#gemrb@irc.freenode.net logs for 24 Mar 2012 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:41:36] <-- edsrzf has left IRC (Ping timeout: 260 seconds)
[01:31:27] --> edsrzf has joined #gemrb
[02:03:26] --> joneirik has joined #gemrb
[02:36:00] --> cfchris6 has joined #gemrb
[02:40:40] <cfchris6> just fyi: the current git tree crashes for me when starting a "new game" (bg2). Backtrace is at http://paste.pocoo.org/show/570308/ git bisect says 87d8674b0387c6f90e4da597 is the first bad commit
[05:44:46] <-- kida_laptop has left IRC (Ping timeout: 246 seconds)
[05:51:30] <-- joneirik has left IRC (Remote host closed the connection)
[06:49:34] --> Yoshimo has joined #gemrb
[07:12:51] --> lynxlynxlynx has joined #gemrb
[07:12:52] <-- lynxlynxlynx has left IRC (Changing host)
[07:12:52] --> lynxlynxlynx has joined #gemrb
[07:12:52] --- ChanServ gives channel operator status to lynxlynxlynx
[08:27:08] <-- lynxlynxlynx has left IRC (Ping timeout: 245 seconds)
[08:30:40] --> lynxlynxlynx has joined #gemrb
[08:30:40] <-- lynxlynxlynx has left IRC (Changing host)
[08:30:40] --> lynxlynxlynx has joined #gemrb
[08:30:40] --- ChanServ gives channel operator status to lynxlynxlynx
[10:27:54] --> suncoss has joined #gemrb
[10:28:40] <suncoss> hi all
[10:31:43] <-- Yoshimo has left IRC (Quit: Yoshimo)
[10:34:19] <-- suncoss has left IRC (Quit: Page closed)
[10:50:37] --> Yoshimo has joined #gemrb
[12:02:57] <-- edsrzf has left IRC (Ping timeout: 245 seconds)
[14:08:56] --> kida_laptop has joined #gemrb
[15:16:19] <-- barraAway has left IRC (Ping timeout: 264 seconds)
[15:18:09] --> barra_home has joined #gemrb
[15:54:08] <CIA-28> GemRB: 03avenger_teambg * rd66f8cdfd189 10gemrb/ (NEWS gemrb/core/System/DataStream.cpp): added TODO
[15:56:42] --> Avenger has joined #gemrb
[15:56:51] --- ChanServ gives channel operator status to Avenger
[15:57:02] <Avenger> hi!
[15:59:45] <Avenger> cfchris6 are you here?
[16:10:50] --> Ty_ has joined #gemrb
[16:11:24] <Ty_> Anyone around lend me a quick hand? I can't get past character creation in BG2 because the keyboard won't work (android)
[16:11:32] <Ty_> Is there any way to bring the soft keyboard up instead?
[16:11:44] --> sirrozha has joined #gemrb
[16:12:31] <Avenger> i'm sure it is possible, but i don't know the specifics. There is probably an sdl menu where you can configure these?
[16:12:33] <tomprince> Our android expert is online right now.
[16:12:37] <sirrozha> Hi, anyone has working config for iGemRB (iOS) for 960x640 screen?
[16:12:58] <Avenger> sirrozha: you got the widescreen mod?
[16:13:17] <sirrozha> @Avenger, yeah
[16:13:40] <Avenger> what is the problem?
[16:14:27] <Avenger> and i'm sure tomprince wanted to say, our android/iphone experts are offline
[16:14:46] <sirrozha> can’t tell right now, because I’ve tried like month ago and I was stuck at menu or something, the game didn’t run fullscreen and was irresponsive to touches
[16:15:09] <Avenger> well, gemrb advanced a lot in the last month regarding touch input
[16:15:36] <Avenger> there are daily ios builds
[16:15:40] <sirrozha> regarding e.g. Baldurs gate, which version should I get?
[16:15:40] <sirrozha> does it matter anyway
[16:15:40] <Avenger> grab the newest
[16:16:11] <Avenger> any from recent 1-2 weeks would do, i think
[16:16:34] <sirrozha> I meant the version of BG :)
[16:16:56] <sirrozha> GOG?
[16:17:33] <Avenger> any version should work, but the config should be properly pointed to the game data
[16:18:07] <sirrozha> okay, thank you, I’ll try now
[16:18:12] <Avenger> i successfully configured (on linux) all gog and non-gog versions
[16:23:39] <Avenger> tomprince: any idea how this commit http://gemrb.git.sourceforge.net/git/gitweb.cgi?p=gemrb/gemrb;a=commitdiff;h=87d8674b0387c6f90e4da597e2a14d681f2bd119 could possibly cause this: http://paste.pocoo.org/show/570308/
[16:27:39] <fuzzie> .. that's a constructor being called with a NULL this?
[16:28:48] <tomprince> That is.
[16:28:56] <tomprince> That is decidely strange.
[16:29:04] <-- sirrozha has left IRC ()
[16:29:28] <fuzzie> well, if it's the first time around then the avatar table init could potentially possibly have corrupted something, but that seems rather unlikely to do that
[16:30:14] <Avenger> hello fuzzie :)
[16:30:40] <Avenger> yes fuzzie, it seems like it is called on a null this, but it is inside a 'new'
[16:31:21] <Avenger> but he found the crasher with bifurcation, and he found a commit that touches exactly that object
[16:32:13] <tomprince> I'm not sure I believe that result.
[16:32:16] <Avenger> all i can do is to add some more parameter checking in the table ini
[16:32:31] <tomprince> I don't think it should be *possible* to have that.
[16:33:00] <fuzzie> well, stack corruption is the only thing I can think of?
[16:33:24] <tomprince> Should perhaps instrument CharAnimations::CharAnimations to print out this on every call.
[16:33:27] <fuzzie> I'd also be more likely to believe that it was miscompiled though.
[16:33:34] <tomprince> Same here.
[16:33:49] <Avenger> but if he did bifurcation he did frequent recompiles
[16:34:57] <Avenger> and i doubt it was out of memory
[16:35:12] <tomprince> I wonder if he was using gcc 4.3?
[16:35:17] --> brad_a has joined #gemrb
[16:37:53] <brad_a> wjp: any hope of sprites getting fixed? :(
[16:38:12] <fuzzie> oh, I might have time for that if not
[16:38:25] <wjp> hm, I did say something about last week, didn't I?
[16:39:19] <tomprince> In any case, a failed allocation will throw a std::bad_alloc.
[16:39:47] <fuzzie> well, or else abort
[16:39:53] <fuzzie> it certainly won't try running the constructor!
[16:40:26] <Avenger> btw the animation id, is stock bg1 player character anim. (0x5300)
[16:40:45] <Avenger> so it should break everytime
[16:41:07] <brad_a> yes yes you did :)
[16:41:59] <wjp> hm, and speaking of CharAnimations, I suppose that Fall from Grace palette issue is also still on the todo list?
[16:42:15] <Avenger> yep ;)
[16:42:56] <CIA-28> GemRB: 03avenger_teambg * r8c5525c0ce34 10gemrb/gemrb/core/CharAnimations.cpp: some hardening on avatars table loader
[16:42:59] <Avenger> btw, i found one possible stack corruption
[16:43:10] <Avenger> if unsigned long and unsigned int are not the same size
[16:44:31] <wjp> that stack trace is from a 64 bit system, so that's likely the case there
[16:44:51] <Avenger> well, i have a 64 bit linux too, so...
[16:45:20] <Avenger> and the long/int problem wasn't introduced with the found commit
[16:45:41] <tomprince> Like I said. I don't believe the bisection.
[16:45:54] <fuzzie> If it's corruption, it could just be coincidence.
[16:46:09] <wjp> trying a clean compile with a fix for that corruption sounds like a good idea regardless
[16:47:01] <Avenger> yeah, lets hope my blindfighting feat works
[16:50:39] <brad_a> avenger should we treat CR and LF the same in DataStream::ReadLine?
[16:50:53] <wjp> ok, time to refactor the bpp logic in the sprite renderer...
[16:51:06] <brad_a> because of people potentially using classic mac data as we saw on the forum
[16:51:18] <brad_a> wjp: :D
[16:59:37] --> alx3apps has joined #gemrb
[17:00:03] <-- alx3apps has left #gemrb
[17:08:05] <Avenger> brad_a yes
[17:10:16] <wjp> hm, it would be convenient to make the blender classes non-templated, but make the operator() a template instead
[17:10:36] <wjp> would there be any issues with explicitly specializing those?
[17:12:57] <Avenger> wjp: i have absolutely no idea ;) Templates already entered gemrb :)
[17:13:42] <tomprince> I suspect that msvc6 might have trouble with them. It has much better support for class templates, as opposed to function templates.
[17:14:13] <Avenger> hmm, in that case please don't make operator templates, hehe
[17:14:54] <wjp> can always make the blender a template argument to yet another class of course :-)
[17:15:29] <Avenger> what's wrong with the good old switch/case or polymorphic function prototypes
[17:15:54] <wjp> they either cause ifs in inner loops or lots of code duplication
[17:16:36] <Avenger> inline functions or macros?
[17:16:37] <tomprince> Or indirect branch.
[17:16:50] <wjp> macros have been done ;-)
[17:16:54] <fuzzie> Avenger: you can think of the template usage here as just being fancy inline functions :)
[17:17:27] <Avenger> yeah, i can. But msvc chokes on some
[17:18:05] <fuzzie> well, msvc is pre-standardization, so you can make it choke on the other things too :/
[17:18:20] <fuzzie> erm, msvc6, ofc
[17:18:26] <fuzzie> and at least this stuff is easy to test
[17:18:33] <Avenger> hmm, i wanted to bring up another bug... ah yes, petrifified characters are targetable?
[17:18:40] <Avenger> -fi
[17:19:13] <Avenger> i mean, they are apparently not, but i don't know why. Last time i checked this worked
[17:23:35] <Avenger> actually, it seems this works for me
[17:24:08] <brad_a> he is probably looking at one of the other stone people?
[17:24:25] <brad_a> iirc there are several around the map
[17:26:01] <Avenger> yeah, probably. It sucks that petrified characters don't have the right color.
[17:32:30] <brad_a> they should be light blue
[17:32:52] <brad_a> or you mean the char itself or the circle?
[17:38:44] <-- Yoshimo has left IRC (Read error: Connection reset by peer)
[17:39:18] --> Yoshimo has joined #gemrb
[17:42:27] <Avenger> the circle is fine. The avatar should be recolored to stone gray
[17:50:23] <CIA-28> GemRB: 03avenger_teambg * re6a112b2f3b9 10gemrb/gemrb/core/Scriptable/Actor.cpp: don't issue verbal constants (except VB_DIE) when incapacitated
[17:52:49] <wjp> oh, instant crash on bg2 load in CharAnimations here
[17:52:50] <-- barra_home has left IRC (Quit: Verlassend)
[17:55:08] <Avenger> cool
[17:55:23] <wjp> same this==0
[17:55:37] <Avenger> did you get my commit?
[17:56:51] <wjp> trying that now
[17:57:41] <wjp> that seems to help
[18:01:14] <-- brad_a has left IRC (Quit: brad_a)
[18:26:52] <wjp> weird; at least one of my savegames is now showing the wrong area after loading
[18:47:18] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[19:12:19] <-- Ty_ has left IRC (Quit: Page closed)
[19:15:56] <wjp> Avenger: caused by 277255caed615c0261760e05f0aadb368eb8d93a ("separated SetMap and Initialization of actor so ApplyFeats can be called in guiscript")
[20:16:10] <Avenger> that's weird
[20:16:50] <Avenger> how wrong is that area :)
[20:17:11] <wjp> it flashes a blue background and then seems to show the inside of a small building, instead of the temple district
[20:17:34] <Avenger> what if you load that game in the original
[20:17:59] <wjp> it's an original savegame
[20:18:12] <wjp> I don't remember ever having seen this
[20:18:15] <Avenger> and what is in the small building O_o
[20:18:39] <Avenger> the temple district is a master area
[20:18:58] <Avenger> the small building is probably not. Where are your actors?
[20:19:06] <wjp> they should all be standing outside
[20:19:18] <wjp> there are no actors in the displayed area
[20:19:31] <Avenger> weird
[20:19:32] <wjp> is there a console command to show the currently displayed area?
[20:19:39] <Avenger> ctrl-x ?
[20:19:50] <wjp> ar1204
[20:21:04] <fuzzie> if it's an original save, maybe this is caused by area loads due to delayed actor moves?
[20:21:08] <Avenger> ctrl-g is even better
[20:21:15] <fuzzie> although I don't see any centering code in InitActor
[20:21:44] <wjp> everyone should be in ar0900
[20:22:33] <wjp> hm, the console does show things about delayed moves
[20:22:48] <wjp> [GameScript]: MoveBetweenAreas: Pheasant to ar1204 [-1.-1] face: 6
[20:23:11] <Avenger> lol, pheasant went through a door?
[20:23:18] <fuzzie> no
[20:23:26] <fuzzie> pheasant scripted to move there
[20:23:30] <Avenger> really?
[20:23:35] <fuzzie> there's *loads* of these
[20:23:45] <fuzzie> which is why I keep complaining about our lack of support here :P
[20:24:02] <fuzzie> although I don't actually remember a pheasant. i remember other animals.
[20:24:03] <Avenger> i never seen any pheasant in the temple district
[20:24:19] <Avenger> especially patrolling ones :)
[20:24:30] <wjp> ctrl-g:
[20:24:35] <wjp> [Unknown]: ar1204
[20:24:36] <wjp> [Unknown]: ar0900
[20:24:36] <wjp> [Game/DEBUG]: Currently loaded areas:
[20:24:36] <wjp> Current area: ar0900 Previous area:
[20:24:37] <fuzzie> sure, they're teleported from somewhere else
[20:24:56] <fuzzie> quest maybe? I don't have scripts here to look
[20:25:18] <Avenger> hmm, i guess i see what happened
[20:25:24] <Avenger> pheasant is a global actor, right?
[20:25:33] <Avenger> it has a movebetweenareas effect on it
[20:25:38] <fuzzie> yes
[20:25:43] <Avenger> and gemrb evaluates the global actor immediately
[20:25:51] <fuzzie> 21:21 <fuzzie> if it's an original save, maybe this is caused by area loads due to delayed actor moves?
[20:26:06] <Avenger> yes, i saw that, but i thought it is a script action
[20:26:09] <fuzzie> I just don't see how your commit changes it.
[20:26:36] <Avenger> it changes the order of setmap on actors
[20:26:41] <Avenger> the last setmap wins
[20:26:47] <Avenger> and it is now for global actors, i guess
[20:27:14] <Avenger> easy fix would be to implement movebetweenareas effect correctly
[20:28:05] <Avenger> at first apply --> make actor to global. anywhere after that, move actor to coordinate if current area equals to the one in the effect.
[20:28:41] <Avenger> maybe tomorrow i'll do this :)
[20:28:49] <Avenger> or just now?
[20:28:56] <fuzzie> (yay)
[20:42:16] <Avenger> hmm, so, a transitioning pheasant also has the party stats (favored weapons, quickslots, most powerful vanquished, etc) ?
[20:42:29] <Avenger> that's so hackish
[20:43:25] <Avenger> those fields (PlayerStats) are saved for everyone in the npc/pc lists in game.
[21:01:52] <Avenger> someone could check the speech channel :(
[21:02:08] <fuzzie> hm?
[21:02:23] <Avenger> speech doesn't work. even in the irenicus intro cutscene
[21:03:26] <Avenger> hmm, maybe i just set voice volume to 0...
[21:03:42] <Avenger> lol
[21:03:53] <fuzzie> so I can stop looking at it now? :P
[21:03:57] <Avenger> yes
[21:04:03] <fuzzie> silly :-)
[21:05:23] <Avenger> by the way, i made the move to area effect change
[21:06:29] <CIA-28> GemRB: 03avenger_teambg * r7c7cffa87f55 10gemrb/gemrb/plugins/FXOpcodes/FXOpcodes.cpp: move to area effect (instead of immediate moving, make actor global and wait till the current area is the target area)
[21:06:45] <Avenger> there may be some hidden flag that makes apart these temporary global actors from real global actors
[21:06:58] <wjp> let's see
[21:07:16] <wjp> map is ok again now; thanks
[21:07:21] --> edsrzf has joined #gemrb
[21:07:23] <Avenger> really? whew
[21:07:39] <Avenger> then we just need to find an actor that moves by this effect
[21:07:55] <Avenger> like, you could check where your pheasants are headed
[21:07:59] <Avenger> i guess, ar1204
[21:08:21] <Avenger> then move there, save game again, and check if the pheasants vanished from global (game) and entered the area
[21:08:43] <wjp> I'll need more detailed instructions :-)
[21:08:52] <Avenger> even more detailed :D
[21:09:01] <Avenger> load the pheasant game
[21:09:29] <wjp> ok
[21:09:30] <Avenger> then MoveToArea("ar1204") it was ar1204, i guess, because ar0900 is the temple district
[21:10:02] <Avenger> do you see pheasants?
[21:10:21] <wjp> well, no
[21:10:25] <Avenger> :(
[21:10:29] <wjp> but it is the inside of a building
[21:10:38] <Avenger> yes, they should be there somewhere
[21:11:08] <wjp> this is the windspear cottage, apparently
[21:11:26] <Avenger> if they still run their effects, the effect realizes the target area is the same as fx->Resource and properly moves them to the target point
[21:12:22] <wjp> but I see no visible birds
[21:12:33] <Avenger> maybe save the game to a different slot
[21:12:58] <wjp> maybe it's hidden somewhere for black magic scripting purposes?
[21:13:01] <Avenger> and check if they are still in the global area of the game (in dltcep look at the npc dropdown menu)
[21:13:49] <Avenger> if we don't fix this, then the genie will be broken in the beginning of bg2
[21:18:56] <-- Avenger has left IRC (Quit: bye!)
[21:19:08] <wjp> hm, no pheasant in that list
[21:19:09] <wjp> (edit->game->load external game->NPC dropdown, right?)
[21:19:17] <wjp> but I don't see it in the previously-broken original save either
[22:58:21] --> Random_ has joined #gemrb
[22:59:01] <Random_> Hi guys, I'm using the Android version of GemRB with BGII and can't get past character creation since it needs a keyboard and soft keyboard can't be used.. is there any way around this? Thanks!
[23:05:47] <-- Yoshimo has left IRC (Quit: Yoshimo)