#exult@irc.freenode.net logs for 23 May 2003 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage


[03:03:51] <-- Fingolfin has left IRC ("42")
[03:26:40] --> ShadwChsr has joined #exult
[04:24:09] --> matto has joined #exult
[04:24:35] <-- Matt_O has left IRC (Read error: 104 (Connection reset by peer))
[06:30:43] --- matto is now known as Matt_O
[07:14:22] <-- ShadwChsr has left IRC ()
[08:16:37] --> `daniel has joined #exult
[09:39:04] <-- `daniel has left IRC ("Client Exiting")
[11:30:19] --> Fingolfin has joined #exult
[11:30:19] --- ChanServ gives channel operator status to Fingolfin
[11:48:19] --> Colourless has joined #Exult
[11:48:19] --- ChanServ gives channel operator status to Colourless
[11:48:43] <Colourless> hi
[11:48:50] <DarkeZzz> Greetings.
[12:02:18] <Fingolfin> yo
[14:53:44] <-- Kirben has left IRC (Read error: 54 (Connection reset by peer))
[14:55:18] --> Kirben has joined #exult
[14:55:18] --- ChanServ gives channel operator status to Kirben
[15:07:18] --> wjp has joined #exult
[15:07:18] --- ChanServ gives channel operator status to wjp
[15:07:52] <wjp> hi
[15:08:16] <Colourless> hi
[15:25:59] <Fingolfin> hiya
[16:09:54] --> Dark-Star has joined #exult
[16:15:19] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[16:38:14] --> sbx has joined #exult
[17:12:45] <wjp> hehe, somebody devoted an entirely email on the SDL ML to a supposed grammatical error in another mail
[17:13:00] <wjp> (of course, the original email was quite correct, and the reply was a commonly made error :-) )
[17:14:21] * Coren_ wavies.
[17:14:30] <Colourless> heh, what was it?
[17:14:36] <wjp> hi Coren_
[17:15:05] <Coren_> Rumors of my death have been vaguely exgagerated.
[17:15:09] * Coren_ smirks.
[17:15:32] <wjp> original email: "It's been lying around for about 18 months" (referring to a piece of software)
[17:15:41] <wjp> Reply: "I think you meant "laying around", but thanks for the chuckle :-)"
[17:15:52] <Colourless> foolish
[17:15:58] <wjp> yes
[17:16:03] <wjp> lay is transitive
[17:16:24] <wjp> s/lay/to lay/ to be pedantic :-)
[17:16:56] <Colourless> my dictionary has lying in it for 'lie'
[17:17:15] <wjp> yeah, lying is from lie, which is correct
[17:17:17] <Colourless> i think laying is an american thing
[17:17:56] <wjp> no, british too, AFAIK. But 'to lay' is a transitive verb, meaning something like "to put down"
[17:19:04] <Colourless> yeah i know, I know, the word laying exists, but saying 'laying around' i think is an american thing
[17:19:16] <wjp> oh, could be
[17:19:43] <wjp> hm, m-w.com has some info about it: http://www.m-w.com/cgi-bin/dictionary?book=Dictionary&va=lay
[17:20:03] <wjp> (The "Usage" section at the bottom)
[17:20:40] <wjp> lol
[17:20:52] <wjp> and the fun part is that _two_ people replied with something like "Yes! Of course!"
[17:20:52] <Colourless> hehe
[17:21:03] <wjp> (3 emails on a ML about this... *sigh*.. talk about off-topic :-) )
[17:21:15] <Colourless> Remember that even though many people do use lay for lie, others will judge you unfavorably if you do.
[17:21:15] <wjp> I'm tempted to reply (off-list)
[17:23:01] <Fingolfin> life must be nice if people have time for such things =)
[17:23:12] <wjp> :-)
[17:24:21] <Colourless> 3 of an inanimate thing : to be or remain in a flat or horizontal position upon a broad support <books lying on the table>
[17:24:30] <Colourless> hmmm
[17:25:19] <-- Fingolfin has left IRC ("42")
[17:25:54] <wjp> just replied... :-)
[17:26:00] <wjp> I wonder what kind of reaction I'll get :-)
[17:27:45] * sbx hops away.
[17:27:46] <-- sbx has left IRC ("X-Chat [1.6.4]")
[17:31:14] <Colourless> a flame?
[17:46:06] <wjp> not so far :-)
[17:46:09] <wjp> bbl, dinner
[21:17:38] --> Fingolfin has joined #exult
[21:17:38] --- ChanServ gives channel operator status to Fingolfin
[21:25:20] --> artaxerxes has joined #exult
[21:25:20] --- ChanServ gives channel operator status to artaxerxes
[21:25:28] <artaxerxes> hi all
[21:25:30] <wjp> hi
[21:25:42] <artaxerxes> how are you!?
[21:26:04] <wjp> busy :-)
[21:26:11] <artaxerxes> more exams?
[21:26:21] <wjp> no
[21:26:33] <wjp> just working on my master's thesis (or whatever the correct english term is)
[21:26:38] * artaxerxes is updating the zaurus port of Exult
[21:26:46] <artaxerxes> nice
[21:26:53] <artaxerxes> what's the topic of your thesis ?
[21:26:57] <wjp> and there's pentagram, of course :-)
[21:27:01] <artaxerxes> course
[21:27:06] <wjp> radical field extensions
[21:27:06] <artaxerxes> ;-)
[21:27:29] <artaxerxes> sounds.... big!
[21:27:32] <Colourless> i told you to do theoretical multi threaded interpreted programming languages :-)
[21:27:48] <wjp> that's more suited for CS, I think :-)
[21:27:55] <wjp> (which I guess I'll have to do sometime too)
[21:28:03] <wjp> of course, it would be strictly theoretical :-)
[21:28:39] <artaxerxes> I realise I haven't updated the zaurus port in ages!
[21:29:06] <Colourless> until yesterday. i haven't compiled exult in almost 4 months!
[21:29:16] <artaxerxes> 5 months for zaurus !
[21:29:31] <artaxerxes> what do you think of party formation?
[21:29:38] <Colourless> that is, at all :-)
[21:29:49] <Colourless> much better :-)
[21:29:51] <artaxerxes> it works really well.
[21:29:55] <Colourless> jeff finally saw the light
[21:29:59] <artaxerxes> ;-)
[21:30:23] <Colourless> a long long time ago, i actually started to implement party stuff myself.
[21:30:35] <Colourless> however I didn't want to have to rup the guts out of the old code to do it
[21:30:39] <artaxerxes> I notice s/o added USE_FMOPL_MIDI in the config.h file!
[21:30:42] <Colourless> so nothing happened
[21:30:59] <artaxerxes> I guess it's pretty disruptive to add that
[21:31:04] <artaxerxes> (party formation)
[21:32:19] <artaxerxes> I'm getting really happy since si-frenchhas staued in the 90% activity for the whole week! :-)
[21:32:47] <wjp> 97% ?!
[21:32:48] <wjp> wow
[21:33:14] * wjp wonders what we're doing wrong with pentagram :-)
[21:33:23] <Fingolfin> hehe
[21:33:24] <artaxerxes> from the beginning, there has been about 380 d/l of the patch!
[21:33:30] <Colourless> we only have 1 or 2 fans wjp, that
[21:33:35] <wjp> ah :-)
[21:33:39] <Colourless> is what we are doing wrong
[21:34:16] <Fingolfin> wjp: radical field extensions, eh? nice. any particular specialisation on that topic? or "just" in general?
[21:34:40] <wjp> Fingolfin: mostly general stuff, atm
[21:34:48] <wjp> (in fact, more general than just field extensions)
[21:35:03] <Fingolfin> Colourless: ah, I am a fan, too! I think you just need more pageviews, downloads, cvs commits, *and* some luck that the SF.net stats RNG favors you =)
[21:35:11] <wjp> Fingolfin: no, you're a team member :-)
[21:35:34] * Colourless see servus and cashman *sudder* as fans :-)
[21:35:40] <Fingolfin> lol
[21:35:43] <Colourless> s/sudder/shudder/
[21:36:09] <Fingolfin> wjp: I didn't do anything for agges, though... I don't even have U8 data files right now, so I don't even know how far you guys are... although what you told me sounds really good, so I need to look at it
[21:36:27] <Fingolfin> maybe it'll get me hooked again and I'll spend time on pentagram instead of ScummVM
[21:36:33] <Fingolfin> (just don't tell anybody in #scummvm ;-)
[21:36:44] <artaxerxes> ?logs
[21:36:44] <exultbot> Logs are available at http://www.math.leidenuniv.nl/~wpalenst/exultlog.php3
[21:36:51] <Colourless> we are almost finished Fingolfin :-)
[21:36:52] <artaxerxes> it's all visible here!
[21:37:02] <Colourless> TGWDS even work now :-)
[21:37:10] <Fingolfin> Colourless: finished as in 100% support for everything?
[21:37:10] <artaxerxes> ;-)
[21:37:37] <wjp> Fingolfin: finished as in the execution scene is running :-)
[21:37:46] <artaxerxes> nice
[21:37:47] * artaxerxes bows
[21:38:12] <Colourless> issues with animations, no camera movement, no sound, not lightning, no splash aside
[21:38:17] <Colourless> s/not/no/
[21:38:29] <wjp> no avatar control...
[21:38:33] * Colourless adds no music to the list as well
[21:38:43] * Colourless is purely taking about the execution scene
[21:39:02] <artaxerxes> how do you start the execution scene if you have no control of the avatar ?
[21:39:12] <wjp> you add a hack :-)
[21:39:22] <Fingolfin> hehe
[21:39:26] <Fingolfin> still, that's pretty cool
[21:39:34] <Colourless> and then you add another hack so it works on anyone elses system other than wjps
[21:39:48] <Fingolfin> hah
[21:40:06] <wjp> the original hack was of the type "hatch obj.id 21183"
[21:40:20] <wjp> the new hack is "see if obj.id 21183 is the right quality. If not, try 21184. Then hatch it" :-)
[21:40:49] <Fingolfin> I still wonder about a nicely defined backend system, I made paper sketches for a new backend systgem for ScummVM: SDL is not sufficient for us, as we support systems SDL does not support, and we need midi / adlib drivers, etc... and ideall want to make it possible to provide "native" backends, too, if desired
[21:40:54] <Fingolfin> maybe one could collaborate there somehow :-)
[21:40:59] <Colourless> if 21184 still isn't the right one, your version of u8 isn't supported by pentagram :-)
[21:42:32] <Fingolfin> <g>
[21:43:07] <artaxerxes> exult is now compiled for zaurus again
[21:43:11] <Colourless> also if your version of u8 isn't the english version, your version of u8 isn't supported... yet
[21:43:12] * artaxerxes is preparing a package
[21:43:54] <Colourless> big problem for u8 is executing the 'first' egg, will be a hack regardless of what we do
[21:48:46] <Fingolfin> Colourless: how comes?
[21:49:09] <Fingolfin> in SCUMM, we have bootscripts... at the start, you run script 1, and that's it, basically
[21:49:12] <Fingolfin> the rest follows from there
[21:49:13] <wjp> no such thing in U8
[21:49:27] <Fingolfin> so you have to hard code the startup sequence?
[21:49:31] <Fingolfin> and it differes between versions?
[21:49:32] <wjp> there's a 'FIRST' egg which you need to hatch, but you also need to give the avatar his inventory
[21:49:37] <wjp> yes, yes
[21:50:17] <wjp> well, it doesn't really differ between versions
[21:50:22] <wjp> not if we do it properly
[21:50:42] <wjp> (search for the egg in the right area, instead of just assuming it has a fixed object id)
[21:52:23] <servus> no shiny box usecode container in the backpack? :)
[21:52:47] <Colourless> thank god, no
[21:52:56] <Fingolfin> haha
[21:53:11] <Fingolfin> yeah, the inventory is also setup by the scripts in SCUMM, luckily
[21:54:04] <Fingolfin> the worst ones are the very old (v1/v2) scumm games, because you have to hardcode the font (bad for non-english versions), a few words (which have to be localized for non-english datafiles... which means you have to figure out a way to recognize those), etc.... all that is solved in V3 and upwards, though =)
[21:54:10] <wjp> luckily the avatar doesn't start with much in the way of inventory :-)
[21:54:30] <Fingolfin> (regarding localization detection, lukcily even that is possible because you can match ona few specific string resources ;-)
[21:54:32] <Fingolfin> wjp: yeah
[21:55:03] <Fingolfin> so what was that trick to get around the OS X internal compiler error?
[21:56:41] <wjp> manually include that .inl file
[21:56:50] * wjp looks up details
[21:57:16] <wjp> in SoftRenderSurface.cpp, manually #include SoftRenderSurface.inl
[21:57:20] <wjp> (5 times)
[21:57:42] <wjp> I kind of wonder what the preprocessor makes of it
[21:58:01] <Colourless> but where do you need to do that?
[21:58:32] <wjp> I mean remove the #include directives, and include the file there manually
[21:58:33] * artaxerxes is testing the new package
[21:59:10] <artaxerxes> gotta go! Good week-end all!
[21:59:14] <wjp> bye
[21:59:17] <-- artaxerxes has left #exult ()
[22:03:14] <Fingolfin> wjp: hm
[22:03:27] <Fingolfin> I'll play a bit with that
[22:05:44] <wjp> I'm really amazed that this fixes it
[22:05:53] <wjp> I would mean the preprocessor is broken
[22:05:57] <wjp> s/I/It/
[22:06:17] <Fingolfin> well I immediatly have to think of #import, which is a one-time-#include, and I wonder if that's related to the bug
[22:06:18] <Colourless> the preprocessor may be doing somehing stupid, such as only including the file once
[22:06:27] <Fingolfin> the prepocessor apple uses *is* modified after all
[22:07:15] <Fingolfin> did yuv look at the preprocessor output which crashes (I did add that to my apple bug report, at their request, if you feed it to the compiler, it crashes, too - but I never really looked at it closely
[22:07:28] <wjp> no, he didn't
[22:07:36] <wjp> (at least, if he did, he didn't tell me)
[22:07:43] * Fingolfin does it now
[22:08:04] <Fingolfin> uhh yeah that's clearly bad
[22:08:15] <wjp> that was quick :-)
[22:12:06] <Fingolfin> hrm, yeah, because I looked at the wrong spot ;-) so far this looks sane. if you want I can put it on my web site.. in fact maybe you could also dump the preprocessor output and we could compare it
[22:12:09] <Fingolfin> i.e. with -E
[22:12:37] <wjp> what's your exact commandline?
[22:13:20] * Colourless decides not to dump his preprocessor output. he doesn't want to know ho large the file would become with windows.h included
[22:14:43] <Fingolfin> g++ -Wp,-MMD,"graphics/.deps/SoftRenderSurface.d",-MQ,"graphics/SoftRenderSurface.o",-MP -Wall -DHAVE_CONFIG_H -DDATA_PATH=\"data\" -I./convert -I./convert/u8 -I./convert/crusader -I./misc -I./filesys -I./tools -I./tools/disasm -I./tools/compile -I./tools/fold -I./tools/shapeconv -I./kernel -I./graphics -I./audio -I./usecode -I./world -I./world/actors -I./gumps -I./conf -I./. -I. -I/sw/include/SDL -D_THREAD_SAFE -E graphics/SoftRenderSurfac
[22:14:46] * Colourless does so anyway
[22:14:58] <Fingolfin> 725034 byte, uncompressed =)
[22:15:32] <wjp> Fingolfin: line got cut off
[22:15:41] <wjp> is the end " -E graphics/SoftRenderSurface.cpp" ?
[22:15:44] <Fingolfin> wjp: ok
[22:16:01] <Fingolfin> g++ -Wp,-MMD,"graphics/.deps/SoftRenderSurface.d",-MQ,"graphics/SoftRenderSurface.o",-MP -Wall -DHAVE_CONFIG_H -DDATA_PATH=\"data\"
[22:16:01] <Fingolfin> -I./convert -I./convert/u8 -I./convert/crusader -I./misc -I./filesys -I./tools -I./tools/disasm -I./tools/compile -I./tools/fold
[22:16:02] <Fingolfin> -I./tools/shapeconv -I./kernel -I./graphics -I./audio -I./usecode -I./world -I./world/actors -I./gumps -I./conf -I./. -I. -I/sw/include/SDL
[22:16:05] <Fingolfin> -D_THREAD_SAFE -E graphics/SoftRenderSurface.cpp > cpp.out
[22:16:09] <Colourless> 1281 kb here :-)
[22:17:09] <wjp> 30288 lines, 787176 bytes
[22:17:18] <Colourless> the amount of that code which is the headers vs the actual sources is tiny :-)
[22:17:45] <Colourless> i have 111675 lines :-)
[22:17:55] <Fingolfin> hrr, stupid web server is acting up on me
[22:18:00] <wjp> http://www.math.leidenuniv.nl/~wpalenst/cpp.log (slowly uploading)
[22:18:07] <wjp> nvm, I'll zip it first
[22:18:08] <Colourless> i do have line numbers turned on, you may not
[22:18:25] <wjp> line numbers are on here too
[22:18:27] <Colourless> of course it may just be all the headers that are included
[22:18:42] <wjp> http://www.math.leidenuniv.nl/~wpalenst/cpp.log.gz
[22:18:55] <wjp> ack
[22:18:57] <wjp> http://www.math.leidenuniv.nl/~wpalenst/cpp.out.gz
[22:20:46] <Colourless> heh, the actual code from SoftRenderSurface.cpp starts at about line 109740 :-)
[22:21:45] * wjp is still scrolling down :-)
[22:22:36] <wjp> around line 29K here
[22:22:43] <Fingolfin> I'll be back in a moment
[22:22:43] <Colourless> seatch for template<class uintX> SoftRenderSurface<uintX>::SoftRenderSurface(SDL_Surface *s)
[22:22:49] <-- Fingolfin has left IRC ("42")
[22:23:21] <Colourless> my guess is your compiler strips away empty lines. msvc does
[22:23:25] <Colourless> n't
[22:23:38] * wjp looks at file
[22:23:41] <wjp> no it doesn't :-)
[22:24:12] <wjp> 20949 lines without empty lines
[22:24:38] <wjp> yikes, over 9000 empty lines?
[22:25:20] <Colourless> there are many many empty lines because of comments being stripped. and when a header is included, even if it's #ifdef'd out, there are still 'n' emtpty lines if the header was 'n' lines long
[22:26:01] <wjp> hm, no really long blocks of empty lines here
[22:27:11] <Colourless> template<class uintX> void SoftRenderSurface<uintX>::Paint(Shape*s, uint32 framenum, sint32 x, sint32 y)
[22:27:18] <Colourless> is on line 109989
[22:27:23] <Colourless> int scrn_width = clip_window.w;
[22:27:35] <Colourless> is on line 110106
[22:27:49] <wjp> 29410
[22:28:05] <wjp> 29414
[22:28:15] <Colourless> :-)
[22:28:31] <Colourless> mine has over a 100 blank lines :-)
[22:28:35] <wjp> that second one shows up several times of course
[22:36:59] --> Fingolfin has joined #exult
[22:36:59] --- ChanServ gives channel operator status to Fingolfin
[22:37:29] <Fingolfin> re
[22:37:39] <Fingolfin> that was longer than planned <sigh>
[22:38:58] <Fingolfin> my CPP output is at http://131.155.225.251/~maxhorn/pentagram/SoftRenderSurface.txt.gz , BTW
[22:39:00] <Fingolfin> 100 KB
[22:40:13] <-- SoulFKOFF has left IRC (Read error: 104 (Connection reset by peer))
[22:40:36] <wjp> lots of c++ header mismatches, obviously :-)
[22:41:28] <Fingolfin> yeah, only, actually
[22:41:33] <Fingolfin> i.e. only the headers differ
[22:41:39] <Fingolfin> the pentagram specific part seems to match 100%
[22:42:01] <Fingolfin> but maybe the compiler is missparsing the "# <linenumber>" statements
[22:42:33] --> SoulFKOFF has joined #exult
[22:46:47] <Fingolfin> if I comment out all the #include "SoftRenderSurface.inl", then it compiles. even one enabled, and it dies
[22:46:56] <Fingolfin> renaming the file to have a .h extensions doesn't change that
[22:47:17] <Colourless> might be template realted
[22:47:22] <wjp> did you try feeding it the preprocessed source without the "# <linenumber>" statements?
[22:47:40] <Fingolfin> yes, but that ggave me multiple
[22:47:41] <Fingolfin> line-map.c: file "graphics/SoftRenderSurface.cpp" left but not entered
[22:47:41] <Fingolfin> l
[22:48:35] <Colourless> my guess it's getting confused by a change of file within a templated function
[22:48:55] <Colourless> or mayby just because it's within a function at all
[22:48:59] <Fingolfin> yeah quite possible
[22:49:02] <Colourless> s/mayby/maybe/
[22:49:08] <Fingolfin> I can find that out easily =)
[22:49:43] <Fingolfin> if I replace the .inl by a "printf("Hello, world!\n");" it also gives a compiler error
[22:49:56] <Fingolfin> but if it's empty, or only contains comments, it works
[22:50:03] <Fingolfin> which supports Colourless' theory
[22:50:10] <Fingolfin> now I try to #include it in a non-template only
[22:50:46] <wjp> interesting
[22:51:15] <Fingolfin> if I only have the "Hello, World" in the .inl, this compiles fine:
[22:51:16] <Fingolfin> void fooBar(int x)
[22:51:16] <Fingolfin> {
[22:51:16] <Fingolfin> #include "SoftRenderSurface.inl"
[22:51:16] <Fingolfin> }
[22:51:27] <Fingolfin> (won't compile fine with the normal .inl, obviously)
[22:52:08] <Colourless> try templating it, then instantiating the templated function
[22:53:19] <Fingolfin> template<class uintX>
[22:53:19] <Fingolfin> void fooBar(int x)
[22:53:19] <Fingolfin> {
[22:53:19] <Fingolfin> #include "SoftRenderSurface.inl"
[22:53:19] <Fingolfin> }
[22:53:25] <Fingolfin> no instantce -> compiles fine
[22:53:59] <Fingolfin> make instance -> boom
[22:54:15] <Colourless> ah ha! now go file your bug report :-)
[22:54:29] <Colourless> to apple that is, not to the pentagram bug tracker :-)
[22:54:31] <wjp> preprocessed output of a couple of lines? :-)
[22:54:34] <Fingolfin> I already did, Colourless, a couple of months ago, and they ssaid it's a known bug and they already fixed it...
[22:55:01] <Fingolfin> but they haven#t made a bug fix release, those suckers <grmbl>
[22:55:17] <Fingolfin> so i guess/hope the new dev tools due to for the WWDC conference in June will have the fix
[22:55:35] <Fingolfin> plus several other bug fixes I am waiting for, like the one disabling exult on OS X
[22:55:47] <Colourless> so what did they change so much from normal gcc to cause that problem
[22:59:07] <Fingolfin> they have lots of changes compared to base GCC (and are pushing many back to mainline GCC, mostly it#s the FSF being slow in accepting the patches - e.g. took quite some time for the precompiled header stuff to get in)
[22:59:42] <Fingolfin> there are various template related bugs, I think they modified a lot of stuff in that area, and my guess would be that this was made to get templates working in Objective-C++
[22:59:48] <Fingolfin> i.e. to be able to mix Objective-C and C++ code
[23:00:04] <Fingolfin> Objective-C and C++ are almost completely orthogonal extensions to C
[23:00:09] <Colourless> ah yes
[23:00:35] <Fingolfin> but for categories, <> are used, too, maybe that caused problems or so
[23:09:33] <wjp> I should go
[23:09:43] <wjp> g'night
[23:09:53] <-- wjp has left IRC ("Zzzz...")
[23:10:25] <Colourless> oddly enough, i think i should go too :-)(
[23:10:27] <-- Colourless has left IRC ("casts invisibility")
[23:50:09] --> Kirben has joined #exult
[23:50:09] --- ChanServ gives channel operator status to Kirben