#jogamp @ irc.freenode.net - 20150817 05:05:30 (UTC)


20150817 05:05:30 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20150816050529.html
20150817 05:05:30 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20150817050530.html
20150817 05:19:21 * monsieur_max (~maxime@anon) has joined #jogamp
20150817 05:31:26 * jvanek (jvanek@anon) has joined #jogamp
20150817 05:45:16 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150817 06:10:06 * badshah400 (~badshah40@anon) has joined #jogamp
20150817 06:46:53 * monsieur_max (~maxime@anon) has joined #jogamp
20150817 06:53:56 * elect_ (~elect@anon) has joined #jogamp
20150817 06:54:19 <elect_> hi
20150817 07:17:22 * elect_ (~elect@anon) Quit (Quit: Leaving)
20150817 07:17:38 * elect (~elect@anon) has joined #jogamp
20150817 07:26:29 * elect (~elect@anon) Quit (Read error: Connection reset by peer)
20150817 07:29:03 * elect (~elect@anon) has joined #jogamp
20150817 07:50:30 * badshah400 (~badshah40@anon) Quit (Ping timeout: 240 seconds)
20150817 08:15:33 * eclesia (~husky@anon) has joined #jogamp
20150817 08:15:45 <eclesia> hi
20150817 08:18:45 <zubzub> heya
20150817 10:35:18 <xranby_> hi hi
20150817 12:10:22 * gouessej (5ee4b442@anon) has joined #jogamp
20150817 12:12:21 <gouessej> Hi
20150817 12:12:47 <gouessej> xranby_: I have a problem with the fat JAR
20150817 12:13:02 <xranby_> gouessej: what kind of problem?
20150817 12:13:08 <gouessej> xranby_: java.io.IOException: Could not determine a temporary executable directory
20150817 12:13:36 <gouessej> xranby_: The initialization of the cache seems to fail
20150817 12:13:53 <xranby_> gouessej: sounds like a new "corner case"
20150817 12:13:53 <gouessej> java.io.IOException: Could not determine a temporary executable directory at com.jogamp.common.util.IOUtil.getTempDir(IOUtil.java:1168) at com.jogamp.common.util.cache.TempFileCache.<clinit>(TempFileCache.java:80) at com.jogamp.common.util.cache.TempJarCache.initSingleton(TempJarCache.java:88) at com.jogamp.common.os.Platform$1.run(Platform.java:309) at java.security.AccessController.doPrivileged(Native Method) at com.jogamp.c
20150817 12:14:07 <xranby_> on which platform is this?
20150817 12:14:23 <gouessej> Windows 8
20150817 12:15:06 <xranby_> hmm.. IOUtil.getTempDir shoudl work regardless if it is used from a fat jar or not
20150817 12:15:09 <gouessej> It's critical, we can't make a release with such a bug
20150817 12:15:22 <xranby_> do non-fat jar work on windows 8 ?
20150817 12:15:24 <gouessej> Yes, you're probably right, I have to investigate
20150817 12:15:45 <gouessej> I haven't tested since I switched to 8 at work
20150817 12:17:35 <xranby_> i tested jogamp on windows 10 using a friends laptop yesterday
20150817 12:17:47 <xranby_> and that initialized fine
20150817 12:18:00 <xranby_> thus i cant see why windows 8 sould be a special case
20150817 12:18:19 <gouessej> I tested in an applet
20150817 12:18:43 <xranby_> ok. so the bug may be constrained to using the java applet plugin
20150817 12:19:24 <xranby_> in that case we need to know the plugin version and java version
20150817 12:19:38 <gouessej> It works in an applet here: http://jogamp.org/deployment/jogamp-current/jogl-applet-version-napplet.html
20150817 12:19:52 <gouessej> I use a fat JAR
20150817 12:20:08 <xranby_> ok so fatjar + applet == broken on windows 8?
20150817 12:20:34 <gouessej> the native libraries are in natives/windows-amd64 in the fat JAR containing the application (and JogAmp itself)
20150817 12:20:34 <xranby_> normal jar deployment + applet == ok on windows 8
20150817 12:20:38 <xranby_> ?
20150817 12:21:27 <gouessej> I have to enable the debug logs to see what happens
20150817 12:21:31 <xranby_> thank you
20150817 12:22:27 <gouessej> I can test with TUER as it isn't an applet
20150817 12:25:59 <gouessej> I obtain the same error
20150817 12:27:05 <gouessej> It's reproducible with OpenJDK 1.7 and Oracle Java 1.8 update 45
20150817 12:35:09 <gouessej> xranby_: How can I enable the debug logs in GlueGen?
20150817 12:36:27 <gouessej> "-Dnewt.debug=all -Dnativewindow.debug=all -Djogl.debug=all" isn't enough
20150817 12:36:33 <xranby_> debugging of gluegen is done using the -Djogamp.debug
20150817 12:38:53 <gouessej> xranby_: Thank you very much
20150817 12:41:02 <xranby_> if you only want to debug IOUtil then use -Djogamp.debug.IOUtil
20150817 12:50:41 <gouessej> I'm going to write a bug report
20150817 12:57:05 <gouessej> Maybe I can run the test fat JAR
20150817 13:01:40 <xranby_> gouessej: first try use the main method of GLWindow java -cp jogamp-fat.jar com.jogamp.newt.opengl.GLWindow
20150817 13:02:56 <gouessej> xranby_: It seems to work
20150817 13:03:59 <xranby_> ok in that case IOUtil work as long as it have permissions to write to the temp folder... how do we test the plugin.. hmm
20150817 13:04:11 <xranby_> maybe the plugin is more restricted hmm
20150817 13:04:20 <gouessej> It doesn't work even with TUER
20150817 13:04:34 <gouessej> with a plain bat file
20150817 13:05:11 <xranby_> java -cp jogamp-fat.jar com.jogamp.newt.opengl.GLWindow work while TUER does not work?
20150817 13:05:33 <gouessej> yes
20150817 13:06:43 <gouessej> I can compare the both
20150817 13:08:16 <gouessej> They have something in common
20150817 13:08:49 <gouessej> java.lang.IllegalArgumentException: Buffer is not an ELF Header at jogamp.common.os.elf.ElfHeaderPart1.<init>(ElfHeaderPart1.java:361) at jogamp.common.os.elf.ElfHeaderPart1.read(ElfHeaderPart1.java:343) at jogamp.common.os.PlatformPropsImpl.readElfHeaderPart1(PlatformPropsImpl.java:420) at jogamp.common.os.PlatformPropsImpl.access$100(PlatformPropsImpl.java:38) at jogamp.common.os.PlatformPro
20150817 13:09:47 <xranby_> i whonder why it tried to use the elf header reader on windoes
20150817 13:10:15 <xranby_> we use that functionality to find out what kind of natives to load on linux
20150817 13:10:55 <xranby_> gouessej: is this the fatal exception that prevented your application to run?
20150817 13:11:09 <gouessej> no
20150817 13:11:29 <gouessej> xranby_: There is something probably wrong on my side
20150817 13:12:30 <xranby_> fun.. if i download both the jogamp-fat.jar jogamp-fat-test.jar into the same directory
20150817 13:12:38 <xranby_> then i can run the tests usin
20150817 13:12:40 <xranby_> using
20150817 13:12:54 <xranby_> java -cp jogamp-fat-test.jar com.jogamp.openal.test.manual.Sound3DTest
20150817 13:13:05 <xranby_> thus by only specifying the jogamp-fat-test.jar jar
20150817 13:14:13 <xranby_> the jogamp jar loader is magic and amazing
20150817 13:18:57 <gouessej> Oops I don't use the very latest version
20150817 13:35:30 <gouessej> xranby_: I think that I have found the culprit
20150817 13:37:17 <gouessej> xranby_: Ant doesn't include bin/exe-windows-i586-268b.bin by default
20150817 13:38:25 <gouessej> xranby_: but the documentation doesn't mention "bin" in the default excludes: https://ant.apache.org/manual/dirtasks.html#defaultexcludes
20150817 13:41:14 <xranby_> so this is belongs to the evaluation / corner cases of https://jogamp.org/bugzilla/show_bug.cgi?id=1108
20150817 13:43:47 <xranby_> gouessej: it looks like you have a reproducing test environment for this bug
20150817 13:44:11 <gouessej> xranby_: Lol is it a good piece of news?
20150817 13:45:30 <xranby_> gouessej: i hope sgother can reappear and guide you how to test this bug
20150817 13:45:39 <xranby_> so that we may fix it proper
20150817 13:46:01 <xranby_> personally i do not know the details good enought
20150817 13:47:02 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20150817 13:57:21 <gouessej> xranby_: It seems to be my fault as I didn't want to include everything
20150817 13:57:39 <xranby_> gouessej: great that you have managed to get it working
20150817 13:58:28 <gouessej> xranby_: I should have carefully excluded OpenCL and the Oculus VR binding instead of trying to use fine-grained inclusion filters.
20150817 14:00:13 <xranby_> awesome you have then verified the fix for 1108
20150817 14:01:16 <xranby_> since you have tested one build with exe-windows-i586-268b.bin included that is working and one build without
20150817 14:03:48 <gouessej> it's still in progress...
20150817 14:04:20 <gouessej> Yes I can test
20150817 14:04:32 <gouessej> Now the JAR contains the ".bin" file
20150817 14:07:29 <gouessej> xranby_: It works :)
20150817 14:07:34 <xranby_> \o/
20150817 14:11:41 <gouessej> I'll have to fix it in my game when I'm back at my home
20150817 14:11:44 <elect> anyone here has any resource about acceleration structures on java?
20150817 14:13:06 <zubzub> structures in what sense?
20150817 14:13:17 <xranby_> elect: are you building a pgu raytracer?
20150817 14:13:21 <xranby_> gpu
20150817 14:13:39 <elect> maybe in the future
20150817 14:13:56 <elect> but right now I am just looking at acceleration structures on cpu
20150817 14:14:13 <eclesia> physic engine ?
20150817 14:14:19 <gouessej> elect: There are already some hardware accelerated raytracers, see Quaternion Julia raytracer or something similar
20150817 14:14:33 <xranby_> maybe this is of help https://graphics.stanford.edu/papers/gpu_kdtree/
20150817 14:15:05 <gouessej> Xerxes has given an example of spatial subdivision, kd-tree
20150817 14:15:37 <gouessej> Actually, I don't understand exactly what you mean by "acceleration structures", it's so vague
20150817 14:15:44 <elect> kd-tree
20150817 14:15:47 <elect> for examples
20150817 14:16:09 <elect> structures that lets you test intersection very fast
20150817 14:16:21 * eclesia has a quadtree
20150817 14:16:48 <xranby_> elect: rmk0 have many in his project
20150817 14:16:51 <gouessej> If your scene evolves a lot, I don't advise you to use any kind of tree
20150817 14:16:58 <xranby_> let me try find his links
20150817 14:16:59 <elect> no, static
20150817 14:17:09 <elect> mostly
20150817 14:17:17 <gouessej> Then, trees are suitable
20150817 14:17:43 <gouessej> otherwise you should use cells and portals.
20150817 14:17:45 <elect> I found this, https://github.com/fpsunflower/sunflow, but it looks dead
20150817 14:18:01 <elect> it's for a static cad scene
20150817 14:18:58 <elect> eclesia, quadtree are suitable for dynamic scenes, arent they?
20150817 14:19:45 <eclesia> more like average for everything
20150817 14:20:21 <elect> xranby_, I'd not like to bother you, you could just past me its blog, I will do the rest
20150817 14:20:51 <elect> but kd-trees should outperform them in static, no?
20150817 14:22:01 <eclesia> depends. quadtree have the benefit of being predictable since it's always a division in the middle of the range.
20150817 14:22:06 <elect> I am kind of disappoint there is so few material on java and acceleration structures
20150817 14:22:42 <xranby_> elect: http://mvn.io7m.com/io7m-jspatial/
20150817 14:22:44 <eclesia> search in physics and 3d engine .
20150817 14:23:00 <xranby_> elect: that is rmk0's work
20150817 14:23:27 <elect> thanks xranby_
20150817 14:23:46 <xranby_> elect: rmk0 has this all mysterious software ecosystem
20150817 14:23:52 <xranby_> http://io7m.com/software/
20150817 14:24:29 <elect> he is a mysterious person
20150817 14:24:35 <xranby_> that somehow is evolving opengl into haskell
20150817 14:24:39 <elect> we need to dig
20150817 14:25:12 <eclesia> rmk0's work are 2D and 3d quadtrees
20150817 14:26:20 <elect> why this?
20150817 14:26:30 <elect> I mean, work or passion?
20150817 14:27:13 <xranby_> making sure that the computer runs deterministic is an philosophical idea
20150817 14:27:21 <xranby_> i would say passion
20150817 14:27:23 <elect> :D
20150817 14:28:56 <eclesia> competition too (want to crush oracle :w )
20150817 14:29:06 <elect> he has good chances
20150817 14:29:48 <elect> anyway, quadtree are for 2D
20150817 14:29:52 <gouessej> elect: it depends on what you mean by dynamic
20150817 14:29:59 <gouessej> "dynamic scenes"
20150817 14:30:01 <elect> for 3D we talk about octree, dont we?
20150817 14:30:10 <eclesia> yes
20150817 14:30:15 <gouessej> yes but there isn't only one solution
20150817 14:30:49 <elect> gouessej, I load a cad model and then I play with it with mirrors and other stuff where I need intersection
20150817 14:31:19 <gouessej> If your geometry is modified for example like in Red Faction 2 in which you can break the walls, use a trick or don't use a tree
20150817 14:31:23 <zubzub> you play with a cat by intersecting it with mirrors?
20150817 14:31:47 <elect> ^^
20150817 14:31:58 <gouessej> The trees are costly to recompute, that's what I mean
20150817 14:32:10 <elect> gouessej, no, I dont have anything like that
20150817 14:32:25 <elect> yeah, I know, you have to spent some time to build them
20150817 14:32:49 <elect> zubzub, "play" in italian has also meaning like "dealing, handling"
20150817 14:32:59 <elect> dunno if also in eng
20150817 14:33:31 <elect> anyway, Im not (so) pervert
20150817 14:33:37 <elect> you have to believe me
20150817 14:33:42 <elect> °.°
20150817 14:35:28 <eclesia> Hide your cats everything !
20150817 14:35:33 <eclesia> everyone*
20150817 14:36:08 <rmk0> .o.
20150817 14:36:30 <elect> rmk0, why so misterious
20150817 14:36:36 <elect> share it with us
20150817 14:36:43 <rmk0> eh?
20150817 14:36:54 <xranby_> elect: you should have asked me
20150817 14:37:01 <elect> :)
20150817 14:37:04 <xranby_> why i think rmk0 has a mysterious ecosystem
20150817 14:37:18 <xranby_> since i was the one saying mysterious
20150817 14:37:23 <elect> xranby_, I also had the same feeling
20150817 14:37:34 <xranby_> ok.. then i ask you
20150817 14:37:37 <rmk0> in the process of moving everything to github... it's a long and painful process
20150817 14:37:43 <xranby_> elect: why do you have the mysterious feeling?
20150817 14:37:48 <xranby_> rmk0: awesome
20150817 14:37:54 <elect> anyway, for static scene, what is best, kd-tree or octree?
20150817 14:37:56 <xranby_> this means you are no longer a mysterious person
20150817 14:38:05 <rmk0> i see
20150817 14:38:07 <xranby_> sunning your own "empire"
20150817 14:38:10 <xranby_> running
20150817 14:38:25 <elect> coz he is one of the gods here, but I know basically nothing about him
20150817 14:38:35 <rmk0> i think back when i started putting all the projects together, github was ... shit
20150817 14:38:46 <eclesia> (still is)
20150817 14:38:50 <rmk0> but i've been using it for the past six months or so for contract work and it's gotten a lot better
20150817 14:38:54 <rmk0> it's at least tolerable now
20150817 14:39:11 <rmk0> reasonable issue tracking and so on
20150817 14:39:23 <zubzub> I like github
20150817 14:39:27 <xranby_> i wonder if the internet archive mirrors git's stored on personal webservers
20150817 14:39:44 <zubzub> thing is, I don't know of anything better than github
20150817 14:39:53 <xranby_> i actually wished people to run their own servers
20150817 14:40:05 <xranby_> and at the same time help mirror other peoples servers
20150817 14:40:13 <elect> xranby_, for small small dev it doesnt make sense
20150817 14:40:21 <rmk0> i'm fine with github doing the hosting only because they allow me to get all my data out
20150817 14:40:22 <xranby_> putting all in the hands of github is not the long term solution
20150817 14:40:34 <rmk0> i run a private mirror here of all the github repositories, and download tickets daily, etc
20150817 14:40:41 <zubzub> github allows you to use it's software privately on your local servers iirc
20150817 14:40:48 <rmk0> does it?
20150817 14:40:58 <zubzub> yes in exchange for $$$ fcrs :)
20150817 14:41:01 <xranby_> i use a github competitor at my premices
20150817 14:41:03 <rmk0> i'm ... not sure i want to. i imagine it's a giant pile of ruby or something
20150817 14:41:04 <xranby_> called gitlab
20150817 14:41:17 <xranby_> and gitlab is indeed a large pile of ruby
20150817 14:41:22 <rmk0> ugh.. yeah, they can keep it
20150817 14:41:31 <zubzub> https://enterprise.github.com/home
20150817 14:42:07 <rmk0> i'll stick with their API
20150817 14:42:12 <xranby_> free as in free 45 day tria
20150817 14:42:14 <xranby_> l
20150817 14:42:23 <elect> rmk0, do you have jspatial on github too?
20150817 14:42:41 <rmk0> elect: not yet, no. i have about 50 repositories to migrate and i've only gotten about ten done so far
20150817 14:42:49 <rmk0> it will be there, it's just a matter of time
20150817 14:43:09 <elect> have you started recently doing that?
20150817 14:43:26 <rmk0> yeah, been doing it on weekends for the past few weeks
20150817 14:43:39 <rmk0> i wrote this as an amusing experiment: https://github.com/io7m/ftgr
20150817 14:43:51 <rmk0> using it to convert the fossil repositories to make it look as if i've been using git all along
20150817 14:43:56 <rmk0> ssssh!
20150817 14:44:58 <elect> mum is the word
20150817 14:46:22 <rmk0> it's nice that travis-ci.org is free for public github repositories
20150817 14:46:44 <rmk0> have been attaching that to each migrated repository for testing on openjdk7, oracle jdk7, and oracle jdk8
20150817 14:46:50 <elect> rmk0, are you planning to drop fossil or keep it as backup?
20150817 14:47:09 <xranby_> zubzub: do your jogl tarvis build scripts work realiablY?
20150817 14:47:13 <rmk0> i imagine the fossil repositories will probably live read-only on io7m.com forever
20150817 14:47:23 <rmk0> no real reason to delete them
20150817 14:47:32 * eclesia send a psychic wave ... bitbucket ... mercurial ...
20150817 14:47:40 * rmk0 vomits
20150817 14:47:57 <elect> eclesia, you have no sex appeal here
20150817 14:47:57 * xranby_ suddently gets an urge to drink coffee
20150817 14:48:12 <rmk0> ...
20150817 14:48:14 <zubzub> xranby_: adunno actually
20150817 14:48:16 * rmk0 nudges sgothel
20150817 14:48:25 <elect> you will join all of us on github, it's just a matter of time and you know
20150817 14:48:39 <zubzub> xranby_: so I guess not, but if jogl can be build on ubuntu 12.04 it should be able to work just fine
20150817 14:49:14 <eclesia> elect: never :p
20150817 14:49:22 <elect> yeah yeah ^^
20150817 14:49:31 <xranby_> zubzub: i think jogl builds on 12.04
20150817 14:49:47 <elect> I tried recently on 15.04
20150817 14:49:50 <xranby_> what is actually required to get started with tarvis?
20150817 14:49:59 <xranby_> do you have to sign up someplace?
20150817 14:50:00 <elect> without success
20150817 14:50:19 <xranby_> elect: impossible
20150817 14:50:26 <xranby_> elect: i build it all the time
20150817 14:51:02 <elect> it's my nubness
20150817 14:51:09 <xranby_> elect: hmm one idea
20150817 14:51:36 <xranby_> elect: apt-get build-dep libjogl2-java
20150817 14:51:40 <zubzub> xranby_: you need a github account
20150817 14:51:47 <xranby_> that will pull in all build dependencies needed
20150817 14:51:56 <zubzub> xranby_: btw it's "travis" ;)
20150817 14:52:03 <xranby_> since ubuntu build jogl2 fine
20150817 14:52:18 <xranby_> zubzub: ok i have a github account
20150817 14:52:22 <xranby_> whats next?
20150817 14:52:27 <zubzub> xranby_: then you just configure the .travis file and enable your project to build on the travis site
20150817 14:52:33 <zubzub> that's it
20150817 14:52:37 <elect> I'll try xranby_
20150817 14:52:37 <rmk0> xranby_: sorry, didn't realize you were talking about travis... yeah, just "sign in with github" and enable it
20150817 14:52:57 <rmk0> it needs read access to the repositories that you have to grant
20150817 14:53:24 <xranby_> always good to have one more machine running the jogl junit tests
20150817 14:53:34 <rmk0> would be interesting to see how it handles building and testing jogl, would assume it's running on a machine without gpu access
20150817 14:53:42 <elect> rmk0, have you run some benchmark with your jspatial?
20150817 14:53:51 <rmk0> elect: nope
20150817 14:53:51 <xranby_> you can always configure it to use mesa3d with software acceleration
20150817 14:53:56 <zubzub> rmk0: you'd probably just have to install mesa
20150817 14:53:57 <xranby_> software renderer i mean
20150817 14:54:03 <zubzub> and a headless X server or something
20150817 14:54:07 <rmk0> zubzub: is that... possible?
20150817 14:54:12 <zubzub> sure
20150817 14:54:17 <zubzub> why not?
20150817 14:54:21 <rmk0> i really don't know much about it, i just enabled it and stared at build logs
20150817 14:54:30 <rmk0> didn't know i had any control over the containers
20150817 14:54:53 <zubzub> well you can configure them
20150817 14:54:57 <xranby_> i belive ubuntu uses a headless X server to test their builds
20150817 14:55:03 <rmk0> right
20150817 14:55:14 <zubzub> each line you add to the configuration file is basically a command that's executed
20150817 14:56:30 <zubzub> right the thing I had with jogl and travis was the build duration
20150817 14:56:35 <zubzub> travis assumed the build got stuck
20150817 14:56:50 <zubzub> while in fact it was just running all tests (very slowly in sw)
20150817 14:56:57 <rmk0> hehe
20150817 14:57:01 <zubzub> so far it did work afaik
20150817 14:57:11 <xranby_> zubzub: we can disable the slowest test i guess
20150817 14:57:22 <rmk0> think the default for the containers is a 32mb heap
20150817 14:57:26 <xranby_> we already disable some tests when running on arm
20150817 14:57:51 <xranby_> most of the cruicial jogl tests is about lifecycle
20150817 14:58:12 <xranby_> and not so much about stressing the hardware
20150817 14:59:26 <zubzub> yeah well with some thinkering it should work on travis
20150817 15:00:31 <xranby_> https://jogamp.org/bugzilla/show_bug.cgi?id=1187 i have one bug that i can only reproduce using the mesa3d using a slow driver
20150817 15:01:00 <xranby_> for example this bug would be excellent if travis could do the testing
20150817 15:02:16 <gouessej> I don't advise you to use Github even though I still have a few projects on it
20150817 15:02:49 <gouessej> It's a matter of freedom. I plan to host my repositories on my own server.
20150817 15:02:54 <elect> gouessej, what happened btw with libgdx and jogl?
20150817 15:03:07 <xranby_> elect: it work
20150817 15:03:12 <xranby_> libgdx and jogl
20150817 15:03:33 <gouessej> I have to update it once again but it works
20150817 15:03:45 <elect> then now the jogl backend is integrated back?
20150817 15:03:52 <elect> I was reading this http://stackoverflow.com/questions/21148389/libgdx-where-is-jogl-package
20150817 15:03:55 <xranby_> elect: we currently maintain the libgdx jogl parts inside our fork
20150817 15:03:57 <gouessej> elect: Are you dreaming awake?
20150817 15:04:05 <elect> why?
20150817 15:04:05 <gouessej> xranby is right
20150817 15:04:09 <xranby_> elect: i think libgdx in general will be split up
20150817 15:04:13 <xranby_> into many gits
20150817 15:04:26 <gouessej> The maintainers wil probably never put our stuff into master
20150817 15:04:37 <elect> political reasons?
20150817 15:05:28 <gouessej> elect: Are you talking about LibGDX or anything else?
20150817 15:05:32 <xranby_> i think a merge may be possible if we condense the git histopry down
20150817 15:05:36 <xranby_> history
20150817 15:05:40 <elect> yes, gouessej
20150817 15:05:53 <xranby_> so that when we send a pill request then they can review it easy
20150817 15:05:58 <xranby_> pull
20150817 15:06:05 <gouessej> elect: I'm less optimistic than Xerxes but maybe he's right
20150817 15:06:51 <zubzub> a pill request
20150817 15:06:57 <zubzub> a blue pill
20150817 15:07:00 <zubzub> a red pill
20150817 15:07:18 <xranby_> i have to convince them to swallow, we must support GNU/Linux arm pill
20150817 15:07:50 <xranby_> only the jogl backend enable GNU/Linux arm support
20150817 15:08:14 <xranby_> thus it should be in their interest to add it
20150817 15:08:44 <gouessej> xranby_: good point but ...
20150817 15:08:45 <xranby_> the raspberry pi is one of the first widely available systems
20150817 15:08:52 <elect> is there any friction between them and jogl?
20150817 15:08:59 <gouessej> L**** supports OpenGL ES
20150817 15:09:23 <elect> the use of the * from gouessej suggests me yes :)
20150817 15:09:25 <xranby_> gouessej: L**** do not plan to support it out of the bo
20150817 15:09:26 <xranby_> x
20150817 15:09:38 <gouessej> elect: I'm a bit suspicious, I can't be peremptory about that
20150817 15:10:28 <xranby_> gouessej: but it is a valid suspicion yes
20150817 15:10:38 <gouessej> xranby_: Yes but it's becoming less difficult if it supports OpenGL ES and I don't see the point of supporting OpenGL ES but not ARM
20150817 15:11:59 <xranby_> one of the reasons why i joined the jogl jogamp project was that we give unified desktop + opengl es 2 support
20150817 15:12:09 <xranby_> hand are good at it
20150817 15:12:13 <xranby_> and
20150817 15:12:27 <gouessej> Don't you remember this? https://www.raspberrypi.org/forums/viewtopic.php?f=34&t=19532
20150817 15:14:56 <xranby_> i have indeed tested to use lwjgl back in 2011
20150817 15:15:13 <xranby_> and help the raspberry pi users who really wanted to try it in 2012
20150817 15:15:41 <xranby_> but all my negotiations with lwjgl ended up with that they would not do any builds
20150817 15:15:52 <xranby_> thus thats the reason i joined jogamp
20150817 15:16:29 <xranby_> where we do provide builds
20150817 15:16:46 <gouessej> they seem to be reluctant to make any effort
20150817 15:17:13 <gouessej> and I hate those nasty comments about OpenJDK. I LIKE OPENJDK!
20150817 15:19:36 <gouessej> I talked about the efforts on ARM and embedded environments in general.
20150817 15:20:23 <gouessej> xranby_: Who did you contact in its team?
20150817 15:21:05 <xranby_> gouessej: many years now... spasi was keen to fix bugs, but the persons running the build servers had no interest
20150817 15:22:40 <xranby_> the libgdx team is different i think, they want to have gnu/linux arm support
20150817 15:23:10 <xranby_> have to think about all the small details
20150817 15:23:29 <gouessej> me too
20150817 15:23:30 <xranby_> as my first step to convince that gnu/linux arm is wokring
20150817 15:23:40 <xranby_> i will start publish the libgdx demos
20150817 15:23:45 <xranby_> for people to use
20150817 15:23:55 <gouessej> Good idea
20150817 15:24:00 <xranby_> thus then the users can help push the idea
20150817 15:24:40 <xranby_> example i published this inside the game forum: https://www.raspberrypi.org/forums/viewtopic.php?f=78&t=118349
20150817 15:24:48 <gouessej> I'd like to fix OpenGL ES 1 support in JogAmp's Ardor3D Conitnuation, maybe I'll have something to show in a few months :)
20150817 15:25:44 <xranby_> you have to build up awareness that java and opengl es 2 hardware acceleration work in general
20150817 15:26:00 <gouessej> xranby_: Good job. It's easy to understand.
20150817 15:27:10 <gouessej> xranby_: OpenGL ES 2 will require a bigger effort for me. Sorry for the silly question, is it possible to run OpenGL ES 1 on a Raspberry Pi?
20150817 15:27:35 <xranby_> yes it is possible to run opengl es 1
20150817 15:27:40 <xranby_> on a raspberry pi
20150817 15:28:20 <xranby_> it simply work
20150817 15:28:26 <xranby_> no hacks needed
20150817 15:30:02 <xranby_> ther is also this raspberry pi store.. that i can flood with demo applications
20150817 15:30:39 <gouessej> Ok then I'll fix it under GNU Linux on my desktop machine and I'll test it on the Raspberry Pi
20150817 15:31:02 <gouessej> xranby_: I hate the centralized stores :)
20150817 15:32:40 <xranby_> at least it is one store that allow free/libre software
20150817 15:32:44 <xranby_> to be published
20150817 15:33:29 <gouessej> Who controls this store?
20150817 15:34:53 <xranby_> indiecity
20150817 15:36:47 * gouessej (5ee4b442@anon) Quit (Quit: Page closed)
20150817 15:36:56 <xranby_> http://blog.indiecity.com/?p=2494
20150817 15:37:04 <xranby_> "Great news everyone: a group of indie developers have stepped up and bought IndieCity!"
20150817 15:38:23 * elect (~elect@anon) Quit (Ping timeout: 265 seconds)
20150817 15:38:31 <xranby_> so it seems indiecity only supply the raspberry pi packages now days
20150817 15:38:33 <rmk0> gouessej: think the key is to not be dependent. allow github to host repositories until they do something objectionable, and then switch to the private mirror you've been maintaining all long
20150817 15:38:59 <rmk0> there's a lot of things that github does that i don't particularly want to have to maintain/host myself (issue tracking being the main one)
20150817 15:39:10 <rmk0> ... he's not actually here, is he
20150817 15:39:14 <rmk0> \o\
20150817 15:40:49 <xranby_> ok i have activated the travis-ci ... next i need to try push zubzub's configureations
20150817 15:42:16 <xranby_> http://blog.indiecity.com/
20150817 15:42:21 <xranby_> hmm maybe it is dead
20150817 15:42:23 <xranby_> i dont know
20150817 15:42:34 <xranby_> the servers are up and running
20150817 15:53:07 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150817 16:02:54 * eclesia (~husky@anon) has left #jogamp
20150817 17:37:21 * sgothel (~sgothel@anon) has joined #jogamp
20150817 17:37:21 * ChanServ sets mode +v sgothel
20150817 17:37:34 <sgothel> ... back
20150817 17:37:48 <xranby_> sgothel: welcome!
20150817 17:37:59 <sgothel> @Xerxes and @Julien: KUDOS for the patches/bug-reports .. ho ho ho ..
20150817 17:38:15 <sgothel> sorting out spam mail .. the diving into it :)
20150817 17:38:21 <sgothel> *then*
20150817 17:38:48 <xranby_> sgothel: how was the vacation?
20150817 17:39:01 <sgothel> work & pleasure .. was great
20150817 17:39:16 <sgothel> quite foggy in Bergen/Norway
20150817 17:39:28 <sgothel> but almost 2 days of sunshine :)
20150817 17:40:10 <xranby_> sounds good, welcome back :)
20150817 17:41:03 <sgothel> the fat-jar -> inflated thingy, well .. it is more a follow-up enhancement I would say :)
20150817 17:41:20 <sgothel> but since you have done it this way already .. great!
20150817 17:41:41 <sgothel> Bug 1194: excellent
20150817 17:43:58 <sgothel> Bug 1192: Big KUDOS for triaging this thing down to the metal!
20150817 17:44:44 <sgothel> Bug 1193: @Julien and others: Please provide a small compute shader example .. (otherwise we have no test)
20150817 17:45:37 <sgothel> Bug 1182: Workaround - sweet!
20150817 17:47:05 <sgothel> @Xerxes: A big hug!
20150817 17:51:43 <xranby_> sgothel: its good to test jogamp on other peoples laptops etc, found that 1183 triggered on intel mesa3d while borrowing a laptop
20150817 17:55:04 <xranby_> i also tried to borrow a laptop running window 10 with dual gpu graphicscards one intel and one nvidia gtx
20150817 17:55:35 <xranby_> windows aparently has a "right" click -> select gpu to use before launching applications
20150817 17:56:14 <xranby_> and a preference tool that contains a list of application names that automatically gets the good gpu
20150817 17:56:29 <xranby_> now the scary part
20150817 17:56:37 <xranby_> things initialized fine
20150817 17:56:58 <xranby_> thus we could use the intel or the nvidia card depending on selected gpu
20150817 17:57:30 <xranby_> BUT when you closed the application then an odd looking windows popup appeared in the corner
20150817 17:57:48 <xranby_> claiming that the gpu had crashed and needed to be resstarted
20150817 17:58:09 <xranby_> information automatically sent to microsoft
20150817 17:58:16 <xranby_> and no logs left on the device
20150817 17:58:22 <xranby_> troubleshooting a Windows PC these days is like ¯\_(ツ)_/¯
20150817 17:59:03 <xranby_> i was not able to see any logs indicating that we had done anything wrong
20150817 17:59:26 <xranby_> the popup in the corner closed automatically and the computer continued humming like nothing odd have eer happened
20150817 18:00:37 <xranby_> so all i am saying is that there might be something in the lifecycle on windows that can be improved
20150817 18:02:52 <xranby_> i can borrow this laptop again in the remote future
20150817 18:07:46 <xranby_> i was testing this using 2.3.1
20150817 18:08:03 <xranby_> thus i need to re-evaluate with all the latest changes on top
20150817 18:31:55 <rmk0> sgothel: lo!
20150817 19:00:28 <sgothel> (finished family bath :)
20150817 19:01:47 <sgothel> NVidia/Intel GPU 'sharing' stuff .. well, works on OSX and on one Windows machine I had ..
20150817 19:01:59 <sgothel> @Mark: Hello!
20150817 19:02:37 * rmk0 waves
20150817 19:03:21 <sgothel> My coffeepot is not a flies bathtub! grr :)
20150817 19:20:31 * elect (~elect@anon) has joined #jogamp
20150817 19:45:10 * Eclesia (~eclesia@anon) has joined #jogamp
20150817 19:47:30 <Eclesia> sgothel: welcome back
20150817 19:53:10 * Eclesia doesn't have sweet jogamp news to give :( but is working on a draft of animation API if needed
20150817 19:55:39 <xranby_> sgothel: Eclesia has written a CPU rasterizer that can generate an image buffer with text within 20ms i am trying to use this buffer to repaint a NEWT nativewindow after reciving a redraw event
20150817 19:56:38 <xranby_> in order to fix bug1178 cornercase 6 https://jogamp.org/bugzilla/show_bug.cgi?id=1178#c3
20150817 19:56:48 <sgothel> bitmapped text w/o AWT usage?
20150817 19:56:55 <xranby_> yes
20150817 19:57:05 <sgothel> howto image->native-colormap ?
20150817 19:58:23 <Eclesia> ?
20150817 19:58:39 <xranby_> redraw event -> get source ** TBD -> get native color byte order ** TBD -> generate image using same byteorder -> paint -> update native window surface ** TBD
20150817 19:59:15 <sgothel> yes, i.e. the native surface must be able to receive such a pixel-rectangle ..
20150817 19:59:30 <sgothel> we have conversion stuff there already, used for icons
20150817 20:00:06 <sgothel> bitmapped text w/o AWT would be great - solving the 'text/message box' idea
20150817 20:00:29 <xranby_> indeed, the nice thing is that eclesia's classes only depend on eclesias classes :)
20150817 20:01:07 <Eclesia> if you want to remove awt. that means you need : TTF reader, geometries, triangulator, image api, and rasterizer , which I all have ^^
20150817 20:01:31 <Eclesia> (quality of the rasterizer isn't great but will improve with time)
20150817 20:01:33 <sgothel> then - @Eclesia - to use your work, you would need to offer it under our (c) header .. w/ an appropriate bug number
20150817 20:01:40 <sgothel> the TTF reader: please use ours
20150817 20:01:49 <sgothel> i.e. please avoid duplicate code
20150817 20:02:16 <xranby_> 20150814 20:37:07 <Eclesia> http://i.imgur.com/mVB9JeU.png
20150817 20:02:25 <sgothel> we might have the old 'AWT based geometry' thingy avail as well ..
20150817 20:02:38 <sgothel> i.e. TTF -> pixels .. ofc w/o AWT
20150817 20:02:58 <sgothel> surely would be nice to compare both TTF classes as well
20150817 20:03:09 <sgothel> (jogamp's current one - and Eclesia's)
20150817 20:03:47 <sgothel> dunno what this has todo w/ Bug 1178 .. well
20150817 20:03:52 <sgothel> dinner - back in a bit ..
20150817 20:04:40 <Eclesia> xranby_: Well then have fun transposing the classes for jogamp ^^
20150817 20:04:48 <xranby_> sgothel: 1178 needs the get source ** TBD -> get native color byte order ** TBD fill image using paint using transparent colour and update native window surface ** TBD
20150817 20:05:17 <xranby_> Eclesia: yes i wish java had an import local dependency
20150817 20:05:44 <xranby_> thus some way to bundle classes that is only accessible by one class
20150817 20:06:31 <xranby_> having to rename everything, do work
20150817 20:06:41 <xranby_> but.. is... the non ideal solution
20150817 20:06:56 <Eclesia> agree, duplicated work
20150817 20:06:58 <xranby_> maybe the pragmatic solution
20150817 20:08:14 <xranby_> a custom classloader maybe can solve it
20150817 20:09:00 <Eclesia> I think it might be a good time for you to decide where is the limit of jogl, how far you want to go. what it does and what is not his job.
20150817 20:10:32 <Eclesia> perhaps limit yourself to the possibility to display a PixelRectangle is the worse case. up to the user to 'fill' this rectangle
20150817 20:10:59 <Eclesia> just giving ideas
20150817 20:13:18 <xranby_> it is the platform independent glue you need in order to hardware accelerate all the things!
20150817 20:14:36 <xranby_> it has indeed grown into a toolkit
20150817 20:16:13 <Eclesia> just saying you will pull a lot of classes and apis.
20150817 20:18:05 <Eclesia> ... or maybe you could trim down the CPURasterizer to support only colors, no blending and basic shapes + text
20150817 20:18:32 <Eclesia> won't have to pull everything this way.
20150817 20:28:17 <sgothel> PixelRectangle -> native surface might be one desirable feature
20150817 20:29:00 <sgothel> then .. our existing TTF parser can be used to generate text ofc .. iff somebody likes to do it - not my prio here ofc
20150817 20:29:32 <Eclesia> iff ?
20150817 20:30:01 <sgothel> https://en.wikipedia.org/wiki/If_and_only_if
20150817 20:30:45 <sgothel> @eclesia: how is your other work going, i.e. bugs assigned to you - are there any? (just catching up .. bug-walk)
20150817 20:33:00 <Eclesia> I'm not following them day by day. haven't tested to frame one you've done yet
20150817 20:33:18 <Eclesia> (frame not displayed in the desktop bar)
20150817 20:34:07 <Eclesia> clipboard/drag&drop are not really bugs, nice features I would say
20150817 20:34:27 <sgothel> Bug 1167 unit test, your 3d port update (maybe add a bug report, for jogl-demos)
20150817 20:34:51 * elect (~elect@anon) Quit (Ping timeout: 246 seconds)
20150817 20:35:04 <xranby_> sgothel: hmm we do not have a platform independent WindowImpl getPixelFormat() .. or do we?
20150817 20:36:22 <xranby_> at @Override
20150817 20:36:22 <xranby_> public void windowRepaint(final WindowUpdateEvent e) {
20150817 20:36:22 <xranby_> i can use e.getSource() and check if it a WindowImpl
20150817 20:36:23 <Eclesia> sgothel: 1167 : that isn't me
20150817 20:36:48 * elect (~elect@anon) has joined #jogamp
20150817 20:37:05 <xranby_> then i have to find what pixelformat this window impl wants
20150817 20:37:14 <xranby_> before i prepare an image
20150817 20:37:17 <sgothel> duh
20150817 20:37:22 <sgothel> good morning sven :)
20150817 20:37:59 <sgothel> ok .. thx Eclesia .. now you have reset my nickname association table :)
20150817 20:38:22 <Eclesia> sgothel: Eclesia = Johann Sorel
20150817 20:38:30 <sgothel> @Eclesia: :) - the unresizable .. well ..
20150817 20:38:34 <sgothel> and so forth ..
20150817 20:38:38 <Eclesia> ^^
20150817 20:38:54 <sgothel> so next release will take 1-2 weeks as it seems .. so go fishing bugs :)
20150817 20:39:05 <xranby_> and pass it down using this new PixelRectangle -> native surface API added to WindowImpl
20150817 20:39:16 <sgothel> @Xerxes: we sort of have that
20150817 20:39:25 <sgothel> i.e. NativeWindow's Capabilities
20150817 20:39:54 <sgothel> but I must refrain on thinking too deeply on this feature for 2.3.2 .. i.e. will not make it
20150817 20:40:16 <sgothel> this must be for next release I am afraid .. or you push it all in like crazy :)
20150817 20:40:39 <xranby_> sure, we can live without it for 2.3.2 because it is only a minor glitch triggering on one of if you request a non opaque caps
20150817 20:40:47 <xranby_> on one os (raspberry pi)
20150817 20:41:03 <sgothel> I thought .. fix all crashes and .. hmm
20150817 20:41:05 <Eclesia> sgothel: I've just list the missing things compare to awt/swing/javafx. it's not to make you angry or anything. I wish Newt will crush them all one day :)
20150817 20:41:12 <xranby_> yes fix crashes and win
20150817 20:41:19 <xranby_> do we have anY?
20150817 20:41:26 <sgothel> sure - all good!
20150817 20:41:34 <sgothel> I guess you fixed a few ..
20150817 20:41:41 <sgothel> and then the new GL updates ..
20150817 20:41:48 * elect (~elect@anon) Quit (Ping timeout: 264 seconds)
20150817 20:41:53 <sgothel> then we shall let it sink for a few days ..
20150817 20:41:58 <sgothel> (testing .. )
20150817 20:42:24 <xranby_> i can report that processing 3's main feature is that they have replaced awt with newt
20150817 20:42:39 <sgothel> HURRAY
20150817 20:42:40 <xranby_> and they are super stonked (happy) that they now have stutter free rendering
20150817 20:42:58 <rmk0> sgothel: could you run your /srv/www/jogamp.org/deployment/maven/fix-uid-mod.sh script on the server? need to deploy to the repos for gouessej and the permissions in the repos are currently wrong
20150817 20:43:05 <xranby_> and they ditch the fixed function pipeline
20150817 20:43:11 <sgothel> the usual goal of 2.3.2 should be IMHO: have time free for graph .. while 2.3.2 works almost flawless ..
20150817 20:43:29 <xranby_> only using shaders
20150817 20:43:36 <xranby_> + they add arm support
20150817 20:43:53 <sgothel> err .. yes .. one sec
20150817 20:44:27 <sgothel> done
20150817 20:44:31 <rmk0> thanks!
20150817 20:47:54 * rmk0 fights gpg-agent
20150817 20:48:55 <sgothel> @Xerxes: XRRGetScreenSizeRange(..) is a 'cheap call" ?
20150817 20:49:07 <sgothel> i.e. does not consume much time?
20150817 20:49:24 <sgothel> i.e. no perf. regression observed [yet]? - I assume so ..
20150817 20:49:38 <sgothel> reason I ask - sometimes those XRR calls are crazy .. :)
20150817 20:50:47 <xranby_> sgothel: i ran the whole jogl junit.run on the intel machine
20150817 20:50:57 <xranby_> felt normal time to launch windows
20150817 20:51:09 <xranby_> sucess after 45 min
20150817 20:51:10 <sgothel> @Julien: I see your commits regarding COMPUTE .. hmm .. do you have a compute example to test? or how did you test?
20150817 20:51:16 <sgothel> great - thx
20150817 20:51:39 <sgothel> and you also had a fix for the inflated fat-jar ?
20150817 20:51:59 <sgothel> and for the .. err .. undesirable extension/jre-system folder?
20150817 20:52:25 <sgothel> (I guess that completes it .. i.e. your work last week as far as I can remember now)
20150817 20:52:44 <sgothel> I fetched: Bug 1183 and Bug1192 branches here ..
20150817 20:52:58 <sgothel> and have read the bugzilla emails
20150817 20:53:12 <sgothel> (sloppy me .. oh well)
20150817 20:53:47 <sgothel> but .. well, feels good to *ask* for things sometimes - *relax* :)
20150817 20:54:40 <sgothel> what a 'back again' day - tons of good news here ..
20150817 20:58:34 <xranby_> sgothel: if you look inside the xrandr tool src then you see that even this tool uses the workaround!
20150817 20:58:38 <xranby_> static void
20150817 20:58:38 <xranby_> get_screen (Bool current)
20150817 20:58:38 <xranby_> {
20150817 20:58:38 <xranby_> if (!has_1_2)
20150817 20:58:38 <xranby_> fatal ("Server RandR version before 1.2\n");
20150817 20:58:38 <xranby_> if (res)
20150817 20:58:38 <xranby_> return;
20150817 20:58:39 <xranby_> XRRGetScreenSizeRange (dpy, root, &minWidth, &minHeight,
20150817 20:58:39 <xranby_> &maxWidth, &maxHeight);
20150817 20:58:40 <xranby_> if (current)
20150817 20:58:40 <xranby_> res = XRRGetScreenResourcesCurrent (dpy, root);
20150817 20:58:41 <xranby_> else
20150817 20:59:08 <xranby_> thus it do not take longer than your xrandr command tool
20150817 20:59:12 <sgothel> 'thank god its free software' :)
20150817 20:59:32 <sgothel> so nobody fixes the root cause .. oh well :)
20150817 20:59:45 <sgothel> taking it .. ofc, thx a very lot
20150817 22:21:00 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20150817 22:40:34 <xranby_> sgothel: i also have a fix for the maven deployment
20150817 22:40:51 <xranby_> and unpack all jars into one one jar deployment
20150817 22:41:52 <xranby_> https://github.com/xranby/gluegen/tree/bug1172
20150817 22:42:20 <xranby_> https://jogamp.org/bugzilla/show_bug.cgi?id=1172
20150817 22:43:41 <xranby_> jogamp-scripting do not need to be modified because, with this 1172 change, then the native jars are at the right position when you repack the fat-jar
20150817 22:44:15 <xranby_> and gluegen already know how to load the natives from the new location thanks to your work with adding support to fat jar
20150817 22:44:18 <sgothel> ah .. so its the new default native/os.arch ..
20150817 22:44:43 <sgothel> sweet! great oversight ..
20150817 22:45:41 <sgothel> hmm .. then we should no more look into the native jar root dir .. but in this new default only ..
20150817 22:46:43 <sgothel> (speed up .. remove redundant search ..)
20150817 22:46:55 <sgothel> right?
20150817 22:47:06 <xranby_> yes, always good to speed up starttime
20150817 22:48:03 * a11mad11 (4c47eb44@anon) has joined #jogamp
20150817 22:52:29 <sgothel> we have to do the same for the tool libs, e.g. openal here .. to make things round ..
20150817 22:54:03 <xranby_> oh so joal do not use gluegen-cpptasks-base.xml ?
20150817 22:54:47 <sgothel> yes it does .. just need to double check .. AFAIK we copy the soft openal lib 'manually' in openal's build.xml
20150817 22:54:57 <sgothel> since it is not the 'glue' native lib
20150817 22:55:22 <sgothel> but I can walk through that .. i.e. validate those cases
20150817 22:55:38 <sgothel> (need to do that anyways for code review)
20150817 23:00:08 <xranby_> libopenal.so gets places at the right location /natives/linux-amd64/libopenal.so when i build openal with the gluegen 1172 applied
20150817 23:00:17 <sgothel> great!
20150817 23:06:35 <sgothel> pushed gluegen merge .. checking code now :)
20150817 23:10:10 <sgothel> 'native.tag.jar' .. right .. they all use that macro for native jar production
20150817 23:10:46 <sgothel> now the underlying file system at build time is different from the native jar content .. hmm
20150817 23:11:15 * a11mad11 (4c47eb44@anon) Quit (Ping timeout: 246 seconds)
20150817 23:11:25 <sgothel> not that this is really important .. just mentioning it to earmark possible sideffects ..
20150817 23:17:04 <sgothel> PlatformPropsImpl.queryElfFile(..) -> 'java' and 'jvm' native libs .. hmm
20150817 23:26:30 <sgothel> I guess we need to search those sys folders just for this case .. so partially bringing them back for above method
20150817 23:43:42 <sgothel> further more .. how could NativeLibrary.open find a system library w/o searching them?
20150817 23:45:10 <sgothel> IMHO: JNILibLoaderBase shall use the non-system-search, where all others shall use the system-search version, w/ systemSearchFirst = true
20150817 23:45:40 <sgothel> I manually revert this .. and we can discuss result .. one sec
20150817 23:55:27 <xranby_> yes https://jogamp.org/bugzilla/show_bug.cgi?id=1194 need much scrutiny
20150818 00:30:47 * rmk0 (~rmk0@anon) Quit (Ping timeout: 252 seconds)
20150818 00:31:24 * rmk0 (~rmk0@anon) has joined #jogamp
20150818 00:31:24 * rmk0 (~rmk0@anon) Quit (Changing host)
20150818 00:31:24 * rmk0 (~rmk0@anon) has joined #jogamp
20150818 00:55:41 <sgothel> @Xerxes .. 'fixed' .. pls review/test
20150818 05:05:30 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20150818050530.html