#gemrb@irc.freenode.net logs for 10 Jun 2011 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage

[03:13:52] <-- barra_home has left IRC (Quit: Verlassend)
[03:26:16] <-- Lingerance has left #gemrb ("I'll be back")
[03:36:41] --> raevol has joined #gemrb
[03:51:31] --> budlust has joined #gemrb
[04:15:57] <-- raevol has left IRC (Quit: Leaving.)
[06:44:33] --> budlust_ has joined #gemrb
[06:47:55] <-- budlust has left IRC (Ping timeout: 260 seconds)
[07:02:54] --> lynxlynxlynx has joined #gemrb
[07:02:54] --- ChanServ gives channel operator status to lynxlynxlynx
[07:11:57] <-- budlust_ has left IRC (Ping timeout: 276 seconds)
[07:12:14] --> lubos has joined #gemrb
[07:13:16] --> budlust has joined #gemrb
[07:14:42] --> adominguez has joined #gemrb
[08:18:29] --> raevol has joined #gemrb
[08:23:57] <-- budlust has left IRC (Quit: Lost terminal)
[08:47:34] <-- |Cable| has left IRC (Quit: Leaving)
[09:01:38] <-- raevol has left IRC (Quit: Leaving.)
[09:16:04] --> barra_home has joined #gemrb
[09:26:39] --> SiENcE has joined #gemrb
[09:42:21] --> budlust has joined #gemrb
[11:37:18] <-- budlust has left IRC (Read error: Connection reset by peer)
[11:38:44] --> budlust has joined #gemrb
[11:40:52] <-- Gekz has left IRC (Read error: Operation timed out)
[11:43:37] --> Gekz has joined #gemrb
[11:43:37] <-- Gekz has left IRC (Changing host)
[11:43:37] --> Gekz has joined #gemrb
[12:17:09] <-- lynxlynxlynx has left IRC (Ping timeout: 258 seconds)
[12:31:13] --> lynxlynxlynx has joined #gemrb
[12:31:13] <-- lynxlynxlynx has left IRC (Changing host)
[12:31:13] --> lynxlynxlynx has joined #gemrb
[12:31:13] --- ChanServ gives channel operator status to lynxlynxlynx
[13:02:30] <-- adominguez has left IRC (Quit: Saliendo)
[13:46:52] --> |Cable| has joined #gemrb
[14:31:20] --> Maighstir has joined #gemrb
[14:35:56] <-- |Cable| has left IRC (Remote host closed the connection)
[14:36:49] <-- budlust has left IRC (Read error: Connection reset by peer)
[14:38:10] --> budlust has joined #gemrb
[14:40:01] --> |Cable| has joined #gemrb
[15:40:48] <-- lubos has left IRC (Quit: Leaving.)
[15:45:15] --> Cable_ has joined #gemrb
[15:45:15] <-- |Cable| has left IRC (Read error: Connection reset by peer)
[16:01:30] <-- SiENcE has left IRC (Quit: @all: cya)
[16:25:19] --> brada has joined #gemrb
[16:25:39] --- brada is now known as brad_a
[16:25:54] <brad_a> hey guys
[16:26:06] <brad_a> finally got gemrb working on ipad
[16:26:23] <brad_a> so thanks to whoever maybe in here that helped push me in the right direction
[16:31:01] <fuzzie> did you find masks which gave you the right colours?
[16:31:22] <fuzzie> if you look in the log, i rambled a bit about it, after having found that the masks for the 8bpp videos were wrong
[16:31:58] <fuzzie> but really nice work :)
[16:38:56] <brad_a> well the colors do appear a bit off but not in a super annoying way
[16:39:20] <brad_a> i dont know how to look at said log
[16:39:32] <brad_a> i just used 0 btw
[16:39:42] <brad_a> so i guees SDL is choosing for me
[16:40:09] <fuzzie> i think it defaults to the wrong way around, with red/blue swpped, but not sure
[16:40:32] <fuzzie> http://log.usecode.org/gemrblog.php?log=9Jun2011 is yesterday's log, but it's probably not very helpful if you don't know masks at all - if you have it working, probably best to leave it alone for now
[16:40:37] <brad_a> I need to separate the python source out so that we can merge this into the codebase
[16:41:00] <fuzzie> and i can always look at the masks myself when i have time, in a week or so
[16:41:40] <brad_a> it looks pretty decent honestly and maybe even is correct its just the logo that looks wrong to me but i may just be wrong
[16:43:16] <fuzzie> well that's great then :)
[16:45:19] <brad_a> hmmm what problems exist for sdl 1.3? I havent played yet just been in the menus but everything works and looks fine thus far
[16:45:43] <brad_a> i had to use 1.3 to get this to work under ios
[16:47:44] <fuzzie> i'm not sure, because sdl 1.3 is completely different to 1.2
[16:48:03] <brad_a> let me start a game real quick...
[16:48:32] <brad_a> the movies ar all fine once you change the masks to 0
[16:49:19] <brad_a> i see the appearance thing the dool is all blue
[16:49:23] <brad_a> doll
[16:49:45] <brad_a> crap
[16:49:49] <brad_a> i have no text input
[16:49:54] <brad_a> i cant start a game
[16:50:16] <brad_a> ill have to do it on my comp and stick it in the ipad folder
[16:51:27] <-- budlust has left IRC (Quit: Lost terminal)
[16:58:42] <brad_a> its interesting reading this chat log that you guys were troubleshooting the exact same problem that i was at the same time. clearly i should have bben on here!
[17:06:38] --> Bo_Thomsen has joined #gemrb
[17:13:09] --> barra_away has joined #gemrb
[17:15:09] <lynxlynxlynx> hehe
[17:15:59] <lynxlynxlynx> i just got another wtf of a crash
[17:16:25] <-- barra_home has left IRC (Ping timeout: 255 seconds)
[17:16:31] <lynxlynxlynx> map::updatescripts tried to work on a null actor
[17:16:43] <fuzzie> hm, didn't you encounter that before?
[17:17:11] <lynxlynxlynx> i don't remember
[17:17:24] <lynxlynxlynx> if so, why didn't we put a guard in there?
[17:17:36] <fuzzie> because you should never get null actors there
[17:17:53] <lynxlynxlynx> the count is 6, the party size, but it tries to work with the 7th member
[17:18:27] <lynxlynxlynx> this is at the end of the going-to-brynlaw cutscene, let me look it up
[17:18:39] <lynxlynxlynx> dhewg went the bodhi way, right?
[17:19:15] <fuzzie> yes
[17:19:32] <fuzzie> hence how we found the issues with the triggering there
[17:19:58] <fuzzie> oh, hm, actually, maybe that was someone else? i don't remember
[17:20:19] <fuzzie> i think a null actor in the *queue* can only be corruption
[17:20:33] <fuzzie> don't know where you're getting one, if it's a queue entry or not
[17:22:06] <lynxlynxlynx> the last thing in the log is a music load, right before that the movetoareas in cut41d.baf
[17:26:51] <lynxlynxlynx> can't see anything interesting, i'll try valgrind
[17:53:19] <lynxlynxlynx> ==26199== Conditional jump or move depends on uninitialised value(s)
[17:53:19] <lynxlynxlynx> ==26199== at 0x4F6959C: Interface::GetPalette(unsigned int, int, Color*) const (Interface.cpp:2680) <-- looks like a false positive
[17:54:36] <lynxlynxlynx> another one is in Actor::RefreshEffects(EffectQueue*) (Actor.cpp:2250) , but that's it
[17:54:53] <fuzzie> but it still segfaults?
[17:56:59] <lynxlynxlynx> yes
[17:57:17] <lynxlynxlynx> before there's also something relating to savegames, but i gotta run now
[17:57:34] <fuzzie> oke
[18:08:51] --> raevol has joined #gemrb
[18:25:58] <brad_a> so about linking against sdl 1.3
[18:26:46] <brad_a> the blue paperdoll is happeing on both 1.2 and 1.3 so other than theat the only graphical glitch i get is 4 lines drawn on the screen
[18:27:20] <brad_a> i cant test much more because the ipad port is crashing the min it tries to play the footstep sound
[18:29:29] <brad_a> so it appears that sdl 1.3 is largely backwards compatible with the 1.2 api tho im sure there is better performance or something to gain by using the 1.3 api
[18:33:35] <brad_a> 0 libSystem.B.dylib 0x94386156 __kill + 10
[18:33:36] <brad_a> 1 libSystem.B.dylib 0x94386148 kill$UNIX2003 + 32
[18:33:36] <brad_a> 2 libSystem.B.dylib 0x94418899 raise + 26
[18:33:36] <brad_a> 3 libSystem.B.dylib 0x9442e94d __abort + 124
[18:33:36] <brad_a> 4 libSystem.B.dylib 0x94411628 release_file_streams_for_task + 0
[18:33:37] <brad_a> 5 GemRB iOS 0x000ad318 Actor::PlayWalkSound() + 460 (Actor.cpp:2917)
[18:33:55] <brad_a> does this mean that the sound is being released prematurely?
[18:34:36] <brad_a> 2917 is just the function return and its a void function so i dont know what to make of it
[18:37:23] <brad_a> commenting out the call to playwalksound makes it so i can play. so yeah there is something going on here. ill try to figure it out
[18:38:21] <fuzzie> huh :-/
[18:39:04] <fuzzie> llvm-gcc or llvm?
[18:39:04] <brad_a> i dont know :( playwalk sound causes a crash on iOS and im not sure why yet
[18:39:11] <brad_a> gcc 4.2
[18:39:52] <brad_a> should i try llvm?
[18:40:03] <brad_a> or llvm-gcc
[18:40:24] <fuzzie> no, i just noticed that a google search for release_file_streams_for_task produces llvm stuff
[18:40:35] <fuzzie> does other sound/music work?
[18:40:38] <brad_a> oh yes
[18:40:46] <brad_a> very nicely 8)
[18:48:39] <brad_a> there is no error in the console either. just that its going to play the walk sound. no error at all
[18:52:44] <brad_a> culprit: strnuprcpy(Sound, anims->GetWalkSound(), sizeof(ieResRef) );
[18:52:48] <brad_a> still dont know why
[19:00:19] <brad_a> the walk sound in the avatartable is full of null bytes
[19:23:31] <brad_a> i rather wish i could start a new game on the ios build. because this table is being populated when loading the game i created using my os x build
[19:23:46] <brad_a> so i dont know if its caused by that or what
[19:27:00] <brad_a> i guess it wouldnt have anything to do with that... it would be silly if that were part of the save
[20:23:25] <lynxlynxlynx> it isn't, it's from gemrb's override
[22:29:36] <-- brad_a has left IRC (Quit: brad_a)
[22:30:08] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:45:17] --> brad_a has joined #gemrb
[22:46:49] <brad_a> about this footstep crash i was on about earlier
[22:47:06] <brad_a> i traced it back to the call to srtnuprcpy
[22:47:37] <brad_a> i have no idea what its doing that ios doesn like but changing the call to a regular strncpy works
[22:47:44] <brad_a> footsteps play and no crash
[22:49:36] <brad_a> the really odd bit is that it was crashing in the system library with a sigabrt and nothing like sigfault or the like
[22:58:35] <brad_a> as im sure you are all aware there are several calls to that function in gemrb and none of them other than the footstep sound cause this... most peculiar.
[23:08:41] <brad_a> running this through gdb and setting it to ignore sigabrt reveals that the compiler has produced a bad instruction somehow
[23:09:17] <brad_a> ud2a
[23:09:21] <brad_a> whatever that means
[23:10:12] <brad_a> well eithr the compiler or i suppose that function could be corrupting something
[23:10:44] <brad_a> but i thought the instruction part would be write protected to prevent this
[23:11:43] <-- Maighstir has left IRC (Quit: ~ Trillian Astra - www.trillian.im ~)
[23:13:26] <brad_a> it appears the last line in that function is causing this
[23:13:31] <brad_a> *dest=0;
[23:13:43] <brad_a> why is it there? i have no idea
[23:14:15] <brad_a> removing it seems to have produced no ill sideeffects and fixes my problem
[23:26:33] <wjp> strnuprcpy is a rather dangerous function
[23:27:24] <wjp> there seems to be some confusion about zero termination of the output
[23:28:17] <wjp> it might be best to get rid of it entirely, or add a special variant for ieResRefs
[23:28:29] <wjp> but I don't know
[23:28:44] <wjp> somebody should think about it properly :-)
[23:29:08] <brad_a> yeah it certainly had me up a wall
[23:55:23] <brad_a> so my next goal is to get the virtual keyboard to show up when needed. anybody know where i ought to implement taht?