#gemrb@irc.freenode.net logs for 11 Sep 2012 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[00:36:15] <-- edheldil_ has left IRC (Ping timeout: 276 seconds)
[00:44:12] <-- CIA-39 has left IRC (Ping timeout: 272 seconds)
[00:58:00] --> CIA-44 has joined #gemrb
[01:08:51] --> edheldil_ has joined #gemrb
[01:38:26] --> finn has joined #gemrb
[01:38:50] --- finn is now known as Guest41434
[01:43:49] <-- Guest41434 has left IRC (Quit: Page closed)
[01:44:48] <-- edheldil_ has left IRC (Ping timeout: 240 seconds)
[02:51:28] <-- duckpunch has left IRC (Quit: leaving)
[06:38:19] --> edheldil_ has joined #gemrb
[07:01:51] --> lynxlynxlynx has joined #gemrb
[07:01:52] <-- lynxlynxlynx has left IRC (Changing host)
[07:01:52] --> lynxlynxlynx has joined #gemrb
[07:01:52] --- ChanServ gives channel operator status to lynxlynxlynx
[07:08:00] <-- edheldil_ has left IRC (Ping timeout: 272 seconds)
[07:09:01] <lynxlynxlynx> wose_: we use the standard cmake findpython module
[07:09:24] <lynxlynxlynx> you must have a pretty odd install for it to get confused
[07:12:19] <fuzzie> they messed with it in 2.8.8 I think
[07:13:26] <fuzzie> you can do FIND_PACKAGE(PythonLibs 2.3) nowadays I think
[07:14:18] <fuzzie> the intention was still that it shouldn't get 3.x by default but I don't really trust cmake upstream to do it right
[07:14:53] <fuzzie> and indeed it looks like they didn't
[07:18:32] <wose_> i'm on debian sid, i can't check the cmake version, but the cmake FindPythonLibs thingy did the wrong thing
[07:20:25] <wose_> i also tried setting the 2 env variables to force a python include path and lib as suggested in the header of the FindPythonLibs cmake plugin, that didn't worked either
[07:21:26] <wose_> so it is a cmake problem, i got it working with hard coded path, just wanted to let you know that there might be problems compiling on some systems
[07:22:14] --> edheldil_ has joined #gemrb
[07:23:50] <fuzzie> hm, sid actually has 2.8.9 which supposeldy has the right variables in the header there :P thanks for the heads-up in any case
[07:40:29] <lynxlynxlynx> that's the latest
[07:40:50] <lynxlynxlynx> we could ifdef the version check for people with older cmake
[07:42:16] <lynxlynxlynx> the vars didn't work, since they're from cmake, not normal shell ones
[07:55:59] <wose_> oic
[08:07:30] <lynxlynxlynx> where are both pythons installed?
[08:14:00] <wose_> i'm at work right now, i can check this in about 8h
[09:01:21] <lynxlynxlynx> ok
[10:12:44] --> rocket_hamster has joined #gemrb
[10:17:50] --> Husar1 has joined #gemrb
[10:33:48] <-- Husar1 has left IRC (Quit: Husar1)
[10:34:30] --> Husar1 has joined #gemrb
[10:41:32] <CIA-44> GemRB: 03bradallred * r8c81cb4a229f 10gemrb/gemrb/GUIScripts/ (GUIOPT.py bg2/GUIOPT.py):
[10:41:32] <CIA-44> GemRB: Begin merging GUIOPT by using BG2/GUIOPT as a base
[10:41:32] <CIA-44> GemRB: Signed-off-by: Jaka Kranjc <lynxlupodian@users.sourceforge.net>
[10:41:33] <CIA-44> GemRB: 03bradallred * r6517068ab0cf 10gemrb/gemrb/GUIScripts/ (bg1/GUIOPT.py iwd2/GUIOPT.py):
[10:41:33] <CIA-44> GemRB: GUIOPT BG1/IWD2: remove unused variables.
[10:41:33] <CIA-44> GemRB: Signed-off-by: Jaka Kranjc <lynxlupodian@users.sourceforge.net>
[11:05:24] <-- rocket_hamster has left IRC (Ping timeout: 248 seconds)
[11:12:14] <CIA-44> GemRB: 03lynxlupodian * rcdaf6ce553bb 10gemrb/gemrb/GUIScripts/ (GUIOPT.py GUIOPTControls.py):
[11:12:14] <CIA-44> GemRB: Move Option Control code to its own module
[11:12:14] <CIA-44> GemRB: based on two brad's patches
[11:14:22] --> rocket_hamster has joined #gemrb
[11:15:52] <-- rocket_hamster has left IRC (Client Quit)
[12:19:38] --> rocket_hamster has joined #gemrb
[13:36:52] --> duckpunch has joined #gemrb
[16:07:57] <-- Husar1 has left IRC (Quit: Husar1)
[16:24:39] <-- rocket_hamster has left IRC (Read error: Connection reset by peer)
[16:34:49] <CIA-44> GemRB: 03lynxlupodian * rf42b15bc5281 10gemrb/gemrb/GUIScripts/GUIOPT.py: GUIOPT: when closing the character sounds window reopen the audio one
[16:34:50] <CIA-44> GemRB: 03lynxlupodian * r30a44f48530e 10gemrb/gemrb/GUIScripts/ (GUIOPT.py bg1/GUIOPT.py):
[16:34:50] <CIA-44> GemRB: Merge BG1 GUIOPT into master
[16:34:50] <CIA-44> GemRB: based on brad's patch
[16:36:44] <-- edheldil_ has left IRC (Ping timeout: 272 seconds)
[16:39:00] <-- nutron has left IRC (Remote host closed the connection)
[16:44:19] --> Yoshimo has joined #gemrb
[17:19:22] <wose_> lynxlynxlynx: https://gist.github.com/3699907 i hope this helps
[17:21:06] <lynxlynxlynx> v2 is also directly under /usr/lib/ ?
[17:22:16] <wose_> yes
[17:23:05] <wose_> semms like it's using the highest version available
[17:26:18] <CIA-44> GemRB: 03lynxlupodian * ra28b943c4a12 10gemrb/gemrb/GUIScripts/GUIOPT.py:
[17:26:18] <CIA-44> GemRB: GUIOPT: make the quit win's exit button the default
[17:26:18] <CIA-44> GemRB: also makes exiting quicker
[17:26:20] <CIA-44> GemRB: 03bradallred * rd0ba314b250d 10gemrb/gemrb/GUIScripts/ (GUIOPT.py iwd/GUIOPT.py):
[17:26:20] <CIA-44> GemRB: Merge IWD1 GUIOPT into master
[17:26:20] <CIA-44> GemRB: cleaned up version
[17:26:20] <CIA-44> GemRB: Signed-off-by: Jaka Kranjc <lynxlupodian@users.sourceforge.net>
[17:26:56] <lynxlynxlynx> that's probably the reason then, they're installed side-by-side without parent dir suffixing
[17:29:20] <wose_> is somewhere in the gemrb cmake files specified which version to use? even if they were in a seperate path what tells cmake to use v2 and not v3?
[17:59:26] <lynxlynxlynx> it's all internal to cmake
[17:59:36] <lynxlynxlynx> so yes, it's their bug
[18:01:55] <fuzzie> they decided that it should prefer 2.x by default if there was no version specified
[18:02:02] <fuzzie> but there's no sign of anyone actually having implemented that in what was committed
[18:02:49] <fuzzie> but I assume FIND_PACKAGE(PythonLibs 2.3) would be fine for us, so not a disaster I hope
[18:07:06] <lynxlynxlynx> if that circumvents the module, i guess so
[18:08:17] <lynxlynxlynx> wose_: try chaning INCLUDE(FindPythonLibs) to what fuzzie suggested
[18:08:25] <lynxlynxlynx> without your hardcoded paths of course
[18:14:58] --> rocket_hamster has joined #gemrb
[18:20:41] <wose_> aye, this works. it uses the installed 2.7
[18:22:33] <lynxlynxlynx> cool
[18:30:12] --> edheldil_ has joined #gemrb
[18:54:37] <-- Yoshimo has left IRC (Quit: Yoshimo)
[18:57:31] <CIA-44> GemRB: 03lynxlupodian * r650de7b50435 10gemrb/CMakeLists.txt:
[18:57:31] <CIA-44> GemRB: cmake: use find_package's version check if available
[18:57:31] <CIA-44> GemRB: fixes debian sid python2 detection
[18:57:34] <-- lynxlynxlynx has left IRC (*.net *.split)
[18:57:34] <-- DrMcCoy has left IRC (*.net *.split)
[18:58:47] --> lynxlynxlynx has joined #gemrb
[18:58:47] --> DrMcCoy has joined #gemrb
[19:00:21] <-- DrMcCoy has left IRC (Remote host closed the connection)
[19:00:38] --> DrMcCoy has joined #gemrb
[19:05:36] <wose_> lynxlynxlynx: i think you need to change gemrb/plugins/GUIScript/CMakeLists.txt as well it also INCLUDEs FindPythonLibs i don't know if it is even necessary to 'find' them again
[19:06:06] --> Yoshimo has joined #gemrb
[19:06:37] <lynxlynxlynx> looks extraneus, yes
[19:07:11] <lynxlynxlynx> it was there from the start
[19:18:01] <lynxlynxlynx> yep, builds for me without it
[19:18:33] <lynxlynxlynx> tomprince: is there any chance of getting gembot back?
[19:19:25] <CIA-44> GemRB: 03bradallred * rf2c7b7cefc29 10gemrb/gemrb/GUIScripts/ (GUIOPT.py iwd2/GUIOPT.py):
[19:19:25] <CIA-44> GemRB: IWD2 GUIOPT: use shared GUIOPTControls module
[19:19:25] <CIA-44> GemRB: Signed-off-by: Jaka Kranjc <lynxlupodian@users.sourceforge.net>
[19:19:26] <CIA-44> GemRB: 03lynxlupodian * r0bffbba88213 10gemrb/gemrb/plugins/GUIScript/CMakeLists.txt:
[19:19:26] <CIA-44> GemRB: cmake: removed extraneus python module include
[19:19:26] <CIA-44> GemRB: thanks wose
[19:19:35] --> wrotek has joined #gemrb
[19:19:37] <lynxlynxlynx> bah!
[19:19:38] <wose_> just checked it too, works fine without the include
[19:19:49] <lynxlynxlynx> forgot i was in the middle of the review
[19:20:43] <CIA-44> GemRB: 03lynxlupodian * r58f3a9b7d723 10gemrb/gemrb/GUIScripts/ (GUIOPT.py iwd2/GUIOPT.py):
[19:20:43] <CIA-44> GemRB: Revert "IWD2 GUIOPT: use shared GUIOPTControls module"
[19:20:44] <CIA-44> GemRB: pushed too early, needs fixing
[19:20:44] <CIA-44> GemRB: This reverts commit f2c7b7cefc2986db53389229835b2104595e51a2.
[20:09:03] <lynxlynxlynx> mves crash again, interesting
[20:09:46] <lynxlynxlynx> free((char*)ptr - ((char*)ptr)[-1]);
[20:09:48] <lynxlynxlynx> -1??
[20:13:14] <rocket_hamster> why would anyone free memory like that :D
[20:17:08] <fuzzie> seems to be ffmpeg's attempt to force malloc to be aligned?
[20:17:30] <fuzzie> just in case your libc is insane or something
[20:27:44] <lynxlynxlynx> (gdb) p ptr
[20:27:44] <lynxlynxlynx> $1 = (void *) 0x40000000000335f
[20:27:49] <lynxlynxlynx> pretty huge address
[20:29:05] <fuzzie> isn't that non-canonical? (assuming you're on amd64)
[20:29:19] <fuzzie> oh, no, I guess 0x0400 at the start
[20:29:41] <fuzzie> oh, that's still non-canonical I suppose :P so corrupt?
[20:37:49] <lynxlynxlynx> i have no idea what non-canonical would mean in this case
[20:38:16] <lynxlynxlynx> it starts in GemRB::BIKPlayer::EndAudio
[20:38:18] <-- Yoshimo has left IRC (Quit: Yoshimo)
[20:38:26] <lynxlynxlynx> i just skip the movie and it crashes reliably
[20:40:35] <fuzzie> when trying to free s_bands?
[20:41:38] <lynxlynxlynx> yes
[20:41:54] <fuzzie> that is quite strange
[20:42:01] <fuzzie> called from Play()?
[20:42:14] <lynxlynxlynx> yep
[20:42:38] <fuzzie> well, s_bands should be set to NULL in the constructor I guess
[20:43:33] <fuzzie> since it seems that it wouldn't get initialised if there's no audio
[20:43:40] <fuzzie> maybe that is your problem, in fct?
[20:44:48] <lynxlynxlynx> will check
[20:45:27] <lynxlynxlynx> BISLOGO.mve - i don't remember if it does have sound or not
[20:45:38] <lynxlynxlynx> isn't it the one with the lightning?
[20:48:40] <lynxlynxlynx> the null didn't help, though a more random value now appears: 0x35ca386ef0643838
[20:50:37] <fuzzie> that certainly looks beautifully corrupt
[20:50:47] <fuzzie> but I meant sound globally, i.e. the mixer returns CanPlay
[20:52:17] <fuzzie> seems like it should crash in that case though
[20:52:57] <fuzzie> due to DecodeAudioFrame -> DecodeBlock -> s_bands
[20:54:53] <fuzzie> rsyncing iwd2 across but it'll take a while
[20:56:23] <lynxlynxlynx> when you can, just go to the options and movie window, it's the quickest way to trigger it
[20:56:36] <lynxlynxlynx> i can try to grind it tommorow
[20:56:47] <lynxlynxlynx> still in the process of merging guiopts
[21:09:56] <fuzzie> hmmmm
[21:11:09] <fuzzie> so who broke PlayMovie?
[21:11:40] <fuzzie> long long time ago
[21:14:39] <fuzzie> so yes, I can reproduce by forcing CanPlay to false
[21:15:50] <fuzzie> and it breaks in DecodeAudioFrame because s_block_size is uninitialised because the audio didn't get inited
[21:19:04] <lynxlynxlynx> likely the same crash all the ios testers reported
[21:19:50] <fuzzie> sf is sloooow
[21:20:27] <CIA-44> GemRB: 03fuzzie * rc21b77c17d95 10gemrb/gemrb/plugins/BIKPlayer/BIKPlayer.cpp:
[21:20:27] <CIA-44> GemRB: BIKPlayer: Don't crash if audio can't play.
[21:20:27] <CIA-44> GemRB: Thanks to lynx for spotting this.
[21:21:41] <lynxlynxlynx> cool, let's see
[21:22:57] <lynxlynxlynx> still triggers here and since i reverted the null init, it has the same 0x40000000000335f value
[21:23:04] <lynxlynxlynx> oh wait
[21:23:08] <lynxlynxlynx> nevermind
[21:23:25] <lynxlynxlynx> i just did a make, forgot to get your change >>
[21:26:23] <lynxlynxlynx> all is fine
[21:32:37] <CIA-44> GemRB: 03lynxlupodian * rdab510e188fe 10gemrb/gemrb/GUIScripts/iwd2/GUIOPT.py: iwd2: removed the now extraneus guiopt
[21:32:39] <CIA-44> GemRB: 03lynxlupodian * r90df739be380 10gemrb/gemrb/GUIScripts/ (GUIOPT.py iwd2/GUIOPT.py):
[21:32:39] <CIA-44> GemRB: IWD2 GUIOPT: use shared GUIOPTControls module
[21:32:39] <CIA-44> GemRB: based on brad's initial patch
[21:32:39] <CIA-44> GemRB: noticed some control ids are wrong in bg2 and iwd2
[21:32:39] <CIA-44> GemRB: 03lynxlupodian * reb4b0ca9296f 10gemrb/gemrb/GUIScripts/ (GUIOPT.py iwd2/GUIOPT.py): guiopt: functionally synced main and iwd2
[21:38:27] <fuzzie> hehe
[21:45:59] <-- edheldil has left IRC (Remote host closed the connection)
[22:14:51] <CIA-44> GemRB: 03lynxlupodian * r36541be786d8 10gemrb/gemrb/GUIScripts/GUIOPT.py: guiopt: added gamma bounds checking, since iwd2 has a larger range
[22:14:51] <CIA-44> GemRB: 03lynxlupodian * r4f761c0dd5c9 10gemrb/gemrb/GUIScripts/GUIOPT.py: guiopt: fixed a few bad bg2 and iwd2 control ids
[22:35:42] --> nutron has joined #gemrb
[22:50:04] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[23:06:35] <-- rocket_hamster has left IRC (Read error: Connection reset by peer)
[23:10:31] --> rocket_hamster has joined #gemrb
[23:26:43] <-- rocket_hamster has left IRC ()
[23:28:15] --> rocket_hamster has joined #gemrb