is much more complicated
[15:50:29] <Kirben> are you doing this for exult or pentagram ?
[15:50:46] <Fingolfin> pentagram
[15:50:47] <Fingolfin> and indeed
[15:50:57] <Fingolfin> Makefile.cygwin and .win32 are very similiar
[15:51:11] <Fingolfin> I hope to "merge" them both into the main Makefile
[15:51:30] <Fingolfin> I only need somebody to test (and to ask question regardin these environments which I never used :-)
[15:51:44] <Fingolfin> first two differences are the command names, and / vs. \
[15:51:53] <Fingolfin> the / vs. \ is the biggest issue I guess, hm
[15:51:54] <Kirben> well for cygwin you need extra flags
[15:52:14] <Fingolfin> extra flags? for what. compilation?
[15:52:21] <Kirben> yes
[15:52:50] <Kirben> If you must do it, do it for msys and nothing less.
[15:53:06] <Colourless> which flags?
[15:53:15] <Colourless> i don't see any differences
[15:54:22] <Fingolfin> the .win32 is for mingw ?
[15:54:58] <Colourless> yeah it's for just mingw by itself
[15:54:58] <Fingolfin> how would I delete a directory with all contents? i.e. like "rm -rf DIR" in Unix
[15:55:28] <Colourless> rd /s will work in WinNT
[15:55:29] <bj0ern> deltree c:\blah
[15:55:38] <Kirben> I would have ot look it up again, I mean a configure under cygwin shell will break under mingw make.
[15:55:39] <Colourless> no, deltree is 'old' dos
[15:55:44] <bj0ern> ah
[15:56:03] <Colourless> doesn't support LFN in WinNT
[15:56:20] <Kirben> why are we even talking about makefile.win32 ? configure doesn't run via win32 console
[15:56:25] <bj0ern> there was a rmdir command too in the older dos versions, maybe they named it like that?
[15:56:55] <Colourless> annoyingly rd /s (rmdir /s) doesn't work on win9x.... but i don't think it would be a big loss if win9x wasn't properly supported :-)
[15:56:57] <Fingolfin> Kirben: oh, ok, that's bad of course. not much I can do then :-/
[15:57:07] <Fingolfin> Colourless: nah :-)
[15:57:56] <Fingolfin> anyway, our configure.in does check for mingw32. is that going to support our configure now or not? I am not sure how to interpret Kirben's remark, is that targeted at mingw32 ?
[15:58:35] <Colourless> also oddly enought somewhere along the was i picked up rm, Microsoft File Removal Utility, which acts exactly the same as *nix rm
[15:58:45] <Colourless> s/was/way/
[15:59:00] <Kirben> only if msys is used in addition to mingw.
[15:59:45] <Fingolfin> BTW, I need to add install targets to Makefile.in I guess =)
[15:59:57] <Kirben> let me see with current cvs
[16:00:23] <Fingolfin> ok, I'd like know what exactly breaks when you try to use configure/Makefile -> then I can try to fix it
[16:00:52] <Kirben> where is configure ?
[16:01:22] <Fingolfin> you have to generate it - but I guess I could DCC/email it to you if you have no autoconf/automake isntalled
[16:01:29] <Fingolfin> if you have those, you can jsut run bootstrap once
[16:01:42] <Kirben> there is no autoconf/automake for mingw by default
[16:02:35] <Kirben> and msys is just an additional shell (official one) for mingw too incase I'm confusing the issue.
[16:05:42] <Kirben> The points are 1. mingw with msys has no build system (autoconf/automake) unless you compile you own versions, Cygwin's build system isn't compatbile with mingw make and only option left (using mingw compiler switch of cygwin) becomes too messy as more includes/lib are needed (Since it doesn't have complete mingw).
[16:06:24] * Fingolfin is confused
[16:07:05] <Kirben> with which points ?
[16:07:06] <Fingolfin> err, for now I thought we'd be talking about Makefile.cygwin
[16:07:21] <Fingolfin> and that seems pretty Unix like
[16:07:47] <Fingolfin> doesn't cygwin "emulate" a unix environemnt to a certain degree? or do I mix it up again?
[16:08:00] <Kirben> makefile.cygwin should really be makefile.mingw
[16:08:16] <Kirben> yes thats true
[16:08:23] <Colourless> haven't we argued that before :-)
[16:08:33] <Kirben> but makefile.cygwin doesn't uses cygwin's compiler at all.
[16:08:45] <Kirben> Colourless: yes which is the less confusing...
[16:08:50] <Fingolfin> Kirben: so it uses the cygwin environment, but the mingw compiler or what?
[16:08:54] <Colourless> .cygwin is for compiling using mingw under the cygwin shell... .win32 is for compiling using mingw and no shell
[16:08:58] <Fingolfin> ah
[16:08:59] <Fingolfin> ok
[16:10:28] <Kirben> yes cygwin enviroment but with mingw compiler still.
[16:11:37] <Colourless> as ive stated before, if .cygwin is renamed .mingw, non cygwin users will get confused about which makefile (.mingw or .win32) to use with mingw and no shell :-)
[16:12:23] <Kirben> I doubt we have any cygiwn users
[16:12:25] <Colourless> since they would be using mingw the logical choice would be .mingw... which would be the wrong one
[16:12:59] <Kirben> .cygwin only started because it used to do a cross compile
[16:13:01] <Colourless> i agree though that .cygwin is the wrong name... but i don't think .mingw is the right name though
[16:13:21] <Kirben> using the mingw switch of the cygiwn compiler
[16:15:51] <Kirben> mingw does have a standard shell now though (msys) so makefile.mingw would be better for that.
[16:17:59] <Fingolfin> Kirben: so do you have cygwin (i.e. can you use Makefile.cygwin) ? In that case, can I DCC/mail you an alternate Makefile and you tell me how it fails?
[16:18:19] <Kirben> yes I have cygwin installed
[16:18:23] <Kirben> dcc away
[16:19:09] <Fingolfin> does the DCC get thru to you?
[16:19:13] <Kirben> no
[16:20:13] <Fingolfin> then I gotta mail
[16:21:54] <Kirben> Colourless: anyway we need to get rid of the makefile.cygwin asap, with odd using trying to compile it cygwin native and the way a cygwin libs could easily be linked in.
[16:22:06] <Kirben> odd =odd user
[16:24:22] <Colourless> yeah i agree
[16:26:36] <Kirben> Just fails to find sdl
[16:28:01] <Fingolfin> Kirben: indeed, I made a stupid mistake :-)
[16:28:19] <Fingolfin> can you edit it and replace the $(SRC) with $(srcdir)
[16:28:34] <Kirben> I just used ./
[16:28:41] <Fingolfin> ok
[16:28:55] <Fingolfin> what else?
[16:29:02] <Kirben> wants mac midi
[16:29:09] <Fingolfin> hmmm?
[16:29:12] <Fingolfin> lemme look
[16:29:21] <Kirben> g++ -Wp,-MMD,"viewer/../audio/midi_drivers/.deps/mac_midi.d2" -DPENTAGRAM -DDEBU
[16:29:21] <Kirben> G -DSIZEOF_SHORT=2 -DSIZEOF_INT=4 -I/tools -I/viewer -I/files -I/audio -I/header
[16:29:21] <Kirben> s -I/conf -I./sdl/include -g -Wno-long-long -DHAVE_MIDI -DHAVE_SNPRINTF -Dsnpr
[16:29:21] <Kirben> intf=_snprintf -I./viewer -I./tools -I./conf -I./files -I./audio -I./audio/midi_
[16:29:21] <Kirben> drivers -Iheaders -Iviewer -g -Wall -I./sdl/include -c -o viewer/../audio/midi_d
[16:29:21] <Kirben> rivers/mac_midi.o viewer/../audio/midi_drivers/mac_midi.cc
[16:29:24] <Kirben> \mingw\bin\..\lib\gcc-lib\mingw32\2.95.3-6\cpp0.exe: viewer/../audio/midi_driver
[16:29:25] <Kirben> s/.deps/mac_midi.d2: No such file or directory
[16:29:46] <Fingolfin> indeed,
[16:29:59] <Fingolfin> but that's because I commited a bad change to viewer/module.mk
[16:30:07] <Fingolfin> sec
[16:31:23] <Fingolfin> for now, please edit viewer/module.mk
[16:31:26] <Fingolfin> remove the mac_midi line from it
[16:33:57] <Kirben> there is no mac_midi
[16:34:05] <Kirben> oops wrong file
[16:34:12] <Fingolfin> I also fixed it in CVS now
[16:34:20] <Fingolfin> this actually doesn't work over here either =)
[16:34:33] <Fingolfin> I need to enable midi_driver support in the build system
[16:59:07] <-- Colourless has left IRC ("restarting...brb...")
[17:02:43] --> Colourless has joined #Exult
[17:02:43] --- ChanServ gives channel operator status to Colourless
[17:02:50] <Colourless> b
[17:29:57] <Fingolfin> from Midi.c, line 95:
[17:29:58] <Fingolfin> err; // stops a warning
[17:30:08] <Fingolfin> -> for me it *causes* a warning ... hm
[17:30:23] <Fingolfin> but I think I know how to fix it properly
[17:31:13] <Colourless> Fingolfin: yeah, that should cause a warning... of course the person who put it was attempting to stop a unused variable warning
[17:31:23] <Fingolfin> yeah the proper fix is
[17:31:30] <Colourless> i'm guessing the warning you get is about a statement with no effect?
[17:31:36] <Fingolfin> to change the catch to use an annonymous parameter, i.e. remove the "err"
[17:31:40] <Fingolfin> indeed
[17:32:16] <Fingolfin> hm, I can't compile unpackshp.cc
[17:33:20] <Fingolfin> ouch
[17:33:29] <Colourless> that doesn't sound good :-)
[17:33:33] <Fingolfin> we I see, we have two different classes Flex and two different Flex.h files
[17:33:38] <Colourless> yeah
[17:33:39] <Fingolfin> sigh
[17:33:44] <Colourless> one from exult one from pentagram
[17:33:54] <Fingolfin> yup
[17:34:14] <Colourless> the exult one should have some #ifndef PENTAGRAM's around it
[17:34:25] <Fingolfin> yeah
[17:34:27] <Fingolfin> will fix
[17:35:01] <Fingolfin> hmmm
[17:35:26] <Fingolfin> but then if it #include "Flex.h" it will still find hte wrong one if one doesn't get the order of the include paths right
[17:35:26] <Colourless> dir search order causing a problem?
[17:35:31] <Fingolfin> exactly :-)
[17:35:40] <Fingolfin> that explains the order they had in Makefile.in
[17:35:42] <Fingolfin> will reverse it again
[17:35:43] <Fingolfin> sigh
[17:36:25] <Colourless> well, you know the real fix would be to get the exult Flex class to work in pentagram.... the only difference between u7 and u8 flexes is the u8 flexes are missing the magic bytes
[17:36:33] <Fingolfin> hmm
[17:36:43] <Fingolfin> ok
[17:37:26] <Fingolfin> though I like the interface of the pentagram one more =)
[17:37:31] <Colourless> that 'will' require substantial changes to pentagram though... exult and pentagram flex classes are fairly different
[17:37:40] <Colourless> yep, i think pentagrams is better too :-)
[17:38:19] <Colourless> another solution is to rename pentagram's class and file
[17:38:43] <Colourless> maybe using a pentagram wide consitant class naming system :-)
[17:39:13] <Colourless> hell, if not, naming it to FlexFile would remove the problems :-)
[17:39:52] <Fingolfin> hehe
[17:53:55] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[18:42:05] <Colourless> hmm, i think i just a lot of preprocessor macros in my interpreter. Example:
[18:42:13] <Colourless> // Subtract word register from r/m word
[18:42:14] <Colourless> case 0x2B:
[18:42:14] <Colourless> OPCODE2REG("sub");
[18:42:14] <Colourless> TO_NATIVE_PRE(v2);
[18:42:14] <Colourless> CALL_OP2_NOFL(Sub, v1.w, v2.w);
[18:42:14] <Colourless> EMU_OP_OVER(printf(" -> %04X",*v1.w));
[18:42:15] <Colourless> TO_LITTLE_POST(v2);
[18:42:17] <Colourless> break;
[18:42:35] <Colourless> each line is a macro :-)
[18:42:53] <Colourless> except for the case, break and comment obviously :-)
[18:44:00] <Colourless> s/just/use/
[19:03:15] <Fingolfin> he
[19:16:16] <Colourless> so what are you doing about flexes?
[19:24:49] <Fingolfin> right now I just fixed the build order
[19:24:54] <Fingolfin> which works fine
[19:24:57] <Colourless> ok
[19:25:04] <Fingolfin> on the long run, I think we should maybe rename it
[19:27:36] <Colourless> hmm, this is interesting... pentagram flex class contains the member char** objects; which isn't actually used as far as i can tell
[19:28:16] <Colourless> looks like that wjp may have inteded to do caching :-)
[19:29:54] <Fingolfin> hehe
[19:34:35] <Colourless> hmm. well, i'm going to expand the functionality of the Pentagram Flex class, and create a new FlexWriter (that inherits Flex). That unused char** objects isn't going to remain unused for too much longer :-)
[19:35:45] <Colourless> basically i'm going to get flexwriter to use objects as an overide of the data in the actual file
[19:50:48] <Fingolfin> hn
[19:51:04] <Fingolfin> so it will write out flex files?
[20:15:52] <Colourless> yeah
[20:15:59] <Colourless> once it's finished
[20:16:15] <Colourless> otherwise what is the point? :-)
[20:17:25] <Fingolfin> well, with that name, it better should "write" something :-)
[20:17:36] <Fingolfin> but being able to override stuff would be nice for fixes/patches I guess
[21:19:42] <-- Fingolfin has left IRC ("42")
[21:30:54] <-- Colourless has left IRC ("no comment")
[21:31:32] --> matto has joined #exult
[22:04:37] --> protoman has joined #exult
[22:09:15] --> b0hnz has joined #exult
[22:09:27] <-- b0hnz has left #exult ()
[22:26:22] --> Dark-Star has joined #exult
[22:26:29] <Dark-Star> HI
[22:27:10] <Dark-Star> ?version
[22:27:10] <exultbot> The latest official version of Exult is: 0.96beta1
[22:28:00] <Dark-Star> exultbot: nah, that's not correct... try again ;-)
[22:28:46] <Dark-Star> ?help
[22:28:59] <Dark-Star> ?log
[22:28:59] <exultbot> Logs are available at http://www.math.leidenuniv.nl/~wpalenst/exultlog.php3
[22:31:10] <protoman> :-) making fun of bot heh?
[22:31:39] <Dark-Star> no, I just try to make him learn ;-)
[22:33:09] --> irwin has joined #exult
[22:41:42] <-- Dark-Star has left IRC ("core meltdown")
[23:11:42] <-- protoman has left #exult ()
[23:11:49] <-- irwin has left #exult ()