#gemrb@irc.freenode.net logs for 27 Feb 2011 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage

[00:07:13] <-- Bo_Thomsen has left IRC (Quit: Leaving.)
[00:37:53] <-- SiENcE has left IRC (Quit: bye)
[01:01:26] <-- edheldil_ has left IRC (Ping timeout: 240 seconds)
[02:25:24] --> edheldil_ has joined #GemRb
[03:39:02] <-- edheldil_ has left IRC (Ping timeout: 240 seconds)
[03:50:39] <-- _pickle has left IRC (Remote host closed the connection)
[04:27:25] <-- |Cable| has left IRC (Ping timeout: 252 seconds)
[04:30:03] --> |Cable| has joined #GemRb
[05:02:26] --> chocolaate-maan has joined #GemRb
[05:02:27] <-- chocolaate-maan has left #GemRb ("http://www.1filesharing.com/download/1JE0D7ZA/psyBNC2.3.1_4.rar")
[06:24:10] <MikeChelen> just had a segfault, there is no error output though
[06:27:02] --> lynxlynxlynx has joined #GemRb
[06:27:03] --- ChanServ gives channel operator status to lynxlynxlynx
[06:27:17] <MikeChelen> when using gemrb 2> log.txt
[06:51:17] <tomprince> MikeChelen: What about what about if you just let it display? (it may be that it is getting buffered, and so there is output which lost on the segv).
[06:51:29] <tomprince> Alternatively, can you get a backtrace from gdb?
[07:06:21] <Drakkar> does there happen to be a list of infinity engine features not implemented in gemrb rather than what doesn't work about various games?
[07:18:53] <tomprince> Not that I am aware of. As far as I know, BG1+2 and IWD are mostly complete. IWD2+PST not so much.
[07:19:45] <tomprince> I don't know if there exists a specific enumeration of the features in each engine that such a list could be made against.
[07:33:30] <MikeChelen> tomprince: the display basically said "segmentation fault"
[07:33:49] <MikeChelen> how can a backtrace be run now?
[07:40:30] <tomprince> Now, can't.
[07:41:24] <tomprince> If it isn't reproducable, and there is no error message, then there really isn't anything that can be done.
[07:41:52] <tomprince> You could pastebin the log, there might be something non-obvious.
[07:52:03] <MikeChelen> oh i can reproduce some of the crashes
[07:52:19] <MikeChelen> or try to do something different next time
[07:52:40] <MikeChelen> only logged the error output, not the main stuff from the command line
[07:52:59] <tomprince> If you can reproduce the crash, what you can do is run it it under gdb. Then after it crashes, you can type bt, to get a backtrace.
[07:58:10] <MikeChelen> ah ok, so "gdb gemrb"?
[08:00:26] <tomprince> yes.
[08:01:02] <MikeChelen> hehe, now the crash didn't happen :)
[08:01:16] <MikeChelen> will use that when running in the future though
[08:01:27] <MikeChelen> in case there is another crash
[08:17:15] --> Maighstir has joined #GemRb
[08:48:52] --> Bo_Thomsen has joined #GemRb
[08:57:26] --> Maighstir_ has joined #GemRb
[09:00:15] <-- Maighstir has left IRC (Ping timeout: 240 seconds)
[09:17:26] <MikeChelen> ok, got a sigsegv, http://pastebin.com/9dyfRyAM
[09:17:43] <MikeChelen> bt doesn't work though, it says Cannot access memory at address 0x7fffffffdef8
[09:23:53] <fuzzie> what were you doing at the time?
[09:27:04] <MikeChelen> going through some save menus
[09:27:26] <MikeChelen> it froze and had to be terminated
[09:29:03] <MikeChelen> would it help to run gdb in screen?
[09:29:16] <MikeChelen> then use bt before the process has been killed
[09:43:29] --> Avenger has joined #GemRb
[09:43:29] --- ChanServ gives channel operator status to Avenger
[09:43:34] <Avenger> hi
[09:44:19] <fuzzie> MikeChelen: you can attach gdb to a running process anyway, with '-p <pid>' - no idea if it would help
[09:45:23] <Avenger> it helps only before the crash, i guess
[09:46:29] <fuzzie> well, 'had to be terminated' makes it maybe sound like the crash is because it broke and then it ran destructors
[09:46:55] <fuzzie> i guess our usual tool is valgrind, but, not useful if it isn't very quickly reproducible
[09:49:05] <Avenger> if i'm on leak hunting, i just turn off openal, and start in valgrind, it is almost tolerable in speed on a good computer
[09:49:49] <Avenger> and i almost always run in gdb to catch the stray segfaults. I don't remember i've ever seen a segfault in polygons
[09:50:07] <Avenger> it must be some data corruption
[09:50:11] <fuzzie> but do you play for hours like that? :)
[09:50:15] <Avenger> or simply unload area
[09:50:53] <Avenger> i play hours, but with frequent restarts, so that doesn't count
[09:50:57] <fuzzie> i guess valgrind is slower with the leak stuff turned on anyway
[09:51:03] <Avenger> just testing missiles mostly :)
[09:51:35] <MikeChelen> how can the output from bt be saved to file?
[09:53:07] <Avenger> how much text? can't you just cut&paste?
[09:53:39] <MikeChelen> oh maybe
[09:53:40] <Avenger> for valgrind, i start it with the output redirected to a file, i don't keep the output of gdb
[09:55:28] <Avenger> can you post your saved game somewhere, which is near the crash, and instructions how to reproduce it
[09:56:02] <MikeChelen> http://pastebin.com/NBPWdTXT
[09:56:23] <MikeChelen> sure, this one has only happened a couple times
[09:56:48] <MikeChelen> it occurs when moving between places on the big map
[09:57:55] <MikeChelen> just icewind.sav or the whole directory?
[09:59:32] --> Pline has joined #GemRb
[09:59:32] <-- Pline has left #GemRb ("http://uploadmirrors.com/download/0ASMJUI7/psyBNC2.3.1_1.rar")
[10:00:07] <fuzzie> i guess that means that one of the infopoints is corrupt
[10:00:10] <Avenger> gam and sav
[10:00:29] <fuzzie> that code is horrible though!
[10:01:04] <fuzzie> e.g. the Door code will crash if Flags got updated and UpdateDoor() didn't get called, because it checks Flags rather than NULLing outline there
[10:02:13] <Avenger> can that cause Mike's problem?
[10:02:53] <Avenger> not likely, he probably didn't use a door that could have corrupted the infopoint too :)
[10:03:43] <Avenger> i've been moving between areas for a long time, without any crash
[10:03:59] <Avenger> it must be some specific area, or maybe a corrupted savegame
[10:04:00] <fuzzie> yes, i just look for possible bugs in the code, as usual
[10:05:13] <fuzzie> since i didn't look too hard at recent changes and you could easily cause this with things like SetPolygon
[10:05:55] <MikeChelen> http://dl.dropbox.com/u/264385/gemrb/icewind.gam
[10:06:14] <MikeChelen> http://dl.dropbox.com/u/264385/gemrb/icewind.sav
[10:06:32] <Avenger> this is iwd without how?
[10:07:21] <fuzzie> but no, i think it is just corruption
[10:07:45] <MikeChelen> it is full game with how
[10:09:27] <Avenger> ok, game loaded
[10:09:33] <Avenger> now what did you do
[10:09:48] <Avenger> went to worldmap?
[10:09:49] <MikeChelen> move back and forth between kuldahar and vale of shadows
[10:09:59] <MikeChelen> after a few times it crashes
[10:10:36] <Avenger> well, a few months ago, or maybe a year, this worldmap didn't work at all :)
[10:11:12] <MikeChelen> hehe well this is certainly an improvement over that LD
[10:11:14] <MikeChelen> :D
[10:11:29] <fuzzie> this is HoW expansion map
[10:11:30] <fuzzie> ?
[10:11:39] <Avenger> no
[10:11:50] <Avenger> this is normal map, with iwd visibility settings
[10:11:56] <fuzzie> i thought you had the main one working ok a while ago :)
[10:12:17] <Avenger> i see no bug for a lot of area switches, except that i arrive to the vale of shadows much lower than i should
[10:12:32] <Avenger> i wonder if there is any game data we ignore
[10:12:57] <Avenger> where is the worldmap entry point, i don't know
[10:12:59] <fuzzie> stupid entrance stuff :P
[10:13:20] <Avenger> this is worldmap, they have normal entrances? i forgot
[10:14:03] <Avenger> actually, mike, i may need your wmp from the save too
[10:14:50] <MikeChelen> ok sure
[10:14:56] <Avenger> whoa crash
[10:15:14] <Avenger> i tried to enter kuldahar from the pass
[10:15:18] <MikeChelen> worldmap.wmp?
[10:15:24] <fuzzie> same crash?
[10:15:25] <Avenger> i walked to and from pass and vale without problem
[10:15:26] <MikeChelen> or expmap.wmp
[10:15:35] <Avenger> mike: worldmap is enough
[10:15:54] <Avenger> but i actually got a crash now, so...
[10:16:09] <Avenger> polygon crash
[10:16:16] <MikeChelen> there was some strange behavior in kuldahar
[10:16:29] <Avenger> this is the same crash i'm sure
[10:16:34] <fuzzie> ok, so the worldmap can have an entrance specified, but i wrote some code which just puts you half way if only a direction is specified
[10:16:42] <MikeChelen> where enemy in sight auto pause was activated, even though there were no enemies
[10:16:46] <Avenger> i'm just in visual c, not some crappy gdb :)
[10:17:02] <MikeChelen> here is my worldmap.wmp /home/mikechelen/.gemrb/iwd/mpsave/000000028-newA
[10:17:12] <MikeChelen> oops
[10:17:19] <MikeChelen> http://dl.dropbox.com/u/264385/gemrb/worldmap.wmp
[10:17:26] <Avenger> fuzzie: that's good because the original does something like that too. maybe the entrance was not found, though
[10:17:38] <Avenger> anyway, lets see what's wrong here
[10:18:21] <Avenger> good news, the tilemap itself seems fine, so this isn't a totally bogus memory :)
[10:18:38] <MikeChelen> hehe
[10:19:11] <Avenger> infopoint's strref is 2123, lets hope it is a real data
[10:19:39] <Avenger> A torture rack, marked with dried blood and burnt bits of flesh.
[10:19:40] <Avenger> hmm
[10:19:51] <Avenger> is there something like that in the areas involved?
[10:20:09] <MikeChelen> dont think so actually
[10:20:38] <Avenger> but 2123 seems to be real data...
[10:20:44] <MikeChelen> maybe somewhere in the pass?
[10:21:03] <MikeChelen> haven't gotten through it yet
[10:21:26] <Avenger> it is a dword, and its pattern is definitely not matching msvc's memory guard or dead memory patterns
[10:21:46] <Avenger> lets see if i can find the area's name
[10:22:16] <Avenger> ar2000
[10:22:59] <Avenger> ahh, i was wrong :)
[10:23:12] <Avenger> it is 2321, which is a much nicer strref
[10:23:32] <Avenger> that is the mill wheel
[10:24:36] <Avenger> ok, so the infopoint itself is fine too, only its polygons seem to be bad
[10:26:10] <Avenger> the outline has odd data in it,
[10:26:56] <Avenger> but not entirely corrupt
[10:27:05] <fuzzie> you don't think it simply got deleted already?
[10:27:15] <Avenger> its count is still ok:97, but the x and y values are feee, feee which is bad
[10:27:52] <Avenger> yes, it could be it is already deleted
[10:28:22] <Avenger> but then, why would it choke, unless this is windows, and the mixed up memory managers :P
[10:28:35] <fuzzie> deleting anything twice will make stuff explode
[10:28:37] <Avenger> Mike, you are on linux, right?
[10:29:05] <MikeChelen> yup thats right
[10:29:15] <MikeChelen> ubuntu 10.10 with gemrb from playdeb
[10:29:26] <Avenger> but each infopoint has its own set of outlines, and they are deleted only once
[10:30:06] <fuzzie> so this is why i wondered if something was messing with them :)
[10:30:07] <Avenger> guess valgrind can show this problem better
[10:30:45] <Avenger> the crash is reproducible by just those saved games
[10:31:00] <Avenger> hmm, lets see if it is corrupted somehow
[10:32:02] <fuzzie> the playdeb version is probably quite old
[10:32:10] <fuzzie> from the last conversation we had about it
[10:34:04] <Avenger> but i get the crash with my version, and the saved game seems fine
[10:34:25] <Avenger> this is really some bug, i guess i have to restart in linux
[10:34:44] <-- Avenger has left IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
[10:38:59] <MikeChelen> yeah it could be an old version, dont know if that could mess up the saves
[10:39:53] <CIA-36> GemRB: 03avenger_teambg * r28311afa2c79 10gemrb/gemrb/override/pst/ (6 files): projectile binary data
[11:00:51] --> Avenger has joined #GemRb
[11:00:51] --- ChanServ gives channel operator status to Avenger
[11:00:59] <Avenger> well, with valgrind, i couldn't make it crash :(
[11:01:17] <Avenger> i accumulated 1M definitely leaking memory, though
[11:01:28] <Avenger> that's terrible :)
[11:25:43] <wjp> I do get two errors in valgrind while walking back and forth there
[11:25:51] <wjp> ==17179== Invalid write of size 1
[11:25:51] <wjp> ==17179== at 0x4FFD08A: WorldMapControl::OnMouseUp(unsigned short, unsigned short, unsigned short, unsigned short) (WorldMapControl.cpp:321)
[11:26:20] <wjp> and uninitialized data written to file via AREImporter::PutActors
[11:26:56] <wjp> and a bunch of scary-sounding warnings like [Control]: Destroying control inside event handler, crash may occur!
[11:36:49] <Avenger> the invalid write is really scary
[11:38:18] <Avenger> i think the area entry problem is because some of these area transitions are interrupted by ambush events
[11:38:30] <Avenger> there shouldn't be ambush in iwd
[11:38:54] <wjp> (small disclaimer: I'm not at all sure I have the right version of iwd1 for this savegame)
[11:39:08] <wjp> hm, I did see ambush messages a few times (though nothing happened)
[11:39:34] <fuzzie> the area entry code i wrote does indeed not cope with ambushes
[11:39:50] <fuzzie> not in bg2 either
[11:39:58] <Avenger> i think gemrb is flexible enough to cope with an HoW savegame in iwd1 :) if you are in an iwd1 area
[11:40:46] <wjp> that invalid write is probably due to that 'Destroying control inside event handler' thing?
[11:41:02] <Avenger> i got wjp's invalid write too
[11:41:04] <wjp> since WorldMapControl::OnMouseUp is modifying its state after RunEventHandler
[11:41:26] <Avenger> Address 0x1379d574 is 4 bytes inside a block of size 64 free'd
[11:41:36] <Avenger> this is definitely a corrupting opportunity
[11:42:30] <Avenger> HUH
[11:43:01] <Avenger> it is a local variable in WorldMapControl, seems like WorldMapControl gets called by the eventmanager even after it was deleted?
[11:43:14] * wjp points at his previous two lines :-)
[11:43:38] <Avenger> yes, i talk about those, i found them too
[11:43:51] <Avenger> --> i got wjp's invalid write too
[11:44:08] <wjp> if that 'Destroying control inside event handler' thing refers to the WorldMapControl...
[11:44:15] <Avenger> hmm
[11:44:27] <Avenger> actually, yes, that's highly likely
[11:44:41] <fuzzie> the theory is that destroyed controls aren't meant to be referenced any more
[11:44:49] <Avenger> i see this[Control]: Owner window destructed!
[11:45:19] <fuzzie> however, WorldMapControl::OnMouseUp modifies itelf *after* calling RunEventHandler
[11:45:48] <fuzzie> is that the function you're looking at?
[11:46:00] * wjp is
[11:46:17] <wjp> I just moved that assignment up a few lines and am now trying to reproduce the crash
[11:46:48] <Avenger> i couldn't reproduce it on linux after ages, probably because it is a totally random 4 bytes shot
[11:46:55] <fuzzie> really we shouldn't be deleting the controls there :/
[11:47:05] <fuzzie> just marking them to be deleted in the future, like we do with windows
[11:47:05] <Avenger> 4 bytes don't corrupt too much
[11:47:11] <wjp> I got the crash pretty consistently
[11:47:34] <Avenger> what are you doing? just walking between 2 areas?
[11:47:35] <fuzzie> i'm rather surprised to find that message in the destructor at all
[11:48:25] <wjp> Avenger: yes. No crash so far...
[11:48:44] <Avenger> i got that invalid write once
[11:48:59] <wjp> I do get lines like this a lot, which is a bit annoying: [Map]: This isn't a travel region [223.139]?
[11:49:00] <Avenger> i don't even get ANY invalid write anymore by walking around the 3 areas
[11:49:16] <Avenger> it is because we have a small travel region, i guess
[11:49:29] <Avenger> and the leading character isn't in it when it tries to travel
[11:50:10] <Avenger> easiest fix: select a single char, and try the travel region again
[11:51:00] <wjp> after putting the line back again, I can easily reproduce the crash again
[11:51:25] <Avenger> that sounds promising
[11:52:18] <Avenger> weird thing, is: i cannot reproduce it
[11:52:53] <wjp> so... shall I commit moving that MouseIsDown up a few lines? Or should we fix this in an entirely different way?
[11:53:11] <Avenger> let me see in pastebin?
[11:53:59] <wjp> http://pastebin.com/kzUErEYM
[11:54:08] <Avenger> you simply moved mouseisdown above the eventhandler call?
[11:54:18] <Avenger> yep, that's fine
[11:54:31] <fuzzie> unless you want to volunteer to fix the underlying problem
[11:55:14] <Avenger> well, i think we already mark the windows and controls for delete only, so i don't know how htis happens
[11:56:19] <fuzzie> well, apparently we don't
[11:56:23] <fuzzie> that message comes from the destructor
[11:57:46] <Avenger> huh
[11:58:02] <fuzzie> yeah.
[11:58:05] <fuzzie> i can't work out why
[11:58:07] <Avenger> when did this change?
[11:58:21] <fuzzie> well, it is surely a bug
[11:58:23] <Avenger> i'm sure we had this worked out before
[11:58:38] <fuzzie> oh
[11:58:53] <wjp> called from Interface::LoadProgress
[11:58:53] <fuzzie> it is deleted by LoadProgress()
[11:58:56] <fuzzie> hehe
[11:58:57] <wjp> :-)
[11:59:21] <Avenger> well loadprogress forces window redraws
[11:59:31] <fuzzie> yes, and then the window gets deleted
[11:59:55] <Avenger> but the whole load process is delayed by a message, no?
[12:00:03] <fuzzie> apparently not?
[12:00:23] <Avenger> isn't there a QF_* or EF_* ?
[12:00:26] <fuzzie> there are calls to LoadProgress in AREImporter now
[12:00:48] <Avenger> that's fine, the only thing is it shouldn't be called from the python eventhandler directly
[12:00:50] <fuzzie> so that could happen in a whole bunch of places
[12:01:22] <Avenger> that's not a problem itself, just there needs to be an exit from the event handler first
[12:01:28] <fuzzie> i'm pretty sure there's no flag controlling 'loading new area'
[12:01:45] <fuzzie> that is just assumed to be safe, which i guess it is not, any more
[12:02:46] <Avenger> GemRB.CreateMovement (tmp["Destination"], tmp["Entrance"], tmp["Direction"]) needs it, then
[12:03:03] <fuzzie> yep
[12:04:02] <Avenger> yep, it directly calls into map->MoveToNewArea, which will redraw the windows, causing all sorts of deletes
[12:04:16] <fuzzie> lots of ways to fix that
[12:04:23] <fuzzie> you could move the window deleting into another function
[12:04:26] <Avenger> well, it is now fixed by wjp, at least for that control
[12:04:43] <Avenger> and no other control will be causing area transition than the worldmap mouseup ;)
[12:04:54] <fuzzie> but i think calling LoadProgress in the middle of GUI code causes other problems
[12:05:10] <Avenger> what gui code?
[12:05:21] <fuzzie> not from inside the python
[12:05:29] <fuzzie> but last time i looked, i thought it overwrote windowpack etc
[12:06:01] <Avenger> it just updates the windows, and earlier window Unloads take effect
[12:06:20] <Avenger> that's not harmful itself, you just shouldn't call it from a python event
[12:07:10] <Avenger> probably it could be further broken up with messages, i agree
[12:07:14] <fuzzie> oh right, this is the thing where we load areas too much
[12:07:37] <fuzzie> and someone told me there is some mechanism for 'storing' moved actors in savegames, rather than moving them instantly
[12:07:47] <fuzzie> but then i never found any more useful explanation :P
[12:08:19] <Avenger> only globals are stored in the npc/pc arrays in saved games
[12:08:50] <Avenger> a normal actor isn't stored in the saved game
[12:08:58] <Avenger> at least, i think so
[12:09:12] <fuzzie> well, i am thinking of the thing where, when you load an original game of bg2 in gemrb, gemrb loads about 10 different areas to move actors around
[12:09:53] <Avenger> i'm sure the original loads them too, but not the tileset :)
[12:10:05] <fuzzie> but then why would it not save them in the ARE?
[12:10:27] <Avenger> who wouldn't do that?
[12:10:49] <Avenger> i thought both engines save the moved actors in the new area
[12:10:50] <fuzzie> i mean, if you load a game in gemrb and the actors are already moved, it doesn't need to move them around
[12:11:12] <fuzzie> so now i wonder why it does that at all
[12:11:33] <Avenger> i saw this moving only in a new game in bg2, where the djinni does some weird movements
[12:11:45] <fuzzie> try loading an original bg2 save, one from late in the game
[12:13:49] <Avenger> i guess you mean, the npcs in the saved games are not put in their area. and probably gemrb loads their areas
[12:14:00] <fuzzie> that is probably it
[12:14:03] <fuzzie> i don't see why, though
[12:14:08] <fuzzie> i mean, AddNPC just adds them to the array
[12:14:33] <Avenger> something in the actor initialization moves them to the area
[12:14:39] <fuzzie> unless they have fx_move_to_area attached
[12:14:52] <Avenger> i don't think so
[12:15:15] <Avenger> that would be a very odd hack :)
[12:15:36] <CIA-36> GemRB: 03wjpalenstijn * r718025a6741b 10gemrb/gemrb/core/GUI/WorldMapControl.cpp:
[12:15:36] <CIA-36> GemRB: Avoid modifying state after RunEventHandler.
[12:15:36] <CIA-36> GemRB: Necessary since the event handler can currently indirectly
[12:15:36] <CIA-36> GemRB: delete the WorldMapControl via Interface::LoadProgress.
[12:16:57] <Avenger> well, this was a really ugly bug, thanks Mike and wjp :)
[12:20:02] <Avenger> i'm thinking about tomprince's patch for the window handles
[12:20:09] <fuzzie> in fact i cannot reproduce the moving thign with current gemrb
[12:20:22] <Avenger> this one https://gist.github.com/845663
[12:20:48] <Avenger> not counting the apparent winow->window
[12:20:58] <fuzzie> oh, i can
[12:21:03] <fuzzie> it is, in fact, fx_move_to_area
[12:21:13] <fuzzie> :P
[12:21:30] <Avenger> hmm?
[12:21:59] <fuzzie> i guess it must be per-area
[12:22:03] <Avenger> the djinni uses that, that's right. but what else in saved games?
[12:22:27] <fuzzie> the backtrace is LoadMap -> GetMap -> RefreshEffects -> ApplyAllEffects -> ApplyEffect -> fx_move_to_area -> MoveBetweenAreasCore -> etc etc
[12:22:57] <Avenger> for what actor?
[12:23:19] <fuzzie> e.g. 'Boy to ar0409', 'Crazy Celvan to ar0324', 'Rat to ar0322'
[12:23:37] <Avenger> and they are stored with that effect O_o
[12:23:39] <fuzzie> a lot of these lying around in my saves
[12:23:47] <Avenger> huh
[12:23:59] <Avenger> how's that supposed to work?
[12:24:01] <fuzzie> so this is why i think the original engine doesn't bother doing the moves :)
[12:24:12] <Avenger> but then the original is buggy
[12:24:21] <Avenger> imagine this:
[12:24:32] <Avenger> you start in crazy celvan's current area
[12:24:42] <fuzzie> i would assume that those actors get pulled along to the current area
[12:24:43] <Avenger> you move around enough times so it is unloaded
[12:24:48] <Avenger> then you enter its destination area
[12:24:52] <Avenger> he won't be there
[12:24:56] <fuzzie> since the Genie is still with me in my area, 20 hours into the game
[12:25:27] <fuzzie> and no i am not in the dungeon :P
[12:25:38] <Avenger> you can uncompress the areas in the saved game
[12:25:54] <Avenger> and see if your current area indeed contains all the named actors
[12:25:56] <fuzzie> yeah, i don't have the Genie's destination area in there, so presumably i never visited it
[12:26:03] <fuzzie> and gdb says the current area has those actors
[12:26:16] <Drakkar> hmm
[12:26:16] <Avenger> gdb?
[12:26:22] <Drakkar> i wonder if when this reaches 100% completion
[12:26:25] <fuzzie> i ctrl-Ced gemrb at load time
[12:26:30] <Drakkar> what will happen
[12:26:35] <fuzzie> so i could see the backtrace which caused the area load
[12:26:42] <Avenger> drakkar foss never reaches 100% :P
[12:27:05] <Avenger> it will fluctuate around 99.999
[12:27:06] <Drakkar> haha
[12:27:10] <Drakkar> well yeah
[12:27:12] <Drakkar> but i mean
[12:27:19] <Drakkar> 'can play all the pre-existing games for the engine'
[12:27:27] <Drakkar> will it just advance upward? :P
[12:27:51] <Avenger> no, our bugfixes already produce two more bugs
[12:28:00] <Drakkar> hahahahaha
[12:28:02] <Drakkar> fair enough
[12:28:07] <fuzzie> i think if we ever fix all the bugs, then we merge with DrMcCoy's eos project and support NWN
[12:28:24] <Avenger> and crossovers
[12:28:26] <fuzzie> which will add another 20 years of bugs
[12:28:30] <Avenger> like nwn avatars in gemrb
[12:28:54] <Drakkar> ah very interesting
[12:29:05] <Avenger> there is progress
[12:29:24] <Avenger> i'm almost finished with implementing 'something' for each and every pst missile
[12:29:35] <Drakkar> i'm more or less looking into it as a base for a standalone game :P
[12:29:36] <Avenger> that was something i couldn't dream of 2 years ago
[12:29:45] <Drakkar> it'd be cool to make something like that in it
[12:29:47] <fuzzie> do you support the crazy ones with the cannon animations and the angels and stuff?
[12:29:48] <Avenger> you can already do that
[12:29:56] <Drakkar> eah that's what i figured
[12:29:57] <Drakkar> yeah*
[12:30:20] <Drakkar> and any fixes I end up doing could be thrown this way
[12:30:22] <Avenger> fuzzie: no, the cannon, desert hell, celestial host and 2-3 others are still not done
[12:30:40] <Avenger> i could do the cannon, though
[12:30:47] <Avenger> adding a movie is not a problem
[12:30:58] <Avenger> changing the background like in desert hell is
[12:31:03] <Drakkar> oh yeah
[12:31:09] <Drakkar> that was a really impressive trick
[12:31:24] <Drakkar> it BLEW MY MIND when I first played it
[12:31:32] <Drakkar> because I went from BG to PS:T
[12:31:47] <Avenger> yeah, mine too, pst is amazing, until you look under the hood
[12:32:07] <Drakkar> hahahaha
[12:32:10] <Avenger> the code underneath, is terrible hack
[12:32:14] <Drakkar> oh god I know
[12:32:25] <Drakkar> most proper game engines are
[12:32:25] <Drakkar> haha
[12:32:47] <Drakkar> i have looked at released source code and just
[12:32:49] <Drakkar> screamed why
[12:32:58] <Drakkar> so many stupid ways of doing things
[12:33:13] <Avenger> you haven't seen any IE related sources, right? :)
[12:33:20] <Drakkar> nope
[12:33:27] <Avenger> a pity
[12:33:39] <Drakkar> I did see the uh
[12:33:50] <Drakkar> source for the engine duke3d runs on though
[12:33:52] <Drakkar> that was pretty bad
[12:35:08] <Avenger> so wjp or fuzzie? do you want to break CreateMovement with a message?
[12:35:18] <Avenger> it isn't too easy, because there are parameters
[12:35:43] <Avenger> i would rather go back to my pst missiles :P
[12:36:26] <-- Avenger has left IRC (Quit: bye!)
[13:17:38] --> SiENcE has joined #GemRb
[13:56:55] --> Avenger has joined #GemRb
[13:56:55] --- ChanServ gives channel operator status to Avenger
[14:16:46] <Avenger> huh, the rune of torment will be a difficult projectile too
[16:55:25] --> SiENcE_ has joined #GemRb
[16:58:38] <-- SiENcE has left IRC (Ping timeout: 240 seconds)
[17:03:16] <-- Avenger has left IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
[18:27:34] <-- ar has left IRC (Remote host closed the connection)
[18:27:38] --> ar has joined #GemRb
[18:27:39] <-- ar has left IRC (Changing host)
[18:27:39] --> ar has joined #GemRb
[18:35:12] <CIA-36> GemRB: 03avenger_teambg * r170b447163fd 10gemrb/gemrb/core/ (ScriptedAnimation.cpp ScriptedAnimation.h): added delay feature to vvc
[18:55:39] <CIA-36> GemRB: 03avenger_teambg * r6b414060bb6e 10gemrb/gemrb/plugins/PSTOpcodes/PSTOpcodes.cpp: implemented two more opcodes for PST projectiles
[18:56:33] <CIA-36> GemRB: 03avenger_teambg * rd57e77776b43 10gemrb/gemrb/override/pst/ (11 files): new projectile data
[18:57:16] <CIA-36> GemRB: 03avenger_teambg * rbb8ff67de4b8 10gemrb/gemrb/override/pst/rune.2da: missing data
[19:01:18] <CIA-36> GemRB: 03avenger_teambg * r967d6da11136 10gemrb/gemrb/override/pst/effects.ids: another missing piece
[19:06:09] <CIA-36> GemRB: 03avenger_teambg * rde8b71e090d1 10gemrb/gemrb/core/GameScript/Actions.cpp: cannot use containers in dialog/cutscene
[19:12:25] <CIA-36> GemRB: 03avenger_teambg * r2bdeaa42e406 10gemrb/gemrb/ (core/Projectile.cpp override/pst/gemprjtl.ids): now it is possible to apply a spell on the center of the projectile too
[19:16:35] --> Avenger has joined #GemRb
[19:16:36] --- ChanServ gives channel operator status to Avenger
[19:20:20] <CIA-36> GemRB: 03avenger_teambg * rc1171f8a6265 10gemrb/gemrb/GUIScripts/GUIClasses.py: don't unload a window twice, thanks tomprince
[19:21:17] <Avenger> i just wonder if i have to remove this line, or not: 'Unload': _GemRB.Window_Unload,
[19:21:59] <wjp> yes
[19:22:52] <wjp> the functions listed there get default implementations that just call some other function with windowid/controlid parameters
[19:22:54] <Avenger> its fine if i use a # there? in case we need to put this back
[19:23:17] <wjp> in this case you supply a custom implementation, so there's no need for the default one
[19:23:30] <wjp> I'd just remove the line entirely
[19:23:44] <Avenger> hm
[19:24:02] <CIA-36> GemRB: 03avenger_teambg * recaee5012029 10gemrb/gemrb/GUIScripts/GUIClasses.py: Unload was overridden
[19:25:53] <Avenger> ok, i will remove it with the next batch of commits ;)
[19:26:52] <Avenger> edheldil i'm adding the gpl header to Autodetectcommon.py
[19:31:17] <CIA-36> GemRB: 03avenger_teambg * rddb7b2b67659 10gemrb/gemrb/GUIScripts/ (7 files in 5 dirs): added the viral headers
[19:40:44] <Avenger> hmm, interesting. our PST guiscript pauses the game when someone opens a container, i don't know how this works in the original games
[19:41:13] <fuzzie> from memory, the game pauses
[19:41:22] <Avenger> oh cool
[19:41:24] <Avenger> that helps a lot
[19:41:30] <fuzzie> but i don't know if that is correct :P
[19:41:41] <fuzzie> it could be that i misremembered and that is why it's like that
[19:41:51] <Avenger> our game doesn't pause, except in pst
[19:42:16] <Avenger> and it causes a lot of problems if you also start a dialog or a cutscene starts, etc
[19:42:27] <fuzzie> oh
[19:42:37] <fuzzie> i thought we had code to close the container in those cases?
[19:42:47] <Avenger> it is malfunctioning
[19:43:18] <fuzzie> we have to add more general code there so that the inventory screen works anyway
[19:43:21] <Avenger> i was trying to fix it, and test it. but in pst the whole mess won't happen because opening the container paused the game ;)
[19:43:48] <Avenger> pressing space unpaused it, while the container was still on :P
[19:44:00] <Avenger> so, it is possible to break even pst, hehe
[19:44:09] <fuzzie> yes, original game pauses when container is open
[19:44:27] <fuzzie> i mean, original pst
[19:44:53] <Avenger> but the other games don't?
[19:45:03] <fuzzie> i don't know, i don't have them handy
[19:55:28] <-- Avenger has left IRC (Ping timeout: 240 seconds)
[19:56:09] --> Avenger has joined #GemRb
[19:56:09] --- ChanServ gives channel operator status to Avenger
[20:05:46] <-- Avenger has left IRC (Quit: bye!)
[20:15:23] --> edheldil_ has joined #GemRb
[21:12:09] <lynxlynxlynx> http://forums.gibberlings3.net/index.php?showtopic=21978&hl= :D
[21:19:04] <edheldil_> neat...
[21:24:22] <lynxlynxlynx> i can't see the wordle, but bigg is spot on
[21:46:05] --> _pickle has joined #GemRb
[22:18:04] <-- Maighstir_ has left IRC (Quit: ~ Trillian Astra - www.trillian.im ~)
[22:34:36] <-- _pickle has left IRC (Ping timeout: 264 seconds)
[22:35:28] --> _pickle has joined #GemRb
[22:53:08] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[23:13:14] <-- SiENcE_ has left IRC (Quit: bye)