#gemrb@irc.freenode.net logs for 7 Apr 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:00:29] --> ratpor has joined #GemRb
[00:02:31] <-- Gekz_ has left IRC (Quit: Leaving)
[01:05:19] <-- raevol has left IRC (Quit: Leaving.)
[02:13:41] <-- _pickle has left IRC (Remote host closed the connection)
[03:09:01] <-- ratpor has left IRC (Quit: Get used to disappointment.)
[04:44:38] <tomprince_loki> or.cz/cleanups: GameData and FindInDir refactorings.
[05:04:26] --> Nomad010 has joined #GemRb
[05:08:42] --> nomad__ has joined #GemRb
[05:08:56] <-- Nomad010 has left IRC (Ping timeout: 258 seconds)
[05:10:07] --> Gekz has joined #GemRb
[05:32:13] --> raevol has joined #GemRb
[05:42:18] <tomprince_loki> The FileGlob path hasn't been tested at all.
[06:07:42] --> lynxlynxlynx has joined #GemRb
[06:07:42] --- ChanServ gives channel operator status to lynxlynxlynx
[06:12:34] <-- nomad__ has left IRC (Ping timeout: 258 seconds)
[07:38:27] --> edheldil has joined #GemRb
[07:41:31] --> edheldil_ has joined #GemRb
[07:42:04] <-- edheldil has left IRC (Client Quit)
[07:42:25] --- edheldil_ is now known as Edheldil
[07:44:45] --- ChanServ gives channel operator status to Edheldil
[07:49:20] <wjp> oops :-)
[07:51:16] <Edheldil> indeed
[07:51:34] <Edheldil> and thanks to exultbot it's all over the net now :)
[07:51:48] <wjp> already gone
[07:52:25] <Edheldil> I do not doubt that logs like this get mined for passwords
[07:52:40] <Edheldil> thanks, I have already changed it :)
[07:54:27] <Edheldil> but it's better to be deleted
[07:54:39] <Edheldil> anything new?
[08:40:56] <-- raevol has left IRC (Remote host closed the connection)
[08:43:22] --> raevol has joined #GemRb
[08:45:59] <-- |Cable| has left IRC (Remote host closed the connection)
[08:48:38] <fuzzie> hello
[08:53:31] <fuzzie> tomprince_loki: ResourcePath a bit of a confusing name?
[08:58:07] <fuzzie> i would think ResourceSource/AddSource but that's still not ideal
[09:02:45] <fuzzie> Otherwise it seems fine. I don't particularly like handing a plugin ID over, but it does help to make it clear that the interfaces are 'private' now.
[09:09:55] <fuzzie> Edheldil: i hope the silly bg2 UI flickering is fixed, that is new? :)
[09:10:17] <fuzzie> I am still fiddling with the patch to fix the dialog flickering.
[09:21:30] <Edheldil> I have only PS:T installed at the moment.
[09:22:00] <fuzzie> Oh, well, poor PS:T. :( I fixed the map, but I think that is all.
[09:22:29] <Edheldil> btw, when I quit gemrb, it hangs somewhere in OpenAL futex. But I haven't tried your yesterday OpenAL patch yet
[09:23:17] <fuzzie> Oh, you should mention these things :)
[09:24:53] <Edheldil> I want to try it on another computer first
[09:29:18] <fuzzie> well, there are some bugs in there which i think won't show up for everyone
[09:29:58] <fuzzie> well, potential bugs :)
[09:30:56] <Edheldil> I will have to check, but wherever I hear lockup in audio futex, Pulse Audio comes to mind
[09:30:59] <fuzzie> OpenALAudio.cpp:799 is weird too - that 'if' should always evaluate to true
[09:35:07] <fuzzie> But we should be taking the mutex much earlier. I would try moving line 204 to the top of the destructor and see if that helps.
[09:36:19] <fuzzie> Oh, can't do that, ResetMusics takes the lock. Just after that call, then.
[09:37:02] * Edheldil can't look at it at the moment
[09:37:44] <fuzzie> well, if you could find the time sometime (no rush!) I would appreciate knowing if it helps - while it might be Pulse causing it, it might well be our bug!
[09:39:19] <Edheldil> I will look at it today, I hope
[09:41:16] <fuzzie> but PS:T is difficult in general, the GUIScripts are very broken and we are missing so many features
[09:41:39] <fuzzie> and I never know whether I want to fix BG2 to be perfect, or fix PS:T to be playable :)
[09:51:06] <fuzzie> File "./GUIScripts/pst/GUICommonWindows.py", line 592, in CheckLevelUp
[09:51:06] <fuzzie> GemRB.SetVar ("CheckLevelUp"+str(pc), CanLevelUp (pc))
[09:51:06] <fuzzie> File "./GUIScripts/LUCommon.py", line 39, in CanLevelUp
[09:51:06] <fuzzie> Class = ClassTable.FindValue (5, Class)
[09:51:06] <fuzzie> AttributeError: 'NoneType' object has no attribute 'FindValue'
[09:53:50] <Gekz> fuzzie: I need a little bit of help
[09:54:02] <Gekz> I'm using BGT withhout GemRb, and I have come across a bug
[09:54:10] <Gekz> that requires I use MoveToArea as one of the triggers is missing
[09:54:19] <Gekz> the issue is, the mod has absolutely fucked up all of the ARs
[09:54:27] <Gekz> I managed to get Aerie into BG1...
[09:54:30] <Gekz> lets not go there
[09:54:41] <Gekz> so my question is: how can I find the correct AR
[09:54:42] <Gekz> haha
[09:54:45] <fuzzie> this bug is maybe a bit self-induced? :)
[09:55:33] <fuzzie> i would just load the save/script in DLTCEP and work it out from there
[09:55:56] <Gekz> aha
[09:56:11] <Gekz> and no, when I typed in the AR it was meant to be
[09:56:13] <Gekz> it took me to Aerie
[09:56:20] <Gekz> she was like HAVE YOU FOUND QUAYLE?!!
[09:58:33] <wjp> maybe try some of the general modding forums instead of #gemrb if it's not a gemrb issue :-)
[10:00:57] <Gekz> never!
[10:02:22] <fuzzie> pretty sure we are not the best people to ask :)
[10:38:27] <Gekz> something you should know though
[10:38:37] <Gekz> fuzzie: what's the best way to find changes in a variable
[10:38:39] <Gekz> in a save game
[10:55:44] <raevol> fuzzie you should fix BG1 to be perfect :P start at the beginning
[10:56:06] <raevol> just my two cents ;)
[10:56:22] <fuzzie> i mostly just made BG1 worse :( even clicking the sign at the start is wrong now
[10:56:29] <raevol> hehehe
[10:56:56] <raevol> i suppose developing for BG2 and then fixing BG1 is a better way to approach the engine design... 2 has more features?
[10:56:59] <fuzzie> my terrible secret is: i never managed to make it past the mines in bg1.
[10:57:12] <raevol> dude, BG1 has such a classic story :(
[10:57:35] <fuzzie> i watched my brother play most of it, but it was a *long* time ago :)
[10:57:42] <raevol> and the world is much more immersive than BG2...
[10:57:53] <raevol> they're both good in their own ways though
[10:59:31] <fuzzie> but if you want to suggest the most annoying BG1 bugs, i can have a look
[10:59:39] <fuzzie> i guess spawn difficulty is still a big one?
[10:59:59] <raevol> well i haven't tested it since i talked to you guys about it last
[11:00:07] <raevol> hadn't heard there was any fix for it
[11:00:21] <raevol> http://sourceforge.net/tracker/?func=detail&aid=2536293&group_id=10122&atid=110122
[11:00:53] <fuzzie> i can definitely try and fix that.
[11:01:02] <raevol> :>
[11:01:33] --> Maighstir_laptop has joined #GemRb
[11:03:50] <fuzzie> A prioritised list would maybe be nice, though. The savegame problems are on my 'important' list.
[11:04:10] <raevol> ah i see
[11:06:02] <raevol> god dammit i can never remember how to move things i select in GIMP
[11:10:29] <raevol> k figured it out :P
[11:10:30] <raevol> heh
[11:16:34] <raevol> hmm i haven't updated to git with you guys
[11:16:58] <raevol> if i remove my gemrb directory am i going to lose my old game data? not saves, but the actual game data?
[11:17:17] <fuzzie> well, where did you put the data?
[11:17:23] <raevol> it looks like yes, i have a bg1 folder in there
[11:17:46] <fuzzie> maybe a good idea to move it out :)
[11:17:55] <fuzzie> i just renamed my old gemrb directory, to be sure i didn't lose anything.
[11:18:21] <raevol> yea i thought i had all the game data saved somewhere else too..
[11:18:25] * wjp still has an old CVS one :-)
[11:19:48] <fuzzie> there is not very much interesting for bg1 in git so far, i think. the bg1 load screen should look better, fog-of-war updates on the map, some random UI fixes, magic missiles.
[11:20:08] <raevol> yea but git is so much nicer to update from
[11:20:25] <raevol> i actually remember the console commands instead of having to look them up -.-
[11:29:04] <raevol> i can't seem to configure or make, is this because i need to chmod +x ?
[11:29:19] <wjp> do you get an error?
[11:29:42] <raevol> well there's no configure file, i guess that's part of it, unless it's configure.in now?
[11:29:44] <wjp> oh, and you have to run autogen.sh before
[11:30:03] <raevol> that'd do it...
[11:30:17] <raevol> man i always feel like such a noob when i'm trying to build this :P
[11:34:16] <raevol> cool, up and running
[11:34:28] <raevol> heh my large resolution thing *still* carries over, pretty cool
[12:06:14] <tomprince_loki> fuzzie: Updated to use ResourceSource.
[12:07:07] <CIA-43> GemRB: 03lynxlupodian * r10658ff8e972 10gemrb/gemrb/GUIScripts/pst/GUICommonWindows.py: pst: try to fix an undefined ClassTable
[12:07:09] <lynxlynxlynx> try with that, whatever you were doing
[12:16:09] <tomprince_loki> fuzzie: Much better name. I didn't like it either, just couldn't think of anything better.
[12:44:16] <fuzzie> Drat, my past sins are catching up with me.
[12:45:05] <fuzzie> I'm a semester late with the homework for one course, so I figured it was a lost cause and I didn't have to worry about it any more. Poked my head in the lecturer's door today - "Hi, Alyssa! You're a bit late with the homework for my course, you know. You'd better hand it in before I start subtracting points".
[12:50:32] <fuzzie> lynxlynxlynx: That seems to have done it, thanks.
[12:50:52] <tomprince_loki> before? :)
[12:52:55] <tomprince_loki> That seems crazy.
[12:53:31] <tomprince_loki> Although ... I've had a couple presentations like that.
[12:58:47] <fuzzie> The patch looks fine, although maybe it would make more sense to apply it in two (s/ResourceMgr/ResourceSource/ first?).
[13:15:49] <fuzzie> I'm not losing reputation from killing random Flaming Fist folken, is that right?
[13:17:37] <tomprince_loki> Done.
[13:20:32] <fuzzie> Oh! :) I didn't expect you to actually do the work!
[13:21:59] <tomprince_loki> Why not? :)
[13:22:15] <fuzzie> The bg1 portrait icons are all broken for me, too. And the 'Bonus Priest Spells' has the '1' and the ':' so close that it looks like they're one letter (a funny t). Pretty sure Imoen shouldn't have bonus Priest spells anyway, really..
[13:42:25] <tomprince_loki> Anyway, you put up with all my nitpicking yesterday. :)
[14:01:58] <fuzzie> We don't use GetGlobalTint() for actual tinting, I notice.
[14:03:25] <fuzzie> tomprince_loki: was appreciated :)
[14:07:34] <-- raevol has left IRC (Quit: Leaving.)
[14:16:06] * fuzzie kicks CIA-43
[14:16:06] <CIA-43> ow
[14:17:38] <lynxlynxlynx> yay
[14:19:31] <fuzzie> i didn't actually test the iwd variant, so that might be a good idea
[14:25:39] <tomprince_loki> One thing to say for cmake is that it is faster. Although I think that might be almost entirely due to libtool.
[14:29:41] <lynxlynxlynx> configure does a bunch of unnecessary checks too
[14:30:28] <tomprince_loki> Yes.
[14:31:00] <tomprince_loki> or.cz/cleanup: Patch to get rid of all explicit uses of TypeIDs, and hide them behind a type-safe template.
[14:32:02] <fuzzie> Do you think that's portable?
[14:32:16] <fuzzie> Oh, wait, ResourceManager lives in Core, nm.
[14:33:01] <tomprince_loki> Why wouldn't it be portable?
[14:33:42] <tomprince_loki> Just curious, in case I go do something else tricky with templates later.
[14:33:53] <fuzzie> I wouldn't be sure that everything coped with template exports across library boundaries.
[14:34:12] <fuzzie> The main rule about templates is "no partial template specialization", I think.
[14:37:05] <fuzzie> (see boost's compiler configs for a pretty good idea, if you're really interested)
[14:37:53] <tomprince_loki> Yeah, I have seen some of the stuff that goes on there. :)
[14:39:38] <tomprince_loki> All the casts that were involved with the Plugin system have been annoying me, and the code is finally simple enough to get rid of them. :)
[14:41:35] <fuzzie> ok. i pushed two of your commits.
[14:41:45] <fuzzie> will bbl.
[14:51:47] <-- tomprince_loki has left IRC (Ping timeout: 265 seconds)
[14:52:12] <CIA-43> GemRB: 03fuzzie * r6def4dea6e82 10gemrb/gemrb/plugins/Core/Actions.cpp: another attempt to fix the DayNight logic
[14:53:10] <CIA-43> GemRB: 03fuzzie * rb2455539d4eb 10gemrb/gemrb/ (6 files in 6 dirs): add NextDialogState functions to GUIScripts and use it to avoid flickering
[14:53:48] <CIA-43> GemRB: 03tom.prince * r4026a1fe4179 10gemrb/gemrb/GemRB.cpp: main: Get rid of extraneous header files.
[14:53:48] <CIA-43> GemRB: 03tom.prince * rc89e96a874a8 10gemrb/gemrb/plugins/ (10 files in 3 dirs):
[14:53:48] <CIA-43> GemRB: ResourceSource: Rename from ResourceMgr.
[14:53:48] <CIA-43> GemRB: This is to free up the name ResourceManager.
[15:20:52] <Edheldil> tomprince: are templates MSVC6-safe?
[15:44:26] --> nomad__ has joined #GemRb
[16:03:28] <fuzzie> they seem to be fine
[16:07:58] <Edheldil> good ... MSVC6 compatibility is always a showstopper for any "advanced" C++ features :)
[16:46:57] --> tomprince_loki has joined #GemRb
[16:48:37] <fuzzie> things like Symbian and WinCE can be a huge pain too
[16:54:24] <tomprince_loki> We already use templates for the plugins, so the new use shouldn't break things more than it already was.
[16:55:53] <fuzzie> plugin templates are used across library bounaries?
[16:59:12] <tomprince_loki> Not really.
[16:59:25] <fuzzie> i mean, i doubt your patch will work as-is
[17:00:09] <fuzzie> but it shouldn't be a problem making it work.
[17:00:33] <tomprince_loki> We instantiate the CreateResource template in each plugin, and take a pointer to it.
[17:01:05] <fuzzie> mhm, but the template usage is entirely internal.
[17:01:27] <fuzzie> but in this new patch you're using templates which are instantiated as members of an imported class.
[17:01:53] <fuzzie> and i suspect we might have to specifically tag each instance of those for exports, but i'm not sure.
[17:01:56] <tomprince_loki> For the patch I posted, any sane compiler should just inline the template function call.
[17:02:17] <fuzzie> i'm not sure that's allowed behaviour. i don't know.
[17:03:00] <fuzzie> you'd be running code in an object marked as "provided by another library" in your own library context, it seems like that would often break things.
[17:06:24] <tomprince_loki> I know here, that the function is inlined, and the symbol doesn't appear in the plugins.
[17:06:31] <fuzzie> by msvc?
[17:07:13] <tomprince_loki> I don't have MSVC :), I don't even have access to a machine w windows.
[17:07:15] <fuzzie> I mean, typical *nix doesn't have these kind of problems - you have the one C library, for example.
[17:07:37] <fuzzie> As far as I know gemrb doesn't even mark symbols as import/export other than on Windows?
[17:08:37] <tomprince_loki> We do now.
[17:08:44] <fuzzie> We seem to only tweak visibility.
[17:09:34] <tomprince_loki> As far as I know, that is exactly what import/export is.
[17:09:43] <tomprince_loki> Perhaps I am mistaken.
[17:10:01] <fuzzie> MSVC import/export marks a module boundary, affecting compilation.
[17:11:06] <fuzzie> I'm not exactly sure how, but I'm fairly sure that it prevents things like inlining.
[17:12:09] <fuzzie> Would be nice if the obvious MSDN docs were actually up-to-date.
[17:13:18] <fuzzie> The answer seems to be "it depends on what the optimiser decides to do".
[17:14:34] <fuzzie> In any case, you can manually instantiate the templates in Core with the dllexport attribute, which is what the documentation recommends to be sure.
[17:17:24] <fuzzie> (I've only had experience fighting modern MSVC with this, where some STL code is in the MSVC C library and so imported, and some of it isn't. Hours of fun.)
[17:25:42] <-- tomprince_loki has left IRC (Read error: Connection reset by peer)
[17:26:10] --> tomprince_loki has joined #GemRb
[17:32:18] <tomprince_loki> Well, I'll leave it to someone who actually has access ot MSVC. I suspect that there is a way to tell MSVC to just inline the thing and be done with it, especially since the function is a single line that adds a paramater, and casts the return.
[17:32:47] <fuzzie> I guess the real difference here is: .so files on Linux are generally just adding to one huge symbol pool for the process, without any clear seperation of any kind? At a glance it seems really just a "let's link these at runtime as an .so, rather than at compile time as an .o" plan?
[17:33:09] <fuzzie> That's certainly quite a crazy model from a Windows point of view.
[17:35:05] <tomprince_loki> Well, that was why I wanted to turn on RTLD_LOCAL on for loading the plugins, so that the plugins don't put anything in the global namespace.
[17:35:46] <fuzzie> Still, that is simply a visibility thing, there's no kind of attempt at seperation at the library boundary, it seems.
[17:36:27] <fuzzie> If you tried linking your application against one C library and an .so against a completely different C library, it seems doomed to failure.
[17:36:51] <fuzzie> (I suspect __forceinline would work to force inlining, but I'd have to try.)
[17:37:36] <tomprince_loki> Well, the Linux Standard Base tried to define a portable ABI.
[17:38:25] <tomprince_loki> The trick is, you need a stub library+headers to link against, to make sure that you don't use any non-standard library features.
[17:39:11] <tomprince_loki> I do find that updating the c library works fairly well, as opposed to switching out the c library.
[17:39:24] <fuzzie> Mhm.
[17:39:43] <fuzzie> But a random application on Windows can have several C libraries in one address space, with very different ABIs.
[17:39:55] <fuzzie> Different heaps, etc.
[17:40:12] <tomprince_loki> I am not entirely sure why you would want that.
[17:40:34] <tomprince_loki> Seems like it would be more trouble that it is worth.
[17:40:41] <fuzzie> Because one C library is written by Microsoft, and one is written by Borland, and one is written by Intel, etc.
[17:41:40] <fuzzie> At the time any of this was designed, there would've been absolutely no chance of having a single "system C library"; if for no other reason than that their ABIs were nowhere near stable.
[17:42:28] <tomprince_loki> I guess part of the difference is that the dynamic linker on linux is part of the C library ....
[17:42:56] <fuzzie> Making it rather required to have your single C library. :)
[17:43:04] <tomprince_loki> And I think that you could probably have multiple C libraries on linux by playing games with RTLD_LOCAL.
[17:43:07] <fuzzie> It leads to some *awful* problems, obviously.
[17:44:31] <fuzzie> Just trying to deploy the right C library on Windows with your app became a bit of a nightmare. They've backed down a bit on that front with MSVC10, you can just bundle the dll in the directory with your app, again.
[17:44:51] <tomprince_loki> I think another difference is that on windows, there is the Win32 api, which is the OS interface, which doesn't change between compilers.
[17:45:09] <tomprince_loki> Where on linux, the C library is essentially the *only* interface to the OS.
[17:45:12] <fuzzie> Yep.
[17:45:28] <fuzzie> But then again, on Linux the system call interface is stable, while it isn't on Windows.
[17:45:53] <tomprince_loki> It's not? How do they manage that?
[17:46:28] <fuzzie> The Windows system calls aren't public API, they're only made from inside the OS libraries.
[17:47:15] <tomprince_loki> I guess I was thinking of the OS library interface as the syscall boundary.
[17:49:59] <fuzzie> I am still more familiar with the Windows model, despite having used Linux as my primary OS for a long time, just because you have to worry about all these things on Windows, while on Linux I can generally assume I have the source to everything and it'll just work anyway. Bit sad.
[17:50:23] <fuzzie> Anyway, this kind of madness is why this export/import stuff is a lot flakier on Windows.
[17:50:41] <fuzzie> And hence why I'm so paranoid about it.
[17:53:00] <tomprince_loki> That isn't a bad thing. (The being paranoid.
[17:53:29] <tomprince_loki> SIlly English langauge that doesn't have a concise way of making the explicit.
[17:53:43] <tomprince_loki> (I am just reading the lojban mailing list in the other window :))
[17:53:46] <fuzzie> I came to gemrb with all kinds of idealistic views about making it actually useful to people and I haven't lost it yet. :)
[17:54:18] <tomprince_loki> :)
[17:56:08] <fuzzie> By the way, I could easily be *wrong* about any of that Windows stuff, it's coming from experience rather than any kind of actual sources.
[17:56:16] <fuzzie> Anyhow, let me peer at this homework.
[17:57:49] --> |Cable| has joined #GemRb
[18:03:38] <fuzzie> tomprince_loki: Your MUSImporter patch is rebased?
[18:04:35] <tomprince_loki> Rebased, and broken.
[18:04:37] <fuzzie> I'm confused about the bit after "// FIXME: Get rid of GetResourceDesc".
[18:04:42] <fuzzie> Ah, okay.
[18:06:04] <tomprince_loki> That should go, I thought I got rid of that FIXME, but I must have missed it.
[18:06:44] <tomprince_loki> The problem with the MUSImporter patch is that it expects to be able to pass a path to DirectoryImporter, but it can no longer do that.
[18:07:44] <tomprince_loki> I do plan to fix it, but that is a more involved change. I plan to switch DirectoryImporter back to ResolveFilePath, but with the effeciceny of not checking every component every time.
[18:08:30] <fuzzie> *nod*
[18:09:02] <fuzzie> My original thought was simply to add a "start at string position X" parameter to ResolveFilePath.
[18:09:24] <tomprince_loki> That is more or less the thought I had in my head. *grin*
[18:09:55] <fuzzie> But then I decided I didn't really want to be messing with that interface quite then, didn't realise you needed subdirectories for MUSImporter until now.
[18:11:52] <tomprince_loki> I was also going to go through all the other callers of ResolveFilePath, and see what they are doing, and if they need to be doing it where they are.
[18:12:25] <fuzzie> Your refactoring patch (next in line) shouldn't be touching ResourceSource.h?
[18:15:02] <fuzzie> AddSource() is also modifying an argument with ResolveFilePath; should either note that in a comment, or make it const and take a copy.
[18:15:46] <fuzzie> would also be nice if you could change the prototype to be 'PluginID type'?
[18:16:39] <fuzzie> and description[3] should be description[2]?
[18:17:54] <tomprince_loki> Yes, and yes and yes.
[18:18:08] <tomprince_loki> I'll do that in a bit (seminar in 12min).
[18:18:20] <fuzzie> I also have no idea if 'char varname[] = "blah";' is valid syntax, but that's just me having no clue.
[18:19:02] <fuzzie> I guess it's all over the codebase, so that's my New Thing For Today.
[18:22:21] <-- tomprince_loki has left IRC (Quit: tomprince_loki)
[18:22:35] <fuzzie> I guess '-pedantic' is the magical flag which stops g++ from happily allowing random non-ISO junk.
[18:31:46] --> edheldil_ has joined #GemRb
[18:32:59] <-- edheldil_ has left IRC (Client Quit)
[18:33:49] --> edheldil_ has joined #GemRb
[18:51:56] <edheldil_> moving the mutex after ResetMusci() did not help, still locking on quit. But I can't hear anything anyway and get "Music in INITIAL state, autostarting" in terminal
[18:55:04] <fuzzie> a lot of those?
[18:56:51] <edheldil_> not a lot, I saw one after load
[18:56:59] <edheldil_> and one after startup
[18:57:16] <edheldil_> did not look much in scrollback, though
[19:01:11] <fuzzie> well, if there are any other errors/notices there, they might be helpful
[19:05:10] <fuzzie> Putting the mutex at the top of the destructor would still be the right thing to do, too.
[19:05:54] <edheldil_> only [Core]: Starting up the Sound Driver...[OpenAL]: Allocated 30 streams.[OK]
[19:06:43] <edheldil_> and one AutoStarting message after loading main_01.acm
[19:09:30] <edheldil_> hmm, looks like I can't get ANY sound at the moment, might be some suspend/resume issue on the notebook
[19:09:44] <edheldil_> sorrt
[19:09:47] <edheldil_> sorry
[19:10:18] <fuzzie> well, i think we should really be handling that situation, if it is our bug :)
[19:10:57] <fuzzie> but nice to know the lack of sound is not gemrb-specific
[19:21:21] <edheldil_> I will try to reboot
[19:24:16] <-- edheldil_ has left IRC (Read error: Connection reset by peer)
[19:26:05] --> edheldil_ has joined #GemRb
[19:26:28] <edheldil_> yes, reboot fixed sound and lockup. Sorry :)
[19:39:58] <fuzzie> meh, where is fist added to actors?
[19:45:14] <edheldil_> hmm, actually, I managed to get lockup now too, so apparently it's not so bugfree :)
[19:45:21] <fuzzie> Ah, it is being added at CRE load, and then immediately destroyed by DestroyAllEquipment().
[19:45:29] <edheldil_> heh
[19:45:36] <fuzzie> Dear gemrb, please do not remove my fists. :(
[19:45:52] <-- Maighstir_laptop has left IRC (Quit: Maighstir_laptop)
[19:47:59] --> tomprince_loki has joined #GemRb
[19:51:58] <fuzzie> that is another question for Avenger, maybe
[19:54:29] <lynxlynxlynx> you don't get any now?
[19:54:42] <lynxlynxlynx> it used to work
[19:57:29] <fuzzie> this is just a script bug, i guess
[19:57:50] <tomprince_loki> or.cz/cleanup: Updated.
[20:02:37] --> Maighstir_laptop has joined #GemRb
[20:06:34] <CIA-43> GemRB: 03benkovsk * r58bc281c1bdb 10gemrb/gemrb/GUIScripts/pst/GUICommonWindows.py: Fix timeformat of PST main window clock
[20:07:21] <fuzzie> edheldil_: fix your email? :)
[20:08:00] <edheldil_> hehe, just cursing at it :)
[20:15:37] <-- tomprince_loki has left IRC (Ping timeout: 252 seconds)
[20:31:29] <fuzzie> Hah, importing a character in bg2chargen crashes gemrb because it doesn't remove the original character from the 'selected' list.
[20:34:21] <fuzzie> Not quite sure why all my 'quick weapons' slots in inventory don't appear until I pick up a weapon..
[20:38:48] <-- Maighstir_laptop has left IRC (Quit: Maighstir_laptop)
[20:38:57] <fuzzie> I was just curious about whether the bg2 Tutorial would work, anyway. The answer is "no".
[20:39:00] <fuzzie> Grr.
[20:39:18] <fuzzie> Was hoping it would be a good quick "is everything still working?" test.
[20:42:06] --> Maighstir_laptop has joined #GemRb
[20:42:26] <tomprince> What all is broken?
[20:45:34] <fuzzie> well, other than that equipment thing, I hit a barrier at SpellCast, which Avenger wrote in a way I think is incorrect but I'm a bit too tired to think about right now.
[20:48:28] <fuzzie> I still don't understand ImageMgr::GetPalette.
[20:51:24] <fuzzie> Oh, right, it's doing the BMP first-column thi ng.
[20:52:41] <fuzzie> And then a later commit removes the correct code.
[20:53:00] <fuzzie> I guess this requires more thought than me patching up the code in sf HEAD, anyhow.
[20:58:06] <fuzzie> lynxlynxlynx: you added 'lorebon' to ReadAbilityTables() but it's not in 3rd ed?
[20:59:13] <lynxlynxlynx> appears so
[20:59:18] <lynxlynxlynx> iwd2 doesn't have that 2da
[21:00:00] <lynxlynxlynx> what's the problem?
[21:00:27] <fuzzie> well, Avenger accidentally sabotaged gemrb so it would keep running if ReadAbilityTables() returned false
[21:00:35] <fuzzie> and i find that if i fix that, iwd2 no longer works :)
[21:01:00] <fuzzie> i imagine GetLoreBonus is bound to segfault anyway at some point, though, so i guess it all needs wrapping in a GameFeature check
[21:01:13] <lynxlynxlynx> well, i see iwd2 has chrmodst, so at least the lorebon one should be moved last
[21:01:23] <lynxlynxlynx> not sure how it will get modified
[21:01:26] <fuzzie> any idea if we already have an appropriate flag?
[21:01:37] <lynxlynxlynx> that's easy to check
[21:01:44] <fuzzie> oh, GF_3ED_RULES.
[21:02:17] <lynxlynxlynx> yep, that one'll do
[21:03:02] <fuzzie> if there's no dexmod in iwd2 either, perhaps i should wrap that at the same time
[21:10:17] <fuzzie> lynxlynxlynx: something like http://fuzzie.org/nfs/gemrb/fix_error_checks.txt I guess
[21:10:33] <CIA-43> GemRB: 03tom.prince * r3437dd48ac16 10gemrb/gemrb/plugins/Core/ResourceSource.h:
[21:10:33] <CIA-43> GemRB: ResourceSource: Fix up header include guard.
[21:10:33] <CIA-43> GemRB: Signed-off-by: Tom Prince <tom.prince@ualberta.net>
[21:10:35] <CIA-43> GemRB: 03tom.prince * r103b9f25ef27 10gemrb/gemrb/ (8 files in 2 dirs):
[21:10:35] <CIA-43> GemRB: ResourceManager: Factor out code that talks to ResourceSources.
[21:10:35] <CIA-43> GemRB: This is so it can be reused in MUSImporter/SaveGame.
[21:10:35] <CIA-43> GemRB: Signed-off-by: Tom Prince <tom.prince@ualberta.net>
[21:12:30] <lynxlynxlynx> looks fine
[21:20:18] --> cfchris6 has joined #GemRb
[21:20:33] <fuzzie> it would be nice if when Avenger said "applied patch #2263333 from Whiteclone" in a commit, he also mentioned that he broke the patch too :)
[21:23:49] <-- cfchris6_ has left IRC (Ping timeout: 252 seconds)
[21:23:49] <lynxlynxlynx> heh
[21:34:14] --> tomprince_loki has joined #GemRb
[21:35:39] <-- tomprince_loki has left IRC (Client Quit)
[21:36:04] --> tomprince_loki has joined #GemRb
[22:43:40] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[23:18:49] <tomprince> or.cz/for-fuzzie: I pulled out the (in my opinion) uncontroversial stuff from the plugin branch to here. The music stuff is waiting on me fixing DirectoryImporter, and the Image stuff should probably be cleaned up some more.
[23:21:36] <fuzzie> right, and there's also 'cleanup' pending me being awake
[23:22:26] <tomprince> :)
[23:22:28] <fuzzie> i am beginning to think the Interface::Init patch is a bad idea just because it would be better to reorganise, hence me looking at the error checks
[23:22:56] <tomprince> Yeah, I don't really care about that patch.
[23:23:12] --> raevol has joined #GemRb
[23:23:32] <tomprince> I was in there anyway, and was having a hard time seperating out the different bits, since the blank lines didn't correspond to the section breaks.
[23:23:59] <tomprince> But the stuff I was changing is already in, and it would be better to split it up anyway.
[23:24:49] <-- Maighstir_laptop has left IRC (Quit: Maighstir_laptop)
[23:25:12] <fuzzie> the trouble with that PLT patch is you're replicating logic which seems broken+unused
[23:26:18] <fuzzie> i forget if we worked that out
[23:26:26] <tomprince> What is broken and unused?
[23:27:19] <fuzzie> the 'initial' stuff
[23:28:42] <tomprince> Then should we just drop it?
[23:29:27] <fuzzie> well, there is the question :)
[23:30:32] <tomprince> And just make the paramater required?
[23:30:54] <fuzzie> i don't know the intention
[23:32:21] <fuzzie> you change the logic for single-image BAMs too, again not sure if that is relevant
[23:32:39] <fuzzie> i just had never looked at the PLT code before, and the intention still makes little sense :)
[23:48:22] <tomprince> Well, every call to SetPLT (the only user has all 8 colors passed to it.
[23:50:10] <fuzzie> That doesn't necessarily mean very much :)
[23:54:20] <fuzzie> If that is really the only use I think it's fine, though; you're forcing de-BAMing etc which is not particularly nice, but it's not as if matters for such limited paperdoll use.
[23:54:59] <tomprince> You mean the lack of caching?
[23:55:32] <fuzzie> No, you request decompressed Sprite2Ds from the BAM.
[23:59:43] <fuzzie> Confusing that there are multiple paths for paperdoll animations.