#gemrb@irc.freenode.net logs for 26 Feb 2013 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:03:53] <-- rocket_hamster has left IRC (Remote host closed the connection)
[00:05:10] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[00:09:25] <Seniorita> [wiki] todo - [Shared] http://www.gemrb.org/wiki/doku.php?id=todo&rev=1361837114&do=diff
[00:10:07] <-- brada has left IRC (Quit: Page closed)
[00:29:00] --> Gekz has joined #gemrb
[00:29:00] <-- Gekz has left IRC (Changing host)
[00:29:00] --> Gekz has joined #gemrb
[01:55:56] --> brada has joined #gemrb
[03:15:11] <-- Drakkar has left IRC (Ping timeout: 260 seconds)
[03:17:33] --> Drakkar has joined #gemrb
[06:41:48] <-- |Cable| has left IRC (Ping timeout: 276 seconds)
[06:55:03] --> |Cable| has joined #gemrb
[07:01:12] <-- brada has left IRC (Quit: brada)
[07:39:17] <-- fuzzie has left IRC (*.net *.split)
[07:40:21] --> edheldil has joined #gemrb
[07:40:21] --- ChanServ gives channel operator status to edheldil
[07:51:06] --> fuzzie has joined #gemrb
[07:53:11] --> rocket_hamster has joined #gemrb
[07:56:36] <-- edheldil has left IRC (Ping timeout: 264 seconds)
[07:58:41] --> edheldil has joined #gemrb
[07:58:41] --- ChanServ gives channel operator status to edheldil
[08:01:26] --> edheldil_ has joined #gemrb
[08:05:00] <-- edheldil has left IRC (Ping timeout: 264 seconds)
[09:05:23] <-- rocket_hamster has left IRC (Quit: bye!)
[09:46:27] --> lynxlynxlynx has joined #gemrb
[09:46:27] --- ChanServ gives channel operator status to lynxlynxlynx
[10:20:45] --> rocket_hamster has joined #gemrb
[10:35:32] <-- rocket_hamster has left IRC (Quit: bye!)
[10:50:55] <-- Drakkar has left IRC (Ping timeout: 240 seconds)
[10:52:22] --> Drakkar has joined #gemrb
[10:55:58] <-- Gekz has left IRC (Ping timeout: 245 seconds)
[10:58:46] <-- Drakkar has left IRC (Read error: Connection reset by peer)
[10:58:50] --> PixelScum has joined #gemrb
[11:09:40] --> Gekz has joined #gemrb
[11:09:40] <-- Gekz has left IRC (Changing host)
[11:09:40] --> Gekz has joined #gemrb
[11:52:12] <psch> with 800x600 set in gemrb.cfg, touch on the top left corner gives me 0,0, on the bottom right 818,604, without widescreen applied, image is upscaled but keeps aspect ratio
[11:52:41] <psch> that's the conf with the difference between where-i-touch and where-cursor-clicks
[11:53:04] <psch> i dont have enough time to wait for the copy of the widescreen-modded bg2 now, ill check that in ~3 hours
[12:36:14] <-- lynxlynxlynx has left IRC (Ping timeout: 255 seconds)
[12:48:22] --> lynxlynxlynx has joined #gemrb
[12:48:22] <-- lynxlynxlynx has left IRC (Changing host)
[12:48:22] --> lynxlynxlynx has joined #gemrb
[12:48:22] --- ChanServ gives channel operator status to lynxlynxlynx
[13:16:55] <-- lynxlynxlynx has left IRC (Ping timeout: 240 seconds)
[13:18:49] --> rocket_hamster has joined #gemrb
[13:25:24] --> lynxlynxlynx has joined #gemrb
[13:25:24] --- ChanServ gives channel operator status to lynxlynxlynx
[13:31:57] <-- lynxlynxlynx has left IRC (Ping timeout: 276 seconds)
[14:29:03] --> lynxlynxlynx has joined #gemrb
[14:29:04] <-- lynxlynxlynx has left IRC (Changing host)
[14:29:04] --> lynxlynxlynx has joined #gemrb
[14:29:04] --- ChanServ gives channel operator status to lynxlynxlynx
[14:31:24] --> kida has joined #gemrb
[14:37:17] <-- rocket_hamster has left IRC (Ping timeout: 255 seconds)
[15:14:17] --> rocket_hamster has joined #gemrb
[15:32:54] --- ermo^ is now known as ermo
[16:02:37] <psch> widescreen-modded to 870x500 i get 0,0 top left and 885,504 at the bottom right
[16:06:15] <-- DrMcCoy has left IRC (Quit: server reboot)
[16:10:52] --> DrMcCoy has joined #gemrb
[16:16:30] --> edheldil has joined #gemrb
[16:16:30] --- ChanServ gives channel operator status to edheldil
[16:19:44] <psch> the coordinates are from touches on the edge of the screen and not the displayed game content btw
[16:21:28] --> brada has joined #gemrb
[16:27:16] <brada> thanks thats what i was asking for
[16:27:56] <brada> sounds like your widescreen mod was a bit off :p
[16:28:16] <brada> did you forget to compensate for the status bar?
[16:30:02] <fuzzie> so gemrb isn't adjusting coords at all?
[16:31:09] <brada> not in that way
[16:31:20] <brada> i only had 4:3 data and a 4:3 screen
[16:31:32] <fuzzie> SDL20Video seems to be adjusting the coords based on SDL_RenderGetLogicalSize
[16:31:36] <brada> as long as my settings are 4:3 everything scales fine
[16:32:04] <brada> the problem occures when your window has a diffrent ratio than your logical render size
[16:32:09] <fuzzie> which is a bit of an odd thing to do
[16:32:18] <brada> yes
[16:32:22] <fuzzie> since the events are already scaled
[16:32:48] <fuzzie> so which coords is psch printing?
[16:33:07] <brada> the corner of the screen
[16:33:10] <brada> corners
[16:33:11] <fuzzie> no, i mean
[16:33:16] <fuzzie> from which source?
[16:33:42] <brada> presumably the value being passed to eventmanager
[16:34:50] <fuzzie> i mean, a better question is, why do you adjust all coordinates by SDL_RenderGetLogicalSize?
[16:35:16] <brada> as opposed to?
[16:35:26] <fuzzie> leaving them as they are
[16:36:30] <brada> i dont know what line/s you are refering to
[16:36:36] <fuzzie> SDL_RenderGetLogicalSize(renderer, &w, &h);
[16:36:37] <fuzzie> xScaleFactor = (state->xres / w);
[16:36:39] <fuzzie> yScaleFactor = (state->yres / h);
[16:37:16] <psch> my widescreen mod off?
[16:37:25] <fuzzie> which is undoing SDL's coordinate transformation
[16:37:29] <psch> because i have coords over the resolution?
[16:37:47] <brada> so what should i divide by?
[16:37:57] <fuzzie> nothing
[16:38:05] <fuzzie> i mean, you understand i know nothing about this, right?
[16:38:06] <brada> because the coordinate surely arent right without :p
[16:38:20] <fuzzie> but the docs and code imply that you're meant to call SDL_RenderSetLogicalSize and then SDL will convert all coords for you.
[16:38:29] <psch> the coords im printing are CursorPos.x CursorPos.y
[16:38:47] <psch> in ProcessFirstTouch()
[16:39:55] <brada> my unaltered coordinates are 15360, 13843… surely not correct
[16:40:16] <brada> altered are 333, 300
[16:40:19] <fuzzie> brada: event.tfinger.x/y?
[16:40:19] <brada> correct
[16:40:21] <brada> yes
[16:40:45] <psch> well CursorPos.x and .y are scaled already, according to comments in the source code
[16:40:50] <brada> yes
[16:42:28] <psch> with the widescreen mod with the aspect ratio of my screen i can touch buttons where they are displayed though
[16:43:01] <brada> its still marginally off according to your coordinates
[16:43:41] <psch> but that occurs for stock game data too, no?
[16:43:48] <brada> no
[16:43:51] <brada> i mean
[16:43:57] <psch> seeing as i had 818,604 on the bottom left corner for stock bg2
[16:43:59] <psch> err, bottom right
[16:44:04] <fuzzie> psch says that 800x600 gives coords of 818,604
[16:44:05] <fuzzie> right
[16:44:09] <fuzzie> so this can't be a data bug
[16:44:17] <brada> if i rund stock 1024x768 with 1024x768 cfg settings then everything is of course as i expect
[16:44:43] <brada> fuzzie: you are correct
[16:45:03] <brada> he jsut has a slight miscalculation (or widescreen mod does)
[16:45:13] <fuzzie> but this isn't widescreen mod
[16:45:21] <fuzzie> there's no widescreen mod at 800x600, that is base unmodified game data
[16:45:23] <brada> or it pourposely rounds
[16:45:33] <fuzzie> and you can't possibly get 818,604 coords from that
[16:45:38] <brada> right
[16:45:53] <fuzzie> so this must be a code bug?
[16:47:03] <brada> it is sdl scaling things to best fit the request but keeping aspect ratio with the screen
[16:47:22] <fuzzie> that doesn't cause coordinates to exceed things in both dimensions
[16:47:44] <brada> i wrote this a very very long time ago
[16:47:51] <brada> i wont say there isnt a bug in it
[16:47:52] <fuzzie> so something is surely broken here
[16:47:58] <brada> but i dont know where to look really
[16:48:05] <fuzzie> yeah, I'm trying to find the cod in sdl
[16:49:02] <brada> i want fish & chips now...
[16:53:57] <fuzzie> so, SDL's android backend just sets native_xres/native_yres to 1.0 and passes the global coords onward
[16:54:07] <brada> oh
[16:54:09] <fuzzie> no no
[16:54:11] <fuzzie> let me finish
[16:54:23] <brada> :)
[16:56:58] <fuzzie> as far as I can tell this means that SDL's touch stuff multiplies the coordinates by the default 2^14 and then truncates the result, which would be surely insane
[16:58:04] <fuzzie> well, I guess they want to stay integer-based
[16:58:11] <brada> insane, but unsuprising given the history with their touch events
[16:58:18] <fuzzie> but that means that you should divide by 2^14, which you do, by dividing by xres or yres
[16:58:42] <fuzzie> and then it would make sense to multiply by width or height to scale from 1.0 to width/height, which you do
[16:58:54] <fuzzie> so that's really dumb
[16:59:01] <fuzzie> and contradictory to the docs
[17:00:32] <fuzzie> i like how there's a IPHONE_TOUCH_EFFICIENT_DANGEROUS define
[17:00:44] <fuzzie> and it's on by default
[17:00:50] <fuzzie> which is why you get insane finger ids
[17:01:08] <brada> ha
[17:01:27] <fuzzie> if you turn it off then you get finger ids starting at 0 too i guess
[17:01:34] <brada> so finger ids are supposed to be 0-4?
[17:01:48] <brada> why is that a 64bit int then?
[17:01:52] <fuzzie> i don't think there's a 'supposed'
[17:01:57] <brada> ah
[17:02:16] <fuzzie> what it does in dangerous mode is cast the UITouch* to an int and pass that
[17:03:44] <fuzzie> ah, GSoC student
[17:05:00] <brada> i assume it is "dangerous
[17:05:12] <brada> "dangerous because it later tries to dereference it?
[17:07:59] <fuzzie> i have no idea
[17:27:48] <-- edheldil has left IRC (Ping timeout: 264 seconds)
[17:32:36] <-- kida has left IRC (Ping timeout: 248 seconds)
[18:18:48] <-- brada has left IRC (Quit: brada)
[18:56:35] --> vampi-the-frog has joined #gemrb
[18:56:45] <vampi-the-frog> hey guys
[18:57:01] <rocket_hamster> heya
[18:57:04] <vampi-the-frog> I'm trying to get Planescape: Torment to run
[18:57:45] --> brada has joined #gemrb
[18:57:48] <vampi-the-frog> I get to the character creation screen, then I get a segfault. The segfault occurs in TextArea.cpp, I fixed the code so the segfault wouldn't happen, then I get a blue screen, let me make a screenie
[18:58:06] <rocket_hamster> is UI visible?
[18:58:11] <rocket_hamster> or whole screen is blue?
[18:58:47] <vampi-the-frog> UI is visible
[18:59:16] <rocket_hamster> check the paths in your gemrb.cfg
[18:59:34] <vampi-the-frog> http://evo.grigoriada.net/pst1.png
[18:59:36] <vampi-the-frog> http://evo.grigoriada.net/pst2.png
[18:59:39] <rocket_hamster> you have to point CD to data directory
[18:59:58] <vampi-the-frog> even if I have the GOG version?
[19:00:07] <vampi-the-frog> let me post my /etc/gemrb/planescape.cfg
[19:00:08] <rocket_hamster> can you paste the gemrb.cfg?
[19:00:10] <rocket_hamster> tes
[19:00:11] <rocket_hamster> yes
[19:00:43] <brada> i doubt you "fixed" anything
[19:00:54] <psch> the segfault probably was missing data
[19:01:11] <vampi-the-frog> http://pastebin.com/jELUc3uc
[19:01:16] <Seniorita> ##################################################### # - Pastebin.com
[19:01:21] <psch> at least judging from the blue screen, cause that's also missing data, isnt it
[19:01:35] <vampi-the-frog> the font pointer was null, so I just added a null check. before that, it would just crash right after the character creation screen
[19:01:53] <rocket_hamster> line 169 et al.
[19:02:04] <psch> you need values for at least one of the CDx variables
[19:02:07] <rocket_hamster> CDx has to point to game data
[19:02:12] <vampi-the-frog> oh alright
[19:02:36] <rocket_hamster> gamedata not necceseraly means game dir
[19:02:46] <rocket_hamster> in bg2 for example its gamedir/data
[19:03:00] <rocket_hamster> depending if you have GOG original etc, or where you unpacked the data
[19:04:07] <psch> copying the GamePath entry to CD1 and commenting the others should work, judging from my pst gog install
[19:04:49] <vampi-the-frog> wel hot dawg, i appended /data to the CD1 path and it works
[19:04:49] <psch> brada: the touch input confusion is somewhere in sdl if i understood you and fuzzie correctly?
[19:05:06] <vampi-the-frog> but it still crashed after some game intro
[19:05:09] <psch> also, did you get and enjoy your fish and chips?
[19:05:13] <vampi-the-frog> any idea how I can configure openal to actually make sound?
[19:05:16] <psch> vampi-the-frog: are you now pointing pst at the bg2 data?
[19:05:31] <vampi-the-frog> GamePath=/media/win/Program Files (x86)/GOG.com/Planescape Torment/
[19:05:32] <vampi-the-frog> CD1=/media/win/Program Files (x86)/GOG.com/Planescape Torment/data
[19:05:54] <psch> alright, i thought you meant appending to what you pasted :)
[19:06:08] <rocket_hamster> openal for windows? o.O
[19:06:10] <vampi-the-frog> but I'm still getting a segfault
[19:06:14] <vampi-the-frog> no, I'm on linux
[19:06:18] <psch> rocket_hamster: judging from his paths he's not on windows
[19:06:30] <rocket_hamster> yeah i saw Program files and i panicked
[19:06:54] <rocket_hamster> did you install from git?
[19:07:00] <vampi-the-frog> yep, latest git
[19:07:08] <rocket_hamster> do you have openal libs and devel?
[19:08:39] <vampi-the-frog> http://pastebin.com/Uy3WQwmH
[19:08:44] <Seniorita> [ResourceManager]: Found 'fatigue.2da' in 'shared GemRB Unhardcoded data'. [Res - Pastebin.com
[19:08:46] <vampi-the-frog> I have openal
[19:09:00] <vampi-the-frog> no dev though
[19:09:02] --> edheldil has joined #gemrb
[19:09:03] --- ChanServ gives channel operator status to edheldil
[19:09:13] <rocket_hamster> how did you compile gemrb then :)
[19:09:19] <lynxlynxlynx> without that plugin
[19:09:24] <rocket_hamster> yup
[19:09:24] --- ermo is now known as ermo^
[19:09:32] <lynxlynxlynx> apparently you also don't have sdl_mixer
[19:09:46] <vampi-the-frog> hm alright
[19:09:48] <vampi-the-frog> I'll recompile
[19:09:52] <rocket_hamster> config will argue about them its just in middle of output
[19:09:59] <rocket_hamster> so its easy to miss
[19:10:13] <lynxlynxlynx> oh, you set the driver to openal
[19:10:21] <lynxlynxlynx> change that if you have all the files for sdl_mixer
[19:10:44] <vampi-the-frog> aren't I "missing features"?
[19:11:02] <lynxlynxlynx> the only plugin we always compile, since it requires no deps, is the one that teleports all the screaming into a desert
[19:11:30] <lynxlynxlynx> i you want to play with sound and music, sure
[19:11:50] <vampi-the-frog> I set the AudioDriver to sdlaudio, and it works
[19:12:04] <vampi-the-frog> but I want to use openal because of the "limited features" of sdlaudio
[19:13:57] <psch> that means getting openal dev and recompiling
[19:14:00] <vampi-the-frog> recompiled && worked ^_^
[19:14:01] <vampi-the-frog> thanks
[19:15:44] <vampi-the-frog> well, it crashes just when the floating skull tries to talk to me
[19:17:51] <vampi-the-frog> aparently there is a Font instance that is 0x0
[19:18:16] --- ermo^ is now known as ermo
[19:18:32] <brada> missing data
[19:19:21] <psch> i have GamePath=$pstdir and CD1=$pstdir and CD2=$pstdir/data for gog.com pst, and the morte dialogue works fine here
[19:19:34] <vampi-the-frog> hm I'll try
[19:19:36] <brada> you also have a full install
[19:19:39] <psch> true
[19:19:45] <psch> but gog doesn't allow partial i think?
[19:19:50] <psch> i dont remember
[19:19:52] <brada> i dont know
[19:20:20] <brada> the log will tell you what its looking for and cant find
[19:21:44] <vampi-the-frog> tried:
[19:21:46] <vampi-the-frog> GamePath=/media/win/Program Files (x86)/GOG.com/Planescape Torment/
[19:21:46] <vampi-the-frog> CD1=/media/win/Program Files (x86)/GOG.com/Planescape Torment/
[19:21:47] <vampi-the-frog> CD2=/media/win/Program Files (x86)/GOG.com/Planescape Torment/data/
[19:21:53] <vampi-the-frog> still segfaulting
[19:21:58] <vampi-the-frog> what should I be looking for in the log?
[19:22:19] <lynxlynxlynx> font init
[19:22:46] <rocket_hamster> vampi-the-frog, give data path for all CDs just to be sure
[19:23:03] <lynxlynxlynx> nah, even those two are useless unless he actually has a data/data dir
[19:23:50] <brada> with gog you should only need 1 cd path
[19:24:20] <vampi-the-frog> I know I'm looking for font init problems, but what are for example the font file extensions?
[19:24:33] <brada> seach the log for font
[19:24:54] <vampi-the-frog> http://pastebin.com/iYkQnjdW
[19:24:58] <Seniorita> [Core]: Loading Fonts... [ResourceManager]: Found 'fonts.2da' in 'GemRB Unhardc - Pastebin.com
[19:25:13] <vampi-the-frog> could that missing STONESML be the issue?
[19:26:30] <brada> whats with finding the font in override?
[19:27:08] <brada> but yes missing STONESML
[19:27:25] <vampi-the-frog> I dunno, I installed some extra stuff in the native windows version
[19:27:37] <vampi-the-frog> ghostdog's UI and a couple of other thingies
[19:27:40] <vampi-the-frog> patches?
[19:27:45] <vampi-the-frog> is that what they're called?
[19:28:07] <brada> mods
[19:28:11] <brada> probably what broke it
[19:28:33] <brada> i get the [Core/WARNING]: MovieFont not loaded: STONESML too
[19:28:46] <vampi-the-frog> yet yours doesn't crash?
[19:28:51] <vampi-the-frog> let's just compare logs, shall we?
[19:28:52] <brada> i didnt say that
[19:29:00] <vampi-the-frog> oh
[19:29:01] <brada> idont know where yours is crashing
[19:29:12] <vampi-the-frog> right when morte approaches nameless one and should speak
[19:29:32] <brada> you clearly assume i know anything at all about pst
[19:29:49] <vampi-the-frog> it's the first line of dialogue ever in the game
[19:30:10] <fuzzie> bad brada!
[19:30:22] <fuzzie> you should stop trying to fix our bugs and play pst :(
[19:30:26] <vampi-the-frog> there is an instance of TextArea, and its ftext pointer is NULL
[19:30:28] <brada> ha ha
[19:30:33] <brada> but no mine does ot crash
[19:30:37] <brada> not
[19:30:48] <fuzzie> (seriously, pst is great)
[19:30:50] <vampi-the-frog> let's just compare logs. can you please pastebin yours?
[19:30:52] <psch> i have stock pst here vampi-the-frog and it doesn't crash
[19:31:00] <brada> i shall like to find time to play it someday
[19:31:38] <vampi-the-frog> I thought Baldur's Gate was kind of lame and generic, and PS:T is a lot cooler
[19:32:07] <brada> how can it be generic?
[19:32:26] <brada> what came berore it that was like it?
[19:32:52] <brada> http://paste.debian.net/238384/
[19:32:54] <vampi-the-frog> it's just the feeling
[19:32:54] <Seniorita> debian Pastezone
[19:33:10] <brada> you would probably be better served tracking with gdb
[19:33:14] <brada> but there you go
[19:33:42] <vampi-the-frog> I did track with gdb, it's just not loading a font
[19:37:50] <brada> interestingly there are several python errors running PST now
[19:38:27] <fuzzie> you're not doing 'make install', right?
[19:38:36] <brada> AttributeError: 'module' object has no attribute 'RestPress'
[19:38:49] <vampi-the-frog> I did make install
[19:38:56] <fuzzie> i mean, brada
[19:39:01] <fuzzie> sorry
[19:39:03] <brada> me?
[19:39:06] <lynxlynxlynx> he did too
[19:39:09] <brada> i just run via xcode
[19:39:12] <lynxlynxlynx> that's the gcw merge
[19:39:17] <brada> but GUICommon.RestPress is clearly wrong :p
[19:39:38] <fuzzie> i will leave you in lynx's capable hands here
[19:40:24] <lynxlynxlynx> you need to remove wherever xcode installs gemrb/GUIScripts/pst/GUICommonWindows.py
[19:41:14] <brada> every build overwrites the old scripts
[19:41:34] <lynxlynxlynx> yeah
[19:41:45] <lynxlynxlynx> imagine what happens when we delete a file
[19:42:08] <brada> ah
[19:42:11] <brada> well
[19:42:18] <brada> no it overwrites the entire directory
[19:42:24] <fuzzie> so you don't have that file?
[19:42:44] <lynxlynxlynx> you can have only worse problems than that
[19:43:05] <lynxlynxlynx> there is no mention of GUICommon.RestPress anywhere anymore
[19:43:17] <fuzzie> and your log clearly says GemRB.app/GUIScripts/pst/GUICommonWindows.py
[19:44:37] <brada> does cp -r merge then?
[19:45:05] <lynxlynxlynx> overwrites, yes
[19:45:21] <lynxlynxlynx> but since the source directory has nothing, it doesn't touch stale files
[19:49:29] --- rocket_hamster is now known as rhamster
[19:49:44] --- rhamster is now known as rocket_hamster
[19:53:36] --> |Blaze|_ has joined #gemrb
[19:55:35] --> Coriander2 has joined #gemrb
[19:55:36] <brada> clearly i misunderstood how cp works
[19:55:44] <brada> how are you supposed to do that?
[19:55:46] <brada> rm first?
[19:56:50] --> kingron has joined #gemrb
[19:56:53] <-- Coriander has left IRC (Ping timeout: 255 seconds)
[19:57:07] <-- |Blaze| has left IRC (Ping timeout: 260 seconds)
[20:00:20] <-- edheldil has left IRC (Ping timeout: 248 seconds)
[20:01:04] <lynxlynxlynx> yes
[20:01:56] <lynxlynxlynx> dir/{a,b,c} tar/{a,b,c,d}
[20:02:04] <lynxlynxlynx> you can't delete stuff with cp :)
[20:03:05] --- ermo is now known as ermo^
[20:04:43] --> thomcom has joined #gemrb
[20:22:29] <-- vampi-the-frog has left IRC (Remote host closed the connection)
[20:23:25] --> vampi-the-frog has joined #gemrb
[20:24:26] <psch> to be pedantic, you can delete file contents with cp, but not existing files
[20:25:06] <thomcom> how do you delete files with cp?
[20:25:24] <thomcom> overwrite an existing file with an empty file?
[20:25:33] <psch> that's what i mean, yes, but the file still exists
[20:25:38] <psch> you're just removing the link to the contents
[20:26:01] <psch> the data probably still does exist somewhere, my filesystem knowledge isn't the best
[20:26:45] <psch> it's not really valueable knowledge, just a lesson in "think about if you're adding anything to the topic" on my side :/
[20:27:36] <vampi-the-frog> how can I import saves from my windows native version?
[20:32:16] <psch> is there a good synopsis of makefile commands anywhere? things like $(subst ...) and $(wildcard ...) with syntax and all
[20:32:38] <vampi-the-frog> http://www.gnu.org/software/make/manual/make.html
[20:32:41] <Seniorita> GNU `make'
[20:33:29] <psch> i guess i deserve that
[20:33:55] <vampi-the-frog> make=the shit
[20:34:16] <psch> by extension it follow that make == bananas
[20:34:37] <-- rocket_hamster has left IRC (Quit: bye!)
[20:36:28] <vampi-the-frog> I'm getting a missing text window background
[20:36:33] <vampi-the-frog> and it flickers when I move my mouse over it
[20:38:14] <brada> you probably shouldnt have installed those mods
[20:38:23] <vampi-the-frog> this is on a fresh install
[20:38:31] <vampi-the-frog> but I set resolution to 800x600
[20:38:39] <vampi-the-frog> now I set it to 640x480 and it works fine
[20:40:40] <psch> pst doesn't do anything but 640x480 on stock
[20:40:45] --- ermo^ is now known as ermo
[20:41:01] <vampi-the-frog> I'll just full screen it
[20:41:14] --> edheldil has joined #gemrb
[20:41:14] --- ChanServ gives channel operator status to edheldil
[20:44:30] <brada> you can install widescreen mod to get 800x600
[20:44:47] <thomcom> make is definitely banas
[20:44:49] <thomcom> bananas
[20:44:50] <thomcom> have fun with that lol
[20:46:30] <psch> yeah, that's gonna take a bit longer than i hoped
[20:46:51] <psch> well, lynxlynxlynx has the build env as i'm using it right now, but im really not happy with the gemrb makefile
[20:46:58] <thomcom> what do you need to do with the makefile?
[20:47:16] <thomcom> That's awesome. If I check out of Git will I be able to build for Android?
[20:47:17] <lynxlynxlynx> vampi-the-frog: you don't need to import save games
[20:47:21] <psch> thomcom: no
[20:47:29] <psch> if you want i can give the build env tar
[20:47:42] <psch> or i can give you a build if you rather not install android-sdk and -ndk
[20:47:43] <thomcom> I'm interested in being your guinea pig
[20:48:02] <psch> performance is rather lackluster though, im getting ~18 fps on native resolution on my nexus 7
[20:48:18] <thomcom> those things I already have. I build android everydayeeeeeeiiiiiii but I don't port a custom modified SDL 1.1 to SDL 2.0 every daiyyeee hell naw
[20:48:38] <psch> im building against the sdl2 hg repo
[20:48:39] <thomcom> If you want to email me I can test out your build env and give you comments to make it more foolproof.
[20:48:58] <thomcom> I don't mind 18fps as long as I can try to play IWD again and go through that one door.
[20:49:20] <thomcom> I'm dead blocked in IWD and didn't have the balls to try to build for Android myself. ;)
[20:49:25] <psch> heh
[20:49:32] <thomcom> so good timing
[20:49:40] <lynxlynxlynx> doesn't look that ugly to me, but i'd try to autogenerate the src list and maybe also the include one
[20:49:57] <psch> yeah, that's what i want to have lynxlynxlynx
[20:50:08] <psch> maybe offer something wrt building against openal OR sdl_mixer too
[20:50:15] <psch> and similarly for the plugins
[20:50:30] <thomcom> did you build all the dependencies from source or use homebrew or such?
[20:50:35] <psch> although i dont exactly know which are always needed and which provide duplicate features
[20:50:37] <lynxlynxlynx> i wouldn't worry too much about it psch
[20:50:50] <psch> thomcom: mostly i googled for existing ports
[20:51:08] <lynxlynxlynx> i think it would be best to leave the dependency stuff as it is and integrate the rest into our cmake infra, so there's less maintenance overhead
[20:51:24] <lynxlynxlynx> i need to look at it a bit more closely though
[20:51:57] <thomcom> so all we should need is just a very specific list of dependencies and their platforms and version numbers?
[20:52:13] <psch> i have all that wrapped up in a script
[20:52:21] <psch> checking out the ports, prebuilding them etc
[20:52:25] <psch> of the dependencies
[20:52:37] <thomcom> rad
[20:52:49] <thomcom> and does gemRB still have some custom modified libs like pelya's old stuff?
[20:53:05] <psch> latest git head still has references to al_android_{pause,resume}_playback though, which don't exist in openal
[20:53:15] <psch> but that seems to be the last thing that doesn't work when trying to build against git
[20:53:26] <psch> thomcom: im currently including pelya python in my build tar
[20:53:28] --- ermo is now known as ermo^
[20:53:35] <psch> because the ports i found don't compile or don't link properly
[20:54:37] <thomcom> stuff like https://code.google.com/p/python-for-android/ has problems?
[20:54:39] <Seniorita> python-for-android - Py4A - Google Project Hosting
[20:55:17] <psch> yeah that doesn't build for me
[20:55:33] <psch> there's another python for android, python2.7.2, which builds fine but segfaults in gemrb
[20:56:02] <psch> http://code.google.com/p/android-python27/ this one
[20:56:04] <Seniorita> android-python27 - Embed Python 2.7 interpreter and your scripts into an Android APK. - Google Project Hosting
[20:56:48] <thomcom> do we know that pelya's python is modified, or could it just be the one he got to work?
[20:57:25] <psch> his python is 2.6, im guessing py4a head isn't in a clean state
[20:57:45] <psch> i tried looking through their tags but didn't find anything that looks like it should be stable
[20:57:52] <psch> but i also didnt check very thoroughly
[20:57:59] <-- vampi-the-frog has left IRC (Quit: Leaving)
[20:58:42] --- ermo^ is now known as ermo
[20:59:12] <thomcom> is your stuff in the commandergenius repo?
[20:59:29] <psch> no
[20:59:34] <psch> i did all this from scratch
[20:59:36] <brada> no he is building for our repo
[21:00:21] <thomcom> sorry, https://github.com/gemrb/gemrb.git ?
[21:00:25] <Seniorita> gemrb/gemrb · GitHub
[21:00:30] <psch> it could go there i guess
[21:00:40] <psch> but im currently just fiddling with it on my comp
[21:00:43] <thomcom> hehe
[21:00:55] <thomcom> oh I thought I read a second ago it was checked in
[21:01:08] <thomcom> because I want to get it and play with it ;)
[21:01:31] <psch> yeah, let me just check my build once more against the latest git and see what breaks
[21:01:34] <psch> so i can write that down
[21:02:17] <brada> do you even need that openal stuff?
[21:02:40] <psch> al_android_{pause,resume}_playback? afaict no
[21:02:49] <psch> it pauses and resumes fine for me with those calls commented out
[21:02:52] <psch> the audio that is
[21:02:55] <brada> yeah
[21:03:05] <brada> thats what i thought
[21:04:10] <psch> from what i can tell, if those calls go it should build perfectly fine against git head
[21:04:27] <psch> im building in a clean env right now to check, so take that as you will
[21:07:35] <fuzzie> which openal are you using?
[21:07:56] <psch> git://repo.or.cz/openal-soft/android.git
[21:09:13] <psch> i remember trying https://github.com/AerialX/openal-soft-android too, but i don't remember why i didnt stick with it
[21:09:15] <Seniorita> AerialX/openal-soft-android · GitHub
[21:10:23] <fuzzie> interesting
[21:10:29] <fuzzie> your music stops when gemrb gets backgrounded?
[21:10:58] <psch> yeah
[21:11:35] <psch> there's a slight delay, maybe half a second or so, but it pauses and resumes just fine
[21:12:01] <fuzzie> ah right
[21:12:09] <fuzzie> brad already coded the calls to Pause()
[21:12:29] <fuzzie> the AudioTrack.pause() thing is just 'nice', otherwise gemrb is wasting battery playing silence maybe
[21:12:55] <fuzzie> what happens if you do it during a long speech (chapter text) for example?
[21:13:06] <fuzzie> i'm just curious if it's really still playing, it's not important
[21:14:05] <psch> ill try that, hold on
[21:16:04] <psch> touching the home button during the intro videos at least pauses the video sound
[21:16:11] <psch> on resume, the video playback is broken, but sound resumes fine
[21:16:25] <fuzzie> right, the video sound is in small chunks
[21:16:26] <psch> as in, video itsself doesn't reappear, but sound play
[21:16:35] <psch> now for chapter intro text
[21:16:37] <fuzzie> so when it stops decoding the video, it'll stop decoding the audio too
[21:16:58] <fuzzie> and the music stops because brad calls Pause() which stops the music/ambients
[21:18:43] <brada> we could pause speech easily if it is not already
[21:18:54] <fuzzie> yeah, but you're still pushing silence
[21:19:03] <fuzzie> if it does keep on mixing
[21:19:32] <fuzzie> but i don't know if it does
[21:20:09] <psch> irenicus keeps finishing his sentences
[21:20:15] <psch> in the bg2 intro sequence
[21:20:43] <fuzzie> it doesn't matter because you can't do anything about it
[21:21:01] <psch> so the question is if after each sentence there's still mixing being done in gemrb?
[21:21:13] <fuzzie> the mixing is being done inside openal itself, the playback thread
[21:21:43] <fuzzie> the not-so-great part is, once there's nothing left to mix, it's still sitting there mixing+playing silence
[21:22:00] <fuzzie> which is what the al_android_{pause,resume}_playback hacks were avoiding
[21:22:17] <fuzzie> but it's a pretty minor problem, I wouldn't care
[21:23:21] <brada> yeah not wort messing with
[21:23:47] <psch> well, not being able to quit gemrb cleanly kinda does make it a bit relevant i think
[21:24:02] <psch> still minor though, im not disputing that
[21:24:02] <fuzzie> you can just quit it, right?
[21:24:07] <psch> no, it hangs
[21:24:22] <fuzzie> can't you just swipe it out of the task list? :P
[21:24:25] <psch> probably
[21:24:35] <psch> im always going to settings -> apps -> force close
[21:24:46] <psch> but swiping likely does the same?
[21:25:02] <fuzzie> i have nooo idea
[21:25:34] <brada> where does it hang?
[21:25:46] <psch> after clicking "quit"
[21:25:49] <psch> or touching rather
[21:25:55] <brada> i mean where in execution?
[21:26:05] <psch> let me get the log
[21:27:27] <psch> D/dalvikvm(32140): threadid=14: thread exiting, not yet detached (count=0)
[21:27:27] <psch> I/SDL (32140): [STUB] GL_DeleteContext
[21:27:33] <psch> that's what i get on touching quit
[21:27:40] <psch> and the screen freezes
[21:27:46] <psch> still reacts to the menubar though
[21:27:57] <psch> so it's not that much of a problem i guess
[21:28:06] <psch> ill fire up gdb if you want
[21:28:43] <brada> its probably because sdl expects us to call exit()
[21:28:49] <brada> so put that at the end
[21:29:02] <brada> just guessing btw
[21:29:07] <brada> i had similar problems on ios
[21:29:23] <psch> put it where?
[21:29:27] <psch> im GemRB.cpp?
[21:30:14] <brada> whereever is works best
[21:30:34] <psch> how would i know :/
[21:30:34] <brada> that will work fine
[21:30:43] <brada> well i dont know what you are doing for android
[21:30:52] <brada> ios has its own special interface
[21:31:03] <psch> i have modified exactly three things currently
[21:31:05] <brada> if you are doing something similar...
[21:31:20] <psch> i commented the al/android.h include and i commented al_android_pause_playback and al_android_resume_playback
[21:31:25] <psch> the rest is exactly as in gt
[21:31:27] <psch> *git
[21:31:54] <brada> we should cahnge those al things to #if pelya
[21:32:04] <brada> or some similar define
[21:32:18] <fuzzie> i suggested just checking the SDL version, as we already do in GemRB.cpp.
[21:32:35] <psch> yeah, i did that previously but i got confused about my local gemrb git
[21:32:39] <psch> so i reset --hard
[21:33:07] <brada> well sdl version has nothing to do with them
[21:33:26] <fuzzie> if you're compiling for SDL 1.2 on android, you'll have them
[21:33:28] <fuzzie> if you're not, you won't
[21:33:37] <brada> those are part of sdl?
[21:33:54] <fuzzie> being picky beyond that seems unproductive :P
[21:34:09] <psch> thomcom: http://filebin.ca/YOlgz2vpIWF/gemrb_android.tar
[21:34:20] <psch> as it says in TODO, there's 3 things you need to modify from gemrb git
[21:34:45] <psch> otherwise, with android-ndk, android-sdk and $NDKPATH/tools/build in $PATH it should build
[21:34:53] <psch> if not, im here for another ~30 minutes
[21:37:04] <psch> wrt to automating, i'd like to clean stuff a bit, but if lynxlynxlynx is fine with how it is, or would rather do that theirselves, it could go on the bs i guess
[21:37:45] <psch> barring the three compile-stoppers
[21:37:45] <brada> psch: put the exit() after the call to GemRB::main
[21:39:42] <brada> can you do #if ANDROID?
[21:39:50] <brada> instead of #ifdef
[21:40:17] <psch> i dont know what you mean
[21:40:20] <psch> on both accounts
[21:40:31] <brada> fuzzie: ?
[21:40:43] <brada> well psch: what is android defined as?
[21:40:46] <fuzzie> brada: doesn't that mean 'never'?
[21:41:09] <brada> i dont know thats why i ask
[21:41:13] <psch> from what i can tell ndk-build defines ANDROID
[21:41:17] <fuzzie> it's -DANDROID
[21:41:19] <psch> i don't know the value it gets defined to
[21:41:28] <fuzzie> so it should work
[21:43:09] <brada> also why cant you use logical operators with #ifdef?
[21:44:58] <fuzzie> because logical operators can't be defined? :P
[21:45:00] <brada> psch: for openal will reblacing #ifdef android with #if ANDROID && SDL_COMPILEDVERSION < SDL_VERSIONNUM(2,0,0) work
[21:45:37] <psch> assuming ANDROID actually gets a value with ndk-build, i would think it would
[21:45:49] <brada> will you check i mean :)
[21:45:51] <psch> oh
[21:45:58] <psch> yeah i can do that i guess
[21:46:13] <fuzzie> i mean, it's just assumed you'll nest them
[21:46:58] <brada> id be more inclined to nest them if ides didnt insist on puting all # macros on the left :/
[21:48:06] <fuzzie> I can sit here and mock the whole concept of irritating unhelpful IDEs if you want.
[21:48:20] <fuzzie> I'm not sure the positioning of preprocessor directives is quite the best reasoning though.
[21:48:37] <brada> right as usual :p
[21:48:48] <psch> in my java course i have people not using autoformat because it breaks some lines
[21:48:52] <psch> but they also dont indent manually
[21:49:35] <psch> again, little relevance, i should talk less
[21:49:51] <brada> and code more!!! ;)
[21:50:19] <psch> jni/src/main/gemrb/plugins/OpenALAudio/OpenALAudio.h:52:51: error: missing binary operator before token "("
[21:50:25] <psch> this is what you #if gives me
[21:55:05] <psch> but i probably typod somewhere, cause with actually pasting it from here it works
[21:55:53] <brada> well too late
[21:56:00] <brada> i already pushed another version
[21:56:16] <psch> the horror
[21:59:21] <brada> did the exit() fix your hang?
[22:00:22] <psch> did you put it somewhere now?
[22:00:28] <psch> because i have no clue where you think i should put it
[22:00:39] <brada> post your sdl_main file
[22:01:27] <psch> SDL_main.h?
[22:04:36] --- ermo is now known as ermo^
[22:04:49] <brada> i dont know what its called
[22:04:55] <brada> sdl_android_main?
[22:05:42] <psch> i have that, yeah
[22:05:58] <psch> ttp://nopaste.info/c8b79f806e.html
[22:06:22] <psch> h at the beginning
[22:10:15] <psch> it has a commented call to exit, probably for a reason?
[22:11:02] <brada> just uncomment that exit
[22:11:22] <brada> in your case the whole application *is* the sdl thread
[22:11:28] <brada> you dont have a wrapper interface
[22:11:32] <brada> at least no yet
[22:11:37] <psch> alright
[22:12:10] <brada> hopefully you will make an interface at some point ;)
[22:12:50] <psch> you do remember the whole thing about me not really having any kind of experience with neither c++ nor developing for android right? :P
[22:15:13] <brada> you wouldnt do it in c++
[22:15:18] <brada> you would do it in java
[22:15:34] <brada> and i happen to know you are in a java calss!
[22:15:38] <brada> class
[22:15:48] <psch> true that
[22:15:52] <psch> im even teaching it
[22:16:29] <psch> and i do have a project that will familiarize me with android to do
[22:16:44] <psch> so i might be capable to actually write something in a month or two
[22:16:53] <psch> but im not promising anything
[22:18:34] <psch> god im forgetting loads of things it seems
[22:19:10] <psch> thomcom: building against git is totally not working, sorry for any hopes
[22:19:32] <psch> there's no call to SDL_AndroidGetExternalStoragePath() anywhere
[22:19:58] <brada> you dont need that to build tho
[22:20:07] <psch> no, not for building
[22:20:23] <psch> imprecise language there, yes
[22:20:40] <psch> but loading any data doesn't work without a rooted device with the .apk you get when you build against git
[22:20:58] <psch> because with a rooted device you can put gemrb.cfg into the cwd of the activity
[22:21:00] <brada> im working on a patch
[22:21:51] <brada> how are the override/guiscripts copied there btw?
[22:21:57] <brada> apk install?
[22:22:17] <psch> they could get bundled into the .apk i guess
[22:22:20] <psch> but i did it manually
[22:22:30] <brada> lets bundle them please
[22:22:40] <psch> yeah, i gotta figure out how to bundle them
[22:22:57] <psch> but accessing them also has to happen via the Activity then
[22:23:09] <psch> which is another "go write the interface" thing
[22:23:26] <psch> im not sure if i can have the apk install put files somewhere in the way we need it
[22:23:50] <psch> i did find loads of stuff of how to load bundled assets, but that'd definitely turn into quite a headache
[22:24:27] <brada> you cant write some simple java that can move them?
[22:25:11] <brada> or can we not access those assets using c++?
[22:25:21] <psch> i think we can, but id have to look up how
[22:25:36] <psch> that kinda messes with existing logic, but then so does SDL_AndroidGetExternalStoragePath()
[22:25:41] <brada> http://stackoverflow.com/questions/10166638/access-android-apk-asset-data-directly-in-c-without-asset-manager-and-copying
[22:25:42] <Seniorita> Access Android APK Asset data directly in c++ without Asset Manager and copying - Stack Overflow
[22:25:43] <psch> well, not with the logic, but the codebase
[22:25:52] <psch> but that's fine i guess
[22:27:10] <psch> the second answer seems to be what we want
[22:27:26] <psch> i think that's what pelya does too
[22:28:21] <psch> but i guess it doesn't matter how we do it, whether it's loading override, unhardcoded and GUIScripts from the apk or downloading them somewhere
[22:28:21] <brada> how about SDL_AndroidGetInternalStoragePath
[22:28:47] <psch> yeah that's where the apk should get extracted to
[22:29:00] <psch> so if i just add the folders directly into the apk they should end up there
[22:29:30] <psch> which has the added bonus of not exposing the user to those files and reducing the risk of people putting game-override files into the gemrb override folder
[22:30:05] <brada> and having them in a known place instead of making users putz around with configs and paths
[22:30:13] <psch> yup
[22:30:33] <psch> the annoying part is, gemrb is actually more interesting than the android project i have to do for uni :/
[22:30:48] <brada> one thing that would be super handy would be if you could implement gui errors for android :p
[22:30:56] <brada> for the fatal errors i mean
[22:31:27] <brada> just bolt it onto the AndroidLogger class
[22:32:48] <psch> makes sense
[22:33:07] <psch> im wondering if those should be translated or explained
[22:33:35] <brada> nah
[22:33:43] <psch> but that depends on what they are when SDL_AndroidGet{External,Internal}StoragePath() is implemented
[22:34:33] <psch> cause what im seeing right - "The cache path couldn't be registered, please check!" - doesn't really point towards "I don't know where to look at all."
[22:34:44] <psch> but that's because there's an obvious overlook in the code currently
[22:34:56] <psch> anyway, midnight, i'll go sleep
[22:42:11] <psch> except not really midnight but a misset clock on the server that runs my irc client...
[22:42:32] <psch> anyway, im done for today
[22:51:26] <-- kingron has left IRC (Quit: Leaving)
[23:21:04] <-- brada has left IRC (Quit: brada)
[23:21:52] <thomcom> @psch: Thanks!
[23:51:18] <-- thomcom has left IRC (Ping timeout: 245 seconds)