#gemrb@irc.freenode.net logs for 5 Mar 2013 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:01:05] --> rocket_hamster has joined #gemrb
[01:10:32] <-- rocket_hamster has left IRC (Quit: bye!)
[01:55:22] --> brada has joined #gemrb
[03:31:34] --- ermo^ is now known as ermo
[04:12:43] <-- brada has left IRC (Quit: brada)
[04:18:50] --> brada has joined #gemrb
[04:27:04] <-- DrMcCoy has left IRC (Ping timeout: 272 seconds)
[04:33:00] --> DrMcCoy has joined #gemrb
[04:59:46] <-- brada has left IRC (Quit: brada)
[05:27:39] --> brada has joined #gemrb
[06:51:02] <-- brada has left IRC (Quit: brada)
[07:44:52] --> lynxlynxlynx has joined #gemrb
[07:44:52] <-- lynxlynxlynx has left IRC (Changing host)
[07:44:52] --> lynxlynxlynx has joined #gemrb
[07:44:52] --- ChanServ gives channel operator status to lynxlynxlynx
[08:10:54] --- ermo is now known as ermo^
[08:32:20] --> WingedHussar has joined #gemrb
[10:15:48] <-- Coriander has left IRC (Quit: Nettalk6 - www.ntalk.de)
[10:16:46] --> Coriander has joined #gemrb
[11:13:19] <edheldil> fizzle: iesh has problems parsing encrypted streams and bg2's bcs files, so not yet
[13:23:48] --> kida has joined #gemrb
[13:26:42] --> kida_ has joined #gemrb
[13:29:35] <-- kida has left IRC (Ping timeout: 256 seconds)
[13:52:12] <lynxlynxlynx> there are many more non-panic versions
[13:54:17] <lynxlynxlynx> iwds and pst don't seem to use the panic version at all
[13:56:07] --> rocket_hamster has joined #gemrb
[14:17:09] <-- rocket_hamster has left IRC (Remote host closed the connection)
[15:07:59] <psch> alright, im packaging a gemrb.cfg with the apk now and check the .cfg for a placeholder on every start, which gets replaced with the path i extract override etc. to
[15:08:47] <psch> i was playing with a dialog for gamepath etc a bit, but i haven't found out yet how to do that sanely and cleanly
[15:55:48] --> brada has joined #gemrb
[15:59:03] <brada> well time to pull in the sdl changes
[16:17:21] --> thomcom has joined #gemrb
[16:31:46] <psch> thomcom: would you try this apk, see if it finds a config?
[16:31:48] <psch> http://filebin.ca/ZAPDD9FJ2kD
[16:32:23] <psch> i mean, it should definitely find a config, because it's extracting a confing from the apk if it doesn't find a config, but i'm curious where it ends up or if it magically finds the config
[16:37:22] <psch> actually, let me expand on that a bit
[16:37:32] <psch> the config it extracts should definitely end up in a different location
[16:38:18] <psch> but it also should definitely be somewhere gemrb can find it AND gemrb actually looks
[16:38:41] <psch> so i'd be just a matter of correcting a few things in the extracted config to get your game going
[16:39:44] <-- lynxlynxlynx has left IRC (Ping timeout: 240 seconds)
[16:40:19] <brada> psch: you need to update your SDL now
[16:40:52] <psch> you're finished in-tree?
[16:40:56] --- ermo^ is now known as ermo
[16:40:56] <brada> yes
[16:40:58] <psch> oh nice
[16:40:59] <brada> afik
[16:41:07] <brada> i did not do extensive testing
[16:41:15] <brada> but it builds and basic touch input works
[16:41:33] <brada> they said they were going to be reworking the advanced gestures tho
[16:41:41] <brada> so im not going to bother with those right now
[16:41:47] <psch> are doing anything by default with advanced gestures?
[16:41:50] <psch> *are we
[16:41:54] <brada> formation rotation
[16:41:59] <brada> i think
[16:42:06] <brada> honestly i forget how i implemented that
[16:42:29] <brada> yes it is a SDL_MULTIGESTURE
[16:42:43] <psch> so that's possibly broken right now
[16:42:48] <psch> or is it not changed yet in sdl
[16:43:56] <brada> i cant effectively test it with the simulator ;)
[16:44:09] <brada> previously i had used a friends iphone to test it
[16:44:15] <psch> oh right
[16:44:28] <psch> ill have a look in a bit
[16:47:42] <brada> oh also we should be compatible with mice
[16:48:01] <brada> obviously i couldnt test that either
[16:48:04] <psch> i have neither an OTG cable nor a bluetooth mouse
[16:48:16] <brada> oh well
[16:48:33] <fuzzie> mice on iOS?
[16:49:02] <brada> i saw somebody at the uni yesterday with a rediculous setup with an ipad mini and a keyboard mouse combo 3x larger than the ipad
[16:49:29] <psch> i did get myself a bluetooth keyboard for my tablet
[16:49:36] <psch> but it has about the same size
[16:49:42] <fuzzie> i have access to a transformer with a builtin trackpad, but it's android of course
[16:49:45] <psch> anything noticeable bigger seems really silly to me
[16:50:07] <brada> yes its super silly
[16:50:21] <brada> why go ultra portable then have a massive keyboard
[16:51:16] <fuzzie> well you can dump the keyboard in an office/locker/etc and still take the tablet anywhere..
[16:51:27] <brada> true
[16:51:29] <brada> but those cases they make with the keyboard builtin are pretty slick
[16:51:30] <fuzzie> does seem silly to not just also have a portable keyb though
[16:51:33] <brada> id rather do that
[16:51:47] <fuzzie> well, compared to the transformer all the solutions seem like hacks
[16:51:51] <psch> i had a few people complaining when typing on my acer aspire one, which has like 86% sized keyboard
[16:51:56] <fuzzie> i'm hoping that transformer-style hardware will become the standard someday
[16:52:02] <psch> i'd estimate the one for my tablet is at 65% or so
[16:52:10] <psch> works for me, but i can see people not getting along with small keyboards
[16:52:34] <psch> the transformer is really nice hardware, from the looks of it
[16:52:54] <fuzzie> and, yes, I guess people sometimes have trouble typing on my aspire one keyboard too
[16:52:58] <-- thomcom has left IRC (Ping timeout: 245 seconds)
[16:53:21] --> thomcom has joined #gemrb
[16:53:28] <fuzzie> although I think mine is actually smaller than the transformer keyboard..
[16:53:40] <thomcom> http://pastie.org/6392911
[16:53:42] <Seniorita> #6392911 - Pastie
[16:53:53] <thomcom> seems like it fails to open a GemRB.cfg, then opens a different GemRB.cfg secretly?
[16:53:57] <fuzzie> anyway by your silence I assume there's no magical android mouse support I can test :P
[16:54:10] <fuzzie> thomcom: not really secretly, it says it right there
[16:54:16] <fuzzie> I/GemRB ( 9518): [Config/]: Trying to open "/storage/sdcard0/Android/data/net.sourceforge.gemrb/files/.GemRB/GemRB.cfg".
[16:54:24] <psch> thomcom: the first try is in PWD
[16:54:35] <brada> fuzzie: did you ask me something?
[16:54:37] <fuzzie> why .GemRB?
[16:54:41] <fuzzie> brada: 'mice on iOS?' :P
[16:54:55] <brada> ah yes it would appear they can be used
[16:55:06] <psch> fuzzie: because im currently setting $HOME in GemRB.cpp
[16:55:14] <brada> hence the story about they dude with the desktop keyboard mouse combo with ipad mini
[16:55:14] <psch> instead of fudging with Interface.cpp
[16:55:29] <brada> yeah we can deal with that later
[16:55:32] <fuzzie> brada: yeah, just you said 'oh also we should be compatible with mice' and i wondered on what platform
[16:55:39] <psch> thomcom: you should be able to get your game running with making edits to /storage/sdcard0/Android/data/net.sourceforge.gemrb/files/.GemRB/GemRB.cfg
[16:55:43] <brada> oh any platform sdl works with i guess
[16:55:47] <psch> at least from the looks of the log
[16:55:52] <fuzzie> because as far as I can tell there's no android mouse support in there
[16:56:02] <brada> in sdl?
[16:56:06] <fuzzie> yeah
[16:56:09] <brada> sad
[16:56:19] <brada> wasnt there before?
[16:56:23] <fuzzie> nope
[16:56:37] <brada> so pelyas was done with his own work?
[16:56:47] <fuzzie> pelyas was /all/ his own work
[16:57:00] <brada> wait
[16:57:03] <brada> are you *sure*?
[16:57:05] <fuzzie> yes?
[16:57:11] <fuzzie> i mean, about which bit?
[16:57:26] <brada> because the whole reason they put that in there was because people were complaining about not being able to distingusih on mobile devices
[16:57:29] <fuzzie> i mean, pelyas code predates any android code in sdl1.3 tree by a long time
[16:57:42] <fuzzie> distinguish?
[16:57:51] <fuzzie> it seems like it would be pretty easy
[16:57:56] <fuzzie> you have touch events, and mouse events
[16:57:59] <fuzzie> they are completely different things
[16:58:12] <brada> and when they first changed the android touch code to issue mouse events is when all the fighting started :p
[16:58:29] <fuzzie> well, as far as I could tell, there's just people who don't want to bother updating their code to check for touch events
[16:58:40] <brada> right
[16:58:44] <brada> mostly
[16:58:44] <fuzzie> so instead they decided to cause huge problems for everyone sane and generate fake mouse events in sdl for touch events
[16:58:52] <brada> but there were people that wanted to support both
[16:59:06] <fuzzie> mouse events on desktop and touch events on android
[16:59:08] <brada> we had to ifdef the mouse events out
[16:59:14] <fuzzie> yes
[16:59:24] <fuzzie> that would be 'cause huge problems for everyone sane'
[16:59:32] <psch> brada: two quick taps segfault gemrb
[16:59:36] <brada> http://forums.libsdl.org/viewtopic.php?t=8793&sid=d3d0ad01b9a8d8613ff0fcb1983ab60c
[16:59:40] <brada> fussie: ^
[16:59:40] <Seniorita> SDL :: View topic - patch: simulated mouse events/duplicate touch events
[16:59:44] <brada> fuzzie
[16:59:58] <fuzzie> brada: right, how is that not completely broken?
[17:00:07] <psch> ill fire up ndk-gdb and look around a bit
[17:00:15] <fuzzie> "On Android and UIKit, when SDL sees a touch/finger event, it also sends a mouse motion or mouse button event." = super fail
[17:00:20] <brada> right
[17:00:28] <-- thomcom has left IRC (Ping timeout: 245 seconds)
[17:00:40] <brada> now you can see that the mouse events weere made by touch input tho
[17:00:46] <fuzzie> yes, but it's still stupid
[17:00:54] <brada> fair enough
[17:00:56] <fuzzie> it's SDL2, it's shiny and new, just fix your event handling darnit :(
[17:01:04] <fuzzie> anyway look there's patch in there for mouse stuff on android!
[17:02:19] <brada> i guess the point is if/when android gets SDL support for mice we are no longer broken :P
[17:02:51] <fuzzie> so that's hopeful
[17:03:19] <fuzzie> the thread seems fairly agreed with me that it's stupid, also
[17:03:57] <fuzzie> i guess the bug for android mouse support is http://bugzilla.libsdl.org/show_bug.cgi?id=1666
[17:04:00] <Seniorita> Bug 1666 – Handle mouse events in Android
[17:04:01] <fuzzie> i shall bookmark it
[17:04:19] <brada> yeah i was actually jsut looking at that
[17:05:35] --> thomcom has joined #gemrb
[17:05:40] <thomcom> psch: thanks, cool
[17:05:46] <thomcom> nice on leaving the old read in
[17:06:05] <psch> what do you mean "leaving the old read in"?
[17:06:16] <thomcom> the old cfg read in
[17:06:23] <thomcom> instead of just replacing it with your hard coded attempt, you attempt both
[17:06:30] <psch> oh
[17:06:31] <psch> yeah
[17:06:43] <psch> i have to, otherwise im always overwriting your configured gamepath
[17:07:09] <psch> so you're telling me it works? :)
[17:07:32] <-- brada has left IRC (Quit: brada)
[17:11:10] --> lynxlynxlynx has joined #gemrb
[17:11:10] --- ChanServ gives channel operator status to lynxlynxlynx
[17:11:16] <thomcom> Oh well it is trying to work. :)
[17:11:27] <thomcom> I'm going to modify the GemRB.cfg when I get the chance and then I'll give you more update hehe.
[17:11:50] <thomcom> I realized this weekend that my IWD save file is going to be corrupt relative to 0.7.2 so these efforts are essentially in vain. :(
[17:11:55] <thomcom> still fun to work on the project with you
[17:12:01] <psch> it should be fairly usual stuff from now on, i.e. GamePath, CD{1,2,3,4,5}, Width, Heigth...
[17:12:40] <psch> just dont touch GemRBOverridePath, GemRBUnhardcodedPath and GUIScriptsPath heh
[17:12:40] <thomcom> yup and that's already on my phone ready to go
[17:21:51] <lynxlynxlynx> why the corruption?
[17:24:48] <thomcom> i'll link you to it in a second
[17:29:38] --> fizzle has joined #gemrb
[17:34:28] --> brada has joined #gemrb
[17:43:47] <psch> brada: finger0 sometimes is null
[17:43:53] <brada> psch: that doesnt tell me anything btw
[17:44:00] <brada> ok thats better
[17:44:12] <psch> yeah i said i wanted to look around a bit
[17:44:41] <thomcom> @lynxlynxlynx http://www.gemrb.org/wiki/doku.php?id=developers:bugs:iwd_playthrough_bugs
[17:44:45] <brada> well i can not replicate of course
[17:44:46] <psch> didnt expect you to magically know what break when i just say "it segfaults"
[17:44:52] <Seniorita> developers:bugs:iwd_playthrough_bugs [GemRB wiki]
[17:44:52] <thomcom> "Saved games cause assertion in original game (load screen)"
[17:45:27] <lynxlynxlynx> ah
[17:45:50] <psch> http://nopaste.info/62722d78c9.html this is what i could think of to look at
[17:45:51] <Seniorita> Nopaste - powered by project-mindstorm IT Services
[17:46:38] <thomcom> not sure when it was fixed or how
[17:46:57] <thomcom> last android build is like 0.7.0.2 and still causes crash in the original game
[17:47:53] <thomcom> traveler tried to load my game into 0.7.2 on his desktop and it crashed
[17:48:10] <thomcom> so when android is up to 0.7.2, my save game will still be corrupt. :( hehe
[17:48:45] <brada> psch: did you remember to update your includes?
[17:50:21] <psch> the makefile include for gemrb?
[17:50:39] <psch> i reran my prep_env script, which clones and symlinks
[17:50:49] <brada> aslo it would help if you printed types in hex
[17:50:54] <psch> the gemrb makefile points to the symlink directory
[17:50:59] <brada> ah you do
[17:51:03] <brada> both it seems
[17:51:21] <psch> yeah i keep forgetting about /x, actually had to look it up again
[17:56:01] <brada> i dont understand SDL_GetTouchFinger is doing exactly what i was doing before
[17:56:09] <brada> gets the SDL_touch
[17:56:18] <brada> then gets fingers[0]
[17:56:47] <brada> but the functions i was using before are private now
[17:58:26] <brada> im totally at a loss for why finger[0] would be null
[17:59:20] <psch> so it might be something that i didn't discover before instead of something new?
[18:00:21] <brada> looks like a bug in sdl
[18:00:41] <brada> i mean it was working berfore
[18:01:00] <brada> i guess i cant say "bug"
[18:01:12] <brada> since there is no documentation....
[18:01:36] <brada> apparently touch->num_fingers is 0
[18:02:16] <brada> hmmm actually there is a comment in our event handling code about that
[18:04:52] <-- kida_ has left IRC (Ping timeout: 276 seconds)
[18:08:21] <psch> swiping the same way a few times in a row also gives me the segfault
[18:08:31] <psch> i fail to see any meaningful pattern though
[18:09:19] --> rocket_hamster has joined #gemrb
[18:09:48] <fizzle> lynxlynxlynx: does bg2 follow every panic instance with a non-panic action, then?
[18:11:09] <lynxlynxlynx> my bg2 dump is a bit corrupted, but i don't know if it would affect this
[18:11:21] <brada> psch: no need
[18:11:23] <lynxlynxlynx> judging from the grep, there's also a case where only the panic variants are used
[18:11:25] <brada> i already fixed it
[18:11:54] <brada> im currently trying to figure out why we have a minor bug with selection rects now
[18:12:00] <psch> oh, okay
[18:12:10] <lynxlynxlynx> and some where only the nonpanic one is used exclusively
[18:12:39] <lynxlynxlynx> i've excluded the leavearealuaentry pair, so the results could be skewed further
[18:14:07] <fizzle> yeah, there are lots of uses with just the non-panic version
[18:14:10] <fizzle> those are fine
[18:14:31] <fizzle> using panic exclusively would be a problem, however
[18:16:27] <brada> apparently i cant tell the diff between < and <= today :/
[18:22:51] <-- rocket_hamster has left IRC (Quit: bye!)
[18:24:19] <brada> the problem with selection rects appears to be the call to MoveMouse is processFirstTouch
[18:24:42] --> rocket_hamster has joined #gemrb
[18:24:59] <lynxlynxlynx> it's actually in a dialogue
[18:25:02] <lynxlynxlynx> dlg/bg2/priilmu.d
[18:25:34] <lynxlynxlynx> could also be unused
[18:26:10] <lynxlynxlynx> no, that's a dudd
[18:26:54] <lynxlynxlynx> i guess you're safe then
[18:27:08] <fizzle> yay!
[18:27:13] <fizzle> thanks for checking
[18:27:57] <brada> well seems to me the sdl2 version of MoveMouse should be converting from render to window coordinates
[18:30:12] <brada> btw this seems to affect me: http://stackoverflow.com/questions/6325119/gldrawarrays-segfaults-when-drawing-too-many-elements
[18:30:16] <Seniorita> android - glDrawArrays() segfaults when drawing too many elements - Stack Overflow
[18:30:17] <brada> dont know about android
[18:30:40] <brada> probably should get working on that proper SDL2 driver :/
[18:33:09] <brada> i havent heard of anybody complaining about random crashes on ios so maybe its a mac/simulator thing
[18:47:27] <brada> i see whay htat bug happened tho
[18:47:31] <brada> the drag rect i mean
[18:47:44] <brada> its because now we allow sdl mouse events as long as they arent touches
[18:47:55] <brada> and so warping the mouse causes problems
[18:48:04] <brada> easy fix im pushing for it tho
[19:38:36] <psch> huh
[19:38:50] <psch> bluetooth reconnect event from my keyboard freezes gemrb
[19:38:58] <psch> and the homebar too
[19:43:18] <psch> guess i'll turn it off
[19:50:55] <psch> dragging two fingers scrolls the viewport and opens a selection box
[19:51:04] <psch> is that the right behavior?
[20:12:20] <brada> i dont know what a selection box is
[20:12:31] <brada> oh the green box?
[20:12:48] <brada> what are your touch settings?
[20:14:10] <psch> the NumFing* ones in GemRB.cfg?
[20:14:27] <psch> those are commented in my .cfg
[20:14:30] <brada> oh
[20:14:34] <brada> thats fine
[20:14:42] <brada> anyway no that is not supposed to happen obviously
[20:14:56] <brada> and it didnt used to
[20:15:02] <psch> the selection box isn't right, is it
[20:15:40] <brada> unfortunately we cant bisect it either
[20:17:33] <brada> i can only assume its the numFingers
[20:17:38] <brada> which is unfortunate
[20:17:53] <brada> it used to work fine and now it doesnt
[20:18:35] <brada> its almost as if you get 2 finger motion events
[20:18:38] <brada> one that is correct
[20:18:42] <brada> then one that has 0 fingers
[20:19:37] <brada> maye if thats the case i should jsut have a separate check for numFingers == 0 that just bails out
[20:25:03] <brada> we need to be able to get the finger like we did before :(
[20:27:43] <psch> i dont really get this tbh
[20:27:59] <psch> i occasionally see events with numFingers == 0 in gdb
[20:28:02] <psch> but not consistently
[20:28:53] <psch> but the selection box is probably not related to that
[20:29:28] <brada> sure it is
[20:29:34] <brada> why do you think it isnt?
[20:29:54] <psch> i don't know, im not thinking well right now
[20:30:06] <psch> it must be related, you're right, cause the if is checking for <= 1
[20:30:28] <brada> which worked previously
[20:30:35] <brada> we will have to do what fuzzie said before
[20:30:44] <brada> and keep track of numfingers ourselves
[20:33:17] <-- rocket_hamster has left IRC (Quit: bye!)
[20:44:29] <brada> ha ha ha
[20:44:48] <brada> stupid movie events being handled separately
[20:44:59] <brada> no wonder the first event i get is a finger up!
[21:06:38] <brada> i think i see another problem too
[21:07:36] <brada> psch: did we decide fingerid on android represents a finger?
[21:08:18] <brada> i mean if you touch with a finger, release, then touch again is the fingerid the same?
[21:08:25] <psch> let me check that
[21:10:49] <psch> event.tfinger.fingerId is always 0 for the first finger, and 1 for the second finger
[21:11:25] <psch> where "always" == 5 tests right now
[21:11:47] <psch> i remember something about fingerIds being really high ints, the likes of 346409 or so, but i don't remember if you told me about that or it happened to me
[21:14:10] <brada> thats what it is on ios
[21:14:19] <brada> and was the basis for some of this logic :/
[21:14:36] <brada> so i have to change the keyboard show/hide gesture logic
[21:45:29] <brada> anyway i should be pushing a fix for the otehr problem momentarily
[21:48:32] <brada> just fyi tracking fingers is non-trivial
[21:48:43] <brada> sometimes you get a single SDL_FINGERDOWN event for multiple fingers
[21:48:59] <brada> i can go from 0 fingers to 2 but only get one SDL_FINGERDOWN
[21:49:29] <brada> im going to assume the same for finger up
[21:50:57] <brada> i have more testing to do before i push this
[21:52:09] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[22:05:34] <brada> psch: that should be fixed now
[22:05:41] <brada> still need to fix keyboard events
[22:05:43] <brada> later tho
[22:05:44] <-- brada has left IRC (Quit: brada)
[22:36:34] --- ermo is now known as ermo^
[22:46:05] <-- fizzle has left #gemrb
[23:05:27] <-- thomcom has left IRC (Ping timeout: 245 seconds)
[23:18:46] <-- Cable_ has left IRC (Ping timeout: 245 seconds)
[23:22:40] --> brada has joined #gemrb
[23:32:18] --> Cable_ has joined #gemrb
[23:36:34] --> gemrb has joined #gemrb
[23:36:36] <gemrb> hello
[23:36:41] <gemrb> ...
[23:36:42] <gemrb> sorry
[23:36:45] <-- gemrb has left #gemrb
[23:37:12] --> traveler has joined #gemrb
[23:37:18] <traveler> hello
[23:37:25] <traveler> how can i avoid
[23:37:31] <traveler> being stunned by spores in iwd?
[23:38:11] <traveler> it's practically impossible to enter dorn's deep
[23:39:32] <traveler> because of respawning myconids if reloaded
[23:44:03] <traveler> http://www.dandwiki.com/wiki/Myconid_(4e_Race) playable myconids? strange idea
[23:44:04] <Seniorita> Myconid (4e Race) - D&D Wiki
[23:45:01] <-- brada has left IRC (Quit: brada)
[23:50:35] <traveler> eh?
[23:50:37] <traveler> http://www.dandwiki.com/wiki/Teddy_Bear_(4e_Creature)
[23:50:38] <Seniorita> Teddy Bear (4e Creature) - D&D Wiki