#gemrb@irc.freenode.net logs for 19 Apr 2011 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:03:34] --> Bo_Thomsen has joined #gemrb
[00:34:42] <-- Bo_Thomsen has left IRC (Quit: Leaving.)
[01:07:38] --> pupnik_ has joined #gemrb
[01:10:25] <-- pupnik has left IRC (Ping timeout: 260 seconds)
[01:11:32] --- pupnik_ is now known as pupnik
[05:59:29] <-- |Cable| has left IRC (Remote host closed the connection)
[06:01:01] --> |Cable| has joined #gemrb
[06:13:24] --> Cable_ has joined #gemrb
[06:13:25] <-- |Cable| has left IRC (Read error: Connection reset by peer)
[06:14:08] <-- Cable_ has left IRC (Client Quit)
[06:14:21] --> |Cable| has joined #gemrb
[06:14:30] <-- |Cable| has left IRC (Remote host closed the connection)
[06:14:41] --> |Cable| has joined #gemrb
[06:21:59] --> lynxlynxlynx has joined #gemrb
[06:22:00] --- ChanServ gives channel operator status to lynxlynxlynx
[06:39:18] <-- |Cable| has left IRC (Remote host closed the connection)
[06:40:38] --> |Cable| has joined #gemrb
[07:01:52] --> Bo_Thomsen has joined #gemrb
[07:13:49] --> edheldil_ has joined #gemrb
[07:20:48] <-- edheldil_ has left IRC (Ping timeout: 248 seconds)
[07:35:30] --> lubos has joined #gemrb
[07:35:53] <-- pupnik has left IRC (Ping timeout: 240 seconds)
[07:58:29] <dhewg> moin
[08:00:33] <dhewg> fuzzie: what does "never had their effects processed" mean?
[08:05:26] <-- Bo_Thomsen has left IRC (Quit: Leaving.)
[08:08:29] --> pupnik has joined #gemrb
[08:08:29] <-- pupnik has left IRC (Changing host)
[08:08:29] --> pupnik has joined #gemrb
[08:17:52] <fuzzie> dhewg: in statis :P
[08:18:02] <fuzzie> e.g. never present in a loaded map
[08:18:22] <fuzzie> your reproduction recipe was v.helpful for knowing where to look :)
[08:19:06] <fuzzie> maybe we can just flip that check, will look shortly, have an hour or two this morning
[08:23:16] <fuzzie> well i guess you could try it, too
[08:23:28] <fuzzie> i'm just worried it might corrupt saves, but should be obvious from a save/load cycle..
[08:25:31] <dhewg> i can try in ~8h
[08:25:41] <fuzzie> also you should commit/push the IE_CLASS thing so i can cherry-pick it :-P
[08:27:28] <dhewg> its a valid fix?
[08:27:43] <fuzzie> it looked like it
[08:27:50] <dhewg> i thought the real problem was that area loading that wasnt supposed to happen at that point
[08:28:07] <fuzzie> no, that just caused the problem to happen too early
[08:30:34] <fuzzie> and honestly, anything which fixes a problem is good to have in-tree, even if it needs HACK warnings all over
[08:31:04] <dhewg> heh
[08:31:16] <fuzzie> i just want to find time to ponder about the regeneration thing
[08:31:27] <dhewg> so what was the real problem then?
[08:31:38] <dhewg> that harper dude has invalid attributes?
[08:31:51] <fuzzie> a skeleton corpse in the new area has invalid attributes, i guess
[08:32:22] <dhewg> oh
[08:33:48] <fuzzie> i tend to assume the whole engine is obvious if you're asking smart questions :P
[08:35:38] <edheldil> it would be even better if somebody wrote about the engine or its parts :-P
[08:36:36] <fuzzie> yeah, but you can ask us questions on irc! how is that not a perfect documentation method?
[08:38:00] <dhewg> :)
[08:38:04] <dhewg> pushed my fix
[08:41:19] <CIA-52> GemRB: 03dhewg * rc7fba9263341 10gemrb/gemrb/core/Scriptable/Actor.cpp:
[08:41:20] <CIA-52> GemRB: extend class check in Actor::RefreshEffects
[08:41:20] <CIA-52> GemRB: fixes a crash in ar502/bg2, which has actors with invalid attributes
[08:49:25] <-- |Cable| has left IRC (Remote host closed the connection)
[09:34:17] <lynxlynxlynx> :)
[09:41:07] <fuzzie> did you see the regen thing, lynx?
[09:41:31] <lynxlynxlynx> few days ago?
[09:41:35] <fuzzie> yeah
[09:41:51] <lynxlynxlynx> i haven't tried it ingame, but yes
[09:44:02] <fuzzie> well, i mean, the fix
[09:46:16] <fuzzie> hmm, never mind actually, i guess the current code has several ways to break anyway
[10:21:09] --> SiENcE has joined #gemrb
[10:28:47] <lynxlynxlynx> i don't remember, is it in the backlog?
[10:29:43] <fuzzie> yeah, but it's easy to summerise: the current code adds AI_UPDATE_TIME onto the last update time, which breaks if the last update time is far behind
[10:51:14] <fuzzie> we should see how gemrb copes with the unfixpacked bg2 Azuredge Axe
[10:56:54] <lynxlynxlynx> bonus against undead?
[10:58:05] <fuzzie> the fixpack notes say it had corrupt effect indices
[11:05:57] <fuzzie> and the SoA map merge stuff is still wonky
[11:06:43] <fuzzie> i end up dumped at completely the wrong entrance, which should never happen
[11:29:27] <lynxlynxlynx> map merge?
[11:29:38] <fuzzie> adding Watcher's Keep to the map
[11:29:57] <fuzzie> i stopped it from crashing some week ago
[11:32:52] <lynxlynxlynx> oh
[11:33:26] <fuzzie> lecture now though
[12:48:19] <fuzzie> ok, hi.
[12:58:15] <CIA-52> GemRB: 03lynxlupodian * rfe37913c3bf5 10gemrb/gemrb/ (3 files in 3 dirs):
[12:58:16] <CIA-52> GemRB: ResolveRandomItem: skip nonexisting items (eg. scrl88 in bg1)
[12:58:16] <CIA-52> GemRB: also added a few guards for old saves
[12:58:29] <fuzzie> nice
[13:18:14] <-- lubos has left IRC (Remote host closed the connection)
[13:22:44] --> ubermad has joined #gemrb
[13:30:29] <edheldil> hi, ubermad
[13:33:53] <ubermad> eloo edheldil
[13:35:46] <ubermad> I have a problem with pst, look http://img35.imageshack.us/i/dsc00782du.jpg/ annoying gray line
[13:36:34] <ubermad> how to get rid of it
[13:37:30] <fuzzie> that is beholder's android-specific code, no?
[13:38:28] <dhewg> isnt that the touchscreen define?
[13:39:12] <fuzzie> yes, it's hardcoded in, nice
[13:39:22] <fuzzie> so complain at the android porter i guess
[13:39:53] <ubermad> so i can change this?
[13:40:08] <fuzzie> only if you can make your own android build
[13:41:15] <ubermad> android build?
[13:41:27] <fuzzie> oh, or whatever build you're using :P
[13:46:11] <ubermad> next question, I have no character animation during the fight, I only see blood when someone is hit
[13:47:12] <fuzzie> huh
[13:48:23] <ubermad> and hear the sounds of fight
[13:50:13] <edheldil> ubermad: as far as I know, the lines are a feature - they show part when you can scroll
[13:50:48] <edheldil> where
[13:50:51] <fuzzie> yes
[13:50:57] <fuzzie> not being able to turn them off is a bit of a crazy misfeature though
[13:56:48] <ubermad> lines are not so bad but what's with the animation
[14:03:22] <lynxlynxlynx> bad type i guess
[14:05:41] <lynxlynxlynx> re regeneration: torgal was the test case?
[14:05:46] <lynxlynxlynx> it seems easy to fix
[14:06:05] <fuzzie> yes, there's a fix in dhewg's repos
[14:06:16] <fuzzie> just s/lastHeal/gameTime/
[14:06:35] <fuzzie> i have a modified one in my local copy which also makes it check fx->FirstApply and not !fx->Parameter3
[14:07:05] <lynxlynxlynx> hehe, i did the same
[14:07:17] <lynxlynxlynx> as dhewg
[14:07:39] <lynxlynxlynx> so it just needs testing or what?
[14:07:52] <fuzzie> i think it's good as-is
[14:08:03] <fuzzie> i just thought you'd better look at it
[14:08:17] <lynxlynxlynx> both changes are fine
[14:09:14] <fuzzie> i would also like to fix that local variables thing
[14:09:22] <fuzzie> but that needs testing
[14:11:44] <ubermad> how can I fix animation durring the fight?
[14:13:10] <fuzzie> they just don't animate at all
[14:13:10] <fuzzie> ?
[14:17:24] <ubermad> during fight no
[14:17:32] <lynxlynxlynx> heh, i went for the azure edge, noticed the slavers were still there, went through it all
[14:18:07] <lynxlynxlynx> now hendak can't kill lehtinen, because he has the sword in the wrong hand and is dealing stunning damage with the fist instead
[14:19:07] <fuzzie> intrsting
[14:20:11] <lynxlynxlynx> he's probably left-handed
[14:20:33] <dhewg> lol
[14:20:55] <lynxlynxlynx> Equipped: 1000 <-- nothing
[14:21:13] <lynxlynxlynx> but it was visible, that he had the sword in his offhand
[14:21:42] <fuzzie> i *did* write all the code for this down
[14:21:47] <lynxlynxlynx> if that's a problem, it should be easily reproducible without any of this
[14:23:10] <fuzzie> if there's nothing equipped in primary it will try the EquipMostDamagingMelee thing
[14:24:00] <lynxlynxlynx> nope, my tob fighter can attack with the offhand alone too
[14:24:25] <lynxlynxlynx> i wonder if hendak just had a too crappy thac0 to hit with it
[14:24:27] <fuzzie> yeah
[14:24:30] <fuzzie> wait
[14:24:32] <fuzzie> which game? :P
[14:24:40] <lynxlynxlynx> soa
[14:24:52] <fuzzie> i mean, i'm saying, original will try EquipMostDamagingMelee if there's nothing in primary slot, and by that i mean Equipped i guess
[14:25:09] <fuzzie> and then bail out for that frame with AP_WEAPON_UNUSABLE
[14:25:12] <lynxlynxlynx> bbl
[14:25:22] <fuzzie> so if you can fight with offhand in original, i guess Equipped should be able to point there?
[14:25:57] <fuzzie> i am curious if we handle that
[14:45:27] <fuzzie> huh, netbook has no bg2 isos on it, which is a bit odd since it has bg2 installed and i have no external optical drive..
[14:52:33] <fuzzie> oh right, fist is a valid weapon, sigh
[15:10:14] --> budlust has joined #gemrb
[15:10:17] <-- Gekz has left IRC (Remote host closed the connection)
[15:13:15] --> Maighstir has joined #gemrb
[15:17:45] --> Gekz has joined #gemrb
[15:33:39] <-- budlust has left IRC (Read error: Connection reset by peer)
[15:34:00] --> budlust has joined #gemrb
[15:50:19] --> Bo_Thomsen has joined #gemrb
[15:52:12] <fuzzie> dhewg: i assume you're not able to test yet :P
[15:55:18] <-- Gekz has left IRC (Remote host closed the connection)
[16:04:46] <-- SiENcE has left IRC (Quit: @all: cya)
[16:11:12] --> Gekz has joined #gemrb
[16:11:12] <-- Gekz has left IRC (Remote host closed the connection)
[16:12:12] --> edheldil_ has joined #gemrb
[16:14:02] <-- budlust has left IRC (Quit: Lost terminal)
[16:16:16] <dhewg> am now
[16:16:51] <-- edheldil_ has left IRC (Ping timeout: 248 seconds)
[16:19:04] <dhewg> what should i test?
[16:21:23] --> Gekz has joined #gemrb
[16:21:41] <fuzzie> change core/EffectQueue.cpp:1773 from 'return false' to 'return true'
[16:22:23] <dhewg> line mismatch on 0.6.4, that 1st return on "/we save this as variable"?
[16:22:30] <fuzzie> yes
[16:22:51] <dhewg> that maybe fixes the locals storage?
[16:22:58] <fuzzie> in theory..
[16:26:03] <dhewg> instead of "00178ch Variable torgaldies" i see "001714h Resource 1 torgaldi"
[16:29:13] <fuzzie> drat.
[16:29:24] <fuzzie> that's what i was worried about
[16:30:33] <fuzzie> hm i must've misunderstood the ielister stuff too
[16:30:41] <fuzzie> that 'Variable?' field is presumably a flag
[16:32:31] <dhewg> wtf
[16:32:50] <dhewg> i just got a rest-video in the middle of a fight
[16:33:01] <fuzzie> with a tent/campfire?
[16:33:05] <dhewg> yeah
[16:35:16] <fuzzie> more concerned about variables :P
[16:35:37] <dhewg> ;)
[16:35:48] <fuzzie> muh
[16:35:54] <fuzzie> i mean i guess 'Variable?' is 0 there?
[16:37:14] <dhewg> yeh
[16:37:19] <fuzzie> ugh
[16:38:12] <dhewg> also at another offset
[16:38:35] <dhewg> no, nevermind
[16:39:36] <fuzzie> hmm
[16:40:59] <fuzzie> well that sucks
[16:41:38] <fuzzie> //isvariable
[16:41:39] <fuzzie> stream->Write( filling,4 );
[16:41:54] <fuzzie> perhaps i should just go through and add an abort() whenever we write filling
[16:41:55] <fuzzie> sheesh
[16:45:33] --> test32894789234u has joined #gemrb
[16:46:54] <fuzzie> this is hiorrible :(
[16:47:17] <dhewg> the writing code?
[16:47:19] <fuzzie> yes
[16:47:25] <fuzzie> http://fuzzie.org/nfs/gemrb/effimporter_hax.txt pls
[16:48:21] <fuzzie> might corrupt your saves though!
[16:48:26] <dhewg> on top of that true/false toggle?
[16:48:30] <fuzzie> yes please
[16:48:42] <fuzzie> stream->Write( filling,40 ); //12+32+8
[16:48:47] <fuzzie> ^- this was pretty much the icing on the cake
[16:49:17] <dhewg> lol
[16:50:20] <dhewg> gemrb: /home/andre/devel/gemrb/gemrb/plugins/CREImporter/CREImporter.cpp:2947: virtual int CREImporter::PutActor(DataStream*, Actor*, bool): Assertion `stream->GetPos() == CREOffset+ItemsOffset' failed.
[16:50:27] <fuzzie> excellent, i miscounted
[16:50:41] <dhewg> see, assert()/abort() is good!
[16:50:47] <dhewg> needs more of those
[16:50:50] <fuzzie> yes guess who put those in there :-P
[16:51:01] <dhewg> :)
[16:52:07] <fuzzie> refreshed, could try that
[16:54:07] <dhewg> 00172ch Variable? 00000001
[16:54:12] <dhewg> 00178ch Variable torgaldies
[16:54:17] <dhewg> 001714h Resource 1 torgaldi
[16:54:30] <fuzzie> yeah i guess i should fix that too :P
[16:54:37] <dhewg> :P
[16:54:49] <fuzzie> but does it work?
[16:54:56] <dhewg> but saves&loads fine
[16:55:06] <dhewg> define work
[16:55:34] <fuzzie> do you lose the local variables
[16:55:55] <-- Bo_Thomsen has left IRC (Remote host closed the connection)
[16:56:21] <dhewg> i consider myself lucky, because i have a save here where nalia is taken away and i changed the map once, but the locals are still there
[16:56:35] <fuzzie> yes, that works
[16:56:35] <dhewg> loading and just saving that loses the locals without your haxx
[16:56:41] <fuzzie> because the locals are proper locals
[16:56:50] <dhewg> those lines where taken from that save+resave
[16:57:02] <dhewg> so that looks better so far
[16:57:18] <fuzzie> try http://fuzzie.org/nfs/gemrb/effimporter_hax.txt I guess?
[16:57:33] <dhewg> another change?
[16:57:35] <fuzzie> if that's ok then i'll commit and wait for Avenger to come complain
[16:57:46] <dhewg> i used the 2nd try already
[16:57:48] <fuzzie> yes, sorry, that patch tried to zero out the Resource fields
[16:57:54] <fuzzie> tries to
[16:58:18] <dhewg> np, i'll run it again
[16:58:32] <dhewg> also, you can always have a save for such situations
[16:59:43] <fuzzie> btw
[16:59:50] <fuzzie> i take it you don't have duplicate locals for other actors?
[17:00:22] <dhewg> didnt see any yet, but maybe i missed something
[17:00:34] <dhewg> that output from ielister is kinda huge
[17:00:53] <fuzzie> well you can grep for 'Variable' i hope
[17:01:00] <dhewg> well, comparing only 'torgaldies' yields a 100% match from a good save vs a resave with the 3rd try
[17:01:19] <fuzzie> yes, that one will work
[17:01:19] <dhewg> all fields good :)
[17:01:30] <fuzzie> i just want to know if the in-map actors get duplicates, because that would be bad
[17:03:13] <dhewg> ielister 000000045-test/baldur.gam|grep Variable|awk '{ print $3 }'|wc -l
[17:03:13] <dhewg> 849
[17:03:16] <dhewg> ielister 000000045-test/baldur.gam|grep Variable|awk '{ print $3 }'|uniq|wc -l
[17:03:16] <dhewg> 848
[17:03:34] <fuzzie> well there's an obvious question at this point :-p
[17:03:43] <dhewg> inowrite
[17:04:40] <dhewg> ah, just empty entried for $3
[17:04:47] <dhewg> *entries
[17:05:55] <dhewg> afaict those are all Effect opcod 000000e9
[17:06:04] <dhewg> no varname set
[17:06:17] <dhewg> so if that is no issue there are no duplicates
[17:07:03] <dhewg> in neither, the old save and the one with that patch
[17:07:17] <fuzzie> um
[17:07:19] <fuzzie> they have IsVariable set?
[17:07:47] <fuzzie> ('Variable?' in ielister)
[17:08:04] <dhewg> no, zero
[17:08:22] <dhewg> did you put junk in my savegame? :P
[17:08:56] <fuzzie> oh i see
[17:09:20] <fuzzie> never mind, didn't realise ielister displayed Variable always
[17:09:47] <fuzzie> no, i still don't see, that's weird, why the difference?
[17:10:30] <fuzzie> oh well
[17:11:43] <CIA-52> GemRB: 03fuzzie * r1029602469fd 10gemrb/gemrb/ (core/EffectQueue.cpp plugins/EFFImporter/EFFImporter.cpp):
[17:11:44] <CIA-52> GemRB: Save variable effects properly.
[17:11:44] <CIA-52> GemRB: This fixes EFFImporter to save variable names properly, and stops
[17:11:44] <CIA-52> GemRB: EffectQueue from sabotaging fx_local_variable. It might still need
[17:11:44] <CIA-52> GemRB: some work for the earlier games to make sure they save the effects
[17:11:44] <CIA-52> GemRB: as the right version?
[17:11:45] <CIA-52> GemRB: This also makes EFFImporter save CasterLevel while I'm futzing with it.
[17:11:46] <fuzzie> if you're feeling spectacularly helpful you could cherry-pick that and make sure i didn't screw it up completely
[17:12:50] <fuzzie> hm i guess i should have thanked you in the commit messasge too
[17:13:04] <fuzzie> feel free to be annoyed at my impoliteness, forgetful
[17:13:32] <dhewg> heh
[17:13:40] <dhewg> well, i dont care about that
[17:13:43] <dhewg> thx for the fix
[17:13:55] <dhewg> i'll just yell in here if i notice some breakage about it :)
[17:14:06] <dhewg> and i failed at the grep
[17:14:15] <dhewg> if i add -w the wc matches
[17:14:29] <dhewg> it yielded Variable? too before
[17:14:52] <fuzzie> ah
[17:15:31] <fuzzie> well, the more bugs like that fixed, the better..
[17:15:40] <fuzzie> it's a lot easier to fix logic bugs than nasty 'oops, we corrupted your save' ones
[17:15:51] <dhewg> yeah
[17:16:05] <dhewg> and i think this could end in broken games
[17:16:07] <dhewg> maybe
[17:16:20] <dhewg> it should happen everytime you dismiss a party member
[17:16:25] <fuzzie> yes
[17:16:26] <fuzzie> bad
[17:16:40] <dhewg> local quest vars with mismatching .are states and fun like that
[17:19:01] <fuzzie> i spent quite a lot of time chasing up vital data we just deleted in saves
[17:22:18] <dhewg> is it normal/ok that spells between saves change in order?
[17:23:40] <fuzzie> um
[17:23:49] <fuzzie> seems strange
[17:24:15] <fuzzie> i think i added paranoia checks for that too though, shouldn't lose any
[17:27:07] <dhewg> seems to be related to your patch
[17:27:25] <dhewg> it looks like its all there, but in another order
[17:30:39] <fuzzie> it shouldn't be that patch
[17:31:53] --> |Cable| has joined #gemrb
[17:32:33] <dhewg> http://static.hackmii.com/dhewg/save.diff
[17:32:54] <dhewg> just a loaded game from before that patch saved again without doing anything with that patch
[17:33:04] <dhewg> if i save that again, only the time differs
[17:33:08] <fuzzie> and you lost HP? :p
[17:33:33] <dhewg> i dunno? :P
[17:33:48] <fuzzie> that seems not so great
[17:38:21] <dhewg> hm no, not related to that patch
[17:38:31] <dhewg> restart+load+save does that
[17:43:04] <fuzzie> good
[17:47:47] <dhewg> sometimes i get this: [Map]: This isn't a travel region [11.77]?
[17:47:47] <dhewg> Travel direction returned: -1
[17:50:38] <fuzzie> when trying to move areas?
[17:51:30] <dhewg> yeah
[17:51:43] <dhewg> when it definitely is a spot to leave the area
[17:52:00] <fuzzie> that code is most definitely in the 'rewrite' column
[17:52:20] <fuzzie> at the moment we have some really weird hack
[17:53:33] --> Bo_Thomsen has joined #gemrb
[17:59:15] <-- dhewg has left IRC (Read error: Operation timed out)
[18:09:10] <fuzzie> *do* dreams trigger in inns, in original?
[18:17:52] --> barra_home has joined #gemrb
[18:23:39] <fuzzie> can't make it happeh
[18:25:08] <Maighstir> don't they *only* trigger in inns?
[18:25:17] <fuzzie> bg2 ones, i mean
[18:25:23] <Maighstir> ah
[18:25:44] <Maighstir> then I don't actually know
[18:29:34] <fuzzie> :o
[18:29:37] <fuzzie> that is unusual :)
[18:40:45] --> dhewg has joined #gemrb
[18:44:21] <fuzzie> nice, i managed to crash original..
[18:52:35] <dhewg> i raise you with one gemrb backtrace: http://pastie.org/1812637
[18:57:38] <fuzzie> ok didn't we fix that one already? :P
[18:57:58] <fuzzie> you fix my original crash and we'll be even!
[18:58:05] <dhewg> i think i have a deja vu too
[19:00:01] --> edheldil_ has joined #gemrb
[19:01:55] <-- test32894789234u has left #gemrb
[19:03:55] <dhewg> also tons of Couldn't load animation: mberg2, cycle 19
[19:04:12] <dhewg> its a summoned bear
[19:04:38] <dhewg> it seems to be stuck in that animation
[19:05:01] <dhewg> if i select it (only see the green circle), we reappears and the warning spam stops
[19:10:31] <fuzzie> mmh
[19:10:58] <fuzzie> well, that is more understandable as a bug
[19:11:40] <fuzzie> but you reported gibberlings being broken the other day, which is weirder
[19:11:42] <fuzzie> shouldn't happen
[19:12:13] <fuzzie> the problem is that it's one of those things which got fixed by piling more and more random code on top
[19:12:28] <fuzzie> and it's really difficult to work out what it should be
[19:14:06] <fuzzie> while bears just have the wrong anim type in the data
[19:16:05] <fuzzie> but e.g. MOGH/MOGN are marked as anim type 3, IE_ANI_TWO_FILES, but they (and the bears) should be anim type 14, same as the gibberlings
[19:16:08] <fuzzie> so, *flail*
[19:18:02] <dhewg> hehe :p
[19:18:04] <dhewg> marked where?
[19:18:11] <dhewg> that avatar file again?
[19:18:16] <dhewg> or bad original data?
[19:18:43] <fuzzie> yes, in the avatar file
[19:19:01] <fuzzie> my problem is that you reported the gibberlings as broken
[19:19:20] <dhewg> but i think that was bg1
[19:19:24] <fuzzie> ah!
[19:19:32] <fuzzie> ok, that makes sense
[19:19:37] <dhewg> yeah, that tutor fighting cellar
[19:19:50] <dhewg> im not sure if they were broken on bg2 anymore
[19:20:03] <dhewg> but definitely in bg1
[19:21:26] <fuzzie> ok, well, anim type 14 is: MOGH, MOGN, MBAS, MBER, MDOG, MDOP, METT, MGHL, MGIB, MSLI, MSPI, MWLF, MWLS, MXVT, MTAS, MZOM, MWER, MGWE
[19:22:53] <fuzzie> plus i suppose MMAX, MKUR and MDOC which are MDK characters and it looks like MEYE, MMST and weirdly enough MMEL.
[19:23:23] <lynxlynxlynx> muttated gibberelings in bg2 worked for me today (they're in the coronet)
[19:24:38] <fuzzie> let me edit our avatars file for that
[19:25:44] <fuzzie> lynxlynxlynx: the weird thing is, MMEL is Melissan
[19:28:15] <lynxlynxlynx> yeah
[19:28:24] <lynxlynxlynx> mmel is horribly broken now btw
[19:28:31] <fuzzie> in gemrb?
[19:28:34] <lynxlynxlynx> yep
[19:28:38] <fuzzie> ok
[19:28:42] <fuzzie> well, then i guess i can change it :P
[19:28:48] <lynxlynxlynx> so if this is due to her secondary attack, this may just fix it :)
[19:28:52] <fuzzie> MMEL is unfortunately hardcoded into the original in a few stupid places
[19:29:26] <fuzzie> so this won't fix it, we need a special 'is melissan anim' flag or equivalent
[19:29:51] <fuzzie> i'm actually surprised at how accurate this avatars file is
[19:32:24] <lynxlynxlynx> you can also push the regen fix while you're at it
[19:32:42] <fuzzie> oh, i thought you'd do it, but ok
[19:32:42] <lynxlynxlynx> tomprince: prodprod gcc
[19:32:55] <fuzzie> i'm afraid i just made the union thing worse :-P
[19:33:05] <fuzzie> didn't think about it
[19:34:23] <tomprince> lynxlynxlynx: ?
[19:34:55] <lynxlynxlynx> please finish the gcc 4.6 patch
[19:35:29] <fuzzie> did we split out the other changes?
[19:38:19] <CIA-52> GemRB: 03fuzzie * r31b207f54811 10gemrb/gemrb/override/bg2/avatars.2da: set a bunch of bg2 anims to IE_ANI_FOUR_FILES_2
[19:38:29] <CIA-52> GemRB: 03fuzzie * ra556fc5307b7 10gemrb/gemrb/plugins/FXOpcodes/FXOpcodes.cpp: fix fx_set_regenerating_state timing (thanks to dhewg)
[19:39:29] <fuzzie> for my future reference looking at the logs, that commit has correct types, my notes are slightly wrong
[19:41:11] <lynxlynxlynx> no, there was no split
[19:41:22] <tomprince> ... so fix your notes?
[19:41:47] <fuzzie> tomprince: yeah, but i typed them in above :P
[19:42:02] <fuzzie> so in the interests of not exploding my head
[19:44:36] <dhewg> wee, much better with the bears :)
[19:44:38] <dhewg> thx
[19:45:57] <fuzzie> yay
[19:46:09] <fuzzie> well at least one useful thing came out of that
[19:46:13] <dhewg> :)
[19:46:42] <dhewg> while all those warnings are gone now, normal trolls are invisible sometimes
[19:47:14] <dhewg> is this maybe related?
[19:47:15] <fuzzie> umm
[19:47:15] <dhewg> [ResourceManager]: Searching for mtrs.2da...[Override]
[19:47:15] <dhewg> [2DAImporter]: Bad signature!
[19:47:15] <dhewg> [Actor]: TODO:Cannot determine 2DA rowcount for index: 67
[19:47:18] <fuzzie> define normal trolls
[19:47:34] <fuzzie> an anim id from ctrl-m on them would be ideal
[19:47:51] <dhewg> what shows up as "Troll" in the message box
[19:47:55] <dhewg> not the small ones
[19:48:02] <dhewg> sec, i'll look for another :)
[19:48:33] <lynxlynxlynx> the signature bit is harmless
[19:48:56] <dhewg> k
[19:48:59] <dhewg> Debugdump of Actor Troll (Troll, Troll):
[19:48:59] <dhewg> Scripts: troll01 <none> <none> <none> <none> <none> <none> wtatroll
[19:48:59] <dhewg> Area: ar1900 Dialog:
[19:49:02] <dhewg> Mod[IE_ANIMATION_ID]: 0x7F00
[19:49:04] <fuzzie> probably 7F00 which is standard extra monster anim which is IE_ANI_FOUR_FILES=2, except they're all 4 in avatars.2da
[19:49:46] <fuzzie> :(
[19:50:12] <tomprince> so .... has anybody ever compiled snprintf.cpp ?
[19:51:11] <fuzzie> tomprince: on one specific platform i think?
[19:52:11] <fuzzie> i imagine old mingw32
[19:52:46] <fuzzie> i think it relies on nutty headers, from memory
[19:52:48] <tomprince> it tries to include system/types.h and data.h !?!
[19:52:52] <fuzzie> yes
[19:52:55] <fuzzie> samba stuff i think?
[19:53:23] <tomprince> I was trying to compile it with msvc, since apparently, snprintf there is the old stupid version that doesn't tell you the size.
[19:54:13] <fuzzie> someone was complaining relatively recently about it not compiling, but i didn't think about it
[19:54:42] <fuzzie> but it's from weird build lands long ago, we already ruined portability beyond the point where snprintf wouldn't be present, you'd think
[19:55:52] <fuzzie> dhewg: could you change the type of the avatars.2da line for 7F00 to 2, and see if that fixes it?
[19:56:02] <dhewg> yes!
[19:56:11] <fuzzie> you could, or it does? :P
[19:56:18] <dhewg> former :P
[19:56:41] <dhewg> just because the druid grove has a few of them
[19:56:52] <fuzzie> a small number :-P
[19:57:19] <fuzzie> lynxlynxlynx: hello! you seem to have been the last one to change all these anim types
[19:57:22] <fuzzie> oh drat, in 2009
[19:57:24] <fuzzie> how time flies
[19:57:48] <lynxlynxlynx> aha
[19:58:12] <fuzzie> oh i sabotaged your MMST i see :-P
[19:58:36] <fuzzie> looking at history just makes me much more terrified to touch these
[19:58:52] <fuzzie> there's really very few distinctions between 0x7xxx types in the original
[19:58:58] <dhewg> [CharAnimations]: Couldn't create animationfactory: mtrog1e (7f00)
[19:59:11] <lynxlynxlynx> i was mainly silencing the every-frame console spam during/after the playthroughs
[19:59:22] <lynxlynxlynx> just looking at the individual bams
[20:00:08] <fuzzie> dhewg: hmm, let me check how this works then
[20:04:56] <dhewg> i think he goes invisible right after: Performattack for Giant Troll, target is: Analo
[20:05:17] <dhewg> and yes, every time a npc mentions that name i giggle
[20:05:20] <fuzzie> oh i forgot some hardcoding
[20:08:58] <fuzzie> so *all* the monster anims are hardcoded at 8 orientations before mirroring, that's helpful
[20:10:07] <fuzzie> but that means i'm super confused about what's with the non-simple anims. let me poke a bit.
[20:21:23] <fuzzie> i think my real question now is, where does the 'e' come from?
[20:22:49] <fuzzie> i wonder if we're hilariously using leftover BAMs
[20:24:39] <fuzzie> we probably are
[20:24:54] <fuzzie> *sigh*
[20:25:05] <dhewg> whats a leftover bam?
[20:25:25] <fuzzie> files which are present, because they were used in bg1, but which aren't used by the original game engine
[20:26:33] <fuzzie> in bg1, the 'e' files contained the east-looking animations for many anim types
[20:27:06] <fuzzie> but in bg2, they handle those by mirroring the west-looking animations at blit time
[20:27:57] <fuzzie> if anyone can easily edit BAMs in the original, i would much appreciate someone scribbling all over the 'e' bams for something (gibberlings?) and seeing if the original engine notices
[20:28:40] <dhewg> and bg2/gemrb now uses the e variant of bg1 bams?
[20:28:51] <fuzzie> yes
[20:29:09] <dhewg> from where?
[20:29:18] <dhewg> they're part of a bg2 install?
[20:29:22] <fuzzie> yes, hence 'leftover'
[20:29:59] <lynxlynxlynx> i don't remember any trolls in bg1, so this is plausible
[20:30:06] <fuzzie> of course this doesn't work for, say, MEYE, because it's not from bg1 to be leftover
[20:30:09] <fuzzie> which means i just broke it
[20:32:15] <fuzzie> i have an 80-page pad here for bg2 notes from the last few weeks, and it has half full already
[20:33:31] <tomprince> I am going to remove snprintf, since it doesn't semm to work. If somebody needs it, they can resurrect it, or add different implementation that works.
[20:33:58] <fuzzie> ok
[20:34:06] <fuzzie> you don't need it? :)
[20:35:17] <-- ubermad has left #gemrb
[20:36:00] <tomprince> Decided to use a stack _MAX_PATH sized buffer. If that truncates ... no loss.
[20:37:33] <fuzzie> just as long as you don't strncpy :P
[20:45:04] <fuzzie> this is interesting because MTRO as a 0x7xxx animation only uses MTROG1 and MTROG2
[20:45:14] <fuzzie> but there's MTROG11, MTROG12, MTROG13, MTROG22, etc
[20:48:32] <fuzzie> and IC_TROLL1 is an icewind character, different BAMs entirely, so it's not that
[20:49:03] <fuzzie> so i guess MTROG11 etc are the bg1 troll anims
[20:50:28] <fuzzie> despite the complete lack of any MTRO files in my bg1 install.
[20:50:32] <fuzzie> wah.
[20:50:56] <dhewg> sounds like fun :P
[20:51:14] <fuzzie> well it's just horrible
[20:51:32] <fuzzie> why does bg2 have Icewind Dale anims anyway?!
[20:52:16] <dhewg> uhm
[20:52:22] <dhewg> they used vs source safe? :P
[20:54:23] <dhewg> http://static.hackmii.com/dhewg/ertof.png
[20:54:40] <dhewg> are they supposed to attack me?
[20:55:04] <fuzzie> they did, in fact, use visual sourcesafe
[20:55:41] <fuzzie> but i'm hoping only on the incompetent publishing side
[20:55:59] <dhewg> hah
[20:56:12] <fuzzie> and oh i have that 'appears busy' check disabled locally for such reasons :P
[20:56:34] <dhewg> well i totally dont recall that scene
[20:56:41] <dhewg> no idea what is supposed to happen
[20:57:06] <dhewg> if i talk to ertof, jaheira says "fight as best you can" and thats it
[20:57:15] <dhewg> they dont go hostile
[20:57:42] <fuzzie> yeah, it's the 'appears busy' which presumably breaks it, one moment
[20:57:51] <fuzzie> i shall apply grep
[20:58:38] <fuzzie> yes, the conversation's meant to continue
[20:58:41] <fuzzie> but the interjection broke it
[20:59:01] <dhewg> there're multiple persons at the exact same spot
[20:59:25] <dhewg> i kill wurm, then tamile stands there. i kill him and turg is there
[20:59:44] <dhewg> then jarden
[21:00:14] <dhewg> thats it, just 4 :P
[21:00:15] <CIA-52> GemRB: 03fuzzie * r524bcbf130e3 10gemrb/gemrb/core/DialogHandler.cpp: disable the IF_NOINT check in DialogChoose until we work out what to do with it
[21:00:15] <fuzzie> go cherry-pick latest commit
[21:00:48] <fuzzie> i only have two changes left in my local tree, my stashing has clearly gone wrong somewhere
[21:00:59] <dhewg> heh
[21:01:03] <dhewg> huh
[21:01:16] <dhewg> if i kill all those dudes, i get "jaheira - has nothing to say to you"
[21:01:26] <dhewg> what are my ladies trying to tell me all the time?
[21:01:54] <fuzzie> i imagine lynx's theory about banter is probably right
[21:02:44] <dhewg> but it seems like she wants to tell me something about the now dead ertof?
[21:02:54] <fuzzie> she didn't just get damaged?
[21:03:06] <fuzzie> or, someone else
[21:03:14] <fuzzie> she'll complain if Aerie is below 10 HP, or Mazzy below 15 HP, etc
[21:03:14] <dhewg> the whole scene is kinda broken
[21:03:26] <dhewg> no, all fully rested
[21:03:32] <fuzzie> i mean, during battle
[21:04:04] <fuzzie> oh
[21:04:06] <fuzzie> no you're right :P
[21:04:12] <dhewg> i rested, they poped up, i attacked, they didnt fight back
[21:04:22] <fuzzie> if Ertof, Jarden, Turg, Tamile and Wurn are dead, she will talk to you
[21:04:24] <dhewg> everyone was at 100%
[21:05:11] <fuzzie> except we suck
[21:05:16] <dhewg> lol
[21:05:29] <fuzzie> and InitDialog calls dlg->FindFirstState
[21:05:31] <fuzzie> dammit
[21:06:06] <dhewg> so imoen wanted to tell me something about that murdering harper soab
[21:06:07] <fuzzie> that doesn't work because the SetGlobal in the script, to set the first state up, is *after* the dialog init
[21:08:15] <dhewg> gah
[21:08:16] <CIA-52> GemRB: 03fuzzie * r38d2126b1848 10gemrb/gemrb/core/GameScript/GSUtils.cpp: execution continues after BeginDialog
[21:08:22] <dhewg> he killed her!
[21:08:30] <fuzzie> (that doesn't help)
[21:08:31] <dhewg> thanks for fixing that :P
[21:08:53] <fuzzie> do you have a save around there?
[21:09:00] <fuzzie> it would help if i had some testcase
[21:09:14] <dhewg> yeah
[21:09:17] <dhewg> sec
[21:10:01] <fuzzie> pretty sure i know what's wrong, i was just hoping no-one would notice
[21:10:58] <dhewg> http://static.hackmii.com/dhewg/ertof.tgz
[21:11:09] <dhewg> just rest to trigger the scene
[21:11:32] <dhewg> there're still 4 ppl at the same spot, and i bet they're supposed to go hostile at one point
[21:16:24] <lynxlynxlynx> we don't do any npc bumping yet, so the same spot thing is not related
[21:24:47] <dhewg> i also cant kill him?
[21:24:51] <dhewg> ertof
[21:24:58] <dhewg> he has 1hp left
[21:25:02] <dhewg> always
[21:32:48] <fuzzie> hi. i return with cookies.
[21:33:24] <dhewg> yummy
[21:33:31] <fuzzie> 'fix the save directory' might be a good one for the todo list
[21:34:51] <dhewg> ?
[21:35:07] <fuzzie> bg2 tries saving ToB games to mpsave/
[21:36:03] <dhewg> heh
[21:36:27] <fuzzie> ok when i click sleep, that really doesn't quite work
[21:38:56] <lynxlynxlynx> dhewg: probably has one of those minhp items
[21:39:04] <fuzzie> yes
[21:39:07] <fuzzie> it is
[21:39:14] <dhewg> a what?
[21:39:26] <fuzzie> an item designed solely to stop someone from dying
[21:40:40] <dhewg> but with fuzzie's dlg fix and a certain dlg path jaheira frees herself and im supposed to fight the guys, right?
[21:40:51] <fuzzie> yes
[21:40:54] <dhewg> then they do turn hostile
[21:40:57] <fuzzie> i reproduced
[21:41:01] <fuzzie> he ends up with minhp1
[21:41:40] <dhewg> ah there
[21:41:48] <dhewg> can i pickpocket that? :P
[21:44:16] <fuzzie> i'm not sure how this is meant to work
[21:45:39] <fuzzie> SAY #20524 /* ~Pssst...psssst. Hey. Hey there.~ */
[21:45:39] <fuzzie> IF ~~ THEN DO ~DestroyItem("MINHP1")~ GOTO 1
[21:45:43] <fuzzie> ^- like that i suppose
[21:46:55] <dhewg> thats his 1st line
[21:47:03] <fuzzie> it is indeed
[21:47:09] <fuzzie> i expect someone sabotaged DestroyItem again, am waiting for git
[21:47:12] <dhewg> whats "if ~~"
[21:47:21] <lynxlynxlynx> empty condition
[21:47:31] <dhewg> heh
[21:47:33] <lynxlynxlynx> i was probably the last to change di
[21:47:39] <lynxlynxlynx> for undroppables
[21:47:52] <lynxlynxlynx> minhp1 should also have it set
[21:48:05] <dhewg> maybe an uppercase issue again? :)
[21:48:06] <fuzzie> no, it was Avenger
[21:48:23] <fuzzie> no, Avenger just moved it
[21:48:42] <fuzzie> probably me, then
[21:48:49] <dhewg> somehow that guy is funny
[21:48:58] <dhewg> RISE AND SHINE! *stab jaheira*
[21:51:58] <fuzzie> the code actually looks fine
[21:54:57] <fuzzie> but flags here is 0x28a9 which means it has IE_INV_ITEM_STACKED set
[21:58:22] <fuzzie> that seems bad from a doing-the-right thing point of view, but it can't be the issue here, hm
[21:59:21] <fuzzie> lynxlynxlynx: your 'Nonexistent random item' happens an awful lot
[21:59:59] <lynxlynxlynx> what items?
[22:00:13] <fuzzie> you're checking in the wrong place, i guess
[22:00:39] <fuzzie> this way you reject gold, because it looks for stuff like '012.itm' and doesn't find
[22:02:20] <fuzzie> but also "[Interface]: Nonexistent random item (bad table entry) detected: rndscr01"
[22:02:57] <fuzzie> the check should be just before the 'return true;' i guess
[22:03:14] <fuzzie> should i mess with it?
[22:03:35] <lynxlynxlynx> sure, i'm off to bed
[22:03:39] <fuzzie> night
[22:04:56] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:07:48] <-- barra_home has left IRC (Quit: Verlassend)
[22:08:03] <fuzzie> ugh that issue is a bit complicated
[22:10:14] <dhewg> found the cause?
[22:10:25] <fuzzie> no, distracted by random item one
[22:12:20] <fuzzie> don't know if you have that applied to your tree
[22:12:30] <fuzzie> but doesn't look good for inventories. waiting for a build now.
[22:12:44] <dhewg> no, dont have that
[22:16:11] <fuzzie> oh dear
[22:16:21] <fuzzie> so DestroyItem isn't an instant
[22:19:37] <fuzzie> i wonder how that works
[22:20:47] <fuzzie> this shouldn't work, according to my knowledge of the original..
[22:21:22] <fuzzie> however ielister segfaults here, so hm
[22:22:01] <dhewg> hah
[22:27:17] <fuzzie> silly bug
[22:27:33] <CIA-52> GemRB: 03fuzzie * r2c8c87a1f2f8 10ielister/ielister.cpp: don't crash on strings longer than VARSIZE
[22:28:21] <fuzzie> but really, that action should get wiped before it has a chance to run, darnit
[22:30:35] <fuzzie> oh i suppose gemrb does, in fact, wipe it before it has a chance to run
[22:31:18] <dhewg> it has to be non-instant?
[22:31:33] <fuzzie> no, it *should* be instant
[22:31:46] <fuzzie> otherwise it should get wiped when the dialog interjection happens
[22:31:52] <-- Bo_Thomsen has left IRC (Quit: Leaving.)
[22:32:09] <fuzzie> however some 'fuzzie' says "when it is set, the "queue wipe" above seems to wipe everything *except* any running (blocking) action"
[22:33:38] <fuzzie> so sigh
[22:33:40] <fuzzie> :P
[22:36:46] <CIA-52> GemRB: 03fuzzie * rb467a3244e09 10gemrb/gemrb/ (6 files in 3 dirs):
[22:36:46] <CIA-52> GemRB: fix ResolveRandomItem nonexistent item check
[22:36:46] <CIA-52> GemRB: also fix all the callers so they check the return result and don't do an
[22:36:46] <CIA-52> GemRB: unnecessary extra check themselves
[22:51:55] <fuzzie> doesn't look so hard to deal with, but sleepy time
[23:14:43] <-- Maighstir has left IRC (Quit: ~ Trillian Astra - www.trillian.im ~)