#gemrb@irc.freenode.net logs for 19 Oct 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:37:10] <-- Maighstir has left IRC (Quit: ~ Trillian Astra - www.trillian.im ~)
[01:44:27] <-- pupnik has left IRC (Quit: Lost terminal)
[02:15:31] --> pupnik has joined #GemRb
[06:02:16] <-- edheldil has left IRC (Read error: Operation timed out)
[06:02:30] --> edheldil has joined #GemRb
[06:02:30] --- ChanServ gives channel operator status to edheldil
[07:09:54] --> lubos has joined #GemRb
[08:03:10] <-- |Cable| has left IRC (Remote host closed the connection)
[08:03:49] --> lynxlynxlynx has joined #GemRb
[08:03:49] --- ChanServ gives channel operator status to lynxlynxlynx
[09:49:54] <-- Lightkey has left IRC (*.net *.split)
[09:51:31] --> Lightkey has joined #GemRb
[10:53:20] <-- Lightkey has left IRC (Ping timeout: 240 seconds)
[11:11:51] --> Lightkey has joined #GemRb
[12:39:36] <-- tomprince has left IRC (Quit: leaving)
[15:22:59] --> Maighstir has joined #GemRb
[15:32:04] <-- lubos has left IRC (Quit: Leaving.)
[15:54:21] --> barra_home has joined #GemRb
[16:42:19] <-- barra_home has left IRC (Ping timeout: 240 seconds)
[16:46:19] --> spike411 has joined #GemRb
[17:06:30] --> |Cable| has joined #GemRb
[17:24:04] --> barra_home has joined #GemRb
[19:15:45] --> nickdaly has joined #GemRb
[19:24:03] <nickdaly> fuzzie, would you mind checking out the patch at http://github.com/NickDaly/GemRB-MultipleConfigs ? I think it's ready for committing, if you're interested.
[19:28:09] <fuzzie> i haven't been committing for quite a while now, so i doubt i am the best to ask right now
[19:43:30] <nickdaly> Do you know who would be good to ask?
[19:49:09] <lynxlynxlynx> burp
[19:49:43] <lynxlynxlynx> my opinion on this hasn't changed
[19:50:30] <fuzzie> well, some of it is definitely worth pulling, but tomprince is the other person good with this stuff, also not committing recently..
[19:54:38] <lynxlynxlynx> is there a simple way of seeing the diff?
[20:02:11] <fuzzie> http://fuzzie.org/nfs/gemrb/multipleconfigs.txt maybe?
[20:02:21] <fuzzie> that being the diff from the time, not up-to-date
[20:03:23] <fuzzie> (output of git diff <last commit merged with>..FETCH_HEAD)
[20:04:36] <lynxlynxlynx> it'd be odd if github had no feature to deal with this
[20:04:58] --> nickdaly` has joined #GemRb
[20:06:13] <fuzzie> it isn't forked from a github repository, so i'm not sure it does
[20:06:38] <-- nickdaly has left IRC (Ping timeout: 264 seconds)
[20:12:18] <-- spike411 has left IRC (Quit: spike411)
[20:19:18] <nickdaly`> it should be possible to diff it with the latest checkout
[20:19:22] <nickdaly`> let me look
[20:20:56] <lynxlynxlynx> no need, it looks like fuzzie's version is pretty recent
[20:21:31] <lynxlynxlynx> probably latest
[20:30:21] <lynxlynxlynx> rereading it just reinforced my view
[20:35:00] <lynxlynxlynx> fuzzie: what do you see as salvageable?
[20:35:53] <fuzzie> refactoring out getConfigDirectory and splitting out FixConfig()
[20:36:02] <fuzzie> latter is not so easy to pull out of the diff though
[20:36:30] <fuzzie> i have no real opinion on the rest
[20:38:27] <fuzzie> although i'm not sure i heard a really better alternative
[20:39:22] <lynxlynxlynx> to me it looks like a confusing fix at a wrong location (it is the installer that has issues)
[20:39:51] <lynxlynxlynx> but i think all was already said in the bug reports
[20:39:57] <fuzzie> yes
[20:40:25] <nickdaly`> LoadConfig(void)'s codepath is also somewhat cleaner - the only special case is a small non-win32 block
[20:40:25] <fuzzie> while to me, it seems that a global config is always going to be wished for by many users
[20:40:44] <fuzzie> and i'm not sure we can avoid needing game-specific config for many users, either
[20:41:26] <lynxlynxlynx> GamePath and the CDs will always be specific
[20:42:25] <lynxlynxlynx> the gemrb paths can already be ignored, GameType can also be deduced and most of the rest are copies from the game inis
[20:42:33] <fuzzie> yes, but the gemrb paths can't really be ignored
[20:42:43] <fuzzie> we don't even have make install/uninstall which work properly
[20:43:01] <lynxlynxlynx> that's news to me
[20:43:41] <fuzzie> but this isn't to say i see the multiple config paths as being helpful
[20:44:29] <fuzzie> just that every time we spend half an hour on irc trying to make someone's gemrb work, i wish the config was more magical
[20:45:01] <lynxlynxlynx> that's the goal, yes
[20:45:21] <lynxlynxlynx> (also what's wrong with un/install?)
[20:45:33] <fuzzie> but by then they've usually already 'make install'ed a tarball and 'make install'ed git, and it is a huge mess because .pyc files are lying around and files got moved and etc
[20:46:00] <fuzzie> and i don't even start to understand how to make cmake do that right..
[20:46:44] <fuzzie> maybe you do
[20:47:24] <lynxlynxlynx> there's not much we can do if people install over installs (without uninstalling)
[20:47:59] <lynxlynxlynx> we could ignore the .pyc though, i doubt they provide much of a speedup and python isn't the bottleneck yet anyway
[20:48:07] <fuzzie> well, it is the first thing everyone does, since there's no obvious way to run gemrb from the install directory
[20:48:43] <fuzzie> well, i can't be right about the .pyc file, you'd think the install dirs would be root-only
[20:48:59] <fuzzie> i guess i must just remember people who didn't get the old .py files removed at all
[20:49:30] <lynxlynxlynx> we also can't be blamed for crappy package managment
[20:49:50] <lynxlynxlynx> ignoring the pycs would also help us when we're cleaning the scripts
[20:50:15] <fuzzie> i think the problem is that we can't easily ignore the pyc files
[20:50:28] <fuzzie> because the setting is only present starting from 2.6 or so?
[20:51:44] <lynxlynxlynx> that would cover many though
[20:52:41] <fuzzie> hm
[20:52:54] <fuzzie> i hadn't even thought about just trying it and not worrying if it fails :)
[20:54:55] <lynxlynxlynx> i'll look it up
[20:55:36] <lynxlynxlynx> yeah, it's from 2.6
[20:58:36] <fuzzie> so an unchecked PyRun_SimpleString call at GUIScript.cpp:9812 would seem the best way to do that
[21:03:00] <CIA-102> GemRB: 03fuzzie * re0ae625c3cd0 10gemrb/gemrb/plugins/GUIScript/GUIScript.cpp: tell python not to write .pyc files, since they cause us a lot of confusion
[21:05:00] <lynxlynxlynx> oh, thanks
[21:05:27] <fuzzie> sorry if you were doing it, i just keep suggesting things and figure someone else will commit them and they get forgotten
[21:05:45] --> nickdaly has joined #GemRb
[21:06:06] <lynxlynxlynx> there's Py_DontWriteBytecodeFlag, but i haven't found if it is public yet
[21:06:56] <fuzzie> well, i figured that this way, it will just silently fail on 2.5, i don't know if anyone uses that
[21:07:12] <fuzzie> but feel free to stomp over it with whatever
[21:07:18] <lynxlynxlynx> nah, it's fine
[21:07:25] <-- nickdaly` has left IRC (Ping timeout: 265 seconds)
[21:10:15] <lynxlynxlynx> oh and regarding the forsaken suggestions - just put them on the wiki when you're unsure if somebody is working on it
[21:12:32] <fuzzie> if only i were so organise :)
[21:12:44] <fuzzie> but yes, i should dig through logs
[21:18:43] <nickdaly> do you think the multiple configs stuff is a separate enough feature request that it has a might be handled separately from the (un)installation issues, or should I close the ticket at some point?
[21:21:23] <lynxlynxlynx> you can help us by splitting out the patches for the two functions fuzzie mentioned
[21:21:43] <lynxlynxlynx> too bad the guy who was doing the gametype autodetection vanished
[21:22:08] <lynxlynxlynx> he had a partial implementation already done iirc
[21:32:39] <nickdaly> mind if I simplify LoadConfig(), too? the code duplication in that function really irks me
[21:33:01] <lynxlynxlynx> sure, just make it a separate commit
[21:33:10] <nickdaly> fair enough
[21:51:12] <pupnik> :
[22:01:07] --> SiENcE has joined #GemRb
[22:01:49] <-- SiENcE has left IRC (Client Quit)
[22:02:34] <Lightkey> ...
[22:05:25] <-- nickdaly has left IRC (Read error: Operation timed out)
[22:06:32] --> nickdaly has joined #GemRb
[22:12:06] <-- Maighstir has left IRC (Quit: ~ Trillian Astra - www.trillian.im ~)
[22:19:13] <CIA-102> GemRB: 03avenger_teambg * ra1393d482246 10gemrb/gemrb/core/GameScript/GSUtils.cpp: use dialoguerange that was already set
[22:24:19] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[23:13:36] <-- Lightkey has left IRC (Ping timeout: 240 seconds)
[23:14:06] --> deepinthewoods has joined #GemRb
[23:31:32] --> Lightkey has joined #GemRb
[23:52:50] <nickdaly> lynx and fuzzie, the fixconfig stuff is available at git://github.com/NickDaly/GemRB-FixConfig-Branch.git - the first commit just breaks out FixConfig and getConfigDirectory, the second greatly simplifies the logic in LoadConfig(), removing duplicate code.