#gemrb@irc.freenode.net logs for 25 Apr 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[05:59:11] <-- Nomad010 has left IRC (Ping timeout: 260 seconds)
[06:17:11] <-- Gekz_ has left IRC (Quit: Leaving)
[06:21:35] --> Gekz_ has joined #GemRb
[09:32:51] --> Brendan__ has joined #GemRb
[09:34:04] <-- Gekz_ has left IRC (Read error: Connection reset by peer)
[10:21:43] --> Maighstir_laptop has joined #GemRb
[11:07:40] --- Brendan__ is now known as Gekz_
[11:07:49] <-- Gekz_ has left IRC (Changing host)
[11:07:49] --> Gekz_ has joined #GemRb
[11:21:03] --> raevol has joined #GemRb
[11:29:45] <Edheldil_> later
[11:29:48] <-- Edheldil_ has left IRC (Quit: Really?)
[11:48:23] <-- raevol has left IRC (Quit: Leaving.)
[12:28:09] --> tomprince_loki has joined #GemRb
[13:46:31] --> Nomad010 has joined #GemRb
[14:20:56] <-- Nomad010 has left IRC (Ping timeout: 246 seconds)
[14:36:33] --> Nomad010 has joined #GemRb
[14:44:52] <-- Gekz_ has left IRC (Quit: Leaving)
[18:07:24] <wjp> hm, so it looks like this TileRenderer_internal did in fact break on MSVC6?
[18:07:48] <wjp> would the workaround be to add an extra 'PixelType = 0' parameter to the end of the param list?
[18:10:05] --> lynxlynxlynx has joined #GemRb
[18:10:05] --- ChanServ gives channel operator status to lynxlynxlynx
[18:15:04] <fuzzie> wjp: that would be my first attempt, yes
[18:16:04] <fuzzie> but haven't had time to try it
[18:16:18] <wjp> I should install MSVC6 myself I guess
[18:16:55] <wjp> but it's just so annoying to install the various libraries in windows :/
[18:17:48] <fuzzie> it's not so bad for vc++, although i haven't tried python
[18:18:15] <fuzzie> i have a vc6 .iso downloaded from MSDNAA and the bits, just no time to set a VM up
[18:26:55] <tomprince_loki> The other choice would be to turn it into struct, like what was done in plugindef.h
[18:30:22] <fuzzie> although if all of you are going to have vc++6 around, i suppose i should install vc++10 instead
[18:31:35] * wjp wonders if msvc6 will work in xp64
[18:35:50] <tomprince_loki> Although, doing it as a struct would probably intail doing a member template inside the template class. :D
[18:36:17] <wjp> I think the dummy argument at the end would be easier :-)
[18:38:06] <tomprince_loki> I guess they are probably both as ugly.
[18:38:23] <fuzzie> It'
[18:38:29] <fuzzie> It's C++, it's all doomed to be ugly.
[18:39:18] * fuzzie draws toruses on you all.
[18:39:46] <tomprince_loki> The current code isn't *too* ugly there. And either msvc6 will make it worse.
[18:40:01] <wjp> well, a dummy argument doesn't really hurt all that much
[18:40:18] <wjp> it's just one extra line in the function header without any further impact on the code
[18:41:28] <tomprince_loki> Is it actually smart enough to pick up the type from default paramater when it is specified explicitly?
[18:41:50] <fuzzie> i'm told that it works fine.
[18:42:06] <fuzzie> it's just a bug in the optimiser which only checks the parameter list before folding.
[18:42:20] <fuzzie> the problem in plugindef.h was clearly something different.
[18:43:08] <wjp> speaking of ugly, any thoughts on how the selection of blender based on alpha/tinting being enabled could be doable with templates? (SDLVideoDriver.inl)
[18:43:30] <wjp> that's the large #ifdef ALPHA, #ifdef TINT, etc... block at the start of the .inl
[18:44:23] <fuzzie> You wouldn't just do it the same way as the TileRenderer?
[18:44:41] <wjp> the blender is a sort of derived argument
[18:45:29] <tomprince_loki> Briefly looking at, and from my attempts to simplify it, it seems like all the code paths do essentialy the same sequence of things.
[18:54:24] <-- Nomad010 has left IRC (Ping timeout: 276 seconds)
[18:54:50] --> Nomad010 has joined #GemRb
[19:19:11] <tomprince_loki> Relatedly, I have a (mostly?) working proof concept branch that compiles everything into gemrb staticly.
[19:20:49] <tomprince_loki> Depends on gnu LD right now, but that should be fixable.
[19:28:51] <fuzzie> and it still works dynamicall?
[19:29:40] <tomprince_loki> Right now? no.
[19:31:14] <fuzzie> but it could?
[19:31:16] <tomprince_loki> What I have right now is a hack-and-slash job, showing that it can be done.
[19:31:21] <tomprince_loki> It should be able to.
[19:33:12] <tomprince_loki> To get them both to work at once, we might need to change the plugin loading again, so that plugins register themselves, when static constructors run, rather than having PluginMgr call in to them.
[19:34:08] <fuzzie> the moment you say "static constructors", alarm bells go off in portability land :)
[19:34:39] <tomprince_loki> :)
[19:35:01] <tomprince_loki> Well, it is basically the trick in the classloading thing that I posted earlier.
[19:35:12] <tomprince_loki> And, we already have one, in CREImporter.
[19:35:48] <tomprince_loki> I ran into trouble with that, since when the plugins are compiled in statically, the constructors run beofre core exists.
[19:36:30] <fuzzie> or before anything exists, of course..
[19:36:44] <fuzzie> what is your thought, a singleton?
[19:37:21] <fuzzie> i have a large set of exams suddenly scarily close by, so i am likely to not be paying a lot of attention for a while, btw
[19:37:28] <tomprince_loki> Yes. That is what I did. PluginMgr exists on that branch as a function local static.
[19:37:53] <tomprince_loki> :)
[19:38:19] <fuzzie> just as well i don't have to merge everything any more :p
[19:38:22] <tomprince_loki> Well, I am gone for a couple of weeks, more or less. Perhaps somewhat less.
[19:38:41] <tomprince_loki> Have you tried the sound branch?
[19:40:24] <fuzzie> i haven't tried it, but i thought it was just reorganising code at this point.
[19:40:41] <tomprince_loki> I have done some testing, but I don't know all the things likely to break, that should be tested.
[19:41:02] <tomprince_loki> Well, it depends on the PathJoin stuff, which isn't just code reorganization.
[21:20:40] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[21:30:27] --> cfchris6 has joined #GemRb
[21:31:34] <-- cfchris6_ has left IRC (Read error: Operation timed out)
[22:55:16] --> raevol has joined #GemRb
[23:23:30] <-- raevol has left IRC (Quit: Leaving.)
[23:46:14] <-- Maighstir_laptop has left IRC (Ping timeout: 240 seconds)