#jogamp @ irc.freenode.net - 20150415 05:05:31 (UTC)


20150415 05:05:31 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20150414050531.html
20150415 05:05:31 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20150415050531.html
20150415 05:43:45 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150415 05:45:54 * badshah400 (~badshah40@anon) has joined #jogamp
20150415 05:54:49 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150415 05:59:45 * badshah400 (~badshah40@anon) has joined #jogamp
20150415 06:13:28 * jvanek (jvanek@anon) has joined #jogamp
20150415 07:11:17 * monsieur_max (~maxime@anon) has joined #jogamp
20150415 07:45:31 * eclesia (~husky@anon) has joined #jogamp
20150415 07:45:36 <eclesia> good morning
20150415 09:43:14 * phao (~phao@anon) has joined #jogamp
20150415 10:00:13 * Schrostfutz_ is now known as Schrostfutz
20150415 12:20:33 <zubzub_> hmmm
20150415 12:20:40 <zubzub_> newt does not seem to support touch input
20150415 12:20:57 <zubzub_> only as mouse input
20150415 12:21:02 <zubzub_> but mouse != touch
20150415 12:35:34 <zubzub_> 22:54 < Eclesia> dirtytrixzzster: what have you made which could be interesting ?
20150415 12:35:39 <zubzub_> the one thing
20150415 12:35:45 <zubzub_> that I have never ever seen so far
20150415 12:35:47 <zubzub_> in any game
20150415 12:36:03 <zubzub_> is proper water collission with a shore/rock/whatever
20150415 12:36:29 <zubzub_> it's always straight lines (intersection fo water with some object)
20150415 12:37:58 <zubzub_> which totally kills the nice water effects like reflection, ripples at whatnot
20150415 12:38:51 <zubzub_> "ooh look what a nice pool"
20150415 12:38:56 <zubzub_> "too bad it's shaped like a hexagon"
20150415 12:49:59 * zubzub_ is now known as zubzub
20150415 13:08:40 * gouessej (5ee4b442@anon) has joined #jogamp
20150415 13:09:07 <gouessej> Hi
20150415 13:09:38 <gouessej> zubzub: Which OS do you use? Touch input is supported only under Android in NEWT, isn't it?
20150415 13:12:07 * gouessej (5ee4b442@anon) Quit (Client Quit)
20150415 13:23:24 <zubzub> linux
20150415 13:23:31 <zubzub> I would like to use libinput
20150415 13:23:38 <zubzub> for my custom newt backend
20150415 13:24:00 <zubzub> thing is that newt makes no differentiation between mouse or touch
20150415 13:25:37 <eclesia> zubzub: when you will have the nice water effect, I'll gladly accept it in the project
20150415 13:46:21 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150415 13:47:01 * monsieur_max (~maxime@anon) has joined #jogamp
20150415 14:04:18 * gouessej (5ee4b442@anon) has joined #jogamp
20150415 14:04:28 <gouessej> zubzub: http://jogamp.org/deployment/jogamp-next/javadoc/jogl/javadoc/com/jogamp/newt/event/MouseEvent.PointerType.html#TouchScreen
20150415 14:05:27 <gouessej> zubzub: The pointer type allows to make a distinction between mouse and touch input
20150415 14:05:52 * gouessej (5ee4b442@anon) Quit (Client Quit)
20150415 14:07:16 <zubzub> oic
20150415 14:07:21 <zubzub> ok cool :)
20150415 14:08:34 <zubzub> once I get egl kms with libinput implemented in newt, I'll see if I can get something wayland related implemented
20150415 14:08:58 <zubzub> they'll most likely live as seperate newt drivers as they depend on jna
20150415 14:24:37 <zubzub> I'm surprised nobody else implemented egl kms for jogl
20150415 14:24:43 <zubzub> it seems so trivial
20150415 14:24:51 <zubzub> and perfect use case for embedded/kiosk stuff
20150415 14:25:39 <zubzub> it's basically the same ad the rpi back-end, just slightly different setup of the egldisplay and egl surface
20150415 15:00:38 <bruce-> eclesia: in a good radial blur you should not be able to see the individual steps
20150415 15:01:04 <bruce-> eclesia: is this with ping-ponging between two buffers?
20150415 15:01:17 <eclesia> number of sampling steps can be increased
20150415 15:02:38 <bruce-> the trick is to perform radial blur with read from buffer A, write to buffer B, then from B to A, this doubles the amount of steps performed in each iteration
20150415 15:02:47 <eclesia> ping-ponging ? it's a post process , with one texture as input that's all
20150415 15:03:54 <eclesia> why not just increase the number of steps and do it in one pass then ?
20150415 15:05:17 <bruce-> because you need way more texture reads
20150415 15:18:59 <rmk0> they call it a "separable blur": http://sunsetlakesoftware.com/2013/10/21/optimizing-gaussian-blurs-mobile-gpu
20150415 15:19:11 <rmk0> i do the horizontal and vertical passes separately in my own renderer
20150415 15:21:06 <eclesia> I have that for gaussian blur, but for radial that's ... different
20150415 15:21:38 <rmk0> oh... missed the "radial" part
20150415 15:21:41 <rmk0> \o\
20150415 15:52:09 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150415 15:59:36 * eclesia (~husky@anon) has left #jogamp
20150415 16:19:42 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20150415 16:30:26 * Eclesia (~eclesia@anon) has joined #jogamp
20150415 16:41:03 * monsieur_max (~maxime@anon) has joined #jogamp
20150415 17:08:09 <Eclesia> zubzub: isn't that what you wanted : http://lelombrik.net/71505 ?
20150415 17:16:09 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150415 17:47:52 * badshah400 (~badshah40@anon) has joined #jogamp
20150415 18:08:38 <zubzub> Eclesia: well that's another way of solving it
20150415 18:09:04 <zubzub> looks like the make the entire water surface '3d'
20150415 18:09:22 <zubzub> most games these days just apply some nice texture and reflection on a 2d water surface
20150415 18:09:32 <zubzub> which is ok for me
20150415 18:09:44 <zubzub> except that this 2d surface collides with the shore in a straight line
20150415 18:10:20 <zubzub> so if they could fix that straight line so as it looks like the water is going up and down, then that would fix my 'problem'
20150415 18:29:43 * dirtytrixzzster (~gingerb3a@anon) has joined #jogamp
20150415 19:32:52 * dirtytrixzzster (~gingerb3a@anon) Quit (Quit: Konversation terminated!)
20150415 19:40:36 <zubzub> sgothel: why does the rpi newt display driver first creates a 'fake' graphics device (a 'default' one)
20150415 19:40:47 <zubzub> and only goes on creating the real one when a window is created?
20150415 19:40:58 <zubzub> seperately
20150415 19:41:29 <sgothel> (back from the lot :) ..
20150415 19:41:34 <sgothel> err ..
20150415 19:42:06 <sgothel> probably b/c we create native resources lazily ..
20150415 19:42:16 <sgothel> so the native device is an EGL display device
20150415 19:42:28 <sgothel> but would need to check code ..
20150415 19:42:42 <sgothel> I glanced over discussion above .. (EGL/JNA ..)
20150415 19:42:51 <sgothel> why not creating the binding w/ GlueGen?
20150415 19:43:10 <sgothel> it shall be fit enough to digest vanilla headers
20150415 19:43:11 <zubzub> because I'm more familiar with jna :p
20150415 19:43:30 <zubzub> I'm mostly still in the poc phase
20150415 19:43:34 <sgothel> then - I have to re-do the work w/ GlueGen :-/
20150415 19:43:39 <zubzub> yeah
20150415 19:43:40 <zubzub> sorry :p
20150415 19:43:44 <sgothel> well
20150415 19:44:03 <sgothel> probably possible to have your interface definitions .. and their usage as a template
20150415 19:44:21 <sgothel> and create GlueGen'ed impl. matching it .. almost
20150415 19:44:34 <zubzub> well it's actually very simple
20150415 19:44:47 <zubzub> the whole kms drm stuff
20150415 19:44:50 <sgothel> if you have test code for your wayland binding .. and the binding .. I can use that
20150415 19:45:01 <sgothel> kms (kernel mode setting) ?
20150415 19:45:03 <zubzub> I have a client example
20150415 19:45:12 <zubzub> yes
20150415 19:45:32 <zubzub> http://virtuousgeek.org/blog/index.php/jbarnes/2011/10/31/writing_stanalone_programs_with_egl_and_
20150415 19:45:46 <sgothel> so: kms + EGL/kms ?
20150415 19:45:52 <zubzub> yes
20150415 19:46:05 <sgothel> i.e. EGL Native<Type> for kms .. nice
20150415 19:46:12 <sgothel> all fullscreen then, I assume
20150415 19:46:20 <zubzub> basically a generic egl setup for all drivers that support kms
20150415 19:46:21 <zubzub> yeah
20150415 19:46:36 <sgothel> yes, a simple binding like Android, RPI, or the other one .. K?
20150415 19:46:53 <sgothel> (i.e. all EGL based)
20150415 19:47:13 <zubzub> well android and dispman don't support kms iirc
20150415 19:47:25 <zubzub> a whole lot of other drivers do though
20150415 19:47:31 <sgothel> no, I mean the native display device handles are EGL
20150415 19:47:58 <zubzub> yeah
20150415 19:48:09 <zubzub> but what is your question?
20150415 19:48:31 <sgothel> .. just creating a mindmap / big-picture ..
20150415 19:48:48 <sgothel> are those things available in parallel .. or either-or
20150415 19:48:59 <sgothel> i.e. X11/EGL, KMS/EGL, .. etc etc
20150415 19:49:07 <zubzub> xor
20150415 19:49:14 <sgothel> if exclusive .. then fine, good
20150415 19:49:37 <zubzub> you basically take exclusive ownership of the graphics card afaik
20150415 19:49:37 <sgothel> jogamp.newt.driver.kd <- the other pure EGL thingy
20150415 19:49:47 <sgothel> ay
20150415 19:50:13 <sgothel> would be great if you can track this w/ bug reports
20150415 19:50:26 <sgothel> and add links to your resources (code, tests)
20150415 19:50:57 <zubzub> well I'm not sure how I can unit test a kms/egl newt driver setup
20150415 19:50:58 <sgothel> my schedule: Now - April 28th: renovations; April 28th: Moving
20150415 19:51:04 <zubzub> but the kms/egl newt driver is pretty stand alone
20150415 19:51:10 <zubzub> if I get it finished
20150415 19:51:16 <sgothel> April 29th ... setup .. (maybe loss of internet for a few days ..
20150415 19:51:17 <zubzub> so I can create a bug and link the source
20150415 19:51:24 <sgothel> back in a bit .. dinner, laters
20150415 19:51:27 <zubzub> bye
20150415 20:02:17 <zubzub> sgothel: the kd driver does not seem to be included in 2.3.1 (?)
20150415 20:02:22 <zubzub> on maven
20150415 20:15:22 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150415 20:17:27 * badshah400 (~badshah40@anon) has joined #jogamp
20150415 20:22:24 * badshah400 (~badshah40@anon) Quit (Ping timeout: 276 seconds)
20150415 20:37:37 * badshah400 (~badshah40@anon) has joined #jogamp
20150415 20:46:27 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20150415 21:01:23 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150415 23:35:21 * rmk0_ (~rmk0@anon) has joined #jogamp
20150415 23:38:27 * rmk0 (~rmk0@anon) Quit (Ping timeout: 250 seconds)
20150415 23:39:24 * phao (~phao@anon) Quit (Ping timeout: 248 seconds)
20150416 00:16:53 <badshah400> Hello helpful jogamp devs! I would be much obliged if someone could help me bring scirenderer up to scratch for use with jogl 2.3.x (I am trying to fix http://forge.scilab.org/index.php/p/scirenderer/issues/1545/ where the use of old deprecated symbols shows up)
20150416 00:17:44 <badshah400> I am not a scilab/scirenderer dev, but I package scilab for openSUSE, and since upstream seems not too active, I might have to fix this myself
20150416 00:19:03 <badshah400> If there is any guide on how to port e.g., com.jogamp.opengl.util.awt.Screenshot and com.jogamp.opengl.GLPbuffer to new jogl, that would be of significant help. Thanks!
20150416 00:19:11 <sgothel> GLPbuffer has been removed, use an offscreen GLAutoDrawable ..
20150416 00:19:30 <sgothel> let me find the git commit
20150416 00:19:56 <sgothel> jogl ec2d94ca26ddab8ec67135ebc5f2d0a43f6a4c25
20150416 00:20:39 <sgothel> http://jogamp.org/git/?p=jogl.git;a=commit;h=ec2d94ca26ddab8ec67135ebc5f2d0a43f6a4c25
20150416 00:21:07 <sgothel> http://jogamp.org/git/?p=jogl.git;a=blobdiff;f=src/test/com/jogamp/opengl/test/junit/jogl/acore/TestPBufferDeadlockAWT.java;h=ec377662ff1b5a61295e965161445289f5301260;hp=7138513d524935a02a1fe1e067d9f19fc413b7c5;hb=ec2d94ca26ddab8ec67135ebc5f2d0a43f6a4c25;hpb=efa5f1110725d41b7ea58010fe34b2a8aacd185b
20150416 00:21:41 <sgothel> ^^ how to replace GLPbuffer w/ GLAutoDrawable
20150416 00:22:29 <sgothel> and Screenshot w/ AWTGLReadBufferUtil
20150416 00:23:20 <sgothel> should fix all your issues
20150416 00:44:52 <badshah400> sgothel: thank you very much, I will ping you back if I face difficulties, but this is a big help
20150416 01:39:51 * phao (~phao@anon) has joined #jogamp
20150416 02:55:42 <badshah400> sgothel: if you are still around, I think I am almost there with the fixes. there is this one problem still lingering: the use of GLAutoDrawable.getWidth()
20150416 02:56:05 <sgothel> -> getSurfaceWidth()
20150416 02:56:07 <badshah400> Should I replace it with GLAutoDrawable.getSurfaceWidth() and so on for getHeight() too?
20150416 02:56:11 <sgothel> :)
20150416 02:56:14 <badshah400> sgothel: ah, thanks
20150416 02:56:28 <sgothel> this is Hi-DPI related ..
20150416 02:57:02 <badshah400> sgothel: kind of fig it out from the extensive javadoc at jogamp's site (comparing v2.0.2 to v2.3.1)
20150416 02:57:35 <sgothel> great (at least somebody is reading it :)
20150416 02:59:05 <badshah400> sgothel: yes, thanks a lot for such extensive documentation, really useful
20150416 03:05:01 <sgothel> can you push your changes upstream to Scilab ?
20150416 03:13:46 <badshah400> sgothel: I will attach it in the bug report, but I don't have commit rights, so someone will have to review and push it himself/herself, which is good I think to get a review
20150416 03:26:57 <badshah400> sgothel: here goes: http://forge.scilab.org/index.php/p/scirenderer/issues/view/attachment/313/0001-Port-to-jogl-2.3.patch
20150416 03:28:04 <badshah400> finally got it building, thanks to a bunch of help from you
20150416 03:28:25 <sgothel> you are vert welcome
20150416 03:28:29 <sgothel> *very* :)
20150416 03:40:20 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150416 03:43:09 * badshah400 (~badshah40@anon) has joined #jogamp
20150416 03:49:31 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150416 03:50:21 * badshah400 (~badshah40@anon) has joined #jogamp
20150416 04:08:51 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150416 04:10:34 * badshah400 (~badshah40@anon) has joined #jogamp
20150416 05:05:31 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20150416050531.html