JogAmp IRC Chat Logs

#jogamp @ irc.freenode.net - 20130425 05:06:01 (UTC)

20130425 05:06:01 -CatOut- Previous @ http://jogamp.org/log/irc/jogamp_20130424050547.html
20130425 05:06:01 -CatOut- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130425050601.html
20130425 05:43:19 * [Mike] (~Mike]@anon) Quit ()
20130425 05:55:54 * DemoscenePassiv (~Lutsche@anon) has joined #jogamp
20130425 06:36:55 * DemoscenePassiv (~Lutsche@anon) Quit (Ping timeout: 256 seconds)
20130425 06:55:00 <sgothel> real 2m19.549s
20130425 06:55:00 <sgothel> user 3m38.681s
20130425 06:55:00 <sgothel> sys 0m8.599s
20130425 06:55:08 <sgothel> local maven deployment on jogamp ..
20130425 06:55:32 <sgothel> using a dummy gpg key and file:/// protocol :)
20130425 07:08:00 <sgothel> http://jogamp.org/deployment/archive/master/gluegen_666-joal_432-jogl_981-jocl_788-signed/jogl-test-applets.html
20130425 07:54:05 <xranby> kudos! jake2 now have sound using joal tested using http://jogamp.org/deployment/maven/org/jogamp/jogl/jogl-all-main/2.0.2-rc-20130425/
20130425 07:58:14 <sgothel> yup .. some stupid mistakes :)
20130425 07:59:01 <sgothel> http://forum.jogamp.org/Re-Our-Road-to-2-0-2-Update-3-td4029026.html
20130425 08:01:25 <sgothel> I guess we need to have expiration dates for all the quirks in our code :)
20130425 08:03:04 <sgothel> now I can update the ATI driver on that buggy test machine .. hehe
20130425 11:23:08 * jk4_ (~jk4@anon) has joined #jogamp
20130425 11:29:17 * jk4 (~jk4@anon) Quit (*.net *.split)
20130425 12:18:50 * rmk0_ (~rmk0@anon) has joined #jogamp
20130425 12:21:23 * rmk0 (~rmk0@anon) Quit (Ping timeout: 245 seconds)
20130425 17:16:25 * [Mike] (~Mike]@anon) has joined #jogamp
20130425 18:35:28 * jk4_ (~jk4@anon) Quit (Remote host closed the connection)
20130425 18:41:51 * [Mike] (~Mike]@anon) Quit ()
20130425 18:51:13 * DemoscenePassiv (~Lutsche@anon) has joined #jogamp
20130425 18:57:56 * [Mike] (~Mike]@anon) has joined #jogamp
20130425 19:36:05 * gouessej (5279c7b1@anon) has joined #jogamp
20130425 19:36:08 <gouessej> Hi
20130425 19:36:19 <gouessej> Two methods have been removed from GLPBuffer
20130425 19:36:29 <gouessej> they were used in Ardor3D :(
20130425 19:36:55 <gouessej> bindTexture() and releaseTexture()
20130425 19:37:02 <sgothel> Hi Julien .. yes, the floating point texture .. - oh
20130425 19:37:07 <sgothel> just those ?
20130425 19:37:12 <sgothel> hmm
20130425 19:37:26 <sgothel> ah .. right .. no more rendering to texture
20130425 19:37:36 <sgothel> would be best if you could use FBO offscreen drawable
20130425 19:37:52 <sgothel> GLPBuffer is a candidate for removal
20130425 19:38:26 <sgothel> the PBuffer offscreen drawable works as well, but not w/ render-to-texture
20130425 19:38:37 <gouessej> I know, I planned to switch to javax.media.opengl.GLOffscreenAutoDrawable
20130425 19:38:47 <gouessej> it's even written in the source code
20130425 19:38:47 <sgothel> RTT and floating-texture caused a huge mess in the GLContextImpl* classes
20130425 19:39:33 <sgothel> we actually need at least a demo for floating point textures .. as a template for users .. hmm
20130425 19:41:06 <gouessej> I hadn't planned to use GLOffscreenAutoDrawable now, Ardor3D is close to the final release
20130425 19:41:29 <sgothel> we are close to 2.0.2 .. as well
20130425 19:41:45 <sgothel> yeah, GLCOntext must be interchangable regardless of the GLDrawable
20130425 19:41:54 <sgothel> otherwise GLEventStateListener would not work
20130425 19:42:04 <sgothel> err GLEventListenerState ..
20130425 19:43:08 <sgothel> if you like, I can help using GLOffscreenAutoDrawable
20130425 19:44:00 <sgothel> btw - I got some t-shirts and mugs fro Cafepress .. nice stuff
20130425 19:44:34 <gouessej> ok yes
20130425 19:44:39 <gouessej> I have some questions
20130425 19:44:43 <sgothel> when we have new design .. we will make a bunch ..
20130425 19:44:46 <sgothel> go ahead
20130425 19:45:20 <gouessej> What is the most straightforward way of replacing calls of GLPBuffer.bindTexture() and releaseTexture by using GLOffscreenAutoDrawable?
20130425 19:45:40 <sgothel> let me find a unit test ..
20130425 19:48:07 <gouessej> thanks
20130425 19:49:12 <sgothel> TestFBOOnThreadSharedContext1DemoES2NEWT
20130425 19:49:32 <sgothel> final GLOffscreenAutoDrawable.FBO fbod1 = (GLOffscreenAutoDrawable.FBO)
20130425 19:49:32 <sgothel> factory.createOffscreenAutoDrawable(null, fbodCaps, null, glWindow.getWidth(), glWindow.getHeight(), glWindow.getContext());
20130425 19:51:06 <sgothel> fbod1.getTextureBuffer(GL.GL_FRONT).getName()
20130425 19:51:45 <gouessej> ok I'm going to look at this one
20130425 19:52:10 <gouessej> I have already modified the source code to use GLOffscreenAutoDrawable, there are just these 2 calls
20130425 19:52:19 <sgothel> great
20130425 19:52:42 <sgothel> with that texure name .. you can bind it, release is obvious
20130425 19:57:13 <gouessej> but this method is only available in GLOffscreenAutoDrawable.FBO, not in GLOffscreenAutoDrawable
20130425 19:57:54 <sgothel> yes .. since only FBO has a texture, not pbuffer. pbuffer render to texture is platform dependent - not supported everywhere
20130425 19:58:25 <sgothel> is your use case pbuffer explicit ?
20130425 19:59:01 <sgothel> i.e. not just any offscreen, but pbuffer is requested
20130425 19:59:05 <gouessej> Ok I will use GLOffscreenAutoDrawable.FBO :(
20130425 19:59:53 <sgothel> I hear you (slow FBO impl. on some drivers) .. but this should be no issue w/ current stuff
20130425 20:00:12 <gouessej> The cast will fail if it is not supported
20130425 20:00:30 <sgothel> ofc you can query [upfront]
20130425 20:00:43 <sgothel> i.e. GLCaps ..
20130425 20:01:22 <sgothel> it is not supported on .. err .. let's say .. GL 1.1 stuff etc, or old Mesa ..
20130425 20:01:56 <gouessej> The concerned class is JoglPbufferTextureRenderer
20130425 20:02:10 <gouessej> If I can't use a PBuffer, it should be removed
20130425 20:02:33 <gouessej> There is already a class using FBOs
20130425 20:02:47 <sgothel> ok I will have a look at it later .. dinner time now - in such case, I would say, remove it - right
20130425 20:04:24 <gouessej> ok enjoy yourself
20130425 20:20:01 <sgothel> back ..
20130425 20:21:05 <sgothel> another thought .. if really desired, we could bring back functionality via GLDrawableImpl .. as a plugin/spi for GLContextImpl.. but I won't do that alone - since such functionality is really some ugly vendor related stuff ..
20130425 20:23:45 <gouessej> ok
20130425 20:24:09 <gouessej> for the moment, I use the former Windows implementation
20130425 20:24:19 <gouessej> I try to do something similar
20130425 20:25:02 <gouessej> I bind a texture to the offscreen drawable
20130425 20:25:08 <gouessej> it is better than nothing
20130425 20:25:59 <gouessej> is it a good idea?
20130425 20:27:06 <sgothel> 'I bind a texture to the offscreen drawable' ? WGL ? FBO GLAD ?
20130425 20:27:50 <gouessej> I use GLOffscreenAutoDrawable
20130425 20:28:08 <gouessej> This class is a fallback, I can't use FBOs
20130425 20:28:50 <sgothel> oh .. well then, I don't know how you can make the pbuffer use textures later on
20130425 20:29:14 <gouessej> then I have a big problem
20130425 20:29:21 <sgothel> it's more about time to do things here .. i.e. regarding my above idea of bringing things back ..
20130425 20:29:28 <sgothel> sometimes .. well, there is a cut-off ..
20130425 20:29:54 <sgothel> so you are saying pbuffer texture is essential to you ?
20130425 20:30:17 <gouessej> The renderer based on our competitor still has this feature
20130425 20:30:32 <gouessej> I can't remove this class
20130425 20:31:03 <gouessej> I have to check in which test cases it is used
20130425 20:31:16 <sgothel> in such case .. I would go the FBO GLOffscreenDrawable route and have it impl. in the new GLOffscreenDrawable.PBuffer mabe
20130425 20:31:18 <sgothel> maybe
20130425 20:31:41 <sgothel> imho high level apps should favor FBO anyways
20130425 20:31:47 <sgothel> (I know)
20130425 20:33:20 <gouessej> Ardor3D UI uses this class, I can't remove it, it would impact my project
20130425 20:35:34 <sgothel> but it only worked on .. Windows / Linux (I guess)
20130425 20:36:04 <gouessej> yes
20130425 20:36:28 <gouessej> I can't rewrite my own UI framework
20130425 20:37:46 <gouessej> is there a simple way of binding a texture to a PBuffer?
20130425 20:38:10 <sgothel> its on that WGL/GLX/EGL level ..
20130425 20:38:22 <sgothel> check commit which removes it .. one sec ..
20130425 20:38:54 <sgothel> http://jogamp.org/git/?p=jogl.git;a=commit;h=fd418a69eca7b8c1bb74244982305fc6004d0a52
20130425 20:39:13 <sgothel> http://jogamp.org/git/?p=jogl.git;a=blob;f=src/jogl/classes/jogamp/opengl/x11/glx/X11PbufferGLXContext.java;h=a34e050cd488726acf931bc20ded588af41ec13f;hb=3cc6fd350fe56ac5b344ee75cfa8bfe4dcc44f1b
20130425 20:39:31 <sgothel> ha - didn't even work on X11 :)
20130425 20:40:15 <gouessej> How is it possible? My crappy machine supports only OpenGL 1.3
20130425 20:40:50 <sgothel> ? It did not work on X11 see last link .. throws exception
20130425 20:42:00 <gouessej> yes I see
20130425 20:42:49 <gouessej> Ardor3D uses GL_EXT_framebuffer_object
20130425 20:43:04 <sgothel> all that was crippled anyways ..
20130425 20:43:11 <sgothel> good -> our FBO stuff :)
20130425 20:43:26 <sgothel> so I guess you safely can drop pbuffer-texture
20130425 20:43:46 <gouessej> yes
20130425 20:43:58 <gouessej> but I'm not satisfied
20130425 20:44:23 <gouessej> PBuffer can still be used in our competitor
20130425 20:44:26 <sgothel> who is when dealing w/ compromises all the time - me too :)
20130425 20:44:39 <sgothel> oh we can offer pbuffer .. plain and simple ones
20130425 20:44:47 <gouessej> This is not a priority but it would be nice to come back later and try to do something
20130425 20:44:54 <sgothel> but pbuffer-texture was only supported on Windows
20130425 20:45:27 <gouessej> Does GLOffscreenDrawable work only with FBOs?
20130425 20:45:27 <sgothel> then - as mentioned - it should be impl. in new GLOffscreen*Drawable like the FBO specialization for at least WGL, GLX and EGL
20130425 20:45:38 <sgothel> no - pbuffer as well, see unit tests
20130425 20:45:57 <sgothel> it's a nice solution .. can produce FBO, PBuffer and Pixmap/Bitmap
20130425 20:46:03 <sgothel> depending on GLCaps ..
20130425 20:46:35 <gouessej> I will probably have to find a solution to avoid RTT in the GUI
20130425 20:46:41 <gouessej> but not now
20130425 20:47:14 <sgothel> TestGLAutoDrawableDelegateOnOffscrnCapsNEWT, TestGLAutoDrawableFactoryES2OffscrnCapsNEWT and TestGLAutoDrawableFactoryGL2OffscrnCapsNEWT
20130425 20:48:09 <gouessej> ok
20130425 20:48:14 <sgothel> for RTT in a UI, I recommend to use FBObject directly or w/o the decoration of GLAD (which needs a dummy window (FBO)!)
20130425 20:49:34 <sgothel> the high level GLOffscreenAutoDrawable stuff is more for simple applications / tests etc IMHO - where you have no other context/drawable/surface
20130425 20:50:07 <sgothel> see: FBOMix2DemosES2 for example - using FBObject directly
20130425 20:50:52 <gouessej> I won't remove JoglPbufferTextureRenderer now but if we have no time to improve the support of PBuffers, I will have to remove it later
20130425 20:52:58 <gouessej> Maybe I will create a bug report about that but it is not a priority, I understand your position
20130425 20:54:26 <sgothel> still it was only working on windows - and IMHO it's more important to validate against latest builds to catch new stuff and regressions ASAP
20130425 20:54:32 <gouessej> Why would it be so difficult to redesign the previous source code to provide the same features in GLOffscreenAutoDrawable.PBuffer?
20130425 20:54:46 <gouessej> Yes you're right
20130425 20:54:50 <sgothel> its all about time ..!
20130425 20:55:02 <sgothel> it's not complicated .. just follow the FBO path ..
20130425 20:55:24 <sgothel> time == resources
20130425 20:55:52 <sgothel> plus ofc .. that today coding path would use FBO anyways .. I know :)
20130425 20:56:33 <sgothel> I am more inclined to add floating point textures to FBO .. etc
20130425 20:57:06 <gouessej> Ok but I don't want to take care of JavaFX now and I would feel better with no RTT
20130425 20:58:17 <sgothel> what has this todo w/ JavaFX ? Besides .. I don't like to take care of it anyways :)
20130425 21:00:57 <gouessej> I will need some hardware accelerated GUI for JFPSM next year
20130425 21:02:23 <gouessej> maybe a bit before
20130425 21:04:29 <sgothel> maybe I find time to enhance graph .. they toy project
20130425 21:04:36 <sgothel> *the* :)
20130425 21:04:51 <sgothel> whats JFPSM ?
20130425 21:05:11 <gouessej> This is a tiny WYSIWYG editor for FPS
20130425 21:05:20 <gouessej> it is very rudimentary
20130425 21:06:30 <gouessej> I need a powerful API for hardware accelerated GUIs, supported on tablets too
20130425 21:07:00 <gouessej> something as complete as Swing
20130425 21:07:49 <gouessej> I'm not a big fan of JavaFX but we would have to provide a limited support even though we don't want to spend some time in fully rewriting its pipeline
20130425 21:09:45 <sgothel> 'but we would have to provide' who says ?
20130425 21:10:18 <gouessej> I can just port Riven's stuff to draw JavaFX components into FBOs
20130425 21:10:46 <gouessej> Several JOGL adopters already asked me when we will support JavaFX
20130425 21:15:40 <sgothel> I just read about this in the forum ..
20130425 21:15:48 <sgothel> but not really much ..
20130425 21:16:12 <sgothel> and nobody really seems to desire it, otherwise I would have seen some more action .. communication etc
20130425 21:16:28 <sgothel> for now, looks like more a 'nice to have for free (no work)' issue :)
20130425 21:17:49 <gouessej> Ok but the graph UI is not able to compete with JavaFX yet
20130425 21:18:03 <gouessej> :(
20130425 21:18:08 <sgothel> besides .. is there a JavaFX 'open source' implementation which is complete ? :)
20130425 21:18:25 <sgothel> yes, it needs love - and it will also not compete w/ JavaFX :)
20130425 21:18:41 * rmk0_ (~rmk0@anon) Quit (Quit: leaving)
20130425 21:18:45 <sgothel> it follows the UI requirement list I created a while ago .. dunno .. forum etc
20130425 21:18:54 <gouessej> Some parts have been released but I'm not sure we can build it
20130425 21:19:09 * rmk0 (~rmk0@anon) has joined #jogamp
20130425 21:19:09 * rmk0 (~rmk0@anon) Quit (Changing host)
20130425 21:19:09 * rmk0 (~rmk0@anon) has joined #jogamp
20130425 21:22:52 <gouessej> Well it seems possible: http://ladstatt.blogspot.fr/2013/01/compile-openjfx-rt-on-macosx.html
20130425 21:23:38 <gouessej> https://wikis.oracle.com/display/OpenJDK/Building+OpenJFX
20130425 21:24:45 <sgothel> and it has all the nuts and bolts like Maja importer .. GL renderer ? or just some 2d pixel ops ?
20130425 21:25:46 <gouessej> I assume you know Prism and Quantum
20130425 21:29:36 <sgothel> the 1st I shall not know about :) - the latter is totally alien to me
20130425 21:29:59 <sgothel> both open sourced ?
20130425 21:30:34 <gouessej> yes
20130425 21:30:37 <gouessej> Glass too
20130425 21:31:12 <sgothel> cute bold name that is .. lol for a 'main thread only' renderer design - not really like massive parallel like quant* would suggest :)
20130425 21:31:38 <sgothel> yes - assuming they use 'Glass' :)
20130425 21:32:12 <sgothel> next tool - we may name it 'it' :)
20130425 21:32:19 <gouessej> lol
20130425 21:35:08 <gouessej> you're right: Exception in thread "Thread-5" java.lang.IllegalStateException: Not on FX application thread; currentThread = Thread-5
20130425 21:35:38 <sgothel> Oh well .. I know my 'friends' style of doing things :)
20130425 21:35:52 <sgothel> fast results - little useful :)
20130425 21:36:22 <gouessej> lol
20130425 21:37:18 <gouessej> Your 'friends' are doing the same mistakes with JavaFX 3D than they did with Java3D
20130425 21:38:45 <sgothel> TBH: Our thread safe impl. was quite expensive to realize and to maintain, i.e. time invested etc - so I understand them (AWT, SWT, OSX, ..) to simply skip such thing. However, today an application want multiple threads having access to a unique resource .. so I guess you have to walk this route if satisfying such natural requirement as it exist today.
20130425 21:39:58 <gouessej> Therefore what they did seems not natural for you, doesn't it?
20130425 21:40:42 <sgothel> it _was_ natural in .. lets say pre 2000 .. but now since multicore and real multithreading becomes natural .. it's just like missing reality :)
20130425 21:40:54 <sgothel> even the drivers utilize multithreading
20130425 21:41:41 <sgothel> it is crippling things .. thats what it is
20130425 21:42:01 <sgothel> look at our AWT/SWT unit test .. then you know what I mean :)
20130425 21:48:12 <gouessej> yes
20130425 21:48:42 <gouessej> but don't feel offended if you use JavaFX. I don't have enough time to write a GUI library
20130425 22:08:29 <gouessej> good night
20130425 22:08:34 * gouessej (5279c7b1@anon) Quit (Quit: Page closed)
20130425 22:18:54 * DemoscenePassiv (~Lutsche@anon) Quit (Ping timeout: 276 seconds)
20130425 22:18:56 <sgothel> No offense at all .. just a little bit fun & bashing - and hey, if somebody likes to add a binding to it - feel welcome
20130425 22:29:19 * jk4 (~jk4@anon) has joined #jogamp
20130426 03:01:32 * [Mike] (~Mike]@anon) Quit ()
20130426 04:20:54 * [Mike] (~Mike]@anon) has joined #jogamp
20130426 05:06:15 -CatOut- Continue @ http://jogamp.org/log/irc/jogamp_20130426050615.html