#gemrb@irc.freenode.net logs for 30 Oct 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:56:05] <-- Maighstir has left IRC (Quit: ~ Trillian Astra - www.trillian.im ~)
[01:42:44] <-- pupnik_ has left IRC (Quit: Lost terminal)
[01:52:05] --> usurper has joined #GemRb
[01:53:50] <usurper> hi - I've encountered an error in the last commit of gemrb/override/(how+iwd)/classes.2da, I was wondering where and how to submit a patch?
[02:17:55] <usurper> nevermind - just emailed Jaka Kranjc the new 2da file
[02:48:50] <-- barra_ has left IRC (Quit: Verlassend)
[03:58:27] <-- usurper has left IRC (Quit: Page closed)
[05:13:50] --> lynxlynxlynx has joined #GemRb
[05:13:50] --- ChanServ gives channel operator status to lynxlynxlynx
[06:38:47] --> edheldil_ has joined #GemRb
[06:44:41] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[06:55:58] <-- edheldil_ has left IRC (Ping timeout: 252 seconds)
[07:12:29] --> edheldil_ has joined #GemRb
[07:47:11] <-- edheldil_ has left IRC (Ping timeout: 250 seconds)
[08:09:49] --> edheldil_ has joined #GemRb
[08:25:22] --> barraAway has joined #GemRb
[08:51:45] <-- edheldil_ has left IRC (Read error: Operation timed out)
[09:26:34] --> Maighstir has joined #GemRb
[09:50:44] --- barraAway is now known as barra_fife
[12:28:36] --> Avenger has joined #GemRb
[12:28:36] --- ChanServ gives channel operator status to Avenger
[12:29:07] <fuzzie> Avenger: i thought i fixed the parsing for those strings, but maybe it was something similar
[12:29:12] <Avenger> hello
[12:29:25] <Avenger> what strings?
[12:29:39] <Avenger> ahh, the dialog bug
[12:30:07] <Avenger> didn't you just fixed bcs reading?
[12:30:11] <fuzzie> the action code is a bit fragile, because there are broken ()s and stuff in other games too
[12:30:53] <fuzzie> so you really have to be careful when changing it
[12:31:16] <Avenger> i didn't plan that, especially if the original doesn't work :)
[12:31:24] <fuzzie> the original does work :P
[12:31:33] <fuzzie> pst in particular accepts all kinds of weird stuff
[12:31:42] <fuzzie> i don't know about this Melicamp case though
[12:31:54] <Avenger> so you don't have to put the end quote on a string?
[12:32:12] <fuzzie> yeah, it does work in the origina
[12:32:13] <fuzzie> l
[12:32:13] <fuzzie> meh
[12:32:14] <Avenger> that's weird, that means, you cannot put , in any strings in script code?
[12:32:38] <fuzzie> well, in the dialog code :P
[12:32:39] <Avenger> so a ',' can terminate strings
[12:32:46] <fuzzie> let me check some scripts
[12:38:02] <fuzzie> bg1/melica.d: IF ~~ THEN DO ~GiveItemCreate("MISC49,LastTalkedToBy,0,0,0)
[12:38:02] <fuzzie> bg1/sharte.d:ChangeAIScript("SHARTEEL,OVERRIDE)
[12:38:27] <fuzzie> and, oh dear, a *lot* in bg2
[12:38:51] <fuzzie> bg2/kielpc.d:Deactivate("Levers 1-3, 6-7")
[12:38:57] <fuzzie> ^- that is the only one which isn't a missing " mark
[12:40:49] <fuzzie> none in pst, i don't have the other dialogs here
[12:41:29] <fuzzie> we could handle both of those, by just having ',' stop a string if it's not the last parameter?
[12:41:39] <fuzzie> but i doubt the kielpc.d one works in the original. got to cook now.
[12:42:27] <Avenger> haha
[12:43:12] <Avenger> it really sucks to be an IE script writer. I wonder if trained monkeys would do any better
[12:43:38] <Avenger> or why didn't they have some script editor that did syntax check on the fly
[12:43:58] <fuzzie> well, maybe they just had their in-game parsing code check it
[12:46:54] <Avenger> i didn't see the dialog compiler yet. But the bcs reader is very simple
[12:47:04] <Avenger> it has definitely no fuzzy logic :P
[12:47:15] <fuzzie> yes, i don't think we found any mistakes in the original bcs files
[12:47:32] <Avenger> dltcep chokes on some
[12:47:38] <Avenger> but i guess those are unused
[12:47:53] <fuzzie> some of the fixpack files have been broken
[12:48:15] <fuzzie> but i think the original engine choked on those too
[12:48:17] <Avenger> how am i supposed to parse Deactivate("xxx , yyy") and ChaingeAIScript("xxx,yyy) to work both ways?
[12:48:25] <fuzzie> like i said
[12:48:39] <fuzzie> just make "," stop a string unless the string is the last parameter?
[12:49:11] <fuzzie> i mean, i doubt the Deactivate thing works, but you could parse both
[12:49:16] <Avenger> that's terribly complicated, because there is also a ) in deactivate
[12:49:38] <Avenger> i could do this: if it didn't find a matching ", it would look for ,
[12:50:00] <Avenger> that's how i would do it, but i don't know how the ie does i
[12:50:01] <Avenger> t
[12:51:09] <fuzzie> you know, i already documented some of this
[12:51:34] <fuzzie> above DLGImporter's GetStrings, i provide three examples
[12:52:26] <fuzzie> i wonder why i wouldn't have fixed the actions at the same time?
[12:52:52] <Avenger> those examples don't cover the melicamp case?
[12:52:58] <fuzzie> they do
[12:53:02] <fuzzie> the middle one is that
[12:53:07] <Avenger> well, missing quote
[12:53:26] <fuzzie> but i wouldn't do anything so complicated
[12:53:31] <fuzzie> just change it to stop strings on commas
[12:53:32] <Avenger> i don't see how you do it, but i guess you don't do it the way i said?
[12:53:35] <fuzzie> so, line 1497
[12:53:52] <Avenger> that would break kield
[12:54:12] <fuzzie> i'm pretty sure kield is just broken
[12:54:34] <Avenger> why
[12:54:37] <fuzzie> yeah, it's not used in the game
[12:54:41] <Avenger> oh
[12:55:26] <Avenger> bg1 has kielpc.d with those levers
[12:55:32] <Avenger> so it is probably used, but in bg1
[12:55:51] <fuzzie> hmph
[12:55:53] <fuzzie> and the same stupid comma?
[12:56:03] <Avenger> it also has OpenDoor(Door09)
[12:56:07] <Avenger> without any quotes
[12:56:11] <Avenger> trained monkeys....
[12:56:18] <Avenger> Deactivate("Levers1-3,6-7")
[12:56:23] <fuzzie> maybe it isn't used there either? :)
[12:56:32] <Avenger> well, maybe
[12:56:50] <fuzzie> i wouldn't add major hacks to GenerateActionCore unless there is actually a bug
[12:57:13] <fuzzie> i mean, you can if you want to, of course! but it is pretty messy already
[12:57:29] <Avenger> yeah
[12:57:36] <Avenger> but melicamp needs to be fixed
[12:57:50] <Avenger> so we need to add something
[12:58:13] <fuzzie> if you just change line 1497
[12:58:16] <Avenger> the kiel dialog is unused i think
[12:58:18] <fuzzie> of GSUtils.cpp
[12:58:28] <fuzzie> to check for a comma as well
[12:58:41] <fuzzie> then i think that should work fine
[12:58:53] <Avenger> ok, so we break on both a " and a ,
[12:59:07] <Avenger> and add a comment
[12:59:11] <fuzzie> sounds good
[12:59:19] <fuzzie> the rest of the code should cope fine with you doing that
[13:29:01] <Avenger> HoW definitely has BG2 blood in it :)
[13:29:09] <Avenger> bg1 opcodes end at 0xbf
[13:29:19] <Avenger> HoW has some more, including familiars
[13:41:15] <devurandom> """how am i supposed to parse Deactivate("xxx , yyy") and ChaingeAIScript("xxx,yyy) to work both ways?""" << can strings wrap multiple lines?
[13:41:39] <devurandom> Otherwise you could find an unterminated string at lineend and search again, this time for ,
[13:42:23] <Avenger> that's what i suggested, but fuzzie convinced me that the other case is not used anywhere
[13:42:56] <devurandom> Ah, I misunderstood yours and his idea to be about the same idea.
[13:43:33] <devurandom> I understood 'matching ",' to be " followed by , in the code. ;)
[13:44:58] <devurandom> Avenger: Did you succeed in debugging the bg1 savegame crash?
[13:45:07] <Avenger> no
[13:45:36] <Avenger> i just mark stuff in the bg1 disassembly with IDA, so eventually i will find it out
[13:50:41] <Avenger> it crashes in some list
[13:52:23] <Avenger> my totsc is not the same as yours, though. Mine crashes at 0x800622.
[13:54:41] <Avenger> hmm more precisely, this is the effects list
[13:54:44] <Avenger> hmm. hmm
[14:03:29] <devurandom> I have the german version, probably its different.
[14:22:37] <fuzzie> i wonder how things like !Dead("uhinn02') work
[14:22:54] <Avenger> yck
[14:23:07] <Avenger> do they?
[14:23:26] <fuzzie> probably not :)
[14:24:11] <fuzzie> i don't worry about it until we find a bug
[14:25:01] <fuzzie> Avenger: your build was 0.6.2, or latest git?
[14:25:13] <Avenger> git
[14:25:25] <fuzzie> so i'm confused by the reported Jaheira thing
[14:25:28] <Avenger> huh, who wrote: SlowPoisonDamageRate in bg1 effect list
[14:25:30] <fuzzie> i tried mixing party order etc
[14:25:47] <fuzzie> our bg1 effect list is just a copy of our bg2 effect list, i think
[14:25:53] <Avenger> omg
[14:25:55] <fuzzie> i complained about it before, it causes weird bugs :P
[14:25:57] <Avenger> bg2 is bad too
[14:26:16] <fuzzie> yes, they are identical
[14:26:24] <Avenger> yes, it makes the 0xa4 effect stuck on actors, which breaks the original game
[14:26:41] <Avenger> this was one obscure bug :)
[14:27:23] <Avenger> it should be: Cure:Intoxication
[14:27:26] <fuzzie> it would be nice if we could fix the effect lists so that they only have the effects from the original + clearly-marked deliberate gemrb additions
[14:27:35] <fuzzie> but i guess that is a lot of work
[14:27:58] <fuzzie> as it is, i never have any idea whether any effect is good or not :)
[14:29:49] <Avenger> devurandom: change SlowPoisonDamageRate to Cure:Intoxication in effects.2da. Load your savegame then save it in gemrb
[14:29:54] <Avenger> then it should work in bg1 :)
[14:30:29] <Avenger> lol it will still crash, but at another point
[14:30:45] <Avenger> but that's fine, we fixed a bug
[14:31:03] <-- Avenger has left IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
[14:42:09] <CIA-29> GemRB: 03avenger_teambg * r7b6850a22e68 10gemrb/gemrb/core/GameScript/GSUtils.cpp: fixed parsing of unterminated strings in dialog scripts (melicamp the chicken)
[14:45:27] <CIA-29> GemRB: 03avenger_teambg * r26df9101fd4c 10gemrb/gemrb/override/ (4 files in 4 dirs): fixes the remove intox effect to not linger around (it crashes original bg1/bg2)
[15:22:04] <devurandom> Thanks so much Avenger!
[16:21:01] --> lynxlynxlynx has joined #GemRb
[16:21:01] --- ChanServ gives channel operator status to lynxlynxlynx
[16:37:02] <lynxlynxlynx> nice, plenty of activity
[16:38:08] --> SiENcE has joined #GemRb
[17:46:38] <fuzzie> lynxlynxlynx: do you have how/iwd2 dialogs?
[17:46:56] <lynxlynxlynx> yes
[17:47:07] <fuzzie> could you grep -E '\("[^"]*$' them?
[17:47:29] <lynxlynxlynx> nothing
[17:48:26] <lynxlynxlynx> yay
[17:48:50] <fuzzie> ok, that is strange, BIS *not* making an error in any of their games? :)
[17:50:24] <fuzzie> re screenshots: we save too-large screenshot bitmaps
[17:51:47] --> Avenger has joined #GemRb
[17:51:48] --- ChanServ gives channel operator status to Avenger
[17:52:28] <Avenger> damnit, another game version inconsistency. the movable flag is not honoured by bg1
[17:52:47] <Avenger> sw1h13 cannot be moved in bg2, but it is movable in bg1
[17:52:57] <Avenger> both has the movable flag unset
[17:53:38] <fuzzie> well, we already knew some flags were different
[17:53:49] <Avenger> this sucks
[17:54:01] <fuzzie> easy enough to add another GF?
[17:54:38] <Avenger> adding the gameflag is easy, but how do i handle the bg1 version. I'm sure it is used for something, most likely it is still a 'no drop' thing
[17:55:03] <Avenger> i'll just test that by ctrl-y-ing xan :D
[17:57:37] <fuzzie> poor xan :(
[17:57:47] <fuzzie> as if bg2 wasn't hurtful enough
[17:58:36] <Avenger> yes, it is still a no-drop item
[17:58:54] <Avenger> so meeh
[17:59:41] <fuzzie> it doesn't really seem so bad compared to things like missing quotemarks :)
[18:01:07] <Avenger> btw, items of dead people are not drawn
[18:01:17] <Avenger> this is something for lynxlynxlynx :)
[18:01:34] <fuzzie> should they be?
[18:01:51] <Avenger> no, i meant, they shouldn't be drawn in gemrb
[18:02:01] <lynxlynxlynx> they already aren't :)
[18:02:05] <Avenger> hmm
[18:02:08] <fuzzie> inventory should be greyed out?
[18:02:11] <lynxlynxlynx> only the ground items are
[18:02:22] <lynxlynxlynx> it should, i added code for it
[18:02:23] <Avenger> they are greyed out, but i swear i saw items on people
[18:02:36] <lynxlynxlynx> yeah, probably what's on the ground
[18:02:54] <lynxlynxlynx> didn't bother with that
[18:03:30] <Avenger> we need this no drop flag, soon we'll reach the 64 bit limit
[18:03:32] <fuzzie> or just CheckOnDeath, where death isn't necessarily instant?
[18:03:39] <lynxlynxlynx> and possibly unmovable items don't get dropped
[18:03:41] <fuzzie> we can add another int for the flags :)
[18:04:34] <Avenger> yep, i will
[18:04:52] <fuzzie> i'd much rather we use as many GF flags as we need, if it's ever a problem then we can add something more efficient
[18:06:08] <Avenger> yes, and remove any gametype checks ;)
[18:07:36] <-- Avenger has left IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
[18:08:12] <lynxlynxlynx> nice find by usurper earlier
[18:09:23] --> Avenger has joined #GemRb
[18:09:23] --- ChanServ gives channel operator status to Avenger
[18:18:04] <Avenger> ok we'll have arbitrary number of game flags :)
[18:18:29] <fuzzie> vector<bool>?
[18:18:41] <fuzzie> would be fast, too
[18:20:41] <Avenger> hehe, handcoded
[18:21:20] <fuzzie> well, i guess that's fine too
[18:21:27] <tomprince> Or perhaps std::bitset?
[18:22:00] <fuzzie> well, it would be nice to have something which is sure to be fast
[18:22:18] <fuzzie> but i don't really care, just a suggestion if Avenger hadn't done anything yet
[18:23:44] <Avenger> it is something like: array[position>>5] & (position&31)
[18:23:55] <Avenger> err
[18:24:10] <Avenger> it is something like: array[position>>5] & (1<<(position&31))
[18:24:23] <fuzzie> yeah, it just seems pointless to do anything complicated :)
[18:24:34] <Avenger> array size is (GF_COUNT+31)/32
[18:24:42] <fuzzie> i am trying to find some annoying reference counting bug again
[18:25:07] <Avenger> how does it manifest? crash or just nasty messages on quit
[18:25:15] <fuzzie> crash
[18:25:17] <Avenger> eep
[18:25:29] <fuzzie> yeah
[18:25:35] <Avenger> should be only on your side?
[18:26:01] <fuzzie> it only happens with some scripts
[18:26:14] --- barra_fife is now known as barraAway
[18:29:09] <Avenger> wrong scripts?
[18:34:32] <fuzzie> broken code
[18:40:25] <fuzzie> i don't see how this ever works
[18:52:40] <-- Avenger has left IRC (Quit: bye!)
[19:25:40] <CIA-29> GemRB: 03lynxlupodian * rd9cfd7cce811 10gemrb/gemrb/override/ (how/classes.2da iwd/classes.2da):
[19:25:41] <CIA-29> GemRB: iwd, how: Eggert J??n Magn??sson fixed a regression from 15745a62 causing
[19:25:41] <CIA-29> GemRB: bad racial class restrictions
[19:26:34] <Lightkey> something does not like thems characters?
[19:26:58] <lynxlynxlynx> cia probably
[19:27:09] <lynxlynxlynx> ó and ú
[19:27:51] <Lightkey> uh, you sure?
[19:28:02] <fuzzie> they look fine in the repos
[19:28:22] <fuzzie> "Eggert Jón Magnússon", utf8 as required
[19:31:48] <Lightkey> oh, islndisch
[19:31:53] <Lightkey> not used to those ;-)
[20:31:05] <CIA-29> GemRB: 03avenger_teambg * r65e70e91e6e6 10gemrb/gemrb/ (core/Inventory.cpp plugins/AREImporter/AREImporter.cpp): update item flags in containers (at least the movable flag counts)
[20:31:07] <CIA-29> GemRB: 03avenger_teambg * r198019ace6af 10gemrb/gemrb/ (11 files in 8 dirs):
[20:31:07] <CIA-29> GemRB: unlimited gameflags
[20:31:07] <CIA-29> GemRB: implemented GF_NO_DROP_CAN_MOVE flag for bg1
[20:31:07] <CIA-29> GemRB: implemented a more elaborate item_use feature
[20:31:09] <CIA-29> GemRB: 03avenger_teambg * r1116b29a39cd 10gemrb/gemrb/override/ (how/classes.2da iwd/classes.2da): Merge branch 'master' of ssh://gemrb.git.sourceforge.net/gitroot/gemrb/gemrb
[20:33:12] <lynxlynxlynx> sigh
[20:33:43] --> Avenger has joined #GemRb
[20:33:43] --- ChanServ gives channel operator status to Avenger
[20:35:48] <Avenger> i changed a little too much, so there could be bugs. The main change was to let Xan remove/equip his sword, let others (in bg1) to carry around his sword, but still Xan won't drop the sword on death, and only he can equip it
[20:36:02] <Avenger> all this while Boo is still unremovable in both games :)
[20:37:07] <Avenger> as an additional feature, i think now annah/grace in pst can swap their armor
[20:38:04] <lynxlynxlynx> please use more commits
[20:38:23] <Avenger> hmm and there was a crasher in gamecontrol: if you controlled a single char and it died, and you tried to move, game crashed
[20:38:39] <lynxlynxlynx> it's much harder to backtrack bugs when one commit does five things of which three are stated in the commit message ;)
[20:38:41] <Avenger> yeah, i know, but they were mixed
[20:38:53] <lynxlynxlynx> you can use git commit --interactive
[20:38:59] <-- SiENcE has left IRC (Quit: cya @all)
[20:39:04] <lynxlynxlynx> or git add -p to even select individual chunks
[20:39:26] <Avenger> hmm i didn't know i can do that
[20:40:31] <lynxlynxlynx> oh, the first invokes the other, so forget about git add
[20:41:58] <CIA-29> GemRB: 03avenger_teambg * r335e77f16ed8 10gemrb/gemrb/core/ (CharAnimations.cpp GameScript/Actions.cpp): cleanup
[20:42:15] <Avenger> well i hope i didn't add more bugs than i fixed :)
[20:42:51] <lynxlynxlynx> the option printing is not really useful
[20:43:06] <Avenger> ah, damn, i knew i left something in
[20:43:20] <Avenger> i needed it to see if the options are still read correctly
[20:43:30] <lynxlynxlynx> yeah
[20:43:34] <Avenger> as i changed the storage
[20:43:38] <Avenger> you can remove it again
[20:44:20] <lynxlynxlynx> please do it yourself, it'll be faster
[20:44:31] <lynxlynxlynx> i'm in the middle of the global id changes
[20:45:34] <CIA-29> GemRB: 03avenger_teambg * r7d531dc872b0 10gemrb/gemrb/core/Interface.cpp: cleanup
[20:45:52] <Avenger> i considered adding some debug option for that, but i wanted the commit out as fast as possible :)
[20:46:26] <Avenger> did you already consider setting the global id of non actors?
[20:46:35] <Avenger> or it is already done?
[20:47:51] <fuzzie> already done
[20:48:12] <fuzzie> i am not yet done with the changes, i want to removal all pointers which could possibly lead to crashes
[20:48:16] <Avenger> oh yeah, this container has a global id
[20:48:19] <Avenger> whoa
[20:48:20] <Avenger> :)
[20:48:54] <lynxlynxlynx> i'm trying to make the summon's actionbar work
[20:49:03] <Avenger> heh?
[20:49:06] <fuzzie> oh, Avenger fiddled with the removeitem code
[20:49:11] <Avenger> wasn't that working?
[20:49:12] <fuzzie> here comes another nightmare :)
[20:49:29] <lynxlynxlynx> it was working in the sense that it was displayed
[20:49:30] <fuzzie> Avenger: you added some hack which didn't work properly :P
[20:49:44] <fuzzie> lynx is just switching everything to use globalid, so it'll all work
[20:50:01] <lynxlynxlynx> it quickly got me into the core though
[20:50:02] <Avenger> i added race dependent default actionbar, how is that connected to globalids?
[20:50:19] <fuzzie> not that bit
[20:50:21] <lynxlynxlynx> we rely on party slots for almost everything
[20:50:29] <Avenger> ah well, party slots
[20:50:34] <lynxlynxlynx> summons, charmed ones etc are not in there
[20:50:39] <Avenger> summons don't have them :)
[20:50:53] <fuzzie> sure, so we have to use globalids for everything
[20:50:58] <lynxlynxlynx> yes
[20:51:01] <Avenger> you don't have to kill all party slots, though
[20:51:09] <Avenger> fyi: the IE uses them too :)
[20:51:23] <fuzzie> no, for the UI they're fine
[20:51:28] <Avenger> not everywhere, though
[20:51:37] <fuzzie> although it seems a bit pointless to have both ssytems
[20:52:15] <Avenger> it uses party slots for portrait icons i guess, and to check if it needs to print some feedback
[20:52:39] <lynxlynxlynx> everywhere, even for selections
[20:53:07] <Avenger> and the other part of selections uses pointer? :)
[20:53:22] <Avenger> that's probably a bit fishy
[20:53:31] <fuzzie> well, we're very careful about the selection stuff
[20:53:51] <fuzzie> i added paranoid checks everywhere too
[20:57:35] <Avenger> looks like pst also has this 'nodropcanmove' flag
[21:01:56] <Avenger> hmm, when grace equips bod_gp, there is a text; effect resisted: PriestSpellSlotsModifier
[21:02:08] <Avenger> i guess that is not good
[21:03:16] <Avenger> meh, its resist flag is 'dispel/not-bypass resistance' it has '1-self' as target though
[21:03:35] <Avenger> so, i guess self effects are never resisted
[21:04:48] <fuzzie> i thought we already did that
[21:05:02] <lynxlynxlynx> i added it recently, but maybe that place was not enough
[21:05:20] <fuzzie> but this isn't the case in bg1, right?
[21:06:05] <fuzzie> oh, right, there's a GF flag
[21:06:21] <fuzzie> not set for pst
[21:09:23] <Avenger> i set it already
[21:09:33] <Avenger> just didnt' commit
[21:09:41] <fuzzie> ok
[21:09:44] <Avenger> the swap check isn't correct yet
[21:09:55] <fuzzie> for effects?
[21:10:18] <fuzzie> i am working on dialog stuff
[21:14:11] <Avenger> pst has some items, that allow removing only if you got a replacement item
[21:17:14] <lynxlynxlynx> like eyes, i guess
[21:17:42] <fuzzie> ah
[21:17:45] <fuzzie> that would be cool to have
[21:26:51] <Avenger> hah, it is so much fun trying to strip Annah :)
[21:27:07] <CIA-29> GemRB: 03avenger_teambg * re4d353184dff 10gemrb/gemrb/ (3 files in 2 dirs): implemented swap-only items in pst
[21:29:27] <Avenger> hmm, inventory items are not protected, so Grace's diary is still movable. iirc it is unmovable in the original
[21:30:17] <fuzzie> mmmh
[21:30:23] <fuzzie> what is this residue action stuff?
[21:30:39] <Avenger> a leak, i think
[21:30:50] <Avenger> action that has no owner?
[21:31:10] <fuzzie> it is a crash, here
[21:31:17] <Avenger> eep
[21:31:25] <fuzzie> it says residue action and then deletes the action!
[21:31:37] <Avenger> shouldn't it do so? :P
[21:31:40] <fuzzie> no
[21:31:43] <fuzzie> this is in the middle of the game :P
[21:32:03] <Avenger> but it thinks the action is not needed anymore
[21:32:42] <fuzzie> yes, it is strange
[21:32:44] <Avenger> what is the refcount?
[21:33:47] <fuzzie> i mean, the residue code is not at fault
[21:33:56] <fuzzie> since it only does release once
[21:34:14] <Avenger> yes, it wouldn't delete, just release
[21:34:17] <Avenger> so refcount is bad
[21:34:22] <fuzzie> but the refcount is ok
[21:34:41] <Avenger> but then the release wouldn't delete it
[21:34:52] <Avenger> it deletes it too early, so the refcount is fewer than it should
[21:34:54] <fuzzie> it doesn't
[21:35:00] <Avenger> hmm then what crashes?
[21:35:01] <fuzzie> but, yes, that
[21:35:20] <fuzzie> it ends up being too early
[21:36:14] <Avenger> this is all strange
[21:36:26] <Avenger> last time i checked, we got actions not freed
[21:36:34] <fuzzie> aha, there is a Cache.h/Cache.cpp
[21:36:36] <Avenger> now you say sometimes they are freed twice
[21:37:06] <fuzzie> the refcounts are just broken
[21:37:22] <fuzzie> it seems that this cache destroyed things which were in use
[21:37:35] <fuzzie> i will try another valgrind run, it seems easier
[21:37:49] <Avenger> yep, it catches problems earlier than the crash
[21:38:07] <Avenger> see you tomorrow!
[21:38:09] <fuzzie> bye!
[21:38:10] <-- Avenger has left IRC (Quit: bye!)
[21:40:01] <CIA-29> GemRB: 03avenger_teambg * r813d0bbc04c1 10gemrb/NEWS: new feature in the news
[21:43:58] <fuzzie> aha, valgrind wins again
[21:46:34] <Lightkey> cue Hot Chocolate
[21:47:20] <fuzzie> something is, indeed, deleting the running script
[21:47:22] <fuzzie> not so clever
[21:47:37] <fuzzie> don't delete the script when you're the script!
[21:49:10] <lynxlynxlynx> brings a new meaning to autonomous
[21:50:10] <fuzzie> i think i am too tired to fix this one :)
[21:50:22] <fuzzie> but it is simple, ChangeAIScript just happily deletes the running script
[21:50:35] <lynxlynxlynx> there's a new day
[21:50:41] <lynxlynxlynx> in five minutes :)
[21:51:09] <fuzzie> so i'm sure something clever can be done with reference counts or magic pixie dust or something
[21:52:55] <lynxlynxlynx> there's an ai script that changes the ai script?
[21:54:26] <fuzzie> pst has a few areas with people walking around who change scripts a lot
[21:54:53] <Lightkey> lynxlynxlynx: your clock is five minutes off :p
[21:54:55] <fuzzie> so you have people wandering around an exhibit, and then some trigger which makes them move to do something else, etc
[21:58:18] <CIA-29> GemRB: 03fuzzie * rdb35c607c23f 10gemrb/gemrb/core/ (4 files in 3 dirs): clean up dialog handling code for non-actors
[21:58:21] <CIA-29> GemRB: 03fuzzie * rf79274a491a0 10gemrb/gemrb/core/Map.cpp: print globalid in Map::DebugDump too
[21:59:52] <Lightkey> tomorrow also has 25 hours as an extra
[22:00:46] <fuzzie> hmm, valgrind complains about something of Avenger's i don't follow
[22:02:46] <lynxlynxlynx> yeah, tommorow's gonna be awesome
[22:03:11] <Lightkey> no no, today
[22:03:36] <fuzzie> you know what SourceFlags does, lynx?
[22:03:49] <fuzzie> oh, checked for SF_HOSTILE, so some kind of .. thing
[22:06:01] <lynxlynxlynx> they either hold the item or spell flags of the effects parent
[22:06:24] <fuzzie> in this case there isn't a parent
[22:06:37] <fuzzie> well, i guess there is
[22:06:40] <lynxlynxlynx> SF_HOSTILE marks spells that are treated as hostile for ea changing and magic resistance purposes (maybe more)
[22:07:16] <fuzzie> well, i am going to just set it to 0 in this commit, and maybe someone who understands it better can fix it if needed
[22:07:45] <lynxlynxlynx> always?
[22:07:57] <fuzzie> for equipping effects
[22:08:18] <lynxlynxlynx> that should be fine
[22:08:19] <fuzzie> right now they get random garbage in that field
[22:08:25] <fuzzie> which has been leading to odd effects :)
[22:09:26] <fuzzie> thanks for explanation
[22:09:36] <CIA-29> GemRB: 03fuzzie * r356a65bc341b 10gemrb/gemrb/core/Item.cpp: don't use RechargeFlags from an extended header which doesn't exist
[22:10:05] <lynxlynxlynx> zzz
[22:10:21] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:10:21] <fuzzie> ninight
[22:10:23] <fuzzie> oh :)
[22:10:37] <fuzzie> well, happy Pumpkin Day everyone
[22:13:48] <Lightkey> oh yes, I will go to Blocksberg today and meet the others
[22:14:24] <fuzzie> my calendar has 'eat cake' in it, so i will do that
[22:15:08] <Lightkey> well, meeting relatives near Stuttgart, close enough
[22:15:33] <fuzzie> cake, Lightkey, cake!
[22:16:18] * Lightkey thows a cream cake into fuzzie's face
[22:18:07] <fuzzie> delicious.
[22:20:36] <Lightkey> http://www.youtube.com/watch?v=Sw8pcxE6U9k
[22:37:13] <Lightkey> http://www.youtube.com/watch?v=IWOcVFLR_DI not as great as I remember it but still fun if you get the references and I just love the end
[22:40:43] <fuzzie> hmm
[22:40:51] <fuzzie> oop
[22:41:37] <fuzzie> i did something bad somewhere
[22:42:23] <fuzzie> can't talk to grace now
[22:46:39] <fuzzie> oh, lynx has faithfully wrapped the bugfix which was only applied to bg2 in a bg2 check :-)
[22:47:33] <fuzzie> oh, because it's a bg2 action :(
[22:47:53] --> barra_ has joined #GemRb
[22:48:01] <fuzzie> ok that is annoying
[22:48:34] <fuzzie> and it's on-the-fly so i can't even put it in our override
[22:51:19] <-- barraAway has left IRC (Ping timeout: 252 seconds)
[23:35:58] <-- Maighstir has left IRC (Ping timeout: 245 seconds)
[23:50:33] <-- deepinthewoods has left IRC (Ping timeout: 245 seconds)