#jogamp @ irc.freenode.net - 20130430 05:06:09 (UTC)
20130430 05:06:09 -CatOut- Previous @ http://jogamp.org/log/irc/jogamp_20130429050556.html
20130430 05:06:09 -CatOut- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130430050609.html
20130430 05:23:16 * [Mike] (~Mike]@anon) has joined #jogamp
20130430 06:34:32 * [Mike] (~Mike]@anon) Quit ()
20130430 09:42:21 * gouessej (546051e4@anon) has joined #jogamp
20130430 09:42:26 <gouessej> hi
20130430 09:42:38 <xranby> gouessej: hello
20130430 09:42:48 <sgothel> hi Julien, the surface change is due to Windows getHDC ... thats normal
20130430 09:43:05 <sgothel> hi Xerxes, hi All
20130430 09:43:12 <gouessej> yes I have looked at the source code
20130430 09:43:29 <gouessej> but I didn't imagine it happens so often
20130430 09:44:22 <gouessej> it's like lockSurfaceImpl never returns SUCCESS
20130430 09:44:30 <gouessej> that's why I was suspicious
20130430 09:44:53 <sgothel> surface handle change has no impact on that one, i.e. lock result
20130430 09:45:17 <gouessej> ok
20130430 09:45:46 <sgothel> wait .. you are right .. hmm
20130430 09:46:34 <sgothel> it does LOCK_SURFACE_CHANGED .. usually that is ok .. let me review ..
20130430 09:48:16 <gouessej> I'm sorry, I'm a bit suspicious because I have some slowdowns on this machine even though I don't move anything
20130430 09:48:35 <gouessej> I will have to use a profiler
20130430 09:49:09 <sgothel> GLDRawableImpl .. public final int lockSurface() - now acts on SURFACE_CHANGED
20130430 09:49:50 <gouessej> Should the surface change even though I don't resize it?
20130430 09:50:09 <gouessej> the X11 implementation always return SUCCESS
20130430 09:50:17 <gouessej> 'returns"
20130430 09:50:47 <sgothel> SURFACE_CHANGED was impl. if the surfaceHandle changes .. and is impl. for EGL only (X11 has one too, but it's disabled)
20130430 09:51:04 <gouessej> ok
20130430 09:51:15 <gouessej> I have reproduced the bug 723
20130430 09:51:42 <gouessej> the key code is 0x0 when I press an arrow key
20130430 09:51:51 <gouessej> under Windows 7
20130430 09:51:54 <sgothel> great, I will read it later - I closed early yesterday - have to complete monitor stuff, still - it's a bigger thing :)
20130430 09:52:06 <sgothel> pls dump the KeyEvent debug lines .. thx
20130430 09:52:09 <sgothel> in the bug report
20130430 09:52:12 <gouessej> ok I can investigate a bit
20130430 09:52:52 <gouessej> I already did that, I used the newt 'all' debug flag
20130430 09:52:54 <sgothel> so we may return LOCK_SUCCESS in Window's NEWT, since there is no semantical impact .. hmm
20130430 09:53:41 <gouessej> I can look at the implementation of key codes under GNU Linux
20130430 09:53:50 <gouessej> I'm not sure I will find the root cause
20130430 09:54:01 <sgothel> the bug is on Windows right ?
20130430 09:54:13 <sgothel> on Linux as well ?
20130430 09:54:45 <gouessej> under Linux too but a bit different
20130430 09:55:14 <sgothel> however, the LOCK_* result shall not really slow down .. 'just' calling NOP methods .. hmm, will earmark this
20130430 09:55:18 <gouessej> It concerns some NUMPAD keys under Linux and arrow keys under Windows
20130430 09:55:44 <gouessej> I will have to use a profiler. Maybe it comes from something completely unrelated
20130430 09:56:12 <gouessej> Do you determine the key codes in the native source code?
20130430 09:56:34 <sgothel> yes
20130430 09:57:33 <gouessej> XCBEvent.c?
20130430 09:58:05 <sgothel> nope .. thats some ole dummy code - X11Display
20130430 09:58:22 <sgothel> it's quite complicated .. well :|
20130430 09:58:32 <gouessej> ok I'm on it
20130430 09:58:38 <gouessej> thanks
20130430 09:58:42 <sgothel> keyCode US keyboard layout, keySym current keyboard layout
20130430 09:59:31 <sgothel> keySym is being used to determine whether it's an action-, printable- etc key
20130430 09:59:44 <sgothel> that is done in the Java code
20130430 10:00:52 <gouessej> Maybe the bug comes from X11KeySym2NewtVKey
20130430 10:01:50 <sgothel> keep in mind, it works on US and DE layout, incl. that mapping function
20130430 10:02:23 <gouessej> that's why it would be fine to have a unit test
20130430 10:02:35 <gouessej> and to be able to change the layout
20130430 10:02:43 <sgothel> switching the layout .. yeah :)
20130430 10:02:48 <gouessej> to check whether I'm going to break everything
20130430 10:03:02 <sgothel> you ?
20130430 10:04:07 <gouessej> I can try to fix it at least under Linux
20130430 10:04:09 <xranby> sgothel: gouessej: while looking at input: have you checked if keyrepeat work under macosx ? found this report on stackexchange http://stackoverflow.com/questions/16160657/java-input-on-os-x-no-key-repeat-but-with-glut-yes
20130430 10:04:46 <sgothel> thats related to NEWT ?
20130430 10:05:13 <sgothel> they shall be all redirected to our forum and bugzilla!
20130430 10:05:29 <gouessej> xranby I rarely use OSX but this feature works in NEWT, the boolean flag is useful
20130430 10:05:47 <sgothel> and .. the unit tests do work!
20130430 10:05:58 <sgothel> they include auto-repeat
20130430 10:08:09 <gouessej> Ok I have to eat. I come back soon
20130430 10:08:15 * gouessej (546051e4@anon) Quit (Quit: Page closed)
20130430 10:09:17 <xranby> sgothel: gouessej: ok good, i will take a look at the unittests and the boolean flag
20130430 10:09:45 <sgothel> so is it even NEWT related ?
20130430 10:10:04 <xranby> sgothel: looks like glcanvas related
20130430 10:10:23 <sgothel> so it is _not_ -> AWT -> not our issue at all
20130430 10:11:08 <xranby> the solution is then to recommend the user to use NEWT
20130430 10:11:22 <xranby> to stay away from AWT issues
20130430 10:11:28 <sgothel> if you bother reading stackoverflow .. IMHO simply direct them to our forum/bugzilla then we can sort things out. Today/Yesterday we had a new bugreport from them, after I asked them to follow our rules. That is good.
20130430 10:12:14 <sgothel> otherwise - we cannot be bothered w/ all the noise .. IMHO not efficient, enough that we have false bugreports we have to validate etc :)
20130430 10:12:50 <sgothel> I know, I once started to look at other places as well .. but .. you know, it takes lots of time
20130430 10:15:24 <xranby> i ran across this one while watching the jogl twitter feed
20130430 10:15:43 <xranby> so it was indirect linked from the jogamp.org homepage
20130430 10:15:57 <sgothel> :)
20130430 10:17:03 <sgothel> really happy that folks try to deliver test cases and proper bug reports within our channels - kudos to all of 'em.
20130430 10:17:44 <sgothel> note: pls show them the Wiki.FAQ.Bugreport and let the produce test_dbg.log file via etc/test_dbg.sh .. before going any further!
20130430 10:18:07 <xranby> src/test/com/jogamp/opengl/test/junit/newt/event/TestNewtKeyEventAutoRepeatAWT.java
20130430 10:18:07 <sgothel> our initialization is pretty intensive and will show most possible bugs
20130430 10:18:29 <sgothel> NEWT+AWT related, yes
20130430 10:18:34 <sgothel> NewtCanvasAWT
20130430 10:19:00 <sgothel> ofc .. all key/mouse tests use AWT Robot
20130430 10:36:35 * gouessej (546051e4@anon) has joined #jogamp
20130430 10:37:37 <gouessej> Sven I agree with your suggestion and it works very well with NEWT
20130430 10:37:46 <gouessej> AWT is not our problem
20130430 11:09:11 <xranby> ok i posted this answer http://stackoverflow.com/questions/16160657/java-input-on-os-x-no-key-repeat-but-with-glut-yes/16298356#16298356
20130430 11:12:03 <gouessej> xranby nice answer
20130430 11:14:05 <sgothel> thumbs up :)
20130430 11:18:11 <gouessej> Sven, how can I enable the debug in the native source code, in X11Display?
20130430 11:18:32 <sgothel> edit the native header files ..
20130430 11:18:40 <sgothel> some VERBOSE macros ..
20130430 11:18:50 <gouessej> ok :s I will have to rebuild JOGL
20130430 11:20:08 <gouessej> I should get 0x82 as a sym code for the numpad key 2, shouldn't I?
20130430 11:21:31 <gouessej> ok I have to leave. I will try to reproduce this bug again tonight
20130430 11:21:32 <sgothel> 0x82 is the ASCII / UTF value ? maybe .. possible numpad has own values
20130430 11:21:35 <gouessej> Bye
20130430 11:21:40 <sgothel> bye
20130430 11:22:00 * gouessej (546051e4@anon) Quit (Quit: Page closed)
20130430 16:26:32 * [Mike] (~Mike]@anon) has joined #jogamp
20130430 18:14:39 * ___m___ (~Mike]@anon) has joined #jogamp
20130430 18:20:58 * [Mike] (~Mike]@anon) Quit (Ping timeout: 268 seconds)
20130430 18:40:37 * gouessej (5279c7b1@anon) has joined #jogamp
20130430 18:41:04 <gouessej> I'm building JOGL with native debug messages right now
20130430 18:45:48 <gouessej> Sven, which test would you like me to run to reproduce the bug 723?
20130430 19:22:01 * [Mike] (~Mike]@anon) has joined #jogamp
20130430 19:22:01 * ___m___ (~Mike]@anon) Quit (Read error: Connection reset by peer)
20130430 20:08:24 <gouessej> I've just made a pull request
20130430 20:08:42 <gouessej> https://github.com/sgothel/jogl/pull/64
20130430 20:45:32 * gouessej (5279c7b1@anon) Quit (Quit: Page closed)
20130430 21:39:10 * void256 (~void@anon) has joined #jogamp
20130430 23:06:20 * void256 (~void@anon) Quit (Quit: ChatZilla 0.9.90 [Firefox 20.0/20130326150557])
20130430 23:51:26 * xranby_ac100 (~xranby@anon) has joined #jogamp
20130501 00:37:54 * xranby_ac100 (~xranby@anon) Quit (Ping timeout: 256 seconds)
20130501 01:57:32 * [Mike] (~Mike]@anon) Quit ()
20130501 02:46:31 * [Mike] (~Mike]@anon) has joined #jogamp
20130501 02:46:33 * ___m___ (~Mike]@anon) has joined #jogamp
20130501 03:08:21 * ___m___ (~Mike]@anon) Quit ()
20130501 05:06:23 -CatOut- Continue @ http://jogamp.org/log/irc/jogamp_20130501050623.html