#exult@irc.freenode.net logs for 7 Mar 2001 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage


[00:11:37] --> chimera|work has joined #exult
[01:50:28] --> DraX has joined #exult
[01:50:31] <-- DraX has left #exult
[03:31:15] --> matt0 has joined #exult
[04:43:07] --> Kirben has joined #exult
[04:54:14] <matt0> kirben!
[05:26:39] <Kirben> Hi
[05:27:38] <Kirben> Any success with that VC++ compile ?
[05:31:14] <matt0> hehe
[05:31:22] <matt0> no
[05:31:28] <matt0> it was just like they said... big STL problems
[05:31:43] <matt0> if exult wasn't using STL it would probably be an easy compile
[05:31:53] <matt0> and I don't ever use STL so I don't know what the VC++ quirks are
[05:56:10] <Kirben> hmm missed a cvs update
[07:55:44] <Kirben> I hope this hack works
[11:57:21] <-- Kirben has left IRC (Read error to Kirben[co3007967-a.brasd1.vic.optushome.com.au]: Connection reset by peer)
[12:17:47] --> mattownby has joined #exult
[12:17:47] <-- matt0 has left IRC (Read error to matt0[1667079149-yippy.dsl.xmission.com]: Connection reset by peer)
[16:09:45] --> Fingolfin has joined #exult
[16:09:52] <mattownby> max!!!
[16:10:23] <Fingolfin> matt!!!
[16:36:19] <-- Fingolfin has left IRC (42)
[16:38:33] --> matt0 has joined #exult
[16:40:32] <-- mattownby has left IRC (barnes.openprojects.net herbert.openprojects.net)
[17:47:29] --> wjp has joined #exult
[17:47:34] <wjp> hi
[18:06:54] <matt0> wjp!
[18:10:26] --> Fingolfin has joined #exult
[18:11:33] <Fingolfin> lo
[18:16:34] <matt0> greetings!
[18:16:47] <Fingolfin> lo matt
[18:26:59] --> freedman has joined #Exult
[18:27:06] <freedman> Hello, all!
[18:27:15] <Fingolfin> hi jeff
[18:27:21] * Fingolfin reads his emails
[18:27:37] <matt0> freedman!
[18:27:42] <freedman> It's a slow day at work - I'm downloading Sailor Moon videos.
[18:28:42] <Fingolfin> lol
[18:28:43] <freedman> Matt: Giving up on the VC++ compile?
[18:28:51] * Fingolfin is downloading Futurama episodeds ;)
[18:28:57] <Fingolfin> vc++: where is the problem?
[18:29:10] <Fingolfin> I know it has quirks, since i had to help some other people converting stuff to it ;)
[18:29:16] <Fingolfin> it chokes on a lot of std C++...
[18:29:37] <wjp> hi all
[18:29:40] <freedman> I hear it has STL problems.
[18:29:46] <freedman> Hi wjp!
[18:29:51] <matt0> freedman: yes I ran into STL problems as you and wjp said I would
[18:29:57] <Fingolfin> yeah
[18:30:03] <Fingolfin> STL problems are typical for VC++
[18:30:09] <Fingolfin> vc++'s STL sucks ass
[18:30:18] <wjp> matt0: told you so :-P
[18:30:28] <matt0> wjp: so you did =]
[18:30:31] <freedman> Is there a way around the problems? I'd like to use STL at work, but it's got to compile with VC++
[18:31:05] <matt0> I compiled some sample STL programs in VC++ and they worked
[18:31:07] <matt0> just simple things
[18:31:17] <matt0> so I don't know where it is failing
[18:31:39] <matt0> it didn't like stuff like "using std::cerr"
[18:32:06] <wjp> hmm.. maybe #ifndef all those using's?
[18:32:20] <matt0> perhaps
[18:32:30] <Fingolfin> hmm
[18:32:42] <Fingolfin> in fact, look at jameson/christophs work
[18:32:44] <freedman> How about "#define using //" :-)
[18:32:49] <Fingolfin> Alpha's CXX has similiar problems
[18:32:56] <wjp> freedman: good one :-)
[18:33:00] <Fingolfin> <g>
[18:33:27] <freedman> This Win32 crashing bug is getting to be a pain. I can't get it to happen on Linux.
[18:33:39] <freedman> Even with Electric Fence
[18:33:51] <matt0> I think it's a matter of knowing which STL headers to include, which not toinclude.. which usings to use, and which not to use hehe
[18:33:54] <matt0> (basically a mess)
[18:34:12] <matt0> is Electric Fence some kind of fancy debugger I should be using? hehe
[18:34:38] <freedman> Might not be worth the effort. What about BorlandC++? It's much closer to gcc.
[18:35:01] <wjp> I've got BC++5 and 5.5 lying around somewhere, and also have C++Builder 3
[18:35:10] <wjp> maybe I should try it
[18:35:18] <matt0> so no one can get gdb working in windows?
[18:35:35] <wjp> I used to build Exult in BC++ back around version 0.10
[18:35:40] <freedman> Electric Fence is a memory debugger that you link with. Takes up HUGE (~200MB to run Exult) amounts of memory, though.
[18:35:50] <matt0> I see
[18:35:59] <Fingolfin> I have Metrowerks for x86....
[18:36:02] <freedman> wjp: Oh yeah, I remember; the very first win32 port!
[18:36:03] <Fingolfin> even the cross compiler
[18:36:06] <Fingolfin> just need to install it ;)
[18:36:43] <freedman> I still think it's got something to do with the audio code... What else has changed?
[18:36:51] <matt0> it might be an SDL bug too
[18:37:07] <freedman> Yep.
[18:37:21] <matt0> ahh electric fence.. this looks cool =]
[18:37:35] <Fingolfin> hm, does it help to turn of sound completly?
[18:38:04] <freedman> EF is cool; just be sure you have lots of RAM and/or swap. BTW, it's by Bruce Perens.
[18:38:20] <matt0> I have 320M
[18:38:28] <matt0> who is Bruce Perens?
[18:38:39] <freedman> Max: I've heard that it doesn't help.
[18:38:42] <wjp> maybe we should add an option that _really_ disables audio
[18:38:51] <Fingolfin> yeah
[18:38:56] <Fingolfin> e.g. it used to crash on mac
[18:38:58] <Fingolfin> audio I mean
[18:39:03] <Fingolfin> actually, it still might...
[18:39:09] <wjp> btw, I think the device detection is broken when no devices are found
[18:39:18] <Fingolfin> because it used malloc/new inside of the audio callback - which is bad anyway, slows down the system
[18:39:25] <freedman> Maybe we should drop the Win32 version:-) Yes, I know, it woudln't be fair to >half the users.
[18:39:40] <Fingolfin> maybe we should blackmail them ;)
[18:39:46] <Fingolfin> "send your mones to this address:"
[18:39:58] <wjp> s/mones/money/ ?
[18:40:05] <Fingolfin> yeah ;)
[18:40:16] <Fingolfin> as I said on #SDL "maybe I should add a spelling checker to this irc client? ;)"
[18:40:23] <freedman> I think Origin should pay us something:-)
[18:40:48] <wjp> :-)
[18:40:55] <Fingolfin> yeah... how about donating to each of us a machine that allows to play U9 smoothly ? ;)
[18:41:14] <freedman> LOL - That would bankrupt them.
[18:41:16] <wjp> that would be nice :-)
[18:41:19] <Fingolfin> exactly ;)
[18:41:37] <Fingolfin> hmmm
[18:41:42] <Fingolfin> compilation error
[18:42:00] <freedman> About the audio - Maybe I'll take a look at the callback. But it hasn
[18:42:13] <Fingolfin> SDL_LoadWAV in Audio::playwave: we pass a Uint32, but it expects an "unsigned long"... hm
[18:42:21] <freedman> 't changed recently.
[18:42:24] <matt0> ok I'm gonna boot to win32 and try this VC++ again hehe
[18:42:35] <matt0> brb
[18:42:37] <-- matt0 has left IRC (Laserdisc Arcade Emulation Project: http://daphne.rulecity.com , get your CVS password today! =])
[18:42:44] <Fingolfin> jeff: in sdl mixer, or in exult?
[18:43:07] <freedman> In exult. It's in Audio.cc.
[18:44:44] <Fingolfin> I know ;)
[18:44:51] <Fingolfin> I already fixed it here ;)
[18:45:10] <wjp> ok, I'm going to hack in some kind of full-audio-disable option
[18:45:43] <freedman> The main change since Alpha is the switch to 16-bit. But I don't see anything wrong with it.
[18:46:42] <freedman> Looking at the code now in my browser...
[18:49:15] --> [JUN]JB has joined #exult
[18:49:25] <wjp> wb
[18:49:29] --- [JUN]JB is now known as matto
[18:49:36] <Fingolfin> hm, I'll try to run speech on macos again...
[18:49:39] <Fingolfin> maybe it works now ;)
[18:49:49] <Fingolfin> since SDL recently fixed some audio endianess issues
[18:50:12] * Fingolfin fasten his seatbelt and gets ready for a crash
[18:51:15] <Fingolfin> haha
[18:51:20] <Fingolfin> it quits immediatly ;)
[18:51:31] <freedman> :-(
[18:51:42] <freedman> ... Looking for ByteBuffer class...
[18:51:57] <Fingolfin> pcb.h
[18:52:03] <Fingolfin> in audio/pcb.h
[18:52:12] <freedman> Thanks...
[18:52:33] <wjp> maybe I put a crossreferenced Exult source online somewhere
[18:52:40] <wjp> s/put/should put/
[18:53:04] <wjp> (see lxr.linux.no)
[18:53:22] <freedman> Thanks!
[18:53:57] <wjp> I looked at it a few weeks ago, but didn't manage to get it setup properly
[18:54:01] <wjp> I'll try again this weekend
[18:54:11] <wjp> should be useful
[18:54:19] <Fingolfin> yeah
[18:54:20] <matto> well... it looks like it would be possible to get VC++ working
[18:54:26] <matto> but it looks like the time it would take would be enormous
[18:54:27] <Fingolfin> although my compiler does that automatically for me ;)
[18:54:28] <freedman> Okay. Consume calls pop_front, but that doesn't allocate/deallocate. However, the callback does mess with an STL list, and that could deallocate memory. Could that be a problem?
[18:54:29] <matto> and it wouldn't be worth it
[18:54:45] <matto> virtually every include file has some STL stuff in it
[18:54:47] <Fingolfin> jeff: I wondered about that, too
[18:54:55] <Fingolfin> matto: well
[18:55:04] <Fingolfin> matto: as I said, jameson had similiar problems
[18:55:15] <Fingolfin> matto: he fixed it using some prefix header...
[18:55:22] <freedman> But it's been like that all along, and the crashing is recent.
[18:55:48] <Fingolfin> matto: check out alpha_kludges.h
[18:55:49] <matto> Fingolfin: well I am not going to spend anymore time with it.. if someone else wants to, that's fine =]
[18:55:57] <Fingolfin> jeff: hm, well, are you sure that everything is the same?
[18:56:02] <Fingolfin> matto: ok ok ;)
[18:56:03] <freedman> matto: Good decision.
[18:56:38] <freedman> Fingolfin: I made lots of changes for combat/weapons. One of those could be the problem.
[18:56:47] <matto> I am perfectly happy using exult in linux =]
[18:57:22] <freedman> Matto: The X in Exult stands for X-windows. That's what I originally wrote it for.
[18:57:38] <freedman> ...but most users seem to be on Win32:-(
[18:58:58] <wjp> ok, I just added a "config/audio/enabled" switch that really kills all audio
[18:59:18] <-- Fingolfin has left IRC (Ping timeout for Fingolfin[pD9504D6C.dip.t-dialin.net])
[19:00:26] --> Fingolfin has joined #exult
[19:00:59] <Fingolfin> wtf
[19:01:27] <wjp> hmm... our configure scripts need some updating... they don't support building from a directory other than the real source dir. :-(
[19:02:17] <wjp> (I want to be able to build two copies from the same source dir, one debug and one normal version)
[19:03:39] <freedman> wjp: Good idea.
[19:04:14] <wjp> the problem seems to be the -I../../.... include flags
[19:04:36] <wjp> hmm...
[19:04:39] <freedman> wjp: Right.
[19:04:41] * wjp browses infopages
[19:05:01] <freedman> ...I'm a bit distracted by a phone call
[19:05:17] <wjp> k
[19:08:49] <wjp> hmm... this looks like it would take quite a lot of changes...
[19:10:04] <freedman> I've not much patience for configuration.
[19:11:19] <freedman> Looking at the mixer: This could be done without the dynamic list of buffers.
[19:11:38] <Fingolfin> hm
[19:11:44] <Fingolfin> jeff: if you could do it, it would be better
[19:11:53] <freedman> I may play with this tonight. It also needs the ability to loop on a single wav.
[19:12:10] <Fingolfin> the audio callback should do as few things as possible
[19:13:06] <freedman> Fingolfin: Okay. There should just be a fixed set of buffers, enough to hold the max. # of simultaneous streams (4-5?)
[19:13:18] <Fingolfin> yeah
[19:13:29] <Fingolfin> dunno how many there...
[19:13:35] <Fingolfin> do they take up a lot of space? if not, just set some high enough number (e.g. 8 or 16 or 32)
[19:15:14] <freedman> They can be around 32K. But we can allocate when a sound is played, then clean up later (but not in the callback, as happens now).
[19:16:24] <freedman> We really could get by with, say, 4: 1 for the music, 3 for simultaneous sound effects.
[19:18:37] * wjp is happily inserting $(srcdir)'s
[19:19:09] <Fingolfin> wjp: go rocking ;)
[19:19:22] <freedman> Have fun!
[19:19:25] <wjp> we have way too many Makefiles :-)
[19:20:00] <freedman> Wish we could automatically generate those Makefile.xxx's. It's hard to remember to update all of them when I add a file.
[19:20:17] <wjp> hmm... Makefile.am.in ? ;-)
[19:24:48] <wjp> phew... that should do it
[19:29:36] <freedman> Time for lunch. See you later.
[19:29:38] <-- freedman has left IRC (Leaving)
[19:31:18] <matto> SDL mixer!
[19:31:19] <-- matto has left IRC (<[JUN]JB> you slippery weasel! <Slagathor-ONTHEROAD> You know it!)
[19:31:34] <wjp> huh?
[19:38:34] <wjp> ok, this seems to be working
[19:38:54] <wjp> I can now build multiple versions of exult from the same source tree
[19:52:49] <wjp> incoming mailbomb ;-)
[19:53:04] <wjp> (well, nothing compared to yours, but the best I can do :-) )
[20:03:18] <wjp> bbl
[20:09:32] <-- Fingolfin has left IRC (Ping timeout for Fingolfin[pD9504D6C.dip.t-dialin.net])
[23:07:37] <-- wjp has left IRC ([x]chat)