#jogamp @ irc.freenode.net - 20130327 05:06:27 (UTC)


20130327 05:06:27 -CatOut- Previous @ http://jogamp.org/log/irc/jogamp_20130326050614.html
20130327 05:06:27 -CatOut- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130327050627.html
20130327 07:06:09 * DemoscenePassiv (~Lutsche@anon) has joined #jogamp
20130327 07:33:01 * [Mike] (~Mike]@anon) Quit ()
20130327 07:42:26 * DemoscenePassiv (~Lutsche@anon) Quit (Ping timeout: 256 seconds)
20130327 09:17:55 * gouessej (546051e4@anon) has joined #jogamp
20130327 09:18:01 <gouessej> Hi
20130327 09:18:11 <sgothel> Good morning
20130327 09:18:12 <gouessej> The forum seems to be down
20130327 09:18:29 <sgothel> so thx to the forum which is down - you are here :)
20130327 09:18:30 <gouessej> What is wrong?
20130327 09:18:50 <sgothel> I don't know .. lets see whats going on w/ nabble
20130327 09:20:11 <sgothel> http://nabble.com/ .. down ..
20130327 09:20:34 <sgothel> so how is it going ?
20130327 09:22:14 <gouessej> I'm fine
20130327 09:22:27 <gouessej> I have looked at your latest commits
20130327 09:22:34 <gouessej> One of my line is useless
20130327 09:22:38 <sgothel> that is great .. yup
20130327 09:22:41 <gouessej> and should be removed
20130327 09:22:50 <sgothel> the isWindows constraint .. yes
20130327 09:22:58 <sgothel> will take your patch :)
20130327 09:23:23 <sgothel> Xerxes here as well ?
20130327 09:23:25 <gouessej> and <property name="c.build.openal.soft.done" value="true" />
20130327 09:23:33 <xranby> gouessej: thank you for fixing with joal!
20130327 09:23:40 <gouessej> you're welcome
20130327 09:23:47 <sgothel> i.e. the remaining Java Sound stuff ..
20130327 09:24:12 <sgothel> i.e. remove dependencies .. and Wav header parsing
20130327 09:24:14 <gouessej> but I don't feel comfortable enough under Windows to setup a compile node, etc
20130327 09:24:26 <sgothel> well .. we have a workaround now
20130327 09:24:49 <sgothel> IMHO the JavaSound removal is higher prio, so we can use it under Android as well
20130327 09:25:22 <gouessej> Where are these dependencies?
20130327 09:25:30 <gouessej> We don't need Java Sound
20130327 09:25:36 <sgothel> I marked it in the bug report AFAIK
20130327 09:25:41 <gouessej> ok
20130327 09:26:08 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=684
20130327 09:26:09 <xranby> i recall there was a dependency inside the wav unpacking of java sound
20130327 09:27:23 <sgothel> how is ardor working on OSX now ?
20130327 09:27:33 <sgothel> i.e. I have changed a lot stuff for that 'evil' platform :)
20130327 09:27:53 <sgothel> the CALayer layout fix should work IMHO
20130327 09:28:05 <xranby> gouessej: quick ador3d test results on OpenGL ES, i have tested it on the N9 phone and my AC100 laptop
20130327 09:28:07 <sgothel> i.e. misplacement of CALayer stuff ..
20130327 09:28:08 <gouessej> It works but when I use the AWT canvas, it is still behind everything by default
20130327 09:28:26 <gouessej> Xerxes, does it work?
20130327 09:28:40 <xranby> not at all currently, both currently complain that you request a profile with fixedfunc(true)
20130327 09:28:48 <sgothel> hmm .. can you file a bugreport .. with a detailed description for me to reproduce ?
20130327 09:28:58 <xranby> thus you get a GLES1 profile on mobile devices that do not work with shaders
20130327 09:28:58 <gouessej> Sven, do we really need a wav loader in JOAL?
20130327 09:29:26 <sgothel> Why not ? It's a simple header, followed by PCM data
20130327 09:29:46 <sgothel> currently I hack a simple java based JPEG decoder .. based on a C impl.
20130327 09:29:51 <gouessej> Ok Xerxes, I'll find a better solution, thanks. Using getDefault() would be better, wouldn't it?
20130327 09:30:03 <xranby> yes Default will at least be shader only
20130327 09:30:16 <xranby> do ador3d support both fixed func and shaders ?
20130327 09:30:19 <gouessej> Sven, it is the same with any AWT thingy
20130327 09:30:57 <gouessej> Xerxes, Ardor3D supports tons of things but I used a lot PMVMatrix...
20130327 09:31:18 <xranby> if so then getDefault will most likely work!
20130327 09:31:20 <gouessej> Ardor3D has no fully shader based pipeline but it might work on ES
20130327 09:31:49 <sgothel> hmm .. isn't that covered by jME3 ?
20130327 09:32:12 <sgothel> i.e. jME3 successor of Ardor3D .. shader based ?
20130327 09:32:14 <gouessej> JMonkeyEngine does that, you're right Sven
20130327 09:32:37 <gouessej> Ardor3D and JMonkeyEngine 3 are 2 distinct projects
20130327 09:33:03 <sgothel> thought the focus is now on jME3 - yes I know, a fork .. but they went through that workload of FFP removal already
20130327 09:33:38 <sgothel> so how is jME3 going ? is your work merged ?
20130327 09:33:53 <gouessej> My work is already on the official repository
20130327 09:34:01 <sgothel> shall we try some Android tests ?
20130327 09:34:13 <gouessej> JMonkeyEngine 3 has a second renderer for the fixed pipeline
20130327 09:34:37 <sgothel> ah, interesting
20130327 09:34:57 <gouessej> JMonkeyEngine 3 & its JOGL 2.0 renderer should work under Android but some Android specific things are missing
20130327 09:34:57 <sgothel> for Android .. we would need to work on the assets, i.e. file access - probably
20130327 09:35:17 <sgothel> I would have to do the same to run Jake2 on Android
20130327 09:35:44 <gouessej> I will give another try to your port, I hope the sound will work this time
20130327 09:35:46 <sgothel> by using IOUtils class, it would work on Android, if the game data is under 'assets'
20130327 09:36:37 <sgothel> same is true for the temp directory, if needed
20130327 09:36:44 <sgothel> IMHO .. thats all
20130327 09:36:47 <gouessej> I use Ardor3D a lot, I can't drop it. I do my best. There is a remaining minor problem with JMonkeyEngine 3 & DDS
20130327 09:37:12 <gouessej> Xerxes, in theory yes
20130327 09:38:33 <gouessej> For Ardor3D, if I call getDefault(), it might pick the crappy Microsoft driver under Windows, the GDI thing :(
20130327 09:39:44 <sgothel> GLProfile.getDefault() ? That is odd - test_dbg.bat -> test_dbg.log should give us some clue here. IMHO bad driver installation or something.
20130327 09:40:07 <sgothel> or .. we have a bug, sure - never happened here
20130327 09:41:24 <gouessej> If I call getMaximum(true), will it work on N9? Xerxes?
20130327 09:42:42 <sgothel> -> GLProfile getMaxProgrammableCore(boolean favorHardwareRasterizer)
20130327 09:44:03 <gouessej> getMaxProgrammableCore wouldn't work with ES1
20130327 09:44:44 <sgothel> thought you wanted GLSL .. ok, don't really know what your target is
20130327 09:44:59 <gouessej> I think I will use getMaximum(true), it will pick up ES2 if available
20130327 09:45:09 <gouessej> My target is everything
20130327 09:45:29 <sgothel> :)
20130327 09:45:37 <gouessej> From the crappy desktop computer under OpenGL 1.2 to the latest machine under OpenGL 4.3
20130327 09:46:32 <gouessej> I'll replace getMaxFixedFunc(true) by getMaximum(true) when I'm back home
20130327 09:47:02 <xranby> gouessej: i will test it for sure, thank you
20130327 09:47:17 <gouessej> I'll fix the bug 697 too
20130327 09:47:25 <gouessej> Xerxes, thank you for testing
20130327 09:47:48 <sgothel> difference of maximum and default is .. sort of non existent, just the profile search list of maximum has a different priority
20130327 09:48:05 <sgothel> GL_PROFILE_LIST_MAX vs GL_PROFILE_LIST_ALL
20130327 09:48:11 <sgothel> but both favor hw-accel
20130327 09:49:44 <gouessej> I see what you mean. I just want to favor hw-accel and I don't need to force the use of the fixed pipeline anymore
20130327 09:49:51 <xranby> Processing 2.0 Beta 8 http://processing.org/download/ now include the JogAmp jogl armhf natives!
20130327 09:49:58 <gouessej> :)
20130327 09:50:18 <gouessej> I'm gonna contact the creators of Salem, the MMO
20130327 09:50:25 <sgothel> @Julien: Pls file the Ardor3D OSX bug report, makes my life easier :)
20130327 09:50:37 <gouessej> Lol the end users think it uses DirectX
20130327 09:50:40 <sgothel> yes .. if you can make an entry on JOGL .. nice
20130327 09:50:54 <gouessej> Sven, it is not our fault, this is an AWT bug
20130327 09:51:01 <sgothel> oh ..
20130327 09:51:08 <gouessej> NEWT works
20130327 09:51:12 <gouessej> :)
20130327 09:51:35 <sgothel> yeah .. I just want to make sure our OSX port gets stable, even the AWT part :)
20130327 09:51:42 <gouessej> Do you want me to enter a bug concerning AWT in the JogAmp bug tracker?
20130327 09:51:49 <gouessej> I see what you mean
20130327 09:52:11 <sgothel> yes pls .. if I can verify it's not our fault, good - but I like to see it
20130327 09:52:13 <sgothel> thx
20130327 09:52:55 <gouessej> It would be better to reproduce it with a more simple test case
20130327 09:52:59 <gouessej> in pure JOGL
20130327 09:53:07 <sgothel> ofc ..
20130327 09:53:18 <sgothel> if you dare .. :)
20130327 09:53:36 <gouessej> I can reuse my Wikipedia example
20130327 09:53:56 <gouessej> Xerxes, have you started to fix the bug 684?
20130327 09:54:30 <sgothel> Well, I doubt it's so easy, since other AWT unit tests work pretty well on OSX.
20130327 09:54:38 <xranby> i currently focus on less impressive stuff like fixing the joal-demos www examples
20130327 09:54:45 <sgothel> If possible, please use one of our unit test to make a new one
20130327 09:54:58 <gouessej> ok
20130327 09:55:28 <sgothel> Xerxes, IMHO Bug 684 is a real blocker for mobile and non J2SE ..
20130327 09:55:34 <gouessej> I remind you that the Mac book pro you gave me is under Mac OS X 10.6.8
20130327 09:55:49 <sgothel> yup
20130327 09:56:19 <gouessej> I need to work on OSM too very soon
20130327 09:56:20 <sgothel> w/ Java6 I can run non CALayer apps as well
20130327 09:56:27 <sgothel> nice
20130327 09:56:53 <xranby> do we know of any cross platform auto unpack library with comatible license?
20130327 09:57:00 <gouessej> I would like to use Ardor3D under Android too, especially on the Ouya
20130327 09:57:34 <gouessej> If I want to support the Ouya, I will need NEWT Controller API too
20130327 09:57:40 <sgothel> Xerxes: You can skip the WavLoader for later IMHO, I can look for a free wav parser
20130327 09:58:29 <sgothel> BTW .. can we have a weekly bug-walk ?
20130327 09:58:29 <gouessej> Personally, I don't need Wav support, it can be done in a third party library
20130327 09:59:21 <xranby> sgothel: sure which days work for you?
20130327 09:59:31 <sgothel> Julien: the Windows Pointer issue (BUg 697) was my hint ok ?
20130327 09:59:34 <xranby> wendsdays morning meetings ?
20130327 09:59:35 <gouessej> I would move WavLoader to JOAL-Utils and fix it later
20130327 09:59:35 <sgothel> any day
20130327 09:59:51 <gouessej> Sven, your hint is ok for me
20130327 09:59:57 <sgothel> you can tell me when you like it .. and I will be pleased
20130327 10:00:16 <sgothel> bug-walk, to prioritize bugs, and to not forget them .. you know it
20130327 10:00:27 <gouessej> Sven, it would be fine to find a better day or to do it in the evening
20130327 10:00:45 <sgothel> plus - maybe we can assign who validates some .. etc
20130327 10:01:16 <gouessej> I'm able to fix the bug 684, I'm almost sure of that, it's a matter of time
20130327 10:01:23 <sgothel> any time / any day - maybe we post this on forum (when it's up again) .. and ask for best schedule
20130327 10:01:33 <gouessej> good idea
20130327 10:02:06 <gouessej> Have you ever looked at my draft for the NEWT controller API?
20130327 10:02:34 <gouessej> I will have to write another draft for the shader builder
20130327 10:02:35 <sgothel> duck .. oops .. no, sorry - I am currently afraid of new features :)
20130327 10:02:46 <sgothel> ok .. which Bug number ?
20130327 10:03:05 <gouessej> I'm looking ofr it...
20130327 10:03:18 <sgothel> shader builder -> ShaderUtil/Code not satisfactory ?
20130327 10:03:37 <gouessej> 595
20130327 10:04:11 <gouessej> I'm lazy, I don't want to write 5 versions of the same shader
20130327 10:04:19 <gouessej> it goes farther
20130327 10:04:35 <xranby> where should we put engine porting bugs? i will gine one example: when i run processing 2.0 b8 on OpenGL ES devices i notice that the processing code unds up trying to use GLX and framebuffer objects, thus stuff that are not supported using EGL..
20130327 10:04:44 <sgothel> of course not, I use ShaderUtil to inject functions etc .. and GL3/ES2 alignement
20130327 10:05:23 <gouessej> Xerxes, it should be put into the bug tracker of the engine except if we directly care of this engine (Java3D)
20130327 10:05:38 <sgothel> if it is not a bug w/ JOGL, but of it's usage - well, we can add them like for .. yes
20130327 10:05:45 <gouessej> Sven, I want to write shaders in Java :)
20130327 10:06:16 <sgothel> so pls try using ShaderUtil/Code .. look at it's usage, e.g. in the GLMediaPlayer/FFMpeg ..
20130327 10:06:22 <gouessej> yes of course
20130327 10:06:32 <sgothel> and if you like some more features .. we can deal w/ it
20130327 10:06:47 <gouessej> but I would like to avoid writing 5 GLSL files per shader
20130327 10:07:04 <gouessej> and using tons of macros is not fun
20130327 10:07:09 <sgothel> well, if you include them .. sometimes nice
20130327 10:07:24 <gouessej> it won't be ready for siggraph
20130327 10:07:25 <sgothel> we support the 'include' directive
20130327 10:07:52 <xranby> sgothel: the exception looks like this... https://gist.github.com/xranby/123f22afd5645ff149c9
20130327 10:08:49 <xranby> gouessej: ok, yes i may start by open a processing issue upstream
20130327 10:08:59 <gouessej> Sven, do you plan to close the bug 706? it's ok for me
20130327 10:10:07 <gouessej> The end user will have to update his bios so that he can use the latest ATI driver
20130327 10:10:10 <sgothel> I would wish we could get a proper debug log file from him .. to be able to create a workaround .. well, if he doesn't response .. nothing we can do then
20130327 10:10:33 <gouessej> yes it would be better
20130327 10:11:22 <sgothel> allow this bug to sit for a while .. maybe we can have a response - better, a machine exposing this bug
20130327 10:11:38 <sgothel> Xerxes: So machine has X11 and GLX installed ?
20130327 10:11:44 <sgothel> I remember this one ..
20130327 10:12:08 <sgothel> AFAIK I removed GLX support on AC100 X11 config ..
20130327 10:12:20 <sgothel> i.e. if there is GLX .. it should work
20130327 10:12:51 <xranby> this machine is now using an xorg.conf thus X11 have auto chosoen some settings
20130327 10:12:55 <xranby> is not using
20130327 10:13:24 <xranby> lets see if i can fix this somehow by adding an xorg.conf
20130327 10:15:52 <gouessej> Why is the forum still down?
20130327 10:16:37 <gouessej> nabble.com is still down
20130327 10:20:42 <gouessej> Ok I have to leave. Bye.
20130327 10:20:49 * gouessej (546051e4@anon) Quit (Quit: Page closed)
20130327 10:21:01 <sgothel> ok - the exception is informal
20130327 10:21:42 <sgothel> it would be a JOGL bug, we need complete debug logs - you know the drill .. then we can see ..
20130327 10:22:24 <sgothel> looks like simply that no FBO is available, or FBO ABI it used wrong ..
20130327 10:23:09 <sgothel> looks like a simple GLWindow w/ FBO unit test should suffice to reproduce this exception
20130327 10:24:06 <xranby> sgothel: ok i solved the missing GLX by adding this xorg.conf to my system: https://gist.github.com/xranby/5253221
20130327 10:24:40 <sgothel> if that works, it's a bug :)
20130327 10:24:54 <sgothel> i.e. we should tolerate that .. and claim all desktop GL profile n/a
20130327 10:25:02 <sgothel> thought we do ..
20130327 10:25:37 <sgothel> guess I have to launch my AC100 ..
20130327 10:25:45 <xranby> now the exception is reduced to : https://gist.github.com/xranby/64263a31af92d3092772
20130327 10:25:58 <xranby> it only update the screen on "resize"
20130327 10:26:21 <sgothel> by 'now' you mean, manually enabled glx
20130327 10:26:30 <sgothel> ok - so it seems to be informal as mentioned
20130327 10:26:36 <xranby> yes manually enabled GLX
20130327 10:26:39 <sgothel> (and as I remember my code :)
20130327 10:27:25 <sgothel> try this unit test: TestGLAutoDrawableGLWindowOnOffscrnCapsNEWT
20130327 10:27:35 <sgothel> .. and maybe the others .. which lay around there ..
20130327 10:28:02 <sgothel> well, we need a bug report anyways, best to write it down before memory resets :)
20130327 10:28:23 <xranby> sgothel: ac100 armhf setup instructions: https://wiki.ubuntu.com/ARM/TEGRA/AC100
20130327 10:30:09 <xranby> i am using the nvidia-tegra binary GLES driver in 12.10 https://wiki.ubuntu.com/ARM/TEGRA/AC100#Graphics
20130327 10:31:11 <sgothel> yeah .. this will take time to do, currently busy w/ OSX, Android and stuff. Pls use latest JOGL as well, exceptions doesn't match git tip :)
20130327 10:32:29 <sgothel> exception is thrown in an GLEventListener at PGL.java:3255 anyways ..
20130327 10:32:51 <sgothel> init(..) method
20130327 10:33:05 <sgothel> processing.opengl.PGL$PGLListener.init(PGL.java:3255)
20130327 10:33:30 <sgothel> do we have a weblink to that code ?
20130327 10:35:07 <sgothel> I also would recommend to use the GLOffscreenAutoDrawable instead of a NEWT offscreen window for FBO usage, since it's like a bit over the top to use NEWT then ..
20130327 10:35:50 <sgothel> Offscreen NEWT window was impl. to have a composited offscreen NEWT window w/ some input injection .. blabla, actually a new Java plugin :)
20130327 10:38:29 <xranby> oh look, they have migrated to git! https://github.com/processing/processing
20130327 10:38:38 <sgothel> ah!
20130327 10:39:56 <xranby> https://github.com/processing/processing/blob/master/core/src/processing/opengl/PGL.java#L3255
20130327 10:40:24 <sgothel> well, that explains it :)
20130327 10:40:49 <xranby> well hmm. i am surprised that the skets renders a frame fine when i resize the window
20130327 10:40:54 <xranby> sketch
20130327 10:41:16 <xranby> but i get no updates for normal operation
20130327 10:42:57 <sgothel> GLContext: hasBasicFBOSupport() and hasFullFBOSupport()
20130327 10:44:13 <sgothel> it's quite a complex query .. hence it's exposed
20130327 10:44:51 <sgothel> GLContextImpl#1447
20130327 10:45:31 <sgothel> API doc in hasBasicFBOSupport() describes it ..
20130327 10:45:48 <sgothel> you cannot simply query extensions
20130327 10:47:02 <sgothel> http://jogamp.org/deployment/jogamp-next/javadoc/jogl/javadoc/javax/media/opengl/GLContext.html#hasBasicFBOSupport%28%29
20130327 10:47:16 <sgothel> http://jogamp.org/deployment/jogamp-next/javadoc/jogl/javadoc/javax/media/opengl/GLContext.html#hasFullFBOSupport%28%29
20130327 10:47:52 <xranby> sgothel: Gl FBO: basic true, full true
20130327 10:48:12 <xranby> System.err.println("GL FBO: basic "+ gl.hasBasicFBOSupport()+", full "+gl.hasFullFBOSupport());
20130327 10:48:31 <sgothel> very good (is printed in normal GLContext.toString())
20130327 10:48:58 <sgothel> so they have to adapt .. best to use our queries, AC actually knows about all of that AFAIK
20130327 10:49:33 <sgothel> coffee .. now .. back in a sec :)
20130327 11:02:18 <xranby> ... the above profiled the mesa driver..
20130327 11:02:23 <xranby> with tegra2 i get
20130327 11:02:32 <xranby> Gl FBO: basic true, full false
20130327 11:02:40 <sgothel> sounds reasonable ..
20130327 11:02:44 <xranby> that more like what i expected
20130327 11:03:01 <sgothel> I refine the GLX query to remove the ugly exception :)
20130327 11:04:23 <xranby> there is someting else as well.. when i run with GLX disabled i get a crash when i try to interact with the window, mouse click.. this crash do not happen when i run with GLX enabled
20130327 11:04:40 <xranby> interesting
20130327 11:04:54 <sgothel> hmm ..
20130327 11:05:04 <sgothel> crash == JVM crash ?
20130327 11:05:12 <sgothel> whats the hotspot trace
20130327 11:05:20 <sgothel> and where is my coffee ? :)
20130327 11:05:26 <xranby> yes let me send you the hs_err_pid*.log
20130327 11:09:51 <xranby> sgothel: https://gist.github.com/xranby/8e4aa633231fa457cb9c
20130327 11:12:06 <xranby> i can try generate a core dump
20130327 11:15:31 <sgothel> C [libX11.so.6+0x20620] XPeekIfEvent+0xcbj sun.awt.X11.XToolkit.run(Z)V+63j sun.awt.X11.XToolkit.run()V+5
20130327 11:15:37 <sgothel> AWT related
20130327 11:20:09 <xranby> hmm it may be my experimental setup then
20130327 11:20:16 <xranby> of openjdk + oracles jit
20130327 11:20:32 <sgothel> oracle on embedded == experimental
20130327 11:20:57 <xranby> i am using the openjdk awt classes and simply unpacked their latest armhf hotspot on top
20130327 11:21:22 <xranby> so sure.. nothing to get upset about
20130327 11:22:06 <xranby> back to debug/fix the VBO init probing for processing
20130327 11:22:14 <xranby> my dev machine have almost checked out their git
20130327 11:22:35 <xranby> 1.35 GiB
20130327 11:24:17 <sgothel> blobbing :)
20130327 11:25:07 <xranby> sometimes i wish there was a way to tell git to forget old checked in blobs without rewriting history
20130327 11:28:52 <sgothel> http://jogamp.org/git/?p=jogl.git;a=commit;h=9a4fcc7ea4ec61e4ceed791acced734ac04ea270
20130327 11:28:55 <sgothel> can you test ?
20130327 11:29:04 <sgothel> i.e. w/ GLX disabled ?
20130327 11:29:19 <sgothel> maybe the test_dbg.sh .. w/o AWT stuff
20130327 11:29:57 <sgothel> DEBUG -> "GLX [is|not] available on device/server: X11GraphicsDevice..." w/o exception
20130327 11:30:14 <sgothel> again, exception was just informal ..
20130327 11:30:20 <sgothel> but .. it's better this way :)
20130327 11:30:51 <xranby> sgothel: i can test, half of my ac100 is willed with jogl and gluegen source
20130327 11:31:09 <xranby> so its setup to recompile quite swiftly
20130327 11:31:16 <xranby> filled
20130327 11:31:19 <sgothel> next aggregated ..
20130327 11:34:38 <sgothel> https://jogamp.org/chuck/job/jogl/944/ .. on it's way
20130327 11:46:40 * gouessej (546051e4@anon) has joined #jogamp
20130327 11:46:54 <gouessej> I'm back
20130327 11:47:15 <gouessej> the forum is still down :s
20130327 11:51:01 <xranby> gouessej: it may be an internet slip
20130327 11:51:02 <xranby> http://www.downforeveryoneorjustme.com/forum.jogamp.org
20130327 11:51:10 <xranby> split
20130327 11:51:20 <xranby> i fail to resolve it from here as well
20130327 11:51:27 <xranby> but that site claim it is up
20130327 11:51:36 <xranby> err
20130327 11:51:50 <xranby> i only see a default apache : It works!
20130327 11:51:52 <xranby> from here
20130327 11:52:21 <xranby> something tellms me that nabble is having big trouble
20130327 11:52:39 <sgothel> .. and we should back it up more often .. (if this is possible)
20130327 11:52:55 <xranby> http://support.nabble.com/Free-Support-f1.html
20130327 11:52:55 <sgothel> mailman nabble .. bridge anybody ?
20130327 11:53:13 <xranby> http://support.nabble.com/Forum-is-not-accessible-Shows-white-screen-with-quot-It-works-quot-td7583747.html
20130327 11:57:36 <sgothel> Xerxes: https://jogamp.org/chuck/job/jogl/944/label=linux-armv7hf-img/
20130327 11:57:51 <sgothel> https://jogamp.org/chuck/job/jogl/944/label=linux-armv7hf-img/artifact/build/jogl-2.0-b944-20130327-linux-armv6hf.7z
20130327 12:02:08 <xranby> sgothel: kudos
20130327 12:09:56 <xranby> testing: ran into an unexpected issue.. looks like AC is using a custom version of JOGL... http://paste.ubuntu.com/5652075/
20130327 12:10:35 <xranby> do we have a patch that adds the GLCapabilities.setDefaultColor(I)V from AC on the mailinglist?
20130327 12:11:31 <sgothel> no - I refused it, too much trouble, discussed it in bug report, no response yet - should not be required anymore
20130327 12:11:54 <xranby> oh lala.. ok i will have to compile my own version of processing then
20130327 12:12:39 <sgothel> i.e. adding a clear property for FBO rendering is bad solution, better render one extra frame, swap - then compose it ..
20130327 12:13:02 <sgothel> pls make this remark (JOGL incompat.) in bug report ..
20130327 12:13:40 <gouessej> There is something I don't understand
20130327 12:14:12 <gouessej> When do we plan to release JOGL 2.0? Will there be another release candidate?
20130327 12:14:23 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=649
20130327 12:14:28 <gouessej> When will we push another version on Maven Central?
20130327 12:14:36 <sgothel> maven ? already done
20130327 12:14:57 <sgothel> I have sent an update to forum a day ago .. or two
20130327 12:15:10 <sgothel> release .. yeah, the ole question :)
20130327 12:15:38 <sgothel> I added remarks on Wiki .. etc, extra link for aggregated builds
20130327 12:15:42 <gouessej> some things can be postponed of course, NEWT controller API can wait
20130327 12:15:59 <gouessej> JInput integration too if any
20130327 12:16:22 <sgothel> not necessarily .. but I focus on stability / usability of core features right now, if done - 2.0
20130327 12:16:46 <sgothel> and how days are going by . .. little things here and there ..
20130327 12:16:57 <gouessej> That's what I meant, major enhancements and new features can wait for the next minor release
20130327 12:17:24 <gouessej> except JOAL support on non j2se
20130327 12:17:29 <sgothel> hopefully I can walk through all OSX tests this week and the next week, then I feel more comfy
20130327 12:17:31 <gouessej> ;) Xerxes
20130327 12:17:47 <sgothel> yeah, you guys pls make sure you know who does this JOAL stuff ..
20130327 12:17:49 <xranby> gouessej: JOAL needs one maintainer / platform i am afraid :/
20130327 12:18:07 <sgothel> Xerxes, you want to resign ?
20130327 12:18:29 <sgothel> well .. it's nothing of a bit title thingy here anyways - who does the work, does the work :)
20130327 12:18:34 <gouessej> Xerxes, are you sure of that?
20130327 12:18:53 <sgothel> pls - just make clear who does the work - the result is what counts :)
20130327 12:19:02 <sgothel> we can fight for titles later :)
20130327 12:19:38 <xranby> sgothel: gouessej: from experience i can tell that i only test joal / jogl on gnu/linux thus windows, solaris and android gets less attention than deserved
20130327 12:19:41 <gouessej> I can work a bit on the bug 684 after I fix the bug 697
20130327 12:20:20 <gouessej> but someone should move WavLoader to another module, JOAL-Utils or something similar
20130327 12:20:22 <xranby> what i feel is on my table is making sure that the www examples work before the 2.0 release
20130327 12:20:45 <xranby> it makes littel sense to have outdated links on the homepage to java.net
20130327 12:20:46 <xranby> et
20130327 12:20:53 <sgothel> Julien, a wav loader is so simple, that it should remain ..
20130327 12:20:59 <gouessej> Xerxes, we can postpone JOAL support for Android to the next minor release if you want
20130327 12:21:07 <gouessej> ok Sven
20130327 12:21:11 <gouessej> as you wish
20130327 12:21:33 <xranby> this i am working on joal www cleanup here: https://github.com/xranby/joal-demos/tree/www
20130327 12:21:54 <sgothel> Xerxes, if possible - put some joal unit test in :)
20130327 12:21:54 <xranby> the only way to get things done is to fix one little thing at a time
20130327 12:22:00 <sgothel> yup
20130327 12:22:26 <sgothel> note that the packages own unit tests have always highest priority when it comes to fixes .. etc
20130327 12:23:44 <gouessej> It would be fine to have a clear roadmap
20130327 12:23:59 <gouessej> with the bugs that will be fixed later, those blocking the release
20130327 12:24:01 <gouessej> etc
20130327 12:24:04 <sgothel> stability stability - usability across all platforms of what we have
20130327 12:24:12 <sgothel> yes
20130327 12:24:34 <sgothel> for example the JPEG decoder might be required for across platform experience, as the PNG loader was
20130327 12:24:46 <sgothel> JOAL, maybe even JOCL
20130327 12:24:58 <gouessej> who maintains JOCL?
20130327 12:25:01 <sgothel> OSX should work w/o problems, hence I need to work on that
20130327 12:25:15 <sgothel> I will play w/ JOCL one month before siggraph .. i.e. test mobile
20130327 12:25:47 <gouessej> ok but we still have no permanent maintainer for this binding
20130327 12:25:59 <sgothel> :)
20130327 12:26:24 <gouessej> JPEG support would be fine but what about licensing issues?
20130327 12:26:57 <sgothel> you ask for directions, ok: your OSX bugs in bugzilla; OSX fixes; JOAL cleanup; ES3/4.3 -> 2.0.2 release
20130327 12:27:11 <sgothel> and while we go - all must work .. jME3, .. etc
20130327 12:27:27 <sgothel> let's push a deadline: 1 month before siggraph
20130327 12:27:55 <sgothel> and 1 week before siggraph we shall have our slides and demos in shape, I actually had a dream about that :)
20130327 12:28:14 <sgothel> no, it was a funny one - empty slides lol :)
20130327 12:28:36 <sgothel> is that roadmap acceptable ?
20130327 12:28:42 <gouessej> lol
20130327 12:29:06 <gouessej> We'll have to make something more detailed and update it each week
20130327 12:29:11 <gouessej> after the meeting
20130327 12:29:36 <sgothel> sounds great
20130327 12:30:23 <xranby> sgothel: i am testing now on the ac100... no GLX warning... and no rendering hmm
20130327 12:30:51 <sgothel> balance of power: one down, one up :)
20130327 12:31:54 <gouessej> Should someone look at all bugs to prepare the roadmap? Should we tag them in bugzilla?
20130327 12:32:19 <sgothel> well, we could use priorities for our next bug-walk
20130327 12:33:24 <gouessej> ok
20130327 12:33:44 <sgothel> don't know how to 'tag' them .. hmm, maybe indeed a wiki page for 2.0.2 ?
20130327 12:33:59 <sgothel> or we would need to add another field .. hmm
20130327 12:33:59 <gouessej> good idea
20130327 12:34:11 <gouessej> a wiki page would be nice
20130327 12:34:13 <gouessej> it's ok for me
20130327 12:34:46 <sgothel> who?
20130327 12:35:06 <sgothel> who volunteers to be the bug-walk and 2.0.2 manager ?
20130327 12:35:14 <sgothel> (best if not me :)
20130327 12:35:28 <gouessej> I can at least prepare the wiki page
20130327 12:35:37 <sgothel> you are the man! thx
20130327 12:35:44 <gouessej> you're welcome
20130327 12:35:53 <sgothel> you are entitled to kick but :)
20130327 12:36:03 <gouessej> lol
20130327 12:36:55 <gouessej> I think we can support JMonkeyEngine 3 and Ardor3D under Linux ARM before Siggraph but it will be more difficult for Android
20130327 12:37:52 <sgothel> well - I will be there to help ofc
20130327 12:38:18 <xranby> gouessej: kudos, thank you for kicking the ball to roll towards the 2.0.2 goal
20130327 12:38:23 <gouessej> a demo with the Ouya would have been really cool
20130327 12:38:55 <gouessej> Can't we clone Sven and Xerxes?
20130327 12:38:59 <sgothel> dunno what it is .. but you know all the stuff, so I trust you
20130327 12:39:08 <sgothel> hehe .. clone us all
20130327 12:39:14 <gouessej> Ouya is a video game machine under Android
20130327 12:39:27 <gouessej> it is very famous
20130327 12:39:38 <sgothel> is there a pac-man and space invaders port ? I love those ole games :)
20130327 12:39:43 <gouessej> it's under Tegra 3
20130327 12:39:59 <sgothel> mame in java :)
20130327 12:40:43 <gouessej> After fixing some nasty bugs, I must work on OSM
20130327 12:41:19 <gouessej> Sven, look at that: http://en.wikipedia.org/wiki/Ouya
20130327 12:45:01 <gouessej> There is already a "priority" field I can use
20130327 12:45:15 <xranby> (13:31:30) sgothel: balance of power: one down, one up :) <--- i m investering.. processing actually fails quite rearly in initialization, they use a NEWT GLWindow and the window.getChosenGLCapabilities(); retuns NULL on the ac100
20130327 12:45:37 <xranby> but their code did not notice that
20130327 12:45:50 <sgothel> test_dbg.log ?
20130327 12:46:06 <xranby> i am getting there...
20130327 12:46:10 <sgothel> i.e. we should pick up ES2 gl-caps ..
20130327 12:48:36 <gouessej> ok I have to leave. Bye
20130327 12:48:42 * gouessej (546051e4@anon) Quit (Quit: Page closed)
20130327 12:51:42 * ebrayet (~Adium@anon) has joined #jogamp
20130327 13:03:22 <xranby> sgothel: test_dbg.log using jogl build 943 and 944 http://labb.zafena.se/jogamp/test/ac100-ubuntu-12.10-armhf/
20130327 13:03:32 <xranby> 944 includes your latest GLX fix
20130327 13:04:27 <xranby> runninng with GLX disabled
20130327 13:04:34 <sgothel> very good!
20130327 13:04:56 <sgothel> looks good, I was afraid I removed the code path where we can run ES2 w/ X11 w/o GLX!
20130327 13:06:15 <sgothel> EGLUpstreamSurfaceHook[ 128x128, jogamp.newt.driver.x11.WindowDriver: 0x2c00002] -> good!
20130327 13:06:24 * ebrayet (~Adium@anon) Quit (Read error: Connection reset by peer)
20130327 13:06:50 <sgothel> so they only query GL2 resources ?
20130327 13:07:51 * ebrayet (~Adium@anon) has joined #jogamp
20130327 13:08:07 <xranby> https://github.com/processing/processing/blob/master/core/src/processing/opengl/PGL.java#L659
20130327 13:08:22 <xranby> this is the line returning null
20130327 13:09:00 <xranby> https://github.com/processing/processing/blob/master/core/src/processing/opengl/PGL.java#L600
20130327 13:09:06 <xranby> these are the caps they select
20130327 13:09:15 <sgothel> AWT or NEWT ?
20130327 13:09:30 <sgothel> NEWT .. ok
20130327 13:10:05 <sgothel> well, they donw' setVisible(true) on window!
20130327 13:10:28 <xranby> i can try add that
20130327 13:10:30 <sgothel> NEWT is lazy, so must be visible ..
20130327 13:11:05 <sgothel> actually all our resource path is lazy by now, i.e. init at the latest possible moment to safe resources
20130327 13:11:44 <sgothel> so even the AWT path is questionable ..
20130327 13:12:09 <xranby> setVisible(true) before creating the NewtCanvasAWT(window) ?
20130327 13:12:24 <sgothel> that would look ugly
20130327 13:12:40 <sgothel> i.e. top-level -> AWT child
20130327 13:12:45 <sgothel> which is visible :)
20130327 13:13:05 <sgothel> just before line 659 would suffice
20130327 13:13:20 <sgothel> oh .. sorry .. right - no :)
20130327 13:13:45 <sgothel> capabilities should be set when the AWT parent gets visible
20130327 13:13:58 <sgothel> i.e. lazy at a later time ..
20130327 13:15:07 <sgothel> if the pg.parent is already visible, good
20130327 13:15:23 <sgothel> the pg.parent.add call should happen on AWT-EDT then
20130327 13:15:38 <sgothel> as all AWT mutable ops shall happen there
20130327 13:15:47 <sgothel> we even have a unit test for AC's code :)
20130327 13:15:57 <sgothel> which looks quite familiar with this PG class
20130327 13:19:50 <xranby> sgothel: i will run the sketch that triggers the NULL capabilities using jogamp.debug=true and see what pops up
20130327 13:20:28 <sgothel> jogl.debug and newt.debug
20130327 13:20:34 <sgothel> jogamp.debug is overkill
20130327 13:21:24 <sgothel> TestGLCanvasAWTActionDeadlock02AWT
20130327 13:23:11 <sgothel> protected void initSurface(int antialias) {: Is is called from the AWT-EDT ?
20130327 13:23:24 <sgothel> if not, extra care has to be taken ..
20130327 13:24:32 <sgothel> ok ..
20130327 13:24:53 <sgothel> you can add the capabilities assignment in PGLListener.init(..)
20130327 13:25:28 <sgothel> that will be called 1st when GLAutoDrawable (whichever) gets realized'
20130327 13:25:54 <sgothel> still dunno how they manage and satisfy the AWT constraints .. the setup code is quite, interesting
20130327 13:26:03 <xranby> sgothel: help me out here.. when you ask me to run individual tests.. what is the most convinient way to do it? i usually run sh scripts/make.jogl.all..linux.armhf.sh junit.run but that runs all tests
20130327 13:26:16 <xranby> from inside the jogl/make directory
20130327 13:26:16 <sgothel> oh
20130327 13:26:45 <sgothel> scripts/tests-linux-armv6.sh -> scripts/tests.sh
20130327 13:26:53 <sgothel> so you want to edit test.sh ..
20130327 13:27:07 <sgothel> the calling script just takes care of the right build env. path etc
20130327 13:29:09 <xranby> ok i have edited scripts/tests.sh and enabled the line..
20130327 13:29:24 <xranby> when i run sh scripts/tests-armv6_armhf.sh
20130327 13:29:34 <xranby> i only get the usage printed
20130327 13:29:37 <xranby> that puzzle me
20130327 13:30:13 <sgothel> `which java` -> null ?
20130327 13:30:31 <xranby> /usr/bin/java
20130327 13:30:50 <xranby> ah bash
20130327 13:30:52 <sgothel> scripts/tests-linux-armv6hf.sh: spath=`dirname $0`
20130327 13:30:52 <sgothel> . $spath/tests.sh `which java` -DummyArg ../build-linux-armv6hf $*
20130327 13:31:00 <xranby> ok ubuntu is using dash
20130327 13:31:08 <xranby> and this dont work with your script
20130327 13:31:36 <xranby> bash scripts/tests-armv6_armhf.sh worked
20130327 13:31:43 <sgothel> :)
20130327 13:32:29 <xranby> the TestGLCanvasAWTActionDeadlock02AWT passed
20130327 13:32:35 <sgothel> even the borne again shell is new to me :)
20130327 13:32:39 <sgothel> ah .. great
20130327 13:32:47 <xranby> or did it
20130327 13:32:49 <xranby> check ing log
20130327 13:32:55 <sgothel> sure, this one takes care of ..
20130327 13:33:37 <sgothel> can't .. it's GL2 :)
20130327 13:34:37 <sgothel> but you see all the "javax.swing.SwingUtilities.invokeAndWait(new Runnable() {..." code .. thats what I meant w/ AWT-EDT ..
20130327 13:35:25 <sgothel> draw(drawable.getGL().getGL2()); -> should fail
20130327 13:36:11 <sgothel> replace it w/ GearsES2 .. and let me merge it, thx
20130327 13:40:46 <xranby> yup,, i am getting the now a GL2 implementation exception... i will make a new unittest and merge
20130327 13:41:09 <xranby> right now all mo coworkers needs me to prepare some experimental intrumentation setup
20130327 13:41:11 <sgothel> you can drop GL2 in that unit test ..
20130327 13:41:28 <xranby> i will be unavailable for at least 1h
20130327 13:41:48 <sgothel> no hurry - have fun, me lucky you are here
20130327 15:57:05 * xranby_ac100 (~xranby@anon) has joined #jogamp
20130327 17:33:51 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20130327 18:12:35 * ebrayet (~Adium@anon) has joined #jogamp
20130327 18:24:21 * DemoscenePassiv (~Lutsche@anon) has joined #jogamp
20130327 18:59:52 * gouessej (5279cd8e@anon) has joined #jogamp
20130327 19:01:12 <gouessej> Hi
20130327 19:01:27 <sgothel> Hi Julien
20130327 19:01:52 <gouessej> I've just replaced GLProfile.getMaxFixedFunc(true) by GLProfile.getMaximum(true) in Ardor3D
20130327 19:02:21 <gouessej> The exception used in Alut from javax.sound is completely useless
20130327 19:02:54 <sgothel> I have started replacing the exceptions already in my last commits (JOAL)
20130327 19:28:52 <sgothel> good evening
20130327 19:28:55 <gouessej> oh
20130327 19:29:16 <gouessej> Is there still something to do?
20130327 19:29:26 <gouessej> I only see 5 occurences of javax.sound
20130327 19:29:28 <sgothel> it's more a validation thingy
20130327 19:29:48 <sgothel> ok - throw them away :) .. keep Wav loader alive until we replace it
20130327 19:30:02 <sgothel> but API should be free of J2SE sound
20130327 19:30:25 <gouessej> I can replace them by ALException instances
20130327 19:30:34 <sgothel> i.e. the one Android JOAL test works w/ PCM
20130327 19:30:48 <sgothel> I guess you heard the sound ..
20130327 19:30:57 <sgothel> yeah, I added those I guess
20130327 19:31:08 <gouessej> there is a problem
20130327 19:31:09 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20130327 19:31:14 <gouessej> it uses AudioSystem
20130327 19:31:27 <gouessej> If I remove it, the whole class is unusable
20130327 19:31:58 <sgothel> ALException ?
20130327 19:32:36 <sgothel> WAVLoader right ?
20130327 19:32:39 <gouessej> exceptions used in JOAL
20130327 19:32:43 <gouessej> Yes
20130327 19:33:01 <gouessej> Maybe I can use IOUtil
20130327 19:33:13 <sgothel> If possible, encapsule the impl. to something invisible, so we can replace it later
20130327 19:33:36 <sgothel> I looked in the wav parsing those days .. it's just a few bytes prior the PCM data
20130327 19:33:49 <sgothel> giving you the freq, channels, etc
20130327 19:35:30 <sgothel> WAVData wd = WAVData.loadFromStream(ResourceLocation.getTestStream2(), ResourceLocation.getTestStream2Size(), 2, 16, 44100, ByteOrder.LITTLE_ENDIAN, true);
20130327 19:35:38 <sgothel> OpenALTest.java
20130327 19:36:01 <sgothel> maybe a better name for WavData, since it's actually type agnostic.
20130327 19:36:07 <sgothel> I guess it's OK to change the API now ..
20130327 19:37:32 <sgothel> so all we need to extract from the header is: size, channels 2, bits 16, f 44100, ByteOrder.LITTLE_ENDIAN, .. you know it
20130327 19:38:15 <sgothel> anybody here w/ JavaScript experience ?
20130327 19:38:27 <gouessej> me, a bit
20130327 19:38:43 <sgothel> if (typeof node === 'number')
20130327 19:38:43 <sgothel> return node;
20130327 19:38:43 <sgothel> if (typeof node !== 'object')
20130327 19:38:43 <sgothel> throw "invalid huffman sequence";
20130327 19:39:06 <sgothel> this is a poor mans reflection system ? ensuring it's a number only ?
20130327 19:39:26 <gouessej> it looks so
20130327 19:39:45 * [Mike] (~Mike]@anon) has joined #jogamp
20130327 19:40:03 <sgothel> yeah .. node is read from an array, and they have no types :)
20130327 19:40:56 <gouessej> lol
20130327 19:41:17 <gouessej> and they want us to drop Java for this language?
20130327 19:41:20 <gouessej> lol
20130327 19:41:51 <sgothel> well, at least it would mean more employment for all of us, since maintenance of such code is at least 'interesting' :)
20130327 19:42:02 <gouessej> The forum is back online :)
20130327 19:42:22 <sgothel> but this doesn't mean that you should leave this chat :)
20130327 19:43:25 <gouessej> I won't leave it now
20130327 19:44:27 <sgothel> great - I guess Harvey will join soon from Vancouver as well, usually at night
20130327 19:45:07 <gouessej> Actually, why do you want to keep WavLoader in JOAL itself?
20130327 19:45:56 <sgothel> to have a simple tool, like PNG/JPEG - so JOAL can be used .. self contained, at least a little bit
20130327 19:46:18 <sgothel> for example, it sucks that OpenMAX has no deplexer for streams etc
20130327 19:46:36 <sgothel> so a beginner can simply start doing something w/ too much troubles
20130327 19:46:44 <sgothel> it's not too personal though :)
20130327 19:46:49 <gouessej> ok
20130327 19:47:09 <gouessej> I use Paul Lamb's Sound Library, it works very well
20130327 19:47:33 <sgothel> sox ? yup
20130327 19:47:54 <gouessej> Paul Lamb's Sound Library has a plugin for JOAL
20130327 19:48:12 <sgothel> ah ..
20130327 19:50:00 * ebrayet (~Adium@anon) has joined #jogamp
20130327 19:50:11 <sgothel> http://www.paulscode.com/docs/SoundSystem/paulscode/sound/libraries/package-summary.html
20130327 19:51:46 <gouessej> yes
20130327 19:52:11 <gouessej> His own OpenAL binding for Android supports Android 1.5
20130327 19:52:24 <sgothel> so we can grab the wav parser from it .. it's permissive - sure, it's a more serious higher level API
20130327 19:52:47 <sgothel> oh .. well, dunno - can't even emulate Android < 2.3
20130327 19:53:15 <sgothel> actually I guess emulation < 4.0 doesn't work properly - last time I checked
20130327 19:53:17 <gouessej> I know you're not very motivated to support Android < 2.3
20130327 19:53:30 <sgothel> I don't even have a device
20130327 19:53:35 <sgothel> or emulator
20130327 19:53:36 <gouessej> I have one
20130327 19:53:56 <gouessej> My HTC Desire Z
20130327 19:54:40 <sgothel> I know ... you like to go through all those errands ? probably takes a while .. one month or so, will risk our release :]
20130327 19:55:13 <sgothel> it's doable .. just tedious work ..
20130327 19:55:47 <sgothel> ok .. cont. porting jpegdecoder ..
20130327 19:58:45 <gouessej> I will take care of that later
20130327 19:58:57 <gouessej> Don't worry
20130327 19:59:02 <gouessej> stability first
20130327 19:59:38 <gouessej> If we want to compete with AndroidGL, we will have to support Android 2.0
20130327 20:02:07 <sgothel> http://developer.android.com/about/dashboards/index.html
20130327 20:03:25 <sgothel> 9.6% < API Level 9 (2.3)
20130327 20:05:08 <gouessej> Yes but if we want to convince engine maintainers to drop AndroidGL or at least to have a look at our stuff, we need Android 2.0
20130327 20:05:13 <gouessej> support
20130327 20:06:52 <sgothel> not necessarily .. but thats a long discussion, sure we can pick that work up as low prio later ; if you like, you can do it w/ your device ofc.
20130327 20:08:12 <sgothel> let's say at least in the professional field, it doesn't matter, i.e. CAD visualization etc - companies simply buy devices
20130327 20:08:21 <gouessej> Yes. In no way I will bother you with that
20130327 20:08:21 <sgothel> for games .. yes, it may matter
20130327 20:08:49 <sgothel> .. oh, I will help you ofc if you do it
20130327 20:09:36 <gouessej> Ok the fourth int or unsigned int contains the number of channels :)
20130327 20:13:48 <gouessej> actually... no
20130327 20:13:49 <gouessej> :s
20130327 20:14:44 <gouessej> I use that
20130327 20:14:49 <gouessej> is it reliable? http://www.sonicspot.com/guide/wavefiles.html
20130327 20:15:51 <sgothel> I don't know - looks good though. why not taking the source code from paul lamb's library ?
20130327 20:17:47 <gouessej> I just want to be able to maintain the source code
20130327 20:17:55 <gouessej> If I don't understand it, it is not worth
20130327 20:18:05 <gouessej> Who will fix its bug?
20130327 20:18:39 <sgothel> sure - good reference. I would start w/ something existing if readable, then digg into it and use specs.
20130327 20:20:14 <gouessej> Do we have a method to read unsigned short in little endian?
20130327 20:20:48 <sgothel> ULEB stuff .. plenty .. one sec
20130327 20:23:13 <sgothel> package jogamp.common.os.elf.IOUtils (but it uses nativeOrder) .. cont ..
20130327 20:26:26 <gouessej> thanks
20130327 20:26:43 <gouessej> Paul Lamb's Wav codec uses Java Sound :s
20130327 20:28:27 <gouessej> Your class is not visible :s
20130327 20:29:18 <sgothel> just an internal .. and actually not very performant due to ByteBuffer wrapping .. ok for ELF parsing, not so OK for streaming
20130327 20:32:55 <sgothel> PngHelperInternal uses network byte order == Big Endian
20130327 20:33:12 <sgothel> handles EOS .. using input stream
20130327 20:33:39 <sgothel> jogamp.opengl.util.pngj.PngHelperInternal
20130327 20:33:40 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20130327 20:33:59 <sgothel> line 55 for example
20130327 20:38:30 <gouessej> ok
20130327 20:39:32 <gouessej> it should work with little endian
20130327 20:39:45 <gouessej> someone else will have to adapt the code for big endian
20130327 20:39:54 <sgothel> then the other way around ..
20130327 20:40:03 <sgothel> i.e. the file header is either - or ..
20130327 20:41:17 <sgothel> so if wav header int16 is LE: b2 << 8 | b1
20130327 20:41:36 <gouessej> I clean up this mess and I commit
20130327 20:42:35 <sgothel> .. always w/ a unit test - please :)
20130327 20:43:54 <gouessej> Where should I put the unit test?
20130327 20:44:00 * ebrayet (~Adium@anon) has joined #jogamp
20130327 20:45:14 <sgothel> ../../../joal/src/test/com/jogamp/openal/test/junit/util for WavParser maybe ?
20130327 20:46:41 <gouessej> ok
20130327 20:55:33 <gouessej> I think that if RIFF is read in the reverse order at the very beginning, we can consider the stream is in bug endian
20130327 20:55:41 <gouessej> big endian
20130327 20:57:40 <sgothel> all data values are stored in Little-Endian
20130327 20:58:02 <sgothel> so the signatures and header data must be LE
20130327 21:01:20 <gouessej> Motorola uses RIFX instead of RIFF
20130327 21:02:48 <gouessej> http://stackoverflow.com/questions/1111539/is-the-endianness-of-format-params-guaranteed-in-riff-wav-files
20130327 21:03:45 <sgothel> I guess RIFF support for now is good enough .. hence LE
20130327 21:11:18 <gouessej> ok
20130327 21:12:20 <gouessej> I have just committed my changes. Maybe Xerxes can have a look at my code and add a unit test, I have added some references too. I'm tired
20130327 21:13:01 <sgothel> sure .. good night, I will git the sack soon as well
20130327 21:13:19 <xranby_ac100> sgothel: ping
20130327 21:13:20 <xranby_ac100> a quick status update
20130327 21:13:32 <sgothel> pong
20130327 21:13:41 <gouessej> yes Xerxes
20130327 21:13:57 <gouessej> Do you have some time to look at my small changes?
20130327 21:14:07 <xranby_ac100> i belive processing have simply forgotten to do an .validate() after adding the NEWTCanvasAWT to the visiable fame
20130327 21:14:14 <gouessej> lol
20130327 21:14:20 <gouessej> good catch
20130327 21:14:46 <sgothel> Julien: Pls fix your eclipse editor setup :) - good to add references - nice
20130327 21:15:00 <gouessej> oh sorry I have forgotten
20130327 21:15:04 <gouessej> I do it now
20130327 21:15:18 <sgothel> 4 spaces
20130327 21:15:40 <gouessej> Where can I change that?
20130327 21:15:57 <sgothel> @Xerxes: It's not so easy .. all those mods shall happen on AWT-EDT
20130327 21:16:14 <gouessej> "Insert spaces for tabs"?
20130327 21:16:17 <sgothel> otherwise it may break on one platform .. or the other
20130327 21:16:24 <xranby_ac100> gouessej: is it joal changes or the ardor3d changed?
20130327 21:16:38 <gouessej> joal first
20130327 21:16:51 <sgothel> Eclipse: Text Editors
20130327 21:16:59 <gouessej> Ardor3D uses getMaximum(true) now
20130327 21:16:59 <sgothel> tab width 4
20130327 21:17:02 <sgothel> use spces
20130327 21:17:45 <gouessej> The closest option is "Insert spaces for tabs"
20130327 21:17:49 <sgothel> yup
20130327 21:19:35 <xranby_ac100> sgothel: ican ad a guard in the processing code to check that it runs initialization on the AWT-EDT
20130327 21:20:06 <sgothel> maybe we can have a chat w/ AC ..
20130327 21:20:17 <sgothel> sure
20130327 21:20:27 <gouessej> You shouldn't assume that the realization succeeded too...
20130327 21:20:42 <sgothel> otherwise, if lazy .. the caps in the GLEventListener.init(..)
20130327 21:20:44 <gouessej> like I do in Ardor3D
20130327 21:20:54 <xranby_ac100> inside the jogl testcase that i prepared today all awt initialization is done inside SwingUtilities.invokeAndWait blocks to be sure
20130327 21:21:09 <xranby_ac100> i managed to replicate the bug that is end up with a null capabilities
20130327 21:21:11 <sgothel> right .. so the caps assignment in GLEventListener.init(..) _and_ mods on AWT-EDT
20130327 21:21:31 <sgothel> great
20130327 21:21:36 <xranby_ac100> adding a .validate() fixed the bug
20130327 21:21:46 <sgothel> but YMMV ..
20130327 21:21:53 <xranby_ac100> tomorrow i will try the same fix on the processing source
20130327 21:22:04 <sgothel> i.e. maybe it's done at a later time / resize .. etc
20130327 21:22:09 <xranby_ac100> right now i only have a fixed testcase
20130327 21:22:23 <sgothel> so the caps assignment in GLEventListener.init(..) _and_ mods on AWT-EDT
20130327 21:22:24 <gouessej> Calling validate() is really mandatory when adding a NewtCanvasAWT into a visible frame???
20130327 21:22:50 <sgothel> we have something in our test as well, yes .. invalidate() change validate()
20130327 21:24:33 <sgothel> TestParenting01cSwingAWT.testWindowParenting05ReparentAWTWinHopFrame2Frame line 326
20130327 21:24:47 <sgothel> but all on AWT-EDT, trust me :)
20130327 21:25:14 <xranby_ac100> gouessej: the processing sourcecode tries to probe and store the choosen capabilites into a local variable directly after adding the NewtCanvasAWT. for this to succeed the NEWT GLWindow have to had initialized fully
20130327 21:25:29 <sgothel> and still .. I would do the chosenCaps on GLEL.init() .. it's guaranteed there
20130327 21:25:54 <sgothel> Xerxes: you grok what I mean ?
20130327 21:26:44 <xranby_ac100> sgothel: to be honest, i dont
20130327 21:26:57 <xranby_ac100> but i try wrap my mind around this
20130327 21:27:31 <xranby_ac100> lets seeif i find the GLEL.init inside the PGL.java code
20130327 21:27:44 <sgothel> PGL has a GLEventListener added
20130327 21:27:59 <sgothel> in that init method, you can safely fetch the chosenCaps ..
20130327 21:28:17 <sgothel> on the bottom :)
20130327 21:29:26 <xranby_ac100> ok got it.. at the place where it
20130327 21:29:27 <xranby_ac100> throw new RuntimeException("No framebuffer objects available"); ;)
20130327 21:30:01 <sgothel> yup
20130327 21:30:16 <xranby_ac100> good then i will try fix this by rewriting that init
20130327 21:30:51 <xranby_ac100> and make it use the checks inside GLContext to look for the half FBO functionality
20130327 21:30:59 <sgothel> great
20130327 21:31:04 <xranby_ac100> that seems sufficient for most of the processing code
20130327 21:31:15 <sgothel> yup, AC is happy w/ it on OSX
20130327 21:31:22 <xranby_ac100> and will add support for opengl es 2 out of the box
20130327 21:31:42 <xranby_ac100> sgothel: take a look at the top of the PGL.java
20130327 21:31:53 <xranby_ac100> looks like we have to check AWT operation on linux
20130327 21:32:17 <sgothel> line ?
20130327 21:33:03 <xranby_ac100> search for } else if (PApplet.platform == PConstants.LINUX) {
20130327 21:33:07 <xranby_ac100> its the first static initialization block
20130327 21:33:29 <sgothel> his use of AWT is extremely broken :)
20130327 21:33:37 <sgothel> read above .. :)
20130327 21:33:44 <xranby_ac100> around line 131
20130327 21:33:50 <sgothel> jaja
20130327 21:34:17 <sgothel> mods on AWT are like frame.add(..) -> do it on AWT-EDT
20130327 21:34:42 <xranby_ac100> oh well if it make him move all platforms to newt
20130327 21:34:48 <sgothel> haha
20130327 21:34:59 <sgothel> good one
20130327 21:35:43 <xranby_ac100> so if newt had a way to customize cursors on windows processing would abandom awt ;)
20130327 21:36:32 <sgothel> resistance is futile :)
20130327 21:37:02 <gouessej> yeah move to NEWT
20130327 21:37:26 <gouessej> Changing the cursor is not trivial to implement
20130327 21:37:36 <gouessej> but it's doable
20130327 21:40:09 <xranby_ac100> gouessej: i will sleep now, i will check your joal pullrequests and your joal repo tomorrow
20130327 21:40:21 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=676
20130327 21:40:24 <gouessej> yes there is no hurry
20130327 21:40:29 <sgothel> good night
20130327 21:41:05 <sgothel> .. a unit test as a 1st validation is always welcome
20130327 21:41:42 <xranby_ac100> good night
20130327 21:44:31 <gouessej> good night
20130327 21:44:49 * gouessej (5279cd8e@anon) Quit (Quit: Page closed)
20130327 22:14:02 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20130327 23:21:32 * xranby (~xranby@anon) Quit (Ping timeout: 256 seconds)
20130327 23:21:52 * xranby (~xranby@anon) has joined #jogamp
20130327 23:26:02 * xranby_ac100 (~xranby@anon) Quit (Remote host closed the connection)
20130328 00:31:14 * DemoscenePassiv (~Lutsche@anon) Quit (Ping timeout: 245 seconds)
20130328 05:06:41 -CatOut- Continue @ http://jogamp.org/log/irc/jogamp_20130328050641.html