#jogamp @ irc.freenode.net - 20131011 05:05:32 (UTC)

20131011 05:05:32 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20131010050532.html
20131011 05:05:32 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20131011050532.html
20131011 05:17:16 * [Mike] (~Mike]@anon) Quit ()
20131011 06:35:40 * monsieur_max (~maxime@anon) has joined #jogamp
20131011 07:18:47 * Eclesia (jsorel@anon) has joined #jogamp
20131011 07:18:54 <Eclesia> hi
20131011 07:23:30 <monsieur_max> so ... dropping out SoundSystem, it was providing an easy way to use openAL, through Joal... but the result of transition effects are just horrible... i'm getting back to joal
20131011 07:23:38 <monsieur_max> Eclesia: hi
20131011 07:35:01 <Eclesia> hi monsieur_max
20131011 07:46:55 * ebrayet (~Adium@anon) has joined #jogamp
20131011 08:13:10 * xranby (~xranby@anon) has joined #jogamp
20131011 08:50:58 <sgothel> @Xerxes: Got the ODROID board .. will hack on it on the side this weekend
20131011 08:51:48 <sgothel> Intel-Atom: No native ARM support .. must be emulated .. ordered a ASUS MeMO Pad FHD 10 (Intel Atom Z2560)
20131011 08:54:42 <xranby> great!
20131011 09:01:39 <sgothel> .. away for a few hours - laters
20131011 10:02:52 * xranby (~xranby@anon) Quit (Ping timeout: 264 seconds)
20131011 10:05:21 * xranby (~xranby@anon) has joined #jogamp
20131011 10:14:20 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20131011 10:36:48 * xranby (~xranby@anon) Quit (Ping timeout: 240 seconds)
20131011 10:39:15 * xranby (~xranby@anon) has joined #jogamp
20131011 11:49:55 * ebrayet (~Adium@anon) has joined #jogamp
20131011 12:21:52 * rmk0 (~rmk0@anon) has joined #jogamp
20131011 12:21:52 * rmk0 (~rmk0@anon) Quit (Changing host)
20131011 12:21:52 * rmk0 (~rmk0@anon) has joined #jogamp
20131011 13:03:29 <masterzen> hmm, I'm just testing 2.1.0 (compared to the 2.1.0-rc from oct 6th), and I can't find anymore glDrawElements (the one with a Buffer for indices) in GL2ES2... What's the proper replacement?
20131011 13:05:00 <monsieur_max> remind me something :)
20131011 13:05:00 <xranby> masterzen: the proper replacement using GL2ES2 profile is to use VBO
20131011 13:05:06 <monsieur_max> xranby: ;)
20131011 13:05:28 <monsieur_max> masterzen: had to figure this out last week, thanks to xranby very precious help :=
20131011 13:05:36 <masterzen> xranby: ok, so VBO for the indices buffer, then :)
20131011 13:05:53 <xranby> masterzen: because, on recent desktop GL3 core and GL4 core systems glDrawElements (the one with a Buffer for indices) is removed
20131011 13:06:12 <xranby> and GL2ES2 is GL3 core compatible
20131011 13:07:03 <masterzen> xranby: thanks for the explanation!
20131011 13:07:12 <xranby> if you use GLES2 profile then you may still use glDrawElements (the one with a Buffer for indices)
20131011 13:07:21 <xranby> or GL2 profile
20131011 13:13:37 <xranby> or GL3bc, GL4bc ....
20131011 13:13:44 <xranby> in short glDrawElements (the one with a Buffer for indices) is removed is removed for all core profiles
20131011 13:14:52 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=852 <- :)
20131011 13:15:11 <xranby> "Properly reflect OpenGL-core >= 3.1 and ES >= 3.0 API functions restricted to Buffer Objects"
20131011 13:15:12 <sgothel> pls read Brice
20131011 13:15:39 <sgothel> it's still in GLES2 and GL2ES1 ..
20131011 13:20:34 <masterzen> BTW, I now have: CPU data sourcing n/a on desktop with a GL2ES2. If I read the thing correctly that means the profile prevents client side data buffer, is that right?
20131011 13:21:58 <masterzen> yep, the bug report tells it
20131011 13:22:20 <sgothel> might be a core GL3 ctx .. the exception should tell you
20131011 13:22:45 <sgothel> also point to the fact that you hard-cast the gl object :)
20131011 13:23:01 <sgothel> since API entry is gone .. right ?
20131011 13:23:51 <sgothel> I owe the bug still unit tests .. but couldn't add them in time - so I pushed the API change 1st .. for the minor update, dirty me :)
20131011 13:23:52 <masterzen> no, it's on a call to GLArrayDataClient.enableBuffer (I already modified the GLDrawElements
20131011 13:24:36 <masterzen> yes, this is a GL3 core ctx: Version 3.3 (Core profile, arb, ES2 compat, FBO, hardware)
20131011 13:24:41 <sgothel> the whole exception .. pls attach to Bug 852 thx .. a few lines maybe here (message, which ctx, 2-3 lines in gl)
20131011 13:24:50 <sgothel> which method ?
20131011 13:25:16 <sgothel> glDraw* gl*Pointer ?
20131011 13:25:21 <masterzen> I think the exception is correct, our code mostly uses client side buffers
20131011 13:25:27 <sgothel> good
20131011 13:25:38 <masterzen> since it's apparently not supported in core context
20131011 13:26:01 <sgothel> yup .. better than SIGSEGV later by customer :)
20131011 13:26:34 <sgothel> but which method ? since a GL2ES2 doesn't expose API entry anymore ..
20131011 13:28:29 <masterzen> sgothel: jogamp.opengl.gl4.GL4bcImpl.glVertexAttribPointer(GL4bcImpl.java:37492)
20131011 13:28:46 <masterzen> from com.jogamp.opengl.util.glsl.ShaderState.vertexAttribPointer
20131011 13:29:45 <sgothel> ha! :)
20131011 13:30:04 <sgothel> ok .. our own stuff avoiding it ..
20131011 13:30:11 <sgothel> yup .. good one
20131011 13:30:16 <sgothel> (no bug)
20131011 13:30:52 <sgothel> maybe we should throw it earlier in the GLArrayData*
20131011 13:31:30 <sgothel> remember the Android devices w/ ES3 (I still have none) .. it could crash there
20131011 13:31:43 <sgothel> *could have crashed* there
20131011 13:32:34 <xranby> sgothel: can we test it by build an android-x86 setup for one of the desktop machines?
20131011 13:33:03 <xranby> then testing using the latest nvidia/AMD drivers
20131011 13:33:51 <sgothel> I don't know how emulator work - but testing Bug 852 can be simply done in a unit test ... no need for this overkill
20131011 13:34:01 <sgothel> however .. yes, intel android is surely coming up
20131011 13:34:06 <sgothel> (orthogonal)
20131011 13:34:15 <xranby> this would be possible to test inside a vitrual machine.. no emulator
20131011 13:34:35 <sgothel> yes .. but really not a Bug 852 test will all those 'middle man' :)
20131011 13:35:07 <sgothel> I will simply do some hard-cast tests w/ GL3 core profile .. expecting the exception (like Brice found)
20131011 13:35:21 <sgothel> not that I don't trust myself - but we want to be test driven :)
20131011 13:35:46 <sgothel> so me happy w/ the findings
20131011 13:35:54 <sgothel> that damn URI thing ..
20131011 13:36:11 <xranby> https://groups.google.com/forum/#!msg/android-x86/RaIP7qcVitw/rRXZw_o-ZxAJ - * The Current Status of Android-x86 Mars / 2013
20131011 13:36:14 <sgothel> so our URI encode/decode must include '\\' hmm
20131011 13:36:22 <xranby> AMD and Intel are supported using Mesa
20131011 13:36:33 <sgothel> oh .. good
20131011 13:36:53 <sgothel> yeah .. month ago .. Mesa changes implied that (intel)
20131011 13:37:25 <sgothel> really good news for the more open-source driver front :)
20131011 13:37:56 <sgothel> AMD ? They like to continue pushing mobile ? Hope they have enough folks left :/
20131011 13:40:47 <sgothel> will be another looong day w/ the Android emulator .. if it is working now, last time I checked was .. err month ago
20131011 13:41:16 * xranby (~xranby@anon) Quit (Ping timeout: 264 seconds)
20131011 13:41:20 <sgothel> but .. maybe we can try that android API impl. for desktop then :)
20131011 13:43:10 * xranby (~xranby@anon) has joined #jogamp
20131011 14:08:03 <xranby> i wonder which android-x86 port to use....
20131011 14:08:05 <xranby> http://www.android-x86.org/
20131011 14:08:25 <xranby> they have some 4.3 images
20131011 14:08:54 <sgothel> if having a simple 'add jar file[s] here' .. great :)
20131011 14:13:31 <xranby> https://groups.google.com/forum/#!msg/android-x86/iQZ7LaX16Vw/B5H97e2TgCcJ - Q: "Is Opengles 3 supported on 4.3 build?" A: "It is not currently, but the drivers are providing OpenGL ES 3 compatibility"
20131011 14:13:45 <xranby> ... that answer looks.. hmm good
20131011 14:15:20 <xranby> (16:08:55) sgothel: if having a simple 'add jar file[s] here' .. great :) <- maybe we can use the android debug adb!? tool to insert and install the apk across the network?
20131011 14:15:39 <xranby> i will download this and test to run it inside vitrualbox
20131011 14:16:06 <sgothel> biggest benefit: enable us running unit tests un-touched :)
20131011 14:16:17 <sgothel> if .. we can tweak the activity ..
20131011 14:16:32 <xranby> http://www.android-x86.org/download
20131011 14:17:05 <xranby> Android-x86-4.3-devel soulds good to test
20131011 14:17:09 <sgothel> other than that .. well, not so excited about it since we are an API abstraction already .. more for devs IMHO
20131011 14:17:52 <xranby> they have a ndk as well
20131011 14:18:02 <xranby> Android x86 ndk
20131011 14:18:13 <sgothel> hmm
20131011 14:18:13 <xranby> err.. hmm removed!?
20131011 14:18:52 <xranby> http://www.android-x86.org/documents/ndk-how-to
20131011 14:18:53 <sgothel> for sure it's another 'middleware' which could cause problems .. so yeah, for development maybe great ..
20131011 14:19:11 <xranby> do google provide an x86 ndk?
20131011 14:19:11 <sgothel> my biggest issue: how to create Activities on the fly for unit tests ..
20131011 14:19:17 <sgothel> sure
20131011 14:19:28 <sgothel> how could you build it otherwise ?
20131011 14:19:54 <xranby> if they only provide a toolchain targeting arm you may have issues building x86 apps
20131011 14:20:03 <sgothel> that is right :)
20131011 14:20:29 <xranby> we need a ndk that can compile for the android x86 bionic
20131011 14:20:32 <sgothel> we had that Intel guy here .. month ago .. he confirmed it's working (he was doing JOCL)
20131011 14:20:43 * Eclesia (jsorel@anon) has left #jogamp
20131011 14:21:39 <xranby> ok http://software.intel.com/en-us/articles/android-ndk-for-intel-architecture "To support the Intel® Atom™ architecture, NDK r6b or later is required."
20131011 14:22:16 <sgothel> android-ndk-r8d <- current version we use
20131011 14:26:52 <sgothel> offthread: a bunch of tweets about BIOS attacks: https://twitter.com/dragosr/status/388511686744764416 https://twitter.com/dragosr/status/388512915742937089 https://twitter.com/dragosr/status/388521551693217792 .. etc
20131011 14:27:09 <sgothel> http://blog.fefe.de/?ts=aca6c75b
20131011 14:28:16 <xranby> interesting indeed
20131011 14:30:11 <sgothel> http://lists.freedesktop.org/archives/xorg/2013-October/056074.html <- Maturity is 'if you fix a 20 year old bug' :)
20131011 14:31:24 <xranby> growing up is to remove bugs older than yourself
20131011 14:32:22 <xranby> the future it security staff will be expected to grow up fast
20131011 14:33:03 <xranby> software culture is to grow bugs into features
20131011 14:34:02 <xranby> it will guarantee mortage-driven development
20131011 14:34:25 <sgothel> security staff (and evil trojan writers) will be historians ..
20131011 14:34:42 <sgothel> yeah .. fun, if you can look back on more than 20 years
20131011 14:34:56 <sgothel> 10 more to go :)
20131011 14:35:53 <sgothel> should look whether I can harness more than 'olympic rings demo' from ole GL4Java :)
20131011 14:36:08 <sgothel> preserving history .. and predate jogl :)
20131011 14:38:29 * xranby (~xranby@anon) Quit (Ping timeout: 248 seconds)
20131011 14:39:03 * xranby (~xranby@anon) has joined #jogamp
20131011 15:57:15 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20131011 16:24:00 * ebrayet (~Adium@anon) Quit (Ping timeout: 252 seconds)
20131011 16:52:09 * monsieur_max (~maxime@anon) has joined #jogamp
20131011 16:58:30 * [Mike] (~Mike]@anon) has joined #jogamp
20131011 18:53:52 * Eclesia (~eclesia@anon) has joined #jogamp
20131011 18:53:58 <Eclesia> hi
20131011 18:59:37 <Eclesia> question : is it possible to make multiple GL objects sharing the same space ? objective is to have one thread for the rendering loop and other threads doing updates, data loading, stuff like that
20131011 19:21:02 <sgothel> so the 'same space' here is 'shared GL objects' -> YES
20131011 19:22:35 <sgothel> I 'just' added GLDrawableFactory.createDummyDrawable(..) for this purpose (dummy drawable w/ real GLContext) ..
20131011 19:22:38 <sgothel> dummyDrawable = factory.createDummyDrawable(device, true, glp); // own device!
20131011 19:22:38 <sgothel> dummyDrawable.setRealized(true);
20131011 19:22:38 <sgothel> sharedGLCtx = dummyDrawable.createContext(glCtx);
20131011 19:22:57 <sgothel> -> GLMediaPlayerImpl.StreamWorker.initGL(GL) example ..
20131011 19:23:15 <sgothel> i.e. this way .. less 'clutter'
20131011 19:24:00 <sgothel> important: the shared ctx shall not use the same drawable/surface .. otherwise they would block each other (sync/locking)
20131011 19:24:16 <Eclesia> logic
20131011 19:24:46 <Eclesia> sgothel: is it possible to share stuffs with opencl too ?
20131011 19:24:57 <sgothel> yes
20131011 19:25:22 <sgothel> there is at least one JOCL/JOGL example .. pls check yourself .. busy w/ hotfixes for 2.1.1 :)
20131011 19:25:54 <Eclesia> sure sorry for bothering you ;)
20131011 19:34:26 * ebrayet (~Adium@anon) has joined #jogamp
20131011 19:35:06 * Eclesia doesn't have a huge objective for this week-end. Billboards and basic shapes.
20131011 19:39:55 <sgothel> not bothering .. just excusing myself
20131011 19:39:59 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20131011 19:40:23 <sgothel> actually .. if you like to jump in contributing to JOCL .. etc .. you are most welcome.
20131011 19:44:56 <Eclesia> sgothel: i must start by learning OpenCL first. there are plenty of image algorithms in my project which would benefit from it.
20131011 19:45:10 <Eclesia> but only 24Hours a day ...
20131011 19:49:12 <Eclesia> before that I would like to make a proper mkv reader with at least mepg-2 and vorbis codecs
20131011 19:50:08 <sgothel> right
20131011 19:50:11 <Eclesia> perhaps a stream media format too. this way I'll be able to finish the media api
20131011 19:50:30 <sgothel> create your own libav java/opencl impl :)
20131011 19:50:42 <sgothel> 24h/d ..
20131011 19:50:58 <Eclesia> libav ?
20131011 19:51:03 <sgothel> ffmpeg/libav
20131011 19:51:47 <Eclesia> lot of work to do as you can see :D
20131011 19:52:13 <Eclesia> have to prepare my presentation at the JUG too
20131011 19:54:00 <sgothel> you are either a genius super power dude .. or a slight little bit overloaded / underestimating things .. yeah, always up for a good joke :)
20131011 19:54:46 <sgothel> [even though we Germans are more renown for not 'getting it']
20131011 19:54:56 <Eclesia> 30% first one, 60% second, 10% crazy
20131011 19:58:58 <Eclesia> sgothel: just for curiosity, you work in which company ?
20131011 20:04:33 <Eclesia> (okay, bad question, won't ask again)
20131011 20:07:52 <sgothel> work for my wife - best company - good question :)
20131011 20:08:13 <sgothel> -> see wiki .. maintainer ..
20131011 20:08:20 <sgothel> nothing to hide :)
20131011 20:09:28 <sgothel> and yourself ? for another company .. using your great results I hope (and supporting you)
20131011 20:09:42 <sgothel> you mentioned it .. memories ..
20131011 20:09:55 * monsieur_max (~maxime@anon) has left #jogamp
20131011 20:10:13 <Eclesia> http://www.geomatys.com
20131011 20:11:06 <Eclesia> unfortunatly, they don't use my project. but we use JOGAMP ;)
20131011 20:14:03 <Eclesia> (can't show you, navy restricted datas)
20131011 20:51:28 <Eclesia> ++
20131011 20:51:36 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20131011 21:16:53 * xranby (~xranby@anon) Quit (Ping timeout: 248 seconds)
20131011 21:21:19 * xranby (~xranby@anon) has joined #jogamp
20131011 21:44:41 * jk4 (~jk4@anon) has joined #jogamp
20131011 21:51:50 * ebrayet (~Adium@anon) has joined #jogamp
20131011 21:55:19 * ebrayet (~Adium@anon) Quit (Client Quit)
20131011 22:10:26 * xranby (~xranby@anon) Quit (Ping timeout: 264 seconds)
20131011 22:15:39 * xranby (~xranby@anon) has joined #jogamp
20131011 22:22:11 * ebrayet (~Adium@anon) has joined #jogamp
20131011 22:22:19 * ebrayet (~Adium@anon) Quit (Client Quit)
20131011 22:31:45 * void256 (~void@anon) has joined #jogamp
20131012 01:05:47 * void256_ (~void@anon) has joined #jogamp
20131012 01:09:08 * void256 (~void@anon) Quit (Ping timeout: 240 seconds)
20131012 01:09:12 * void256_ is now known as void256
20131012 01:58:46 * void256 (~void@anon) Quit (Quit: ChatZilla 0.9.90.1 [Firefox 24.0/20130910160258])
20131012 05:05:32 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20131012050532.html