#jogamp @ irc.freenode.net - 20150908 05:05:09 (UTC)

20150908 05:05:09 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20150907050509.html
20150908 05:05:09 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20150908050509.html
20150908 06:22:42 * xranby_f22 (~familjen@anon) Quit (Quit: Leaving.)
20150908 07:04:25 * monsieur_max (~maxime@anon) has joined #jogamp
20150908 07:18:05 <xranby> interesting.. it is possible for robovm to generate "fat" binarys supporting many CPU's example robovm -archs x86_64:x86:thumbv7:arm64 -verbose -dynamic-jni -forcelinkclasses jogamp.**:com.jogamp.** -cp jogamp-fat.jar com.jogamp.newt.opengl.GLWindow
20150908 07:39:26 <monsieur_max> Hello people
20150908 07:53:34 <xranby> monsieur_max: hello!
20150908 07:55:07 <monsieur_max> Using Jogl MediaPlayer in my game, i have fluidity issue with videos in my game, what should i do to try to improve this or find the reason ? the video is playing right on vlc, it's a webm video, and results are the same on Linux and Windows
20150908 07:55:32 <monsieur_max> lowering the resolution of the video or the bitrate did not improve things
20150908 07:56:30 <xranby> monsieur_max: can these fluidity issues be observed using the jogamp MovieSimple and MovieCube junit tests?
20150908 07:56:40 <xranby> monsieur_max: you can pass a custom uri to the tests
20150908 07:56:43 <xranby> to test it
20150908 07:56:55 <monsieur_max> that's something i have yet to tests :)
20150908 07:58:11 <xranby> monsieur_max: one moment and i can dig up the syntax and a oneliner to test
20150908 07:58:31 <monsieur_max> oh great :)
20150908 08:01:43 <xranby> java -cp jogl-test.jar:gluegen-rt.jar:jogl-all.jar com/jogamp/opengl/test/junit/jogl/demos/es2/av/MovieCube
20150908 08:02:16 <monsieur_max> great, you're the man !
20150908 08:02:50 <monsieur_max> xranby: i'll test that during lunch time, i'll let you know. Thanks a lot
20150908 08:03:45 <xranby> to use a file on your own system then pass -url file://
20150908 08:04:40 <xranby> example: java -cp jogl/build/jar/jogl-test.jar:gluegen/build/gluegen-rt.jar:jogl/build/jar/jogl-all.jar com/jogamp/opengl/test/junit/jogl/demos/es2/av/MovieCube
20150908 08:04:45 <xranby> java -cp jogl/build/jar/jogl-test.jar:gluegen/build/gluegen-rt.jar:jogl/build/jar/jogl-all.jar com/jogamp/opengl/test/junit/jogl/demos/es2/av/MovieCube -url file:///home/xranby/jogamp/BigBuckBunny_512kb.mp4
20150908 08:05:42 <xranby> sgothel: headsup MovieCube work while MovieSimple appears to be broken
20150908 08:05:54 <xranby> Exception in thread "main-null" com.jogamp.opengl.util.AnimatorBase$UncaughtAnimatorException: com.jogamp.opengl.GLException: Caught IllegalFormatConversionException: d != java.lang.Boolean on thread main-null
20150908 08:05:54 <xranby> at com.jogamp.opengl.util.AWTAnimatorImpl.display(AWTAnimatorImpl.java:92)
20150908 08:05:54 <xranby> at com.jogamp.opengl.util.AnimatorBase.display(AnimatorBase.java:451)
20150908 08:05:54 <xranby> at com.jogamp.opengl.util.Animator$MainLoop.run(Animator.java:198)
20150908 08:05:54 <xranby> at java.lang.Thread.run(Thread.java:745)
20150908 08:05:54 <xranby> Caused by: com.jogamp.opengl.GLException: Caught IllegalFormatConversionException: d != java.lang.Boolean on thread main-null
20150908 08:05:54 <xranby> at com.jogamp.opengl.GLException.newGLException(GLException.java:76)
20150908 08:05:55 <xranby> at jogamp.opengl.GLDrawableHelper.invokeGLImpl(GLDrawableHelper.java:1311)
20150908 08:05:55 <xranby> at jogamp.opengl.GLDrawableHelper.invokeGL(GLDrawableHelper.java:1131)
20150908 08:05:56 <xranby> at com.jogamp.newt.opengl.GLWindow.display(GLWindow.java:759)
20150908 08:05:56 <xranby> at com.jogamp.opengl.util.AWTAnimatorImpl.display(AWTAnimatorImpl.java:81)
20150908 08:05:57 <xranby> ... 3 more
20150908 08:07:00 <xranby> the above was seen using: java -cp jogl/build/jar/jogl-test.jar:gluegen/build/gluegen-rt.jar:jogl/build/jar/jogl-all.jar com/jogamp/opengl/test/junit/jogl/demos/es2/av/MovieSimple
20150908 08:08:21 <xranby> if i instead run the movieCube then everything work as normal
20150908 08:08:25 <xranby> java -cp jogl/build/jar/jogl-test.jar:gluegen/build/gluegen-rt.jar:jogl/build/jar/jogl-all.jar com/jogamp/opengl/test/junit/jogl/demos/es2/av/MovieCube
20150908 08:33:29 * elect (~GBarbieri@anon) has joined #jogamp
20150908 08:54:45 * elect (~GBarbieri@anon) Quit (Read error: Connection reset by peer)
20150908 09:37:00 * jvanek (jvanek@anon) has joined #jogamp
20150908 10:57:18 <xranby> jvanek: have someone demonstrated icedtea-web use on windows?
20150908 10:57:31 <xranby> and mac?
20150908 10:58:08 <jvanek> xranby, javaws (netx -jar) is known to work on win
20150908 10:58:14 <jvanek> .. with some issues formt ime to time
20150908 10:58:21 <jvanek> on mac, imho never tried
20150908 10:58:36 <jvanek> Plugin - Jacob wrote rpoof of concept for winfows npapi plugin
20150908 10:58:43 <jvanek> But Imho its dead end
20150908 10:58:57 <jvanek> the netx jar shoudl work also on mac
20150908 10:59:12 <xranby> is there an accepted plugin api these days?
20150908 10:59:14 <jvanek> with simialr issues as on windows (from tie to time wrong file path, as it is not regularly tested)
20150908 10:59:27 <jvanek> xranby, thats the issue. NPAPI si sying
20150908 10:59:28 <xranby> or do someone have to compile the plugin into javascript?
20150908 10:59:29 <jvanek> *dying
20150908 10:59:42 <jvanek> what?? :-)
20150908 10:59:54 <jvanek> I have no idea what it will do whan you will compile it to JS
20150908 10:59:56 <xranby> guillaum1: here have a metacyclic jvm
20150908 11:00:22 <xranby> jvanek: idea, use guillaum1's jainja jvm to have a jvm running using javascript
20150908 11:00:37 <xranby> use this javascript powered jvm to run the plugin etc
20150908 11:00:56 <jvanek> like pure changing netx.jar and plugin.jar to pure JS?
20150908 11:01:01 <jvanek> and run directly form browser?
20150908 11:01:08 <jvanek> Idea is definitly interesting
20150908 11:01:20 <jvanek> And I'm unable to predcit how it ends
20150908 11:01:34 <xranby> http://jainja.thenesis.org/
20150908 11:01:45 <xranby> i think it iwll be impossible to run native code :/
20150908 11:02:03 <jvanek> sajll drawback, but stilld rawback
20150908 11:03:26 <xranby> so npapi is dying .. is there any plugin api replacing it ?
20150908 11:03:35 <jvanek> I doupt
20150908 11:03:37 <jvanek> *doubt
20150908 11:03:51 <jvanek> I had soem small sucess with ppapi, but DO nto wont ot continue with it
20150908 11:04:05 <jvanek> Questionis different - have java applet still something to offer?
20150908 11:04:15 <jvanek> I would say no, with one exception - card readers
20150908 11:04:25 <xranby> i would say yes. everyone hates webgl
20150908 11:04:35 <jvanek> And those cardreaders have very often native classes
20150908 11:04:35 <xranby> well they have a mixed feeling situation
20150908 11:04:38 <jvanek> *code
20150908 11:04:45 <jvanek> So do I :(
20150908 11:05:10 <jvanek> If you with idea of canvas only applets
20150908 11:05:17 <xranby> people like webgl because it require no plugins and actually give you access to opengl
20150908 11:05:20 <jvanek> then the drawback of not-running native code is big
20150908 11:06:54 <xranby> its kind of sad. that deployment have become harder over the years
20150908 11:07:06 <xranby> deployment of native applications
20150908 11:07:35 <jvanek> unless you think in NaCL direction
20150908 11:07:40 <jvanek> But IO dont like NaCl
20150908 11:07:41 <xranby> well I for one will continue using platforms running natives directly
20150908 11:08:10 <xranby> thus for me i will have eventually to live without a browser deployment option
20150908 11:08:24 <jvanek> => javaws O:)
20150908 11:08:33 <xranby> if the browser people decide that deployment using a browser is a big nono
20150908 11:08:42 <xranby> javaws still work
20150908 11:08:47 <xranby> kind of
20150908 11:08:53 <jvanek> or keep supporting Npapi inbrowsers
20150908 11:09:56 <xranby> jvanek: i am using icedtea-web during deployment at work, however the verification done to check certificates, use quite a lot of time nowdays :/
20150908 11:09:57 <jvanek> anyway , yes. in-browser deployment of natives is fading
20150908 11:10:28 <jvanek> you mean those yes click, yes, click, yes, click ...? Orsomething more compelx?
20150908 11:10:55 <xranby> jvanek: something more complex, its like quite a lot of time is spent in the classloader
20150908 11:11:12 <xranby> an application that should take "5second to startup takes 30seconds on a rasperry pi
20150908 11:11:13 <jvanek> true
20150908 11:11:46 <xranby> jvanek: thus i would liek to brainstorm ways to make it faster
20150908 11:12:34 <jvanek> any inital thought? The classlaoder is soemthing, I already gave up refactoring
20150908 11:13:16 <xranby> jvanek: the only solution i can think of require replacing the whole certificate checks with something faster
20150908 11:13:35 <xranby> as i dont like the certificate validation idea in whole
20150908 11:13:39 <jvanek> faster mean less certificate checks
20150908 11:13:49 <jvanek> Still it is very widely used
20150908 11:13:53 <jvanek> and can not be abandoned
20150908 11:14:04 <jvanek> if tere will be another verification mechanism
20150908 11:14:09 <xranby> like i dont like the idea that if i have one application with a flaw that i must revoke my certificate and thus break all applications that i have deployed
20150908 11:14:12 <jvanek> it have to stay side by side with current
20150908 11:14:28 <jvanek> yes, thats cool feature :D
20150908 11:14:29 <jvanek> :(
20150908 11:15:06 <xranby> i do want a way for the end user to verify the origin of the sourcecode
20150908 11:15:26 <xranby> and a way for the developer to tag safe git commits
20150908 11:15:35 <xranby> and revoke git commit that are unsafe
20150908 11:16:35 <xranby> i think the whole git history should be available and a mechanism to say this git commit is tested and is used to produce these binarys
20150908 11:17:58 <xranby> jvanek: we have discussed these ideas here before and came up with the idea of Source Certification Contract (SCC) http://jogamp.org/wiki/index.php/SCC_Overview
20150908 11:18:12 <xranby> SCC answers: This binary is produced by this set of source code, which is trusted by these people.
20150908 11:18:17 <xranby> SCC verifies whether a binary object's signature matches its assumed source code signatures.
20150908 11:18:26 <xranby> SCC provides a list of personal signatures who trust this source code, as well as a list of personal signatures who trust this binary object. Hence SCC incorporates a chain of trust.
20150908 11:19:08 <jvanek> yy. I recallthat
20150908 11:20:16 <jvanek> I have nothing agaisnt inclusion to ITW
20150908 11:20:50 <jvanek> just.. it wil not be fun to include it inside iutw smoothy (and together with cers)
20150908 11:20:54 <jvanek> *certs
20150908 11:20:58 * jvanek shold get cofee
20150908 11:21:32 * xranby gets the same
20150908 11:38:05 * gouessej (5ee4b442@anon) has joined #jogamp
20150908 11:38:50 <gouessej> Hi
20150908 11:39:40 <gouessej> I'm happy with my tool but if the major browsers block RPM, DEB, MSI, APP, ..., my efforts will be useless
20150908 11:41:30 <gouessej> Have anybody tried to build JOGL with a compact profile?
20150908 11:43:46 <xranby> gouessej: we have tested to use JOGl with a compact profile
20150908 11:44:00 <xranby> it work fine, the jvm is also quite small
20150908 11:44:23 <gouessej> There are still a few dependencies on Java Sound :s
20150908 11:44:41 <xranby> http://jogamp.org/git/?p=jigong/jigong.git;a=commit;h=a94b0e3c62110f077ea391ccdc158b2c804a5bb8
20150908 11:44:48 <xranby> OpenJDK 8 Profiles ..
20150908 11:44:48 <xranby> Stripped binary size GNU/Linux x64:
20150908 11:44:48 <xranby> 30M export/linux-x86_64-normal-server-release/j2re-compact1-image
20150908 11:44:48 <xranby> 42M export/linux-x86_64-normal-server-release/j2re-compact2-image
20150908 11:44:48 <xranby> 48M export/linux-x86_64-normal-server-release/j2re-compact3-image
20150908 11:44:48 <xranby> 94M export/linux-x86_64-normal-server-release/j2re-image
20150908 11:45:46 <xranby> you generate these using: http://jogamp.org/git/?p=jigong/jigong.git;a=blob;f=doc/README.JEP161.txt;h=ac389577a5d5b47eca6dc1f48f0680455612fc43;hb=ece12e39d219f76cdf8b6f71239cf7f3d663401a
20150908 11:47:07 <gouessej> jdeps finds some dependencies beyond compact1
20150908 11:47:16 <xranby> gouessej: today i tested how much i could reduce a jogl NEWT GLWindow application deployed using robovm
20150908 11:47:43 <xranby> gouessej: i ended up with a total footprint of 6.7Mb including all natives required to run on x86_64 on a linux X11 system
20150908 11:48:04 <gouessej> xranby: excellent
20150908 11:48:31 <gouessej> xranby: is it possible to use RoboVM to build an iOS application without a Mac?
20150908 11:48:57 <xranby> it is not. you need to tell robovm the path to the ios sdk
20150908 11:49:04 <xranby> if you want to build for mac
20150908 11:49:12 <xranby> build for ios
20150908 11:50:31 <gouessej> ok :s
20150908 11:50:32 <xranby> https://github.com/sgothel/jogl/compare/master...xranby:robovm <- build instructions how to produce a 6.3Mb x86_64 binary
20150908 11:51:43 <xranby> the memory use is excellent
20150908 11:51:52 <xranby> only consume 750kb of ram
20150908 11:52:00 <xranby> when the application is running
20150908 11:53:32 <xranby> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
20150908 11:53:42 <xranby> 9216 xranby 20 0 715288 93072 32056 S 3,0 2,5 0:01.25 demos.es2.RawGL2ES2demo
20150908 11:54:19 <xranby> like that.. when run top to check the robovm compiled application
20150908 11:56:49 <xranby> oh.. silly me.. those numbers are on Kb hmm .. ok it uses 93Mb ram then
20150908 12:04:24 <xranby> http://robovm.com/robovm-1-7-released-updates-bug-fixes/
20150908 12:04:37 <xranby> If you want to take part in shaping the future of RoboVM, please give our survey a few minutes of your time!
20150908 12:04:46 <xranby> http://robovm.com/stop-survey-time/
20150908 12:05:20 <xranby> take the survey if you want to deploy w/ AOT on mobile, desktop & ARM SoC devices using RoboVM
20150908 12:26:14 * gouessej (5ee4b442@anon) Quit (Quit: Page closed)
20150908 12:27:33 * elect (~GBarbieri@anon) has joined #jogamp
20150908 14:24:04 <rmk0> does robovm target anything outside of android?
20150908 14:24:19 <rmk0> programming for a crippled version of a ten year old version of java is ... horrible
20150908 14:26:59 <monsieur_max> rmk0: yeah, especially when you tasted streams
20150908 14:33:29 <xranby> rmk0: you can use robovm for GNU/Linux, android and ios
20150908 14:33:47 <rmk0> xranby: i mean the class library it targets, really
20150908 14:34:18 <xranby> rmk0: personally i do not care that much of the classlibrary, robovm have enough classes to run jogamp
20150908 14:35:15 <xranby> rmk0: you have a similar situation if you use the compact profiles with openjdk
20150908 14:35:32 <xranby> rmk0: is there some class in particular that you want to use?
20150908 14:37:47 <xranby> my thinkning is that if the java jvm was more modular then you could use the openjdk rt classes with robovm
20150908 14:48:53 <guillaum1> hi all
20150908 14:52:33 <rmk0> xranby: nothing in particular. it's more that you try to use any third party libraries on android and find you can't because they use annotations and other things from the standard jdk
20150908 14:52:48 <rmk0> i've been fighting with it since february, i absolutely hate it
20150908 14:53:10 <rmk0> it's a broken subset of java 6 with frivolous incompatibilities
20150908 14:53:22 <rmk0> see anything that specifically targets android as an anti-feature
20150908 14:54:51 <rmk0> *use annotations and other things from the standard jdk that are missing
20150908 15:01:59 * elect (~GBarbieri@anon) Quit (Read error: Connection reset by peer)
20150908 15:07:21 <rmk0> i don't think the compact profiles are the same, because in the case of the profiles, you get to declare exactly what you depend upon (and third party libraries do the same)
20150908 15:23:21 * elect (~elect@anon) has joined #jogamp
20150908 15:45:23 * elect (~elect@anon) Quit (Ping timeout: 256 seconds)
20150908 15:53:35 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150908 16:22:36 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20150908 16:33:54 * elect (~elect@anon) has joined #jogamp
20150908 18:29:15 * elect (~elect@anon) Quit (Ping timeout: 252 seconds)
20150908 20:38:41 * elect (~elect@anon) has joined #jogamp
20150908 21:18:01 * elect (~elect@anon) Quit (Ping timeout: 256 seconds)
20150909 05:05:09 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20150909050509.html