#gemrb@irc.freenode.net logs for 18 Apr 2011 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage

[00:23:01] <-- Maighstir has left IRC (Quit: ~ Trillian Astra - www.trillian.im ~)
[00:39:54] <-- edheldil_ has left IRC (Read error: Operation timed out)
[01:07:17] --> pupnik_ has joined #gemrb
[01:10:02] <-- pupnik has left IRC (Read error: Operation timed out)
[01:12:15] --- pupnik_ is now known as pupnik
[03:36:21] --> raevol has joined #gemrb
[04:43:14] <-- raevol has left IRC (Quit: Leaving.)
[05:53:22] --> edheldil_ has joined #gemrb
[06:10:19] <-- edheldil_ has left IRC (Ping timeout: 248 seconds)
[06:51:09] --> Bo_Thomsen has joined #gemrb
[07:01:05] --> lynxlynxlynx has joined #gemrb
[07:01:05] --- ChanServ gives channel operator status to lynxlynxlynx
[07:14:35] --> lubos has joined #gemrb
[07:19:23] <-- Bo_Thomsen has left IRC (Quit: Leaving.)
[07:24:03] <edheldil> pupnik: I have n900 now
[07:26:18] <pupnik> yeey! :D
[07:26:27] <pupnik> isn't linux great?
[07:27:19] <pupnik> what's your impression of gemrb on it?
[07:28:42] <pupnik> better/worse than expected? did inspire ideas for things to change?
[07:38:24] --> SiENcE has joined #gemrb
[07:55:49] <edheldil> I haven't tried gemrb on it yet :). And I feel like Don Quixote buying this phone :)
[08:31:50] <fuzzie> ok, morning
[08:36:18] <SiENcE> hey. morning
[08:36:50] <SiENcE> finally i have an ipad2... so now i'm ready to think about an gemrb port to ipad
[08:37:27] <fuzzie> you have some way of building things for it?
[08:56:58] <pupnik> expensive monkey-press-button slate
[08:57:26] <fuzzie> yes.
[08:57:47] <fuzzie> but the alternative is *slow, buggy* expensive monkey-press-button slates. :P
[08:59:08] <SiENcE> i have an macosx vm with xcode 3.x
[08:59:32] <fuzzie> although apparently you can deal with the Xoom I/O issues pretty well by not putting anything on the SD card, ever.
[08:59:34] <SiENcE> but yet no certificate to sign things...so only simulator for testing
[08:59:36] <-- |Cable| has left IRC (Remote host closed the connection)
[08:59:45] <pupnik> my tablet has a keyboard and is called thinkpad x200t
[08:59:49] <fuzzie> can't jailbreak the ipad?
[08:59:59] <SiENcE> i dont know...maybe
[09:00:34] <fuzzie> pupnik: i looked at those, but then there's no software to run on them
[09:02:29] <fuzzie> plus a stylus never survives for more than about five minutes around here before someone manages to lose it
[09:02:46] <pupnik> hehe
[09:03:33] <pupnik> i had three tablets already so i'm not against them :)
[09:03:41] <pupnik> but they feel very limited
[09:04:17] <fuzzie> well, the lower-end stuff is awful
[09:04:28] <fuzzie> i mean, for friendly big touchy stuff
[09:05:04] <fuzzie> i haven't played with any touch devices i liked except the iPad and the Galaxy Tab
[09:05:36] <pupnik> they need omap3 or better class cpu imo
[09:05:57] <pupnik> for video + web
[09:05:59] <fuzzie> and the Galaxy Tab is a bit ruined by Android
[09:06:21] <pupnik> of course, it should run maemo or meego :)
[09:06:27] <fuzzie> i think dhewg actually owns one
[09:06:53] <pupnik> i have an archos a70s for 1/3 the price
[09:07:01] <pupnik> maybe 2/5ths
[09:07:13] <pupnik> just as fast
[09:07:36] <fuzzie> did they fix their driver situation?
[09:07:57] <pupnik> works here
[09:08:09] <pupnik> you mean wifi?
[09:08:26] <fuzzie> i imagine so
[09:12:01] <fuzzie> difficult to get past the marketing-speak for me
[09:28:16] <-- SiENcE has left IRC (Ping timeout: 264 seconds)
[09:34:40] --> SiENcE has joined #gemrb
[09:35:04] <pupnik> apparently so :)
[10:45:56] <fuzzie> so I went to a probability workgroup and learnt about matrix diagonalization.
[10:47:54] <fuzzie> and this afternoon I get to learn about the halting problem.
[10:48:42] <fuzzie> clearly made some terrible mistakes in my life somewhere.
[10:50:55] <dhewg> heh
[10:51:05] <dhewg> and yes, i do own a galaxy tab
[10:52:09] <fuzzie> you're a scary person.
[10:53:06] <pupnik> it's cool to have highend stuff
[10:53:24] <fuzzie> the Galaxy Tab isn't that highend, which is part of the issue
[10:53:26] <fuzzie> just expensive :P
[10:55:52] <fuzzie> but my highest-end mobile device has a 195mhz omap850
[10:56:00] <fuzzie> actually that's not true, i suppose the ADP1 is still here somewhere
[10:58:18] <dhewg> well, if you want something cheaper the nook looks pretty nice
[10:58:55] <fuzzie> i should really buy a kindle
[10:59:10] <dhewg> anyway, who wants to hear about another gemrb issue i ran into? :P
[10:59:20] <dhewg> anyone? bueller?
[11:00:01] <dhewg> its that dude on the bridge map killing ppl
[11:00:18] <fuzzie> go on :P
[11:00:18] <dhewg> if you confront him, he admits the murders and runs down the stairs
[11:00:29] <fuzzie> yes
[11:00:32] <dhewg> if he changes maps, gemrb segfaults
[11:00:37] <fuzzie> nice
[11:00:46] <dhewg> its because if some HP extra i assume
[11:01:04] <fuzzie> HP extra?
[11:01:04] <dhewg> i think he has either an insame number or is unkillable
[11:01:29] <fuzzie> i mean, he's meant to admit the murders, run down the stairs and escape
[11:02:03] <fuzzie> our map code is incredibly stupid though
[11:02:05] <dhewg> yes, but something is happening to his char then
[11:02:07] <dhewg> sec
[11:02:08] <fuzzie> we replace the most recently used map(!)
[11:02:17] <dhewg> http://pastie.org/private/fpjegidndq6mppzs6o5ooa
[11:02:23] <dhewg> i could work around that there
[11:02:49] <dhewg> i forgot what bonlevel was, but something insane like MAX_INT
[11:02:54] <dhewg> or -1
[11:03:14] <fuzzie> um.
[11:03:36] <dhewg> that doesnt make it crash, but i run straight into the next issue :P
[11:03:48] <dhewg> so dunno if something else is broken earlier on
[11:04:35] <fuzzie> must remember, the map stuff got broken with a "//starting from 0, so we see the most recent master area first" comment in ad22a8fa
[11:04:45] <fuzzie> which is incredibly stupid as far as i can tell
[11:05:34] <fuzzie> that patch is kinda worrying
[11:05:48] <dhewg> its prolly totally wrong
[11:05:55] <fuzzie> i mean, i guess BaseStats[IE_CLASS] must be 0
[11:06:09] <dhewg> then, with that patch, as soon as he changes maps, nalia gives me the torgal death dialog
[11:06:16] <fuzzie> but then function shouldn't be called at all
[11:06:20] <dhewg> and i get a new journal entry about the fort
[11:06:30] <dhewg> so something is totally borked there
[11:06:55] <fuzzie> hmph
[11:06:57] <fuzzie> got a save?
[11:06:59] <-- lynxlynxlynx has left IRC (Read error: Connection reset by peer)
[11:07:02] <fuzzie> i can't really look at it now
[11:07:12] <fuzzie> i have to get lunch and then go learn about the halting problem
[11:07:13] <dhewg> i do, but cant access it atm :)
[11:07:19] <fuzzie> but later, maybe
[11:07:41] <fuzzie> if i'd thought ahead i could've gone and written the halting problem poem on the blackboard before anyone got there, ah well
[11:19:03] <edheldil> as a haiku, perhaps? :)
[11:19:41] <edheldil> reminds me of spanning tree protocol poem :)
[11:19:59] <fuzzie> http://www.bemuzed.com/lucasd/halting-poem.html if you haven't seen it
[11:20:15] <fuzzie> probably rather more difficult to haiku it :P
[11:23:41] <fuzzie> though i suppose there's some others, heh
[11:31:05] <edheldil> For eternity // a program may run or more // but not halt for sure
[11:31:07] <edheldil> ;-)
[11:31:44] <fuzzie> :)
[11:33:37] <edheldil> btw, have you already read GEB-EGB? :)
[11:33:59] <fuzzie> yes
[11:34:15] <fuzzie> has been a while since, though
[12:52:24] <fuzzie> ok hi
[12:52:31] <fuzzie> turns out it wasn't halting problem, even
[12:52:40] <fuzzie> so i made a hasty escape
[13:18:21] --> ubermad has joined #gemrb
[13:36:09] <dhewg> lucky you, with berserk that map change would not work
[13:51:51] --> lynxlynxlynx has joined #gemrb
[13:51:51] <-- lynxlynxlynx has left IRC (Changing host)
[13:51:51] --> lynxlynxlynx has joined #gemrb
[13:51:51] --- ChanServ gives channel operator status to lynxlynxlynx
[13:58:49] <edheldil> heh
[15:02:06] <pupnik> nice fuzzie...
[15:02:17] <pupnik> copyspamming your link
[15:02:51] <fuzzie> v.old one
[15:03:19] <fuzzie> instead of fixing gemrb's trigger stuff i am doing optimisations for scummvm
[15:05:16] <dhewg> but... but... but... :'(
[15:19:50] --> budlust has joined #gemrb
[15:20:21] <edheldil> reading a gemrb thread at maemo.org is revealing
[15:20:47] <tomprince> in what wise?
[15:21:55] <edheldil> that some things which seem to be natural to us are nonobvious (multiple configs) or missing (rmb) to users
[15:24:48] <edheldil> I will have to look at your logging changes, hopefully they will open a road to curbing the console spam :)
[15:24:52] <fuzzie> it is frustrating in the sense that it's difficult to get anyone there to *talk*, though
[15:25:03] <fuzzie> the android stuff has been nice in the sense that they use our forum
[15:25:33] <edheldil> exactly. Revealing was not perhaps the right word
[15:25:46] <fuzzie> and there's willingness to help
[15:25:55] <fuzzie> the maemo stuff was always very much minimal-effort when i asked
[15:26:38] <fuzzie> although there's someone new building stuff since i last grumped, to be fair
[15:27:00] <edheldil> why not, though - maemo should not be too different from vanilla, they have better chance with nagging us to do their work :)
[15:28:25] <fuzzie> you'd think
[15:28:53] <fuzzie> although us sabotaging build systems doesn't really make developers like us :P
[15:29:25] <edheldil> was tghis a problem on Maemo? Can't remember
[15:30:14] <edheldil> big part of that thread is about missing rmb and shortcuts in general
[15:30:17] <fuzzie> they just patched autotools back up, i think
[15:30:21] <edheldil> ah
[15:32:12] <tomprince> Does autotools not work?
[15:32:18] <fuzzie> it didn't, at some point
[15:32:29] <edheldil> do they now?
[15:32:30] <fuzzie> afaik it does currently
[15:32:34] <edheldil> ah
[15:32:46] <pupnik> i forwarded your frump fuzzie
[15:32:59] <tomprince> according to gemrb.hocat.ca/buildbot it does.
[15:33:20] <edheldil> I got used to cmake's percents
[15:33:20] <tomprince> I even have it running the test now. Still now exec on windows.
[15:33:27] <fuzzie> hmm
[15:33:30] <tomprince> s/now/no/
[15:33:48] <fuzzie> ok, they have one change to autotools
[15:34:29] <fuzzie> - AC_PATH_PROGS(PYTHON, python python2.3 python2.2 python2.1 python2.0 python1.6 python1.5)
[15:34:32] <fuzzie> + AC_PATH_PROGS(PYTHON, python2.5 python2.3 python2.2 python2.1 python2.0 python1.6 python1.5)
[15:34:42] <fuzzie> someone did complain to me about that i think, and i forgot
[15:35:24] <edheldil> we could merge that and also get rid of old pythons
[15:35:43] <fuzzie> we should probably do 'python2.7 python2.6 python2.5 python2.4 python' or something
[15:35:53] <fuzzie> cmake does the same thing, *sigh*
[15:37:04] <fuzzie> oh wow the maemo patches are horrible :-P
[15:37:55] --> budlust_ has joined #gemrb
[15:38:02] <fuzzie> it opens /sys/devices/platform/gpio-switch/proximity/state, reads the whole thing, then checks if the first byte is 'c', in the middle of EventMgr
[15:38:19] <pupnik> aaarrrghhh
[15:38:35] <-- budlust has left IRC (Read error: Operation timed out)
[15:38:39] <pupnik> just testing new build and wondering 'why so slow'
[15:39:57] <fuzzie> define 'slow'
[15:40:15] <edheldil> it's their modifier for rmb. It's probably not supported by sdl
[15:40:20] <fuzzie> some stuff in master is limited to 15fps right now
[15:40:31] <fuzzie> edheldil: yes, it's just a bit sad that they couldn't put it in SDLVideo
[15:40:36] <pupnik> party is selected. tap on playfield. animation of party walking is jerky.
[15:41:02] <fuzzie> that is quite possibly the framerate thing
[15:41:19] <fuzzie> gemrb tries for 60fps for some reason
[15:41:47] <fuzzie> and bad things ensue
[15:41:54] <pupnik> let me be more precise, there's something arond 12+fps walking animation, then about 150ms pause, about twice a second
[15:42:08] <fuzzie> slow device?
[15:42:32] <edheldil> fuzzie: they were apparently lacking the event marshalling class we talked about :-D
[15:42:40] <pupnik> i'll try oprofile
[15:42:47] <fuzzie> edheldil: as if you need one for a hack :P
[15:42:58] <fuzzie> pupnik: easier is to check CPU usage
[15:43:11] <fuzzie> or, well
[15:43:18] <fuzzie> is it just the walking animation which stops? UI is fine?
[15:43:31] <fuzzie> also i am going to switch gemrb to 30fps now
[15:43:39] <pupnik> if you could do that, that would be great
[15:44:26] <edheldil> fuzzie: I think it was just a random number chosen by Avenger. I remember GemRB running on 100fps
[15:44:33] <fuzzie> yes
[15:44:39] <fuzzie> it should be 30fps, darnit
[15:45:00] <pupnik> should i be seeing DOS linebreaks in gemrb.cfg?
[15:45:25] <edheldil> it was added as a cap to let other processes in the OS run
[15:45:38] <fuzzie> yes, it was broken as a cap at some point, i remember fixing it
[15:46:02] <fuzzie> the problem is that the game's timekeeping is really difficult if you don't run at a multiple of 15fps
[15:48:35] <fuzzie> the original engine aims for 30fps and updates half the game objects every frame
[15:48:51] <pupnik> not a bad target for modern mobiles either
[15:49:36] <fuzzie> but i'm not sure if we actually have a problem with framerate in the original game being too low or something :)
[15:52:40] <fuzzie> imo it is quite elegant
[15:53:22] <-- lubos has left IRC (Quit: Leaving.)
[15:56:27] --> test32894789234u has joined #gemrb
[15:56:31] <-- SiENcE has left IRC (Quit: @all: cya)
[16:18:55] --> barra_home has joined #gemrb
[16:27:31] <-- budlust_ has left IRC (Quit: Lost terminal)
[16:41:23] * pupnik resists temptation to look into gemrb profile
[16:41:44] <fuzzie> you do have it set for 32bpp, right?
[16:41:54] <fuzzie> or whatever your native X depth is
[16:57:34] <-- ubermad has left IRC (Ping timeout: 246 seconds)
[17:04:44] <pupnik> going back to production libc6 helped
[17:05:04] <pupnik> i am using the current build available from maemo.org fuzzie
[17:06:03] <fuzzie> doesn't answer question :p
[17:06:19] <pupnik> ah i see bpp=16
[17:06:31] <fuzzie> does the X server run at 16bpp?
[17:07:29] <pupnik> yes
[17:07:32] <fuzzie> does look like it
[17:07:37] <fuzzie> ok then, should be good
[17:07:43] <pupnik> (==) FBDEV(0): Depth 16, (==) framebuffer bpp 16
[17:10:31] <pupnik> i love having gqview on phone
[17:13:34] <dhewg> what are .mos files?
[17:13:48] <fuzzie> dhewg: tiled UI graphics
[17:14:04] <dhewg> k
[17:14:08] <fuzzie> why? :)
[17:14:14] <dhewg> just poking
[17:14:21] <dhewg> .are are mapfiles?
[17:14:31] <fuzzie> yes
[17:15:06] <dhewg> about that fail here...
[17:15:07] <fuzzie> they refer to .wed files which define the area graphics, and .tis files which have the actual graphical tiles
[17:15:09] <dhewg> [GSUtils]: Executing MoveBetweenAreas("ar0502",[307.444],0) in EscapeAreaCore
[17:15:14] <dhewg> [GameScript]: MoveBetweenAreas: Rejiek Hidesman to ar0502 [307.444] face: 0
[17:15:21] <dhewg> and then boom
[17:15:36] <fuzzie> no backtrace? :P
[17:15:49] <fuzzie> the class thing is definitely fail on our part
[17:16:02] <dhewg> i should create an alias to get bt's from gemrb core files :P
[17:17:36] <dhewg> http://pastie.org/1807916
[17:17:55] <dhewg> see bonindex and my prev hack patch
[17:17:59] <fuzzie> yes
[17:18:02] <fuzzie> but i mean, if you have that applied
[17:18:12] <dhewg> no, thats without obvious :)
[17:18:17] <dhewg> want savegame?
[17:18:20] <fuzzie> i'm pretty sure the solution *there* is that RefreshPCStats shouldn't be called if IE_CLASS is 0
[17:19:08] <dhewg> i can try
[17:19:16] <dhewg> where/what is IE_CLASS?
[17:19:16] <fuzzie> let me just make sure that class is 0 for that
[17:19:27] <fuzzie> you can just look at the caller :)
[17:19:47] <dhewg> Inventory::AddSlotEffects?
[17:20:21] <fuzzie> no, in RefreshEffects
[17:20:42] <dhewg> oh, the other refresh
[17:20:54] <fuzzie> $ ~/src/ielister/ielister skcorp01.cre | grep "0273h Class"
[17:20:54] <fuzzie> 0273h Class 00
[17:21:01] <fuzzie> ^- fail fail is fail
[17:22:15] <fuzzie> should be 0x87, lazy bioware
[17:22:25] <dhewg> seems like no coincidence that this is the last line before the segfault: [ResourceManager]: Searching for skcorp01.cre...[chitin.key]
[17:23:33] <fuzzie> mm
[17:23:38] <fuzzie> maybe the original fixes up on load, i don't know
[17:24:28] <dhewg> what does that mean? is that the char file for rejiek?
[17:24:41] <fuzzie> no
[17:24:46] <fuzzie> it's an actor in the new area
[17:24:57] <dhewg> 502?
[17:25:03] <fuzzie> i assume
[17:25:20] <dhewg> gemrb loads 502 stuff as that dude runs downstairs
[17:25:25] <fuzzie> it shouldn't
[17:25:36] <fuzzie> but it'd just crash as you followed otherwise, so not a big difference here
[17:25:58] <dhewg> from "end dialog" to crash: http://pastie.org/1807943
[17:26:12] --> |Cable| has joined #gemrb
[17:26:48] <fuzzie> [ResourceManager]: Searching for ar0502.are...[chitin.key]
[17:26:54] <fuzzie> ^- this is your area load
[17:26:59] <dhewg> i dont seem to have 501 nor 502 in my savegame
[17:27:06] <fuzzie> mhm
[17:27:07] <dhewg> so both are fresh
[17:27:11] <fuzzie> it's bad original game data
[17:27:31] <dhewg> thats the approach there?
[17:27:42] <dhewg> patch in code or path file+overwrite?
[17:27:45] <fuzzie> but not much you can do about that, but you can stop it from misdetecting the class
[17:28:07] <fuzzie> by changing the IE_CLASS check before RefreshPCStats is called, to make sure to exclude class 0
[17:28:59] <fuzzie> makes some sense or no sense?
[17:30:19] <dhewg> makes sense, but whats 'class'?
[17:30:24] <dhewg> BaseStats[IE_CLASS]?
[17:30:39] <fuzzie> yes
[17:30:57] <fuzzie> i'm pretty sure there's a check right there anyway, which i thought you'd just have to change
[17:31:11] <dhewg> '>0 && <classcount' then?
[17:31:14] <fuzzie> there is indeed
[17:31:24] <fuzzie> yes, that sounds ideal :) change comment too if you can
[17:31:37] <fuzzie> i mean, you could just bother me into doing it
[17:31:44] <dhewg> i'll try if it fixes stuff first :P
[17:31:47] <fuzzie> but deliberately away from machine atm
[17:32:04] <dhewg> well, im here, got a savegame, and know how to compile, so...
[17:33:59] <dhewg> well
[17:34:12] <dhewg> that fixes it the same way as my hackpatch
[17:34:20] <dhewg> but the engine still is confused
[17:34:25] <fuzzie> but without the subtle breakage
[17:34:31] <dhewg> yeah
[17:34:33] <fuzzie> so what happens with it?
[17:34:57] <dhewg> but he runs away, and it consuses that action with a dead torgal
[17:35:14] <dhewg> nalia says 'dead! die you beast!'
[17:35:25] --> Beh0lder has joined #gemrb
[17:35:30] <dhewg> and i get a arnise keep journal entry
[17:35:33] <fuzzie> oh, right
[17:35:37] <fuzzie> see, that makes no sense at all
[17:35:39] <dhewg> which i already did
[17:35:44] <dhewg> the quest i mean
[17:35:45] <Beh0lder> hello all
[17:35:59] <dhewg> fuzzie: lol, right, i see. so now what? :P
[17:36:35] <dhewg> i also get this 4 times: [Inventory]: Invalid item: rndmag03!
[17:36:46] <fuzzie> that is not so good
[17:37:35] <Beh0lder> I posted save for Drizzt bug in BG1 Bugs thread
[17:37:43] <fuzzie> thanks
[17:38:02] <dhewg> gemrb seems to load all the data related to the lower level
[17:38:03] <fuzzie> probably something stupid :/ will look at it later, or lynx will
[17:38:09] <dhewg> while i didnt move at all
[17:38:14] <fuzzie> dhewg: yes, it shouldn't load the area at all, but it does
[17:38:25] <dhewg> glad we agree on that :)
[17:38:35] <fuzzie> the original engine attaches a "move to area" effect to the actor and hides it somewhere
[17:39:05] <fuzzie> so if you load a late-game original engine save into gemrb, it promptly goes off and loads about 15 areas, trying to dump a bunch of actors in their new areas
[17:39:10] <fuzzie> very irritating
[17:39:15] <Beh0lder> This savegame created before talking with Drizzt, but some gnolls are killed
[17:39:17] <dhewg> on the bright side, this is less silly than the shade lord putting the finger of death on himself
[17:39:39] <dhewg> and then yelling 'this can not be!!!1'
[17:39:40] <fuzzie> well, *that* is on the known-horrible list at least
[17:40:08] <fuzzie> Beh0lder: but easy to reproduce it? i can't try now
[17:40:42] <fuzzie> dhewg: ResolveRandomItem in Interface.cpp is responsible for random items
[17:41:41] <dhewg> is that related here?
[17:41:49] <fuzzie> 19:36 <dhewg> i also get this 4 times: [Inventory]: Invalid item: rndmag03!
[17:41:52] <dhewg> oh, that inv error
[17:41:53] <dhewg> k
[17:41:54] <fuzzie> ^- it should've resolved that
[17:42:02] <dhewg> but well
[17:42:28] <dhewg> that all may be related to gemrb loading area stuff thats not related to my party
[17:42:31] <dhewg> i guess?
[17:42:34] <fuzzie> yes
[17:42:40] <fuzzie> i just don't want it forgotten if possible :P
[17:42:48] <fuzzie> i have nalia script open too
[17:43:03] <dhewg> k, i missed the part where irc backlogs are another TODO :P
[17:43:11] <fuzzie> i'm just wondering, did you ever *get* that torgal dialog?
[17:43:17] <dhewg> yes
[17:43:26] <fuzzie> irc backlogs are the TODO where everything is forgotten forever
[17:43:33] <fuzzie> so bad idea to say something here and then not look into it
[17:43:33] <dhewg> :)
[17:44:01] <dhewg> for the record, i saved my backlog and still to plan to craft tracker items for all that stuff
[17:45:05] <fuzzie> ok. bugging us here more effective, but tracker items get done eventually :P
[17:45:07] <lynxlynxlynx> why would you get random treasure on a transition?
[17:45:24] <fuzzie> lynxlynxlynx: resolved at load time
[17:45:43] <fuzzie> so not in a party member's inventory, in some inventory in the new area
[17:45:54] <lynxlynxlynx> ah, right
[17:47:19] <fuzzie> that "Dead! Die you beast!" entry should really only trigger once, ever
[17:47:31] <fuzzie> oh, no
[17:47:36] <fuzzie> hm
[17:48:16] <fuzzie> it's stored in locals
[17:49:59] <pupnik> i'm only seeing the game map + characters, not the bordering windows of the main game screen
[17:50:33] <lynxlynxlynx> check the console, they died
[17:50:34] <pupnik> menu/whatever graphics briefly flicker when i tap playfield
[17:50:39] <fuzzie> dhewg: CREImporter.cpp:2790 should probably have 'true' as extra last parameter
[17:51:24] <fuzzie> however i don't think that matters
[17:52:59] <fuzzie> since the importer does strnlwrcpy so it will break more with that
[17:53:00] <fuzzie> guh
[17:53:24] <fuzzie> ok, better plan, can i have a save before this all breaks?
[17:54:19] <dhewg> sure
[17:54:49] <fuzzie> i really have no clue what Avenger was trying to do with the uppercase thing, it makes no sense
[17:54:59] <dhewg> http://static.hackmii.com/dhewg/rejiek.tgz
[17:56:30] <dhewg> there was the other uppercase breakage for pst
[17:57:29] <dhewg> thats with the fixpack btw, dunno if it matters in this case
[17:57:58] <fuzzie> ok, Nalia has no local variables attached there
[17:58:09] <dhewg> oh
[17:58:16] <dhewg> i just rescued her
[17:58:22] <dhewg> maybe related?
[17:58:34] <fuzzie> she wasn't with you when you killed Torgal?
[17:58:43] <fuzzie> or is there some other thing i forget with Nalia
[17:58:57] <dhewg> she was, but she was abducted and i just got her back
[17:59:07] <dhewg> torgal was before all that
[17:59:30] <fuzzie> interesting
[17:59:43] <dhewg> maybe that wiped the locals, dont look at me
[17:59:45] <pupnik> here's the error i'm getting on latest maemo build - menu icons flicker for a moment then get overdrawn/disappear http://i.imgur.com/WQYHS.jpg
[17:59:49] <fuzzie> you did get your actual Nalia back? :P
[18:00:05] <fuzzie> pupnik: after you save a game using the gui?
[18:00:19] <pupnik> these are from games saved in earlier versions
[18:00:20] <dhewg> poor nalia, dont take her away from me!
[18:00:37] <fuzzie> yes, but did you use the save gui?
[18:00:46] <pupnik> yes
[18:00:52] <fuzzie> that is probably known bug
[18:01:05] <fuzzie> not v.helpful for 0.6.4 users i know :P
[18:01:13] <fuzzie> but i didn't find time to actually check yet
[18:01:13] <pupnik> ok i'll try starting a new game ty
[18:01:22] <fuzzie> but you can just quit and load the game
[18:01:31] <fuzzie> it should be session-specific
[18:01:45] <fuzzie> i didn't actually see that before though
[18:01:50] <fuzzie> maybe different bug
[18:01:51] <lynxlynxlynx> nalia was abducted?
[18:01:53] * fuzzie flail
[18:01:57] <pupnik> i'll try all saves thx
[18:02:10] <lynxlynxlynx> by renal someone?
[18:02:38] <dhewg> yeah that funural scene with that moron in the government district
[18:02:58] <lynxlynxlynx> but that should all play out after you kill torgal, not before
[18:03:13] <dhewg> it did
[18:03:36] <fuzzie> you're being inconsistent here :-P
[18:03:39] <lynxlynxlynx> so how is it that she wasn't there when you killed him
[18:03:58] <dhewg> 19:59:07 <dhewg> torgal was before all that
[18:04:14] <fuzzie> so not 'it did', then ;p
[18:04:56] <dhewg> well, i mean it did play out after killing torgal
[18:05:32] <lynxlynxlynx> tomprince: i get unused var warnings due to the logging changes; was everything merged?
[18:05:54] <lynxlynxlynx> dhewg: so was she there in the keep or not!?
[18:06:08] <dhewg> hehe, yes she was!
[18:06:22] <lynxlynxlynx> ok
[18:06:53] <lynxlynxlynx> my fear is that we loaded a fresh nalia when you got her back
[18:07:08] <fuzzie> hence my question
[18:07:11] <lynxlynxlynx> did she have her inventory?
[18:07:24] <lynxlynxlynx> i hope the scripts don't autodump that on you like with imoen
[18:07:30] <dhewg> no, i got all her items when she was taken
[18:07:40] <dhewg> well not all, most. but she still had a couple left
[18:08:03] <lynxlynxlynx> were they still there when you got her back?
[18:08:16] <dhewg> i think so
[18:08:19] <lynxlynxlynx> besides her ring, which is unremovable
[18:08:38] <dhewg> i think the bow was still in the secondary weapon slot
[18:09:30] <dhewg> but even if something is wrong with her, does it make sense that she gives me that torgal line at that moment?
[18:09:38] <fuzzie> no
[18:09:51] <dhewg> uh oh
[18:09:52] <fuzzie> but most likely something is triggering a dialog
[18:10:01] <fuzzie> and the first matching entry is that one
[18:10:56] <dhewg> hm
[18:11:17] <dhewg> when i enter rejiek's house alone, i get 23k exp when he runs
[18:11:32] <dhewg> its the house of horrors!
[18:12:42] <fuzzie> sigh
[18:13:08] <fuzzie> she's killed 61 creatures, which is not bad
[18:13:12] <CIA-52> GemRB: 03lynxlupodian * r73c85fb200e4 10gemrb/gemrb/core/System/Logging.cpp: fixed compilation with NOCOLOR set
[18:13:15] <dhewg> looks like his running to 502 triggers a death?
[18:13:30] <fuzzie> that would be annoying
[18:13:34] <fuzzie> any idea who? :P
[18:14:30] <dhewg> :P
[18:15:47] <lynxlynxlynx> Beh0lder: i got a crash too, before even seeing drizzt
[18:16:01] <fuzzie> oh well i can leave it in lynx's hands :P
[18:16:08] <lynxlynxlynx> fuzzie: perhaps captain brego or whowasit
[18:16:18] <fuzzie> i really can't look at this too closely right now
[18:16:24] <lynxlynxlynx> depending on your choice, he goes into the house before you
[18:16:36] <lynxlynxlynx> dhewg: the dude that gave you the quest
[18:17:18] <dhewg> this murder quest?
[18:17:43] <dhewg> aegisfield
[18:18:07] <lynxlynxlynx> ah, yes
[18:18:36] <lynxlynxlynx> did he go check it out himself after you brought all the evidence?
[18:19:03] <dhewg> i didnt show him any evidence so far, i think
[18:19:14] <lynxlynxlynx> the crash was in Factory::IsLoaded btw, looked like a bad resref
[18:20:33] <lynxlynxlynx> again, with the same item :)
[18:20:37] <pupnik> why put fullscreen in gemrb.cfg if it's specified by baldur.ini?
[18:21:29] <fuzzie> because we don't deal nicely with the .ini yet
[18:21:43] <pupnik> k, so planned for later stuff
[18:25:01] <dhewg> if i enter the house with nalia alone, she doesnt give me that torgal line
[18:26:50] <fuzzie> yes, it's bound to be her talking to you that triggers it
[18:28:02] <lynxlynxlynx> maybe we just fire off the wrong banter?
[18:29:16] <fuzzie> does it work like that?
[18:29:30] <fuzzie> i mean, if she has no local variables, her dialog file is going to see Torgal is dead
[18:29:59] <lynxlynxlynx> sure
[18:30:06] <fuzzie> and so it seems legitimate that this entry gets shown
[18:36:46] <dhewg> where do you see 'no locals'?
[18:38:26] <fuzzie> there aren't any local variable effects attached to Nalia in your baldur.gam
[18:38:51] <fuzzie> i have no idea how to see it otherwise
[18:39:31] <dhewg> hm
[18:39:49] <dhewg> i assume the loss of those is bad for my game?
[18:40:01] <lynxlynxlynx> the crash is probably due to bad random treasure generation
[18:40:22] <fuzzie> well yes, the loss of them is bad for your game if you actually had any
[18:40:24] <lynxlynxlynx> we want a scrl88.itm, but there is none
[18:41:00] <lynxlynxlynx> http://sprunge.us/EfFj?diff <-- this fixes it, but i'll try to get to the root cause
[18:42:01] <lynxlynxlynx> even better
[18:42:06] <dhewg> fuzzie: how do i dump those from the .gam?
[18:42:07] <lynxlynxlynx> it's from their own 2da
[18:42:42] <fuzzie> dhewg: i am using the 'ielister' tool the 'ielister' gemrb repos
[18:45:59] <dhewg> hm
[18:46:23] <lynxlynxlynx> do you think it's ok to pick another item in ResolveRandomItem if the chosen one doesn't exist?
[18:47:07] <dhewg> fuzzie: what now? can i restore those locals from an older savegame?
[18:47:23] <fuzzie> are they *in* an older savegame?
[18:47:35] <dhewg> i dont know, i assume so?
[18:48:47] <fuzzie> well, good to find out
[18:49:02] <dhewg> yeah, well wine urgh
[18:49:15] <fuzzie> i mean, use ielister
[18:49:20] <fuzzie> find out :P
[18:49:28] <dhewg> its a totalcmd plugin?
[18:49:41] <fuzzie> nooo
[18:49:52] <fuzzie> really, the one in our repository, undocumented anywhere else
[18:50:08] <fuzzie> oh i see, you're reading the readme :-P
[18:50:15] <fuzzie> just run the 'build' shell script
[18:51:40] <dhewg> by repository, you mean the sf.net downloads?
[18:52:10] <lynxlynxlynx> git
[18:52:39] <fuzzie> gemrb.git.sourceforge.net/gitroot/gemrb/ielister
[18:53:02] <fuzzie> i don't quite know what that is as a real url but i assume you can check the .git/config from your gemrb checkout
[18:58:13] <dhewg> what am i looking for?
[18:58:25] <dhewg> for these locals
[18:58:33] <fuzzie> "Effect opcod 000000bb" is a local
[18:58:53] <fuzzie> (the name is under that, e.g. "Variable minscwilltalk")
[18:59:02] <fuzzie> so you want one of those, in Nalia's section
[18:59:41] <fuzzie> 'torgaldies' is the relevant one for after torgal
[19:00:00] <fuzzie> possibly saved as TORGALDIES
[19:01:07] <dhewg> yeah, i see that in the save before the abduction
[19:01:19] <fuzzie> guh
[19:01:39] <dhewg> even in the savegame while she's abducted
[19:02:09] <-- test32894789234u has left #gemrb
[19:03:36] <fuzzie> that's pretty bad then
[19:04:43] <dhewg> ohnoes
[19:05:01] <fuzzie> if you hover over Nalia, you can set them from gemrb's python console with SetGlobal("torgaldies", "LOCALS", 1)
[19:05:18] <fuzzie> where 1 is the 'Variable?' entry and "torgaldies" is the "Variable" entry, obviously
[19:05:29] <fuzzie> but i am pretty worried about why that would happen
[19:05:48] <dhewg> is her abduction state itself a local?
[19:05:54] <fuzzie> quite possibly
[19:05:56] <fuzzie> not sure
[19:07:02] <fuzzie> shouldn't matter as long as you don't copy it
[19:08:47] <dhewg> the last savegame i have with the locals has exactly 7 of those
[19:10:19] <dhewg> naliatalkedto, complain, naliacomplainsisaea, torgaldies, naliaar1302, naliaar1303, bnalia4
[19:13:23] <fuzzie> ask lynx :P really busy atm
[19:18:06] <dhewg> hm
[19:18:08] <dhewg> well
[19:18:24] <dhewg> its probably not related to the abduction itself
[19:18:50] <dhewg> i have a savegame here where she is not in the party (taken away)
[19:18:54] <dhewg> the locals are in there
[19:19:04] <dhewg> i load it, just save it again, and the locals are gone
[19:19:09] <fuzzie> drat.
[19:19:16] <fuzzie> what case are the vars in the save?
[19:19:24] <dhewg> all lower
[19:20:45] <dhewg> does that maybe explain why i see imoen standing at the dungeon exit?
[19:22:40] <dhewg> heh, 'aeriehasproperportrait'?
[19:25:35] <lynxlynxlynx> i guess play that part again and save right before you get her back, so we can confirm if the problem starts on reentry
[19:25:55] <lynxlynxlynx> the aerie thing is probably due to her initial ogre nature
[19:26:54] <lynxlynxlynx> oh, i missed some
[19:29:54] <dhewg> what are all those nalia locals?
[19:30:53] <lynxlynxlynx> dialog and quest stuff
[19:31:06] <lynxlynxlynx> stat
[19:31:10] <lynxlynxlynx> state even
[19:36:35] <Beh0lder> bye
[19:36:58] <-- Beh0lder has left IRC (Quit: Beh0lder)
[19:37:34] <dhewg> so i tried a few things
[19:38:18] <dhewg> afaict the loss of locals happen when a party member leaves, you change the area, save, restart gemrb, save again
[19:38:42] <dhewg> i tried resaving with restarting without map change, and that seems to work
[19:39:00] <fuzzie> hmph
[19:39:09] <fuzzie> maybe we're doing something stupid like saving NPCs with the map
[19:43:17] <dhewg> lynxlynxlynx: is it safe to restore those 7 locals, or is maybe one of those related to the abduction quest?
[19:44:04] <lynxlynxlynx> probably safe
[19:44:09] <lynxlynxlynx> let me check naliacomplainsisaea
[19:44:43] <lynxlynxlynx> looks fine
[19:45:36] <fuzzie> the actor load/save code looks ok
[19:48:59] <dhewg> SetGlobal("torgaldies", "LOCALS", 1) + save doesnt make that local appear in my savegame
[19:49:12] <dhewg> nalia is selected and the cursor points at her
[19:52:35] <dhewg> i see this on the console though: KeyRelease:32 - 2
[19:55:45] <dhewg> sorry, wrote LOCAL, works with plural :)
[20:08:06] --> edheldil_ has joined #gemrb
[20:09:19] <dhewg> heh, now i see "nalia - has nothing to say to you"
[20:09:51] <dhewg> what is it with this house?
[20:11:25] <fuzzie> so indeed, a dialog is being triggered
[20:17:10] <lynxlynxlynx> dlg/bg2/naliaj.d:2038
[20:19:45] <dhewg> is this maybe related to the loading of the area while im still on the upper floor?
[20:20:19] <dhewg> jaheira talks to me when entering 502, and warns me
[20:23:45] <lynxlynxlynx> jaheira has the same dialog chunk, so i wonder why nalia's wasn't found
[20:24:19] <lynxlynxlynx> the way it is coded it seems normal though
[20:24:22] <lynxlynxlynx> only one can comment
[21:37:21] <-- barra_home has left IRC (Quit: Verlassend)
[22:07:23] --> barra_home has joined #gemrb
[22:08:36] <-- barra_home has left IRC (Read error: Connection reset by peer)
[22:18:45] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:27:29] <-- edheldil_ has left IRC (Ping timeout: 240 seconds)
[22:29:20] <fuzzie> ok the local variables thing is still worrying
[22:29:28] <fuzzie> but unfortunately my homework seems to have eaten my whole evening
[22:56:55] <fuzzie> ok
[22:56:59] <fuzzie> so the problem is, Avenger sabotaged it :P
[22:57:25] <fuzzie> EffectQueue::Persistent doesn't store StoreLocalVariable effects
[22:57:56] <fuzzie> but that destroys local variables on global actors who never had their effects processed.
[23:01:09] <fuzzie> it's from 2006 and the effects appear to self-destruct fine, so that should probably be removed, but better ask Avenger