JogAmp IRC Chat Logs

#jogamp @ irc.freenode.net - 20130705 05:05:23 (UTC)

20130705 05:05:23 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20130704050523.html
20130705 05:05:23 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130705050523.html
20130705 06:07:03 * [Mike] (~Mike]@anon) Quit ()
20130705 06:49:42 * Eclesia (~jsorel@anon) has joined #jogamp
20130705 06:56:41 * [Mike] (~Mike]@anon) has joined #jogamp
20130705 07:29:12 * ebrayet (~Adium@anon) has joined #jogamp
20130705 07:57:32 * [Mike] (~Mike]@anon) Quit ()
20130705 08:51:15 * odin_ (~Odin@anon) Quit (Ping timeout: 260 seconds)
20130705 09:01:12 * odin_ (~Odin@anon) has joined #jogamp
20130705 09:52:58 * xranby (~familjen@anon) has joined #jogamp
20130705 09:53:28 <xranby> can someone explain to me why chuck complains for a www/index.html change?
20130705 09:55:33 <xranby> something went wrong in a recent build. Build #209: https://jogamp.org/chuck/job/jogl-demos/./label=win7-x86_64-amd/209/
20130705 09:57:39 <xranby> ok the build error was unrelated to the last change.. https://jogamp.org/chuck/job/jogl-demos/label=win7-x86_64-amd/209/console
20130705 09:59:24 <xranby> redbook-javac examples need to be updated to use NIO Buffers
20130705 12:15:36 * ebrayet1 (~Adium@anon) has joined #jogamp
20130705 12:18:09 * ebrayet (~Adium@anon) Quit (Ping timeout: 264 seconds)
20130705 12:44:54 <rmk0-alt> ow
20130705 12:45:02 <rmk0-alt> netsplit... half of me's over there...
20130705 12:46:46 * rmk0 (~rmk0@anon) Quit (*.net *.split)
20130705 12:53:35 * odin_ (~Odin@anon) Quit (*.net *.split)
20130705 12:53:50 * jk4_ (~jk4@anon) Quit (*.net *.split)
20130705 13:05:20 * xranby (~familjen@anon) Quit (*.net *.split)
20130705 13:05:20 * Eclesia (~jsorel@anon) Quit (*.net *.split)
20130705 13:05:24 * olamedia (~olamedia@anon) Quit (*.net *.split)
20130705 13:05:38 * ebrayet1 (~Adium@anon) Quit (*.net *.split)
20130705 13:05:49 * sgothel (~sven@anon) Quit (*.net *.split)
20130705 13:21:07 * sgothel (~sven@anon) has joined #jogamp
20130705 13:21:07 * Eclesia (~jsorel@anon) has joined #jogamp
20130705 13:21:07 * xranby (~familjen@anon) has joined #jogamp
20130705 13:21:07 * olamedia (~olamedia@anon) has joined #jogamp
20130705 13:21:07 * ebrayet1 (~Adium@anon) has joined #jogamp
20130705 13:21:07 * jk4_ (~jk4@anon) has joined #jogamp
20130705 13:21:07 * Guest51232 (~rmk0@anon) has joined #jogamp
20130705 13:21:07 * Guest20444 (~Odin@anon) has joined #jogamp
20130705 13:22:09 <xranby> sgothel: ping, i have pushed a fix for the chuck jogl-demos issue.. that is glredbook1314/mvarray: glMultiDrawElements require direct NIO buffer
20130705 13:22:47 <xranby> https://github.com/xranby/jogl-demos/commits/master
20130705 13:23:22 -mquin- [Global Notice] As you've probably observed, we're experiencing a bit of network disruption. Please bear with us while we look into it.
20130705 13:23:25 <xranby> i also added a fix allowing me to build jogl-demos with an empty user.swt.jar property
20130705 13:32:49 <sgothel> Hi Xerxes
20130705 13:32:50 <Eclesia> hi
20130705 13:32:57 <sgothel> Hi all!
20130705 13:33:04 <Eclesia> qeustion on GlCapabilities
20130705 13:33:22 <Eclesia> GLCapabilities caps = GLG2DCanvas.getDefaultCapabalities();
20130705 13:33:22 <Eclesia> caps.setFBO(true);
20130705 13:33:22 <Eclesia> caps.setOnscreen(false);
20130705 13:33:31 <Eclesia> how do get back the fbo ?
20130705 13:33:36 <sgothel> @Xerxes: Q: I was thinking the MultiDraw commands shall need PointerBuffer to be NIO - what do you think ? Is that too much ?
20130705 13:33:56 <sgothel> @Eclesia ..
20130705 13:34:17 <sgothel> at init/display/.. ask drawable.getChosen...() whether it's FBO
20130705 13:34:42 <sgothel> then you may cast it to the GLFBODrawable / or similar, please looks at unit tests
20130705 13:34:43 * Guest20444 is now known as odin_
20130705 13:35:13 * odin_ is now known as Guest83277
20130705 13:35:15 * ebrayet (~Adium@anon) has joined #jogamp
20130705 13:36:14 <xranby> sgothel: i cant comment on performance unless we add a performance benchmark test to be run after each build
20130705 13:36:19 <xranby> that would allow us to track it
20130705 13:36:44 <xranby> imho using pinned nio buffers is a sane move to prevent the GC shuffle around data used by the GPU
20130705 13:36:52 * ebrayet1 (~Adium@anon) Quit (Ping timeout: 276 seconds)
20130705 13:37:00 <sgothel> more like a usage / reference of the buffer beyond the call .. you may remember the NIO issues
20130705 13:37:11 <sgothel> i.e. usage by GPU
20130705 13:37:25 <sgothel> yes -> GC .. etc
20130705 13:37:47 <sgothel> also have to re-read the backed array PointerBuffer impl. .. hmm
20130705 13:38:45 <sgothel> right .. PointerBuffer.referenceBuffer only works w/ NIO .. done
20130705 13:41:32 <xranby> Crases inside the GPU driver can be caused by A) the GC decided to move the memory and can be fixed using pinned NIO buffers and B) the memory became eligable for collection (harder to solve without a costly reference count)
20130705 13:42:26 <xranby> to my knowledge B also depends a bit on the JVM implementation
20130705 13:42:41 <sgothel> http://jogamp.org/git/?p=jogl.git;a=commit;h=b9a24308f3ebc6fae9ca79f6020970945936feab <- see comment / example
20130705 13:42:47 <sgothel> I was more like the above ..
20130705 13:43:12 <sgothel> i.e. provide a proper 'int **indices' 2-dim-array
20130705 13:43:15 <xranby> for example the avian JVM GC is one of the most agressive GC implementations i have seen to trigger issue B
20130705 13:43:19 <sgothel> so this needs PointerBuffer
20130705 13:43:22 <sgothel> w/ NIO
20130705 13:44:18 <sgothel> yes .. if primitive memory pointer is used _after_ call (release) by GL you cannot use it and must use NIO
20130705 13:46:57 <xranby> and if primitive memory pointer is used _after_ call(release) by GL someone must make sure there is still a reference held by non end user code
20130705 13:46:59 <sgothel> Harvey said all will go well w/ T-Shirts last night .. and they will send us an invoice. If the latter doesn't arrive till Monday, I will give them a call.
20130705 13:47:32 <xranby> to prevent the GC to free up the NIO memory
20130705 13:47:36 <sgothel> yes .. but w/ reference I actually mean the get-primitive-*-critical lock
20130705 13:47:46 <sgothel> otherwise it may float around in real memory
20130705 13:48:31 <sgothel> you fixed jogl-demos - thx
20130705 13:48:57 <sgothel> now I have to beat down all my regressions from code changing .. hope I am through soon :)
20130705 13:54:58 <Eclesia> Is there a utility class somewhere to convert an FBObject to Bufferedimage ?
20130705 13:56:41 <sgothel> There actually is ..
20130705 13:56:58 <sgothel> let me find the unit test .. hmm
20130705 14:01:48 <sgothel> AWTGLReadBufferUtil
20130705 14:02:01 <sgothel> using readPixels ..
20130705 14:02:47 <sgothel> TestTexture01AWT <- used in
20130705 14:02:50 <sgothel> etc
20130705 14:03:13 <sgothel> deprecated Screenshot uses it as well
20130705 14:04:22 <sgothel> AWTGLPixelBuffer used by GLJPanel in a fancy singleton way ..
20130705 14:05:14 <sgothel> GLPixelBuffer and GLPixelBufferProvider are new classes / interfaces allowing to hook into the pixel stream ..
20130705 14:14:36 <Eclesia> can't find this class
20130705 14:14:55 <Eclesia> AWTGLReadBufferUtil
20130705 14:20:48 <xranby> jogl/src/jogl/classes/com/jogamp/opengl/util/GLReadBufferUtil.java and jogl/src/jogl/classes/com/jogamp/opengl/util/awt/AWTGLReadBufferUtil.java
20130705 14:31:04 <sgothel> you should use eclipse or another IDE and the source attachments to all jogamp jars (if not using the source itself)
20130705 14:31:26 <sgothel> this helps .. i.e. you can find all calltraces .. references etc ..
20130705 14:35:43 <Eclesia> I'm using maven
20130705 14:35:57 <Eclesia> added a dependency to jogl-all
20130705 14:36:05 <Eclesia> but he does found this class
20130705 14:36:09 <Eclesia> does not*
20130705 14:36:26 <Eclesia> 2.0-rc11
20130705 14:36:32 <sgothel> rc12
20130705 14:36:37 <sgothel> all brand new stuff
20130705 14:36:48 <Eclesia> this class too ?
20130705 14:37:15 <sgothel> GLPixelBuffer and GLPixelBufferProvider are new classes / interfaces allowing to hook into the pixel stream .. <- yes, the AWT.. uses it now
20130705 14:38:08 -tomaw- [Global notice] Quick services restart incoming...
20130705 14:40:17 * sgothel (~sven@anon) Quit (Quit: Leaving.)
20130705 14:40:32 * sgothel (~sven@anon) has joined #jogamp
20130705 14:40:32 * ChanServ sets mode +v sgothel
20130705 14:51:07 * jk4_ (~jk4@anon) Quit (Ping timeout: 256 seconds)
20130705 14:51:30 * jk4 (~jk4@anon) has joined #jogamp
20130705 15:58:35 * Eclesia (~jsorel@anon) has left #jogamp
20130705 16:05:12 * Guest83277 is now known as odin_
20130705 16:40:15 * ebrayet (~Adium@anon) Quit (Ping timeout: 256 seconds)
20130705 17:09:37 * [Mike] (~Mike]@anon) has joined #jogamp
20130705 18:39:18 <olamedia> again, about classloading... java.lang.reflect.InvocationTargetException
20130705 18:39:18 <olamedia> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
20130705 18:39:18 <olamedia> at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
20130705 18:39:18 <olamedia> at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
20130705 18:39:18 <olamedia> at java.lang.reflect.Method.invoke(Method.java:597)
20130705 18:39:19 <olamedia> at ru.olamedia.modloader.ModDictionary.start(ModDictionary.java:54)
20130705 18:39:21 <olamedia> at ru.olamedia.modloader.ModLoader.start(ModLoader.java:189)
20130705 18:39:25 <olamedia> at ru.olamedia.launcher.Preloader.loadMods(Preloader.java:65)
20130705 18:39:27 <olamedia> at ru.olamedia.launcher.Launcher.main(Launcher.java:29)
20130705 18:39:29 <olamedia> Caused by: java.lang.UnsatisfiedLinkError: Can't load library: /home/olamedia/workspace/ola-modloader/libgluegen-rt.so
20130705 18:39:32 <olamedia> at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1706)
20130705 18:42:02 <olamedia> which classloader it tries to use?
20130705 18:42:07 <olamedia> at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1706)
20130705 18:42:07 <olamedia> at java.lang.Runtime.load0(Runtime.java:770)
20130705 18:42:07 <olamedia> at java.lang.System.load(System.java:1003)
20130705 18:42:07 <olamedia> at com.jogamp.common.jvm.JNILibLoaderBase.loadLibraryInternal(JNILibLoaderBase.java:468)
20130705 18:47:35 <olamedia> hm.. DynamicLibraryBundle.GlueJNILibLoader.loadLibrary(libBaseName, false, Platform.class.getClassLoader());
20130705 18:57:07 <olamedia> >_< So, u're using system classloader
20130705 19:00:40 <olamedia> Not sure, but using Thread's classloader seems good too
20130705 19:00:49 <olamedia> (like in nitfy-gui)
20130705 20:06:48 * Eclesia (~eclesia@anon) has joined #jogamp
20130705 20:06:52 <Eclesia> hi
20130705 20:07:11 <Eclesia> question : is there a way to create a modal frame with newt ?
20130705 20:07:37 <olamedia> u mean iside of parent?
20130705 20:07:40 <olamedia> *inside
20130705 20:09:20 <Eclesia> something like a dialog
20130705 20:09:30 <Eclesia> outside the main frame
20130705 20:09:55 <olamedia> wont opengl paint over all windows?
20130705 20:10:32 <olamedia> my ati legacy drivers seems to fail in many ways
20130705 20:11:30 <olamedia> :/
20130705 20:11:51 <olamedia> mb it's easier to use gui library to paint dialog?
20130705 20:12:42 <Eclesia> mb?
20130705 20:12:47 <olamedia> may be
20130705 20:12:58 <Eclesia> I make my own lib ^^
20130705 20:13:04 <Eclesia> over newt
20130705 20:13:51 <Eclesia> http://jsorel.ftp-developpez.com/projetun/sceneeditor.png
20130705 20:13:59 <olamedia> u probably can create new NEWT & listen for move events
20130705 20:14:17 <olamedia> to check if it's still inside of parent
20130705 20:15:19 <Eclesia> I just found a way, with proper synchronized
20130705 20:33:34 <Eclesia> first high level widget : http://jsorel.ftp-developpez.com/projetun/filechooser.png WFileChooser :D
20130705 21:09:50 <olamedia> :/ looks like gluegen tries to load library using absolute path, ignoring classloaders
20130705 21:11:55 * Guest51232 eyes self suspiciously
20130705 21:12:10 * ebrayet (~Adium@anon) has joined #jogamp
20130705 21:12:25 * Guest51232 is now known as rmk0
20130705 21:12:30 <rmk0> better
20130705 21:28:58 <olamedia> JNILibLoaderBase tries to use only base ClassLoader.class to findLibrary() :/
20130705 21:29:15 <olamedia> that a fail
20130705 21:31:03 <olamedia> even if system classloader is different, it will still try to use jdk's classloader
20130705 21:31:54 * xranby (~familjen@anon) Quit (Quit: Leaving.)
20130705 21:32:54 * ebrayet (~Adium@anon) Quit (Ping timeout: 252 seconds)
20130705 21:42:41 <olamedia> can't see a workaround
20130705 21:46:25 <olamedia> hmm... it's working with a system classloader and native jars at the same folder
20130705 21:47:13 <olamedia> strange thing it din't tries to find jars using any classloader
20130705 21:47:20 <olamedia> *didn't
20130705 21:56:16 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20130706 00:12:35 * [Mike] (~Mike]@anon) Quit ()
20130706 00:53:43 * [Mike] (~Mike]@anon) has joined #jogamp
20130706 00:53:44 * ___m___ (~Mike]@anon) has joined #jogamp
20130706 02:53:09 <olamedia> https://jogamp.org/bugzilla/show_bug.cgi?id=775
20130706 04:49:49 * ___m___ (~Mike]@anon) Quit ()
20130706 05:05:23 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20130706050523.html