#gemrb@irc.freenode.net logs for 29 May 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:56:59] <-- Maighstir has left #GemRb
[01:30:38] --> budlust has joined #GemRb
[01:44:51] --> mvbarracuda_libr has joined #GemRb
[01:48:17] <-- barra_library has left IRC (Ping timeout: 265 seconds)
[02:04:13] <-- budlust has left IRC (Ping timeout: 264 seconds)
[02:18:42] --> budlust has joined #GemRb
[02:23:46] <-- budlust has left IRC (Ping timeout: 264 seconds)
[02:51:02] --> _pickle has joined #GemRb
[03:07:34] --> budlust has joined #GemRb
[03:15:07] <-- budlust has left #GemRb
[03:22:46] <-- mvbarracuda_libr has left IRC (Read error: Connection reset by peer)
[03:52:50] --> budlust has joined #GemRb
[03:53:49] <-- budlust has left #GemRb
[04:28:56] <-- _pickle has left IRC (Remote host closed the connection)
[06:05:45] --> lynxlynxlynx has joined #GemRb
[06:05:45] --- ChanServ gives channel operator status to lynxlynxlynx
[08:57:27] <lynxlynxlynx> the -D_GLIBCXX_PROFILE run turned up pretty useless :)
[09:12:31] <fuzzie> ah well
[09:12:49] <fuzzie> if you look at the main profile there's not really any hints of the C++ lib being a problem :)
[09:39:30] --> cubathy has joined #GemRb
[09:40:53] --> budlust has joined #GemRb
[09:52:36] <-- cubathy has left IRC (Ping timeout: 245 seconds)
[10:00:24] --> cubathy has joined #GemRb
[10:54:10] --> Gekz_ has joined #GemRb
[11:10:10] <-- budlust has left #GemRb
[11:55:07] <-- Gekz_ has left IRC (Read error: Connection reset by peer)
[11:55:19] --> Gekz_ has joined #GemRb
[12:01:36] <-- Gekz_ has left IRC (Changing host)
[12:01:36] --> Gekz_ has joined #GemRb
[13:12:44] --> cubathy_ has joined #GemRb
[13:13:01] <-- cubathy has left IRC (Ping timeout: 245 seconds)
[14:56:42] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[16:35:00] <-- |Cable| has left IRC (Quit: Leaving)
[16:35:09] --> |Cable| has joined #GemRb
[16:35:15] <-- |Cable| has left IRC (Remote host closed the connection)
[16:36:23] --> |Cable| has joined #GemRb
[16:40:02] <-- cubathy_ has left IRC (Remote host closed the connection)
[16:55:33] <-- Gekz_ has left IRC (Quit: Leaving)
[17:30:24] --> cubathy has joined #GemRb
[17:42:05] <-- cubathy has left IRC (Ping timeout: 240 seconds)
[17:43:29] --> cubathy has joined #GemRb
[17:47:32] <-- cubathy has left IRC (Remote host closed the connection)
[17:59:09] --> cubathy has joined #GemRb
[18:30:23] <tomprince_loki> Any reason we have MYASSERT, rather than just using assert everywhere?
[18:46:10] <fuzzie> nope
[18:47:00] <tomprince_loki> Then it goes. :) Along with TRUE and FALSE.
[18:48:06] <fuzzie> i think MYASSERT just came from the old map code
[18:48:34] <tomprince_loki> That looks to be the case, yes.
[18:48:38] <fuzzie> but then it became apparent that you couldn't rely on STL having a useful map class.
[18:49:13] <fuzzie> because hashes are a completely new invention and they couldn't possibly have thought of such a thing at the time, or.. something.
[18:49:13] <tomprince_loki> couldn't? or could.
[18:49:38] <fuzzie> well, STL's 'map' is awful slow
[18:50:44] <fuzzie> we should steal scummvm's hash map, maybe. or just a random one.
[18:54:13] <tomprince_loki> If we are doing that, we should use unordered_map, if it exists.
[18:54:21] <fuzzie> but as far as I can tell the lack of hash_map on Windows (and the fact the previous code reverted to plain STL map) is the only reason any of the crazy Variables stuff is there.
[18:54:43] <-- cubathy has left IRC (Remote host closed the connection)
[18:57:25] <fuzzie> is there a portable implementation of unordered_map out there?
[18:58:07] <tomprince_loki> boost. :)
[18:58:23] <fuzzie> is it stealable?
[18:59:05] <tomprince_loki> It looks like it isn't entirely trivial to steal it.
[18:59:55] <fuzzie> Hum.
[19:02:37] <tomprince_loki> llvm includes a StringMap, that looks fairly easy to extract.
[19:03:24] <fuzzie> libc++?
[19:03:45] <fuzzie> that was my next thought, but someone phoned in the meantime :-)
[19:04:00] <tomprince_loki> I was talking of llvm itself.
[19:04:57] <fuzzie> oh, i hadn't realised they had a custom library
[19:05:01] <fuzzie> interesting
[19:05:51] <fuzzie> don't suppose you have any idea how usable cmake-for-boost is? :)
[19:06:11] <fuzzie> the above just set me thinking about yet more recent battles with boost.build, and i thought you might have anidea
[19:07:26] <tomprince_loki> No, I don't know.
[19:08:17] <tomprince_loki> I did mange to build with mingw under wine, awhile ago, I think.
[19:13:21] --> cubathy has joined #GemRb
[19:14:26] <tomprince_loki> llvm is meant to be the lost level in the toolchain, so I think they try to avoi depending on anything nonstandard/nonportable.
[19:15:35] <tomprince_loki> And they seem to care about size+speed effeciency.
[19:16:09] <tomprince_loki> They actually tend to use the low bits of pointers to store data. :)
[19:18:56] <fuzzie> i'm not sure that's caring about speed efficiency, really ;p
[19:20:29] <-- cubathy has left IRC (Ping timeout: 240 seconds)
[19:23:48] <tomprince_loki> Well, if you believe http://clang.llvm.org/features.html#performance
[19:26:49] <fuzzie> i think all LLVM's performance claims come with "on post-P4 Intel processors, under OS X" disclaimers
[19:26:58] <fuzzie> "and woe betide you if you don't have really fast bitshifting"
[19:27:30] <tomprince_loki> That may be true.
[19:27:31] <fuzzie> which, of course, doesn't matter at all for a compiler, since everyone's going to have some modern x86_64 processor to do their compiling on :)
[19:29:53] <fuzzie> it has been nice having some motivation for the gcc devs to speed up their compiler, though; gccspeed has actually been improving, a change of direction
[20:24:00] <CIA-24> GemRB: 03tom.prince * re80726cd836b 10gemrb/gemrb/ (5 files in 2 dirs): Get rid of TRUE and FALSE macros.
[20:24:01] <CIA-24> GemRB: 03tom.prince * r7857d4bfd27a 10gemrb/gemrb/ (6 files in 5 dirs):
[20:24:01] <CIA-24> GemRB: Variables: Move Variables iterator out of win32def.h.
[20:24:01] <CIA-24> GemRB: Signed-off-by: Tom Prince <tom.prince@ualberta.net>
[20:24:04] <CIA-24> GemRB: 03tom.prince * r3e35f3b3b85c 10gemrb/gemrb/core/ (Actor.cpp GameScript.h):
[20:24:04] <CIA-24> GemRB: GameScript: Make member variables private.
[20:24:04] <CIA-24> GemRB: Signed-off-by: Tom Prince <tom.prince@ualberta.net>
[20:24:06] <CIA-24> GemRB: 03tom.prince * r891854efc8fc 10gemrb/gemrb/ (5 files in 3 dirs):
[20:24:07] <CIA-24> GemRB: s/MYASSERT/assert/
[20:24:07] <CIA-24> GemRB: Signed-off-by: Tom Prince <tom.prince@ualberta.net>
[20:24:09] <CIA-24> GemRB: 03tom.prince * r8328913512d7 10gemrb/gemrb/includes/ (globals.h win32def.h): Move MAX_VARIABLE_LENGTH to globals.h.
[20:24:10] <CIA-24> GemRB: 03tom.prince * r75635261b15b 10gemrb/gemrb/includes/ (logging.h win32def.h):
[20:24:10] <CIA-24> GemRB: logging: Move logging definitions out of win32def.h.
[20:24:10] <CIA-24> GemRB: Right now, it just gets included from win32def.h, but that should be changed eventually.
[20:41:08] <tomprince_loki> Apparently our autotools build doesn't like being build in a sepearte directory: this is because of all the globbing we do.
[20:41:29] <tomprince_loki> Any reason we still need the autotools build?
[20:48:58] <fuzzie> i'm not sure
[20:52:23] <fuzzie> almost everyone is still using it, which doesn't help
[20:58:36] <fuzzie> maybe that could be tackled somehow if people (Ed, lynx, wjp?) think it's ok, sabotage it with a warning if a var isn't set
[20:59:59] <tomprince_loki> Change autogen to call cmake instead?
[21:01:06] <fuzzie> well, that too :)
[21:01:18] <fuzzie> but most people are not going to have it installed, etc
[21:03:12] <fuzzie> and i guess more to the point, no-one is going to run autogen
[21:51:05] --> cubathy has joined #GemRb
[21:56:17] --> cubathy_ has joined #GemRb
[21:57:10] <-- cubathy has left IRC (Read error: Connection reset by peer)
[22:00:46] --> Maighstir has joined #GemRb
[23:13:20] <tomprince_loki> I'd like to move all the Actions/Triggers/Objects etc... out of class GameScript, since they don't need any access to the internals.
[23:24:33] <fuzzie> oh?
[23:24:37] --> cubathy has joined #GemRb
[23:24:57] <-- cubathy_ has left IRC (Read error: Connection reset by peer)
[23:24:57] <fuzzie> i mean, what internals?
[23:25:07] <fuzzie> GameScript doesn't have much point other than to host those.
[23:25:27] <fuzzie> well, i suppose also to execute them, so ok.
[23:25:52] <fuzzie> you already tried moving them and nothing turns out to need access?
[23:26:14] <fuzzie> i can't think of any lurking unimplemented stuff that couldn't be handled as methods on GameScript
[23:33:21] <fuzzie> i don't see anything after glancing through IESDP either
[23:34:21] <fuzzie> so that sounds fine
[23:38:18] <fuzzie> really got to finish the work on blocking and fix cutscenes, though..
[23:41:02] <fuzzie> tomprince_loki: would you be opposed to trying to sort bits of core/ into some subdirectories?
[23:51:12] <-- cubathy has left IRC (Remote host closed the connection)