#gemrb@irc.freenode.net logs for 14 Jan 2014 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:53:42] --> nutron has joined #gemrb
[00:57:29] <-- raevol has left IRC (Quit: Leaving.)
[01:11:41] <-- edheldil_ has left IRC (Ping timeout: 252 seconds)
[02:25:57] <-- brada has left IRC (Quit: brada)
[02:37:43] --> brada has joined #gemrb
[03:28:14] <-- PKodon has left IRC (Quit: The Rodent Tracker 8000, just like on TV ... Because household pests never build up an immunity to bullets. (Tex Murphy))
[03:41:19] <-- dolio has left IRC (Quit: Leaving)
[03:44:38] --> dolio has joined #gemrb
[04:20:14] <-- brada has left IRC (Quit: brada)
[04:46:01] --> Canageek has joined #gemrb
[06:45:27] --> Yahve has joined #gemrb
[06:45:34] <Yahve> Good Morning!
[06:50:05] <lynxlynxlynx> oj
[06:59:22] <Yahve> Could someone possibly help me on to topics with GemRB for Android?
[06:59:46] <lynxlynxlynx> later, have to go to a meeting
[06:59:54] <Yahve> Ok, will check in againg
[06:59:57] <Yahve> *again
[08:14:50] <-- Canageek has left IRC (Quit: So long, suckers)
[09:16:47] <-- chiv has left IRC (Ping timeout: 272 seconds)
[09:22:18] --> Yahve_ has joined #gemrb
[09:22:26] <Yahve_> Heyya all!
[09:25:53] <Yahve_> Hmm.. my home mIRC stayed online... oh well
[09:28:53] <Yahve_> Seing that I'm somewhat alone here, I'll just ask my first question and hope for an answer at some point. So: Have any of you tried using Easytutu version of BG1 on GemRB for Android?
[09:29:59] <fuzzie> I would expect not
[09:31:45] <fuzzie> I know some people on the forums tried it, because they had resolution issues on lower-res android devices
[09:33:13] <Yahve_> My starts up, then ctd-s after around 15 seconds in Candlekeep
[09:33:26] <Yahve_> http://www.youtube.com/watch?v=jUv8gVWW34Y&hd=1
[09:33:27] <Pepelka> Baldur's Gate on Android - YouTube
[09:33:28] <Pepelka> »So I got Baldur's Gate Tutu to almost work on Android. (Sound's perfect, but I couldn't capture it :( ) I'll contact the great guys at GemRB to see if there'...«
[09:33:34] <Yahve_> That's me
[09:35:01] <fuzzie> do you have log output?
[09:35:07] <Yahve_> http://pastebin.com/aKHZr1e3
[09:35:09] <Pepelka> GemRB BG Tutu log - Pastebin.com
[09:35:32] <fuzzie> I don't have a clue about gemrb on android specifically
[09:36:09] <Yahve_> There are quite a few errors in the log, i think it might be a BGTutu - GemRB issue, but I'm not sure
[09:36:28] <fuzzie> and, oh, I meant android log (to see the crash)
[09:37:06] <Yahve_> I was looking for it but have yet to find it
[09:37:16] <fuzzie> the corruption is real?
[09:37:36] <fuzzie> if so, that looks like we're not clearing the color buffer
[09:38:13] <fuzzie> but that wouldn't be relevant to a crash
[09:38:57] <Yahve_> Do you know where i could find the android log?
[09:39:10] <fuzzie> the android log you can extract using 'adb logcat' on host machine
[09:39:16] <Yahve_> ok
[09:39:43] <fuzzie> or using alogcat/catlog on android 4.1 or below, or by rooting the device
[09:43:56] <Yahve_> http://pastebin.com/n1WVXZGy
[09:43:58] <Pepelka> Android logcat - Pastebin.com
[09:44:52] <Yahve_> E/GemRB ( 6489): [GameScript/ERROR]: Target object: IDS Targeting: 2 0 0 0 0 0 0 0 0 0 E/GemRB ( 6489): Filters: 0 0 0 0 0
[09:48:58] <fuzzie> huh
[09:49:11] <fuzzie> SDL2 bug
[09:50:02] <Yahve_> Is there anything i can do about it?
[09:50:14] <fuzzie> did you build this yourself?
[09:50:47] <Yahve_> Umm.. Now I'm goint to total n00b mode: I don't think so
[09:50:53] <fuzzie> then, no
[09:51:02] <Yahve_> I downloaded the apk from gemrb.org
[09:51:12] <fuzzie> I don't even know who makes those
[09:51:17] <fuzzie> but whoever does will have to deal with it
[09:51:50] <Yahve_> and installed the bg on the pc with all fixes, then copied it to the folder
[09:52:37] <fuzzie> I don't understand how the error strings would leak through to the JNI world
[09:53:51] <Yahve_> don't look at me, I had to google to know what JNI means :)
[09:55:28] <fuzzie> the log is really helpful anyway
[09:56:19] <Yahve_> Glad to know
[09:56:42] <Yahve_> I'll try using the play version of GemRB, as it is only 7.0 but is developed by a different group... might work
[09:56:49] <fuzzie> it is very old I think
[09:57:16] <fuzzie> yes, last updated feb 2012, and it was buggy then..
[09:58:08] <Yahve_> Ow :(
[09:58:46] <fuzzie> from your log it clearly looks like we're leaking all the error messages but I don't see how that can possibly happen :) it'll have to wait until one of the other devs wakes up
[09:59:40] <Yahve_> Ok, I'll hang around in case you need any other info.. I tried using the other version earlier, should I create a fresh log?
[09:59:51] <Yahve_> (know that i know about adb logcat -c )
[10:06:02] <fuzzie> don't think so
[10:10:49] --> Eli2_ has joined #gemrb
[10:13:27] <-- Eli2 has left IRC (Ping timeout: 245 seconds)
[10:17:45] <fuzzie> Yahve_: which android version is this?
[10:18:16] <Yahve_> 4.4.2 (KOT49H)
[10:18:30] <Yahve_> On a nexus 4 16gb (Mako)
[10:24:09] <Yahve_> Rooted, but otherwise original
[10:26:43] <fuzzie> well, I'll bother people for a new build and maybe someone fixed it
[10:28:55] <Yahve_> :)
[10:30:20] <Yahve_> Thank you so much
[10:30:55] <Yahve_> Should I be able to use GemRB with a different game (eg BG2), or this is a universal bug in the build
[10:31:04] <Yahve_> ?
[10:43:08] <lynxlynxlynx> universal
[10:43:16] <Yahve_> :(
[10:43:29] <lynxlynxlynx> and you're probably the first 4.4 tester we know about
[10:43:43] <fuzzie> it's very odd
[10:44:00] <fuzzie> the log reads like it's leaking the error text into the jni layer
[10:44:17] <Yahve_> Glad to help :D
[10:44:19] <fuzzie> it's an old build (I think?), maybe it tries to do popups for errors or something?
[10:44:35] <fuzzie> who should be bothered about the android builds atm, anyone?
[10:45:09] <Yahve_> It's the gemrb-0.8.0-git-5aa84783.apk version
[11:44:55] <lynxlynxlynx> 528 commits ago
[11:46:33] <fuzzie> but has this code changed in the meantime?
[11:55:02] <-- Lightkey has left IRC (Ping timeout: 272 seconds)
[12:00:09] <fuzzie> if it has, I don't see it
[12:00:26] <fuzzie> but I don't see how in any circumstance we'd leak the error strings to the jni layer so I'm missing something
[12:14:18] <Yahve_> If I can help anyhow, I'd gladly do so :)
[12:19:03] <lynxlynxlynx> have you ever built anything from source?
[12:19:32] <Yahve_> I'm a poweruser on windows, but sadly that more or less sums it up.
[12:20:38] <Yahve_> (In short:: no)
[12:21:14] <lynxlynxlynx> ok, nevermind
[12:38:30] <-- nutron has left IRC (Ping timeout: 276 seconds)
[12:51:09] --> nutron has joined #gemrb
[13:17:37] <-- nutron has left IRC (Ping timeout: 245 seconds)
[13:20:51] --> nutron has joined #gemrb
[14:32:11] <-- Yahve_ has left IRC (Ping timeout: 272 seconds)
[14:36:05] --> Yahve_ has joined #gemrb
[15:29:09] <-- Yahve_ has left IRC (Quit: Page closed)
[15:31:59] --> brada has joined #gemrb
[15:32:20] <brada> unfortunately our android build is lacking several fixes that have been done since
[15:37:45] <fuzzie> hey
[15:37:48] <brada> hello
[15:37:53] <fuzzie> do you know why the JNI would be seeing error strings?
[15:38:18] <brada> i didnt do the android stuff, but psch added a special android logger class
[15:38:24] <brada> that is probably responsible
[15:38:25] <fuzzie> > JNI local reference table (0x7b82b1a8) dump:
[15:38:29] <fuzzie> > 507: 0x429bef20 java.lang.String "ERROR"
[15:38:33] <fuzzie> > 506: 0x429d82d8 java.lang.String "Target object: I... (71 chars)
[15:38:37] <fuzzie> > 503: 0x429d8298 java.lang.String "ERROR"
[15:38:41] <fuzzie> > 502: 0x429d81e8 java.lang.String "Target for dialo... (60 chars)
[15:39:23] <fuzzie> so it can't possibly be the logger, because that doesn't split out the strings like that
[15:40:26] <brada> odd
[15:40:37] <brada> the only other android specific code is the java
[15:40:43] <brada> afik
[15:40:45] <fuzzie> it's not that
[15:40:57] <fuzzie> to me it looks like it would be something from psch's branch
[15:41:05] <brada> oh
[15:41:06] <brada> yes
[15:41:08] <fuzzie> but the git id from the above is a commit from *our* tree
[15:41:15] <brada> he built from his source
[15:41:20] <brada> right
[15:41:32] <fuzzie> https://github.com/peschwa/gemrb-android/commit/4eb0cc366ca1d98559f5d633430b1297a470d677 obviously leaks for the level==FATAL case for example
[15:41:34] <Pepelka> Android: added logging of FATAL messages to Android notifications · 4eb0cc3 · peschwa/gemrb-android · GitHub
[15:41:35] <Pepelka> »gemrb-android - Engine Made with preRendered Background«
[15:41:37] <brada> lynx just put that in the file name tho right?
[15:41:42] <fuzzie> I have no idea
[15:41:47] <brada> he did
[15:42:12] <brada> because lynx had to upload it. i guess he just put the hash for the tip at the time it was built
[15:42:46] <fuzzie> ah yes
[15:42:49] <fuzzie> so ok
[15:42:49] <lynxlynxlynx> he usually gave me versioned tarballs
[15:42:56] <fuzzie> that build should be marked broken I guess
[15:43:03] <lynxlynxlynx> so probably the last thing he rebased on
[15:43:10] <fuzzie> the gemrb-android tree has an android logger which leaks JNI objects
[15:43:28] <fuzzie> so it will inevitably be killed after ~120 reported errors
[15:44:30] <brada> nice
[15:46:32] <brada> you should comment on that commit about the leak
[15:46:45] <brada> maybe it will roust him into making a more recent build :p
[17:08:52] <-- kpederse1 has left IRC (Ping timeout: 246 seconds)
[17:10:47] --> kpedersen has joined #gemrb
[17:15:48] --> kpederse1 has joined #gemrb
[17:16:19] <-- kpedersen has left IRC (Ping timeout: 260 seconds)
[17:17:34] <lynxlynxlynx> error: no matching function for call to ‘GemRB::Actor::SetEquippedQuickSlot(int, GemRB::ITMExtHeader*)’
[17:17:34] <lynxlynxlynx> SetEquippedQuickSlot(inventory.GetEquipped(), inventory.GetEquippedHeader());
[17:18:09] <lynxlynxlynx> i thought the compiler would figure out which override to use :s
[17:18:37] <lynxlynxlynx> err overload
[17:19:28] <lynxlynxlynx> it gives precedence to the non-const version
[17:21:05] <lynxlynxlynx> but no, it's that
[17:21:29] <lynxlynxlynx> bbl
[17:24:05] <-- |Cable| has left IRC (Ping timeout: 272 seconds)
[17:28:21] --> Yoshimo has joined #gemrb
[17:33:54] --> raevol has joined #gemrb
[17:37:03] --> |Cable| has joined #gemrb
[17:54:58] --> fizzle has joined #gemrb
[17:58:12] <lynxlynxlynx> ok, so there's only parameter based overloading
[18:02:14] <fizzle> heya
[18:05:14] <fizzle> good to see things pick up a bit of steam again lately
[18:05:42] <lynxlynxlynx> wb
[18:06:14] <lynxlynxlynx> we even have a starting opengl renderer (not merged yet)
[18:06:36] <fizzle> sounds great
[18:07:05] <lynxlynxlynx> currently still a resource hog, but most of the drawing stuff is in place
[18:07:25] <fizzle> although from scrolling back a bit everything sounds a tad broken right now (not just ogl, that is) ;-)
[18:08:15] <fizzle> anyway, don't expect me to do anything useful any time soon
[18:08:37] <fizzle> got a rare evening off today but that's not going to last
[18:09:09] <lynxlynxlynx> nice of you to stop by
[18:09:24] <fizzle> "life is what happens while you're busy with other things" :-P
[18:09:31] <lynxlynxlynx> mhm
[18:09:52] <-- Coriander has left IRC (Quit: Nettalk6 - www.ntalk.de)
[18:11:11] --> Coriander has joined #gemrb
[18:12:58] --> Beholder has joined #gemrb
[18:13:08] <Beholder> hi
[18:13:56] <lynxlynxlynx> oj
[18:23:04] <Beholder> i changed palettemanager code, but i still have small leak in text area somewere
[18:23:22] <fuzzie> you release() the palettes?
[18:23:30] <fuzzie> I can take a look later
[18:23:37] <Pepelka> [wiki] pst_bugs - [Gameplay] http://www.gemrb.org/wiki/doku.php?id=pst_bugs&rev=1389723711&do=diff
[18:23:55] <Beholder> i monitored textures, leak is outside renderer
[18:25:49] <Beholder> i releasead attached only if it's shared (-1 refcounter), if it's not shared i delete texture only
[18:27:38] <Beholder> if i release a last palette, i get a error in ~BAMImporter()
[18:28:03] <Beholder> look at code
[18:33:17] <fuzzie> you should release only when you erase from the mp
[18:33:19] <fuzzie> map
[18:33:24] <fuzzie> since you acquire only when you add to the map
[18:34:33] <fuzzie> acquire = refcount++, release = refcount--, so they have to be balanced
[18:36:35] <lynxlynxlynx> while we're drawing, chiv got a crash in AreaAnimation::Draw because the next frame was null
[18:36:46] <lynxlynxlynx> should we just skip the drawing in that case?
[18:37:17] <lynxlynxlynx> http://pastebin.com/qk4gTq4n @ 6, then it triggers the assert
[18:37:18] <Pepelka> backtrace gemrb torment godsmen foundry - Pastebin.com
[18:37:32] <brada> lynx: its not that simple
[18:37:38] <Beholder> i need to release only if palette !IsShared() ?
[18:37:54] <fuzzie> Beholder: no
[18:38:00] <fuzzie> Beholder: you need to release when you erase
[18:38:00] <brada> 1/10 times it is not simply null and results in a seg fault
[18:38:30] <lynxlynxlynx> oh, so it's corrupted somewhere before
[18:38:33] <brada> so something much worse is at paly, and if you look at the code where the frame comes from you can see that it shouldnt return null
[18:38:35] <brada> yes
[18:39:00] <fuzzie> Beholder: but if you only erase when !IsShared() then yes, put it there too
[18:39:11] <fuzzie> erm, put it there. not 'too'.
[18:39:41] <fuzzie> but you should also put it in Clear(), since you erase all the map entries there
[18:41:51] <Pepelka> [wiki] pst_quests - [Godsmen Foundry] http://www.gemrb.org/wiki/doku.php?id=pst_quests&rev=1389724753&do=diff
[18:53:53] <-- Yoshimo has left IRC (Read error: Connection reset by peer)
[19:05:23] <lynxlynxlynx> hmm
[19:05:41] <lynxlynxlynx> Game.h:284: ieDword interval; // 1000/AI_UPDATE (a tenth of a round in ms) <-- something doesn't match
[19:10:44] <lynxlynxlynx> appears to be only used for upping Ticks in AdvanceTime
[19:12:06] <Beholder> why i need release last ref of palette? I get all palettes with refcount >=1
[19:12:36] <Beholder> and i release if refcount>1
[19:14:19] <Beholder> code that produce palette should release last ref
[19:16:06] <brada> you may be under the mistaken impression that what created the palette hasn't already released its reference
[19:16:59] <brada> ex I create a palette then pass it to you then i release the palette. palette has ref count of 1 and its yours, not mine even tho i created the palette.
[19:18:20] <Beholder> if i release a lat instance i get error in BAMFont destructor
[19:18:26] <Beholder> *last
[19:19:20] <Beholder> destructor deletes palette without previously checking pointer
[19:20:10] <brada> Beholder: you should never delete a palette
[19:22:12] <Beholder> I mean release last ref, not delete pointer. BAMImporter try to release too. But refcount 0 causes a error
[19:22:20] <brada> ah
[19:27:56] <Beholder> i cant decrease ref to 0, it cause error in any places. I get palettes with ref>=1, and increase ref by 1. But if the code outside decrease ref, i dont know anything about it
[19:31:39] <brada> you arent supposed to know or care about code outside increasing references
[19:31:46] <fuzzie> yes
[19:32:05] <fuzzie> BAMImporter keeps a reference to the palette
[19:33:15] <fuzzie> and it calls CreateSprite8 with palette
[19:33:56] <brada> which presumably also increases the count
[19:34:21] <Beholder> right, but i get attached from BAMimporter too
[19:34:41] <fuzzie> but you acquire() when adding that one to the map, right?
[19:34:48] <Beholder> right
[19:35:06] <Beholder> and delete texture if refcount=1
[19:35:21] <Beholder> or decrease refcount ifnot
[19:35:26] <fuzzie> right, but that is bad
[19:35:36] <Beholder> why
[19:35:46] <fuzzie> because you increase refcount once and you decrease refcount many times
[19:36:05] <brada> you shouldnt know or care about the refcount for starters. isnt that private?!
[19:36:15] <fuzzie> brada: yes but there's an IsShared() to allow cleaning up
[19:36:53] <Beholder> i dont understand any
[19:36:54] <fuzzie> so you can do "if this palette has no other references, then delete the texture and release our copy" behaviour
[19:37:06] <fuzzie> Beholder: you must release only when you erase/clear the entry from the map
[19:37:22] <fuzzie> the problem is
[19:37:23] <fuzzie> right now
[19:37:28] <fuzzie> textareas release the palette
[19:37:34] <fuzzie> and you still have a ref!
[19:37:45] <fuzzie> so now *your* code is the owner, and you must release the last ref
[19:37:57] <Beholder> textareas generates new palettes each frame
[19:37:58] <fuzzie> if you don't want that, then you have to stop keeping a copy
[19:38:07] <fuzzie> but then you have to re-create the palette texture every time you need one
[19:38:23] <fuzzie> so that is silly, and the only solution is, you have to release the refs when you're done
[19:38:24] <Beholder> in each blit?
[19:38:36] <fuzzie> which means you have to release when you erase from your map
[19:39:25] <Beholder> and i dont check for shared, right?
[19:39:30] <fuzzie> if the refcounts break then you're counting them wrong somewhere
[19:40:30] <Beholder> it do not break in code from git
[19:40:42] <fuzzie> the code in your git is wrong, though
[19:40:43] <Beholder> and i have no texture leaks
[19:41:00] <fuzzie> because you don't release() when you erase
[19:41:23] <Beholder> if i release, i get refcount error in BamImporter
[19:42:04] <fuzzie> let me try it
[19:44:18] <fuzzie> where is crash?
[19:44:56] <Beholder> in destructor
[19:44:59] <fuzzie> yes
[19:45:02] <fuzzie> I mean, how do I get it?
[19:45:11] <Beholder> ClearUnused
[19:45:15] <fuzzie> no
[19:45:16] <fuzzie> in-game
[19:45:21] <fuzzie> I changed the code and bg1 seems fine for me
[19:45:34] <Beholder> hm
[19:45:45] <Beholder> what about leaks?
[19:45:50] <fuzzie> checking
[19:46:00] <Beholder> open the dialog
[19:46:03] <Beholder> and stay
[19:46:45] <fuzzie> CreatePalettedSprite leaks
[19:47:29] <fuzzie> also missing a release
[19:47:57] <fuzzie> trying again
[19:49:30] <fuzzie> oh, GLTextureSprite2D leaks it too
[19:49:56] <fuzzie> trying again
[19:51:22] <fuzzie> Beholder: http://pastebin.com/mzter5uF
[19:51:22] <Pepelka> diff --git a/gemrb/plugins/SDLVideo/GLPaletteManager.cpp b/gemrb/plugins/SDLVide - Pastebin.com
[19:52:02] <Beholder> pastebin now is banned for Russia :####
[19:52:21] <fuzzie> oh. which pastebin is good? pastebin.ca?
[19:52:48] <Beholder> ca is right
[19:52:58] <fuzzie> http://pastebin.ca/2539579
[19:53:01] <Pepelka> pastebin - Miscellany - post number 2539579
[19:53:57] <fuzzie> fixes all leaks in video for me
[19:55:07] <Beholder> i'll try
[19:58:26] <fuzzie> you really have to wait for us to fix the textarea thing though :)
[20:01:09] <Beholder> and i do not need to cleanup every frame?
[20:01:18] <Beholder> works fine for me
[20:01:28] <Beholder> thank you
[20:13:12] <Beholder> what method fills red on portrait
[20:16:21] <wjp> I think that's BlitGameSprite with a tint?
[20:16:48] <wjp> yes
[20:16:59] <wjp> look for the BlitGameSprite call in Button::DrawInternal()
[20:18:34] <Beholder> hm, not worked for me, need to inspect
[20:19:41] <wjp> what do you see?
[20:20:41] <Beholder> nothing
[20:20:51] <Beholder> no tint
[20:23:07] <wjp> you don't handle BLIT_TINTED in all the branches in BlitGameSprite at the moment
[20:23:43] <wjp> it's an anchored sprite, so it'll do one of the "else blitSprite(...)" right now, probably
[20:26:19] <-- brada has left IRC (Quit: brada)
[20:43:25] --> brada has joined #gemrb
[20:56:08] <fuzzie> we should probably do something about the worldmap MOS loading
[20:56:18] <fuzzie> it is such a huge memory consumer
[20:56:45] <fuzzie> the labels on the world map appear at completely the wrong place too, but I assume that's an opengl issue
[21:00:18] <fuzzie> it is
[21:36:00] <Beholder> hm
[21:36:06] <Beholder> worldmap?
[21:36:12] <Beholder> looks fine for me
[21:36:46] <Beholder> i see only clipping problem
[21:37:03] <Beholder> not misplaced labels
[21:37:40] <fuzzie> bg12
[21:37:41] <fuzzie> bg2
[21:38:19] <Beholder> a
[21:39:43] <Beholder> http://i-fotki.info/16/5aacd525edfc07ce8b4df2265c2862a6bc8616172395629.jpg.html
[21:39:45] <Pepelka> Ifotki.info - бесплатный хостинг фотографий для форумов, дневников, гостевых книг
[21:39:46] <Pepelka> »Ресурс для публикации фотографий на форумах, дневника, личных страницах«
[21:39:47] <Beholder> bg1
[21:40:13] <Beholder> and for area map
[21:40:20] <Beholder> too slow to redraw all
[21:40:35] <Beholder> each frame
[21:40:45] <Beholder> 2-3 fps for me
[21:41:10] <fuzzie> hm maybe the worldmap isn't your fault :)
[21:41:29] <fuzzie> worldmap gives me 30fps (cap)
[21:41:54] <Beholder> woldmap is fast
[21:41:59] <Beholder> areamap is slow
[21:42:00] <fuzzie> ah
[21:42:08] <fuzzie> in first bg1 area?
[21:42:12] <Beholder> yes
[21:42:24] <fuzzie> odd
[21:42:37] <Beholder> 2-3 fps)
[21:42:41] <Beholder> slideshow
[21:44:31] <fuzzie> hm
[21:44:35] <fuzzie> if I uncap it, I get 84fps
[21:44:37] <fuzzie> that is really bad
[21:44:50] <fuzzie> oh I guess it's probably drawing all the fog as individual squares :)
[21:45:00] <brada> heh
[21:45:22] <fuzzie> yep
[21:45:24] <fuzzie> :)
[21:46:06] <-- fizzle has left #gemrb
[21:53:33] <lynxlynxlynx> Beholder: your worldmap isn't fine, the map icons should be colored
[21:53:40] <lynxlynxlynx> we switch their palette manually
[21:59:43] <Beholder> BlitSprite or BlitGameSprite?
[22:01:13] <lynxlynxlynx> for drawing? I don't know, but the sprite swap happens in WorldMapControl.cpp iirc
[22:01:55] <fuzzie> it's BlitSprite
[22:02:00] <fuzzie> with a SetPalette just beforehand
[22:02:38] <fuzzie> oh, that's not how it works
[22:02:42] <fuzzie> anyway it's BlitSprite
[22:03:42] <-- brada has left IRC (Quit: brada)
[22:05:11] <fuzzie> doesn't this code leak the palette?
[22:06:11] <fuzzie> sigh
[22:08:39] --> edheldil_ has joined #gemrb
[22:09:55] <Beholder> palette is in sprite, not external, right?
[22:09:58] <fuzzie> yes
[22:10:03] <fuzzie> BlitSprite gets NULL palette
[22:11:37] <Beholder> SetPalette is never called
[22:12:43] <fuzzie> for me it's all red
[22:13:09] <fuzzie> or kind of brownish for visited
[22:13:11] <fuzzie> or blue for active
[22:13:32] <fuzzie> with opengl backend
[22:14:11] <Beholder> bg1?
[22:14:14] <fuzzie> yes
[22:14:55] <fuzzie> your screenshot looks not good for the icons
[22:15:50] <Beholder> i see grayscale palette in debugger
[22:16:40] <Beholder> in sprites
[22:16:49] <fuzzie> intrsting
[22:17:06] <Beholder> r=g=b
[22:18:03] <fuzzie> http://i.imgur.com/BOvfq1A.png <- latest beholder branch for me
[22:18:15] <lynxlynxlynx> greyscale is what is stored in the bam
[22:18:59] <fuzzie> hm
[22:19:04] <fuzzie> Beholder: do you have old GUIScripts maybe?
[22:19:18] <lynxlynxlynx> that's fine fuzzie
[22:19:30] <Beholder> unlikely
[22:19:40] <lynxlynxlynx> do you have any uncommitted changes?
[22:19:58] <Beholder> no
[22:20:01] <fuzzie> the most likely cause of this is Beholder's GUIMA.py missing the last param to CreateWorldMapControl
[22:21:32] <lynxlynxlynx> this was done what, a year ago?
[22:21:40] <fuzzie> jan 5th 2014
[22:21:51] <lynxlynxlynx> oh, the merge
[22:21:53] <fuzzie> no
[22:22:00] <fuzzie> 826bab7acfeb14f9b738bea9a519fad72062f29a
[22:22:25] <fuzzie> but yes my first thought would be, maybe Beholder has an old python which makes annoying .pyc files everywhere?
[22:22:37] <Beholder> 2.7
[22:22:43] <fuzzie> then, no
[22:22:47] <fuzzie> you have no pyc files?
[22:22:51] <lynxlynxlynx> how many GUIMA.py do you have?
[22:22:55] <fuzzie> you use git to manage the tree?
[22:23:01] <fuzzie> you should *not* have a bg1/GUIMA.py
[22:24:36] <Beholder> no GUIMA in bg1 folder
[22:24:57] <Beholder> merge with latest, no changes
[22:24:59] <fuzzie> interesting then
[22:25:08] <Beholder> i can try to use old driver
[22:25:10] <fuzzie> break in CreateWorldMapControl, see parameter?
[22:25:34] <fuzzie> lynxlynxlynx: you didn't update the doc :o
[22:26:20] <lynxlynxlynx> on it
[22:26:21] <fuzzie> Beholder: WorldMapControl::SetOverrideIconPalette should be called, to set OverrideIconPalette to true
[22:27:03] <lynxlynxlynx> i did, do you mean the txt or something else?
[22:27:49] <lynxlynxlynx> py doc
[22:28:36] <fuzzie> I mean the text before the function
[22:29:23] <Beholder> dont see SetOverrideIconPalette in WorldMapControl.cpp
[22:29:53] <lynxlynxlynx> fixed
[22:30:13] <Beholder> in h
[22:30:21] <lynxlynxlynx> yeah
[22:31:07] <fuzzie> thanks I think
[22:32:58] <Beholder> false here
[22:36:24] <-- Yahve has left IRC ()
[22:37:16] <fuzzie> so, yes, break in plugins/GUIScript/GUIScript.cpp and check CreateWorldMapControl
[22:37:43] <fuzzie> after the PyArg_ParseTuple, recolor should be 1
[22:38:16] <Beholder> i cant
[22:38:27] <Beholder> python prevent me for debuggin
[22:39:37] <Beholder> drawing solid rects with glClear a bit faster than with primitive
[22:40:12] <fuzzie> well, it sounds like guiscript issue, but who knows
[22:41:17] <Beholder> python lib for windows built in VS2005
[22:41:43] <Beholder> i cant build new one from source
[22:47:03] --> Lightkey has joined #gemrb
[22:55:47] <lynxlynxlynx> that would explain it then
[23:21:44] --> brada has joined #gemrb
[23:23:39] <-- Beholder has left IRC (Ping timeout: 265 seconds)
[23:39:27] <-- lynxlynxlynx has left IRC (Ping timeout: 252 seconds)
[23:59:38] <-- brada has left IRC (Quit: brada)