#gemrb@irc.freenode.net logs for 10 Dec 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[02:58:11] --- barra_home is now known as barraAway
[04:24:18] --> barra_home has joined #GemRb
[04:25:15] <-- barraAway has left IRC (Ping timeout: 245 seconds)
[04:30:35] <-- deepinthewoods has left IRC (Quit: Leaving)
[07:47:24] --> jschall has joined #GemRb
[07:50:47] <jschall> Hi, love BG2, trying to get it working on my netbook. I run it under wine on my desktop, but on the netbook it just doesn't perform well enough to be playable. So I'm trying to set up gemrb, and i immediately notice that i can't alt-tab out of it. it really won't work for me if i can't, because the primary reason i want it at native res is so that i can multitask well.
[07:52:01] <jschall> is there any solution? i don't think i can handle a windowed mode, and i actually tried setting fullscreen to 0 in gemrb.cfg and it still comes up fullscreen and i can't alt-tab
[07:52:41] <jschall> windowed mode generally makes it really hard to scroll around the map with the mouse
[08:10:35] <jschall> it's ignoring the fullscreen option anyway. i set it to 1, it's fullscreen. i set it to 0, it's fullscreen
[08:13:09] <jschall> oh, maybe i need to set bg to windowed mode.
[08:14:18] <jschall> winner
[08:18:57] <jschall> well, i made it windowed with no border and made it the size of my screen, so it's basically fullscreen. except, my KDE panel wants to be on top of it.
[08:36:48] --> lubos has joined #GemRb
[08:39:57] <edheldil> jschall: fullscreen mode is controlled by baldur.ini in the original game data. You can change it there
[08:40:13] <jschall> edheldil: i already figured that out =P
[08:40:27] <edheldil> If it still works, you can confine mouse to gemrb window by pressing Ctrl-G
[08:41:02] <jschall> edheldil: i have it working the way i want, i just don't think i can handle the little bugs. i really wish i could get the original engine working properly in wine
[08:41:31] <edheldil> no idea whu Alt-TAB does not work - this should be done by your Desktop manager, i.e. KDE
[08:42:05] <jschall> edheldil: because fullscreen sdl games grab the keyboard and keep it.
[08:42:07] <tomprince> jschall: we want your bugs (so we can fix them)
[08:42:40] <edheldil> well, it's still developed. If you have found bugs we do not know about, feel free to report them in G3 thread
[08:43:27] <tomprince> Or here, or the wiki ... we aren't really very formal around here. :)
[08:43:41] <edheldil> jschall: or, better, in SF bugtracker
[08:44:14] <edheldil> tomprince: I thought about it and the bugtracker should be the preferred way to go
[08:44:46] <tomprince> Well, maybe we should find a better bug tracker?
[08:45:09] <jschall> edheldil: every time i bitch at a DE dev about games being able to prevent alt-tab from working, i get this: "the program shouldn't do that." i respond, "no, the window manager shouldn't LET it." and it's all downhill from there.
[08:46:46] <jschall> edheldil: i want to write a wrapper that just forces the game into a window, and then i can just remove the border from the window and force it on top of the panel and then alt-tab will work. that's what i had to do to gemrb.
[08:47:16] <jschall> edheldil: i had to use always-on-top to force it on top of the panel, but i have a key combination for that so i just un-always-on-top and then alt-tab.
[08:48:41] <jschall> edheldil: well, i haven't delved very deep into this, but i can't see it being quite the same as the original engine, so i think i'm back to square 1: get wine to stop having bad performance.
[08:49:01] <tomprince> I am not saying that gemrb does the right thing, but it does depend on the window manger. I just tested it here, and I can switch away from fullscreen.
[08:49:17] <jschall> tomprince: what wm are you using?
[08:49:25] <tomprince> awesome.
[08:49:42] <tomprince> It might have something to do with it using a different key.
[08:50:23] <jschall> tomprince: it probably has a workaround built in.
[08:50:26] <tomprince> It might be that since GemRB cares about alt, it decides to eat it before the WM does anything with it. Or the WM decides that since GemRB cares about it, it ignores it.
[08:50:40] <jschall> tomprince: because both metacity and kwin have trouble with fullscreen sdl games
[08:51:23] <jschall> tomprince: doesn't matter what the game is, if it uses sdl and its fullscreen, it doesn't let you alt-tab out of it.
[08:53:43] --> lynxlynxlynx has joined #GemRb
[08:53:43] --- ChanServ gives channel operator status to lynxlynxlynx
[08:54:31] <jschall> tomprince: i love kde4, though. i wouldn't go back to gnome.
[09:01:26] <tomprince> Well, awesome is neither gnome nor kde.
[09:02:14] <jschall> tomprince: i know
[09:02:44] <jschall> tomprince: does it have compositing?
[09:02:52] <jschall> tomprince: i loves me some compositing.
[09:03:35] <tomprince> No, but you can use something like xcompmgr.
[09:03:51] <tomprince> It does know how to set the opacity for a compositor.
[09:04:55] <jschall> tomprince: i see
[09:10:16] <tomprince> I think it is probably one the more configurable WMs around.
[09:12:50] <-- |Cable| has left IRC (Remote host closed the connection)
[09:16:04] <edheldil> hehe, this is exactly your problem : https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/537137
[09:16:35] <edheldil> I especially liked the way the first idiot reassigned it to fglrx-installer
[09:18:19] <edheldil> oh my, it's even better as I read on :)
[09:54:21] --> Bo_Thomsen has joined #GemRb
[10:00:04] <-- Bo_Thomsen has left IRC (Quit: Leaving.)
[11:54:36] --> barra_away has joined #GemRb
[11:57:45] <-- barra_home has left IRC (Ping timeout: 245 seconds)
[12:16:12] --> barra__away has joined #GemRb
[12:19:25] <-- barra_away has left IRC (Ping timeout: 245 seconds)
[13:33:46] --- barra__away is now known as barra_home
[13:48:21] --> Bo_Thomsen has joined #GemRb
[14:22:23] <-- Bo_Thomsen has left IRC (Quit: Leaving.)
[14:53:52] --> SiENcE has joined #GemRb
[16:08:31] --> barra_away has joined #GemRb
[16:11:55] <-- barra_home has left IRC (Ping timeout: 245 seconds)
[16:30:33] --- barra_away is now known as barra_home
[16:50:55] <-- lubos has left IRC (Quit: Leaving.)
[16:58:56] <-- SiENcE has left IRC (Quit: @all: cya)
[17:01:09] --> |Cable| has joined #GemRb
[18:45:11] <lynxlynxlynx> huh, crash on intro
[18:46:01] <lynxlynxlynx> related to the arm sdlaudio change
[19:04:24] <fuzzie> yes
[19:04:32] <fuzzie> - //multiply always by 2 because it is in 16 bits
[19:04:32] <fuzzie> - int rawsize = cnt * 2;
[19:04:33] <fuzzie> - unsigned char * memory = (unsigned char*) malloc(rawsize);
[19:04:38] <fuzzie> + // Use 16-bit word for memory allocation because read_samples takes a 16 bit alignment
[19:04:41] <fuzzie> + short * memory = (short *) malloc(cnt);
[19:04:48] <fuzzie> ^- that should be malloc(rawsize); still
[19:04:49] <fuzzie> sorry
[19:05:06] <fuzzie> not really been in the best state for patch reviewing recently. just cnt*2 should work fine.
[19:07:57] <lynxlynxlynx> it's my fault too
[19:08:25] <lynxlynxlynx> i noticed that bit but thought it was related to the smaller type and didn't test
[19:08:31] <fuzzie> just, you would hope that one of us would have caught it :-)
[19:08:35] <fuzzie> oh well
[19:10:00] <lynxlynxlynx> oh well
[19:22:29] <CIA-5> GemRB: 03lynxlupodian * r3dcdade8f62d 10gemrb/gemrb/plugins/SDLAudio/SDLAudio.cpp: SDLAudio: fixed a crash from 23914315
[19:43:49] --> Avenger has joined #GemRb
[19:43:50] --- ChanServ gives channel operator status to Avenger
[19:44:43] <Avenger> hi everyone, especially fuzzie, wjp --> jschall found the same bug that ate my mouse last week. It is good to see it isn't just my system.
[19:45:46] <fuzzie> didn't you lose the mouse entirely, even after gemrb died?
[19:45:46] <lynxlynxlynx> you're both using ubunut? ;P
[19:45:48] <Avenger> looks like repeatedly starting sdl in full screen mode, then trying to run it in windowed mode causes some schrödinger effect
[19:46:03] <Avenger> yes, i lost it
[19:46:11] <Avenger> i use ubunut, yep
[19:50:39] <Avenger> also thanks for the ida upgrade warning, i still have to get used to the new viewing mode, but i hope it is less crashy now
[19:51:41] <fuzzie> well, i doubt the debugger is better :P
[19:54:41] <lynxlynxlynx> we found one more animation complication
[19:55:45] <lynxlynxlynx> the game uses the same animations for gnomes and halflings (and some others), but has different inventory images for them
[19:57:00] <lynxlynxlynx> so our avatars table is correct when it comes to drawing the actors ingame, but in the inventory, the base (wrong) animation is used
[19:57:36] <Avenger> well, then paperdolls need a new table :(
[19:57:44] <lynxlynxlynx> luckily you can see the inventory of only a handful of creatures, so it should be doable with just one more column
[19:57:54] <lynxlynxlynx> we add the armor level manually anyway
[19:58:14] <Avenger> actually, pdolls.2da is there
[19:58:36] <Avenger> it should be just changing a pdolls.2da entry?
[19:58:50] <lynxlynxlynx> oh, nice
[19:59:04] <lynxlynxlynx> it's just missing iirc
[19:59:45] <lynxlynxlynx> i hope it won't be a problem that the resource is a plt, not a bam
[20:00:17] <Avenger> well, i hope the plt and bam has the same resref
[20:00:49] <lynxlynxlynx> there is no bam for it, but if there was, it would, yes
[20:01:18] <Avenger> i guess, we can handle both, transparently. the original engine can
[20:02:18] <lynxlynxlynx> from what i found on the forum, it first tries the plt
[20:15:56] <lynxlynxlynx> hmpf, i forgot how i found the correct one
[20:31:20] <CIA-5> GemRB: 03lynxlupodian * r0d20930a9ec9 10gemrb/gemrb/override/bg2/pdolls.2da: bg2: fixed Jan's inventory doll
[20:46:29] <Avenger> hmm, gemrb doesn't exit properly in Windows
[20:48:47] <fuzzie> crash during destruction?
[20:48:57] <Avenger> only if i compile it in release mode, yep
[20:49:47] <Avenger> this may be a problem because of the separate memory managers in dll's
[20:50:55] <fuzzie> you're not linking against the dll library?
[20:51:02] <fuzzie> if not, you should :P
[20:51:23] <Avenger> i don't know what are the options here
[20:51:26] <fuzzie> but i think destructor order just screws everything up, we need to destroy things manually and very carefully
[20:51:41] <fuzzie> there's a 'multi-threaded release dll' runtime library
[20:51:47] <fuzzie> as opposed to something with 'static'
[20:53:24] <Avenger> hmm, odd, gemrb.cpp is compiled with single threaded ... (at least in debug mode), but that works
[20:53:58] <fuzzie> i doubt it matters
[20:54:54] <fuzzie> just, don't link static :)
[20:55:09] <fuzzie> oh, right, mingw on Windows in release mode breaks with destructors too, i think?
[20:55:31] <Avenger> really? cool :P
[20:55:43] <Avenger> so i'm not alone
[20:56:04] <fuzzie> i think it's just that, when things are destroyed, they still try and access the core, which already got deleted
[20:56:07] <Avenger> btw, i don't know what did you mean with destroying things manually
[20:56:13] <Avenger> i thought we do that
[20:56:23] <fuzzie> i don't think we do
[20:56:26] <fuzzie> too much is automatic
[20:57:22] <fuzzie> yeah, like SDLVideoDriver does core->FreeString(subtitletext)..
[20:58:41] <fuzzie> you could just do some 'video.release();' on Interface.cpp:416 and that should be fixed automatically though
[20:58:55] <fuzzie> but i think there are more annoying problems. no backtrace i guess, in release :(
[20:59:23] <Avenger> do you think the destruction order wouldn't cause problem in debug mode?
[20:59:48] <fuzzie> i think it is a different order :)
[20:59:58] <fuzzie> but, it is just my best guess!
[21:00:00] <fuzzie> i didn't look into it
[21:19:31] <Avenger> there is a window graphics bug, if i unload a window that covers another grayed window (modal shadow), then the unloaded window still remains drawn, because the grayed one isn't redrawn
[21:20:00] <Avenger> this may be some regression
[21:23:10] <fuzzie> odd
[21:25:18] <Avenger> another graphics bug: scrolling a textarea with mouse wheel while the mouse isn't in the textarea --> the window isn't cleared before redraw, causing garbage
[21:26:27] <fuzzie> hmm
[21:26:40] <fuzzie> while you're here: do you know if the textarea in the bg2 inventory is broken simply because the CHU has the scrollbar over it?
[21:26:59] <fuzzie> it looked that way from some poking, but i can't run DLTCEP or something to make sure
[21:27:08] <lynxlynxlynx> i can confirm the garbage (text)
[21:27:44] --> Bo_Thomsen has joined #GemRb
[21:29:31] <lynxlynxlynx> no, there's a 5px gap
[21:30:25] <fuzzie> erm, not inventory. record screen?
[21:30:28] <lynxlynxlynx> the textarea starts at 208,161; the scrollbar at 467,161; 467 > 208+254 (hsize)
[21:30:33] <fuzzie> oh, no
[21:30:38] <fuzzie> but, well, i guess you know the one i mean
[21:30:46] <lynxlynxlynx> i was looking at the item description window
[21:30:50] <fuzzie> yes :)
[21:31:47] <fuzzie> thanks
[21:35:48] <Avenger> whoa, crash
[21:36:35] <Avenger> something about strings
[21:37:11] <Avenger> just copied the biography editor stuff from the bg2 guiscript to iwd2
[21:37:21] <Avenger> maybe bg2 crashes too... lets see
[21:38:56] <Avenger> no crash there,
[21:41:05] <Avenger> hmm, the strref value is 0xcdcdcdcd ...
[21:41:16] <Avenger> that is freed memory
[21:44:48] <Avenger> hah, i see the bug
[21:45:15] <Avenger> %u is the unsigned format string?
[21:45:56] <fuzzie> yes
[22:10:42] <-- Avenger has left IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
[22:15:10] <CIA-5> GemRB: 03avenger_teambg * rf6fb300ec63e 10gemrb/gemrb/plugins/TLKImporter/TLKImporter.cpp: fix buffer overflow in case of huge strrefs
[23:04:40] <CIA-5> GemRB: 03avenger_teambg * r598ac54a89e2 10gemrb/gemrb/GUIScripts/bg2/GUIREC.py: guirec:added missing function in bg2
[23:04:43] <CIA-5> GemRB: 03avenger_teambg * rce0b3cb363b0 10gemrb/gemrb/ (5 files in 4 dirs):
[23:04:43] <CIA-5> GemRB: iwd2 customisation windows
[23:04:43] <CIA-5> GemRB: Support iwd2 SoundSets a bit more by returning the filename prefix, if needed
[23:09:06] --- barra_home is now known as barraAway
[23:18:44] <-- lynxlynxlynx has left IRC (Read error: Connection reset by peer)
[23:35:30] <-- Bo_Thomsen has left IRC (Quit: Leaving.)