#jogamp @ irc.freenode.net - 20130619 19:08:27 (UTC)


20130619 19:08:27 -PircBot- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130619190827.html
20130619 19:08:27 -cameron.freenode.net- *** Looking up your hostname...
20130619 19:08:27 -cameron.freenode.net- *** Checking Ident
20130619 19:08:27 -cameron.freenode.net- *** No Ident response
20130619 19:08:27 -cameron.freenode.net- *** Found your hostname
20130619 19:08:33 * CatOut (~jogamp@anon) has joined #jogamp
20130619 19:08:33 * Topic is 'http://jogamp.org | Hacking 3D Graphics, Multimedia and Processing across Devices'
20130619 19:08:33 * Set by rmk0 on 20130116 23:58:04
20130619 19:32:30 <sgothel> jenkins: uses 7u25 now ..
20130619 19:32:41 <sgothel> jenkins-slaves: moving to 7u25
20130619 19:33:19 <sgothel> after that change and applying a few more JAR constraints and gluegen finalizations .. - we will have another RC due to security .. :|
20130619 19:33:37 <sgothel> i.e. build w/ java7 using jre6 rt.jar .. etc
20130619 20:08:01 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20130619 20:32:07 * ebrayet (~Adium@anon) has joined #jogamp
20130619 20:33:24 <sgothel> one more thing .. how is OSX retina support planned ? i.e. currently we are a bit 'off' in this regard.
20130619 20:35:03 <sgothel> anybody here has one ?
20130619 20:35:54 <sgothel> We have not planned to purchase one on our behalf .. so if this becomes an issue, we need to solve it somewhat.
20130619 20:36:22 <sgothel> dinner :)
20130619 20:36:30 <[Mike]> I don't have one!
20130619 20:36:34 <[Mike]> Enjoy!
20130619 20:36:45 <[Mike]> :)
20130619 20:41:13 * void256 (~void@anon) has joined #jogamp
20130619 20:58:09 * ebrayet (~Adium@anon) Quit (Ping timeout: 248 seconds)
20130619 21:01:42 * ebrayet (~Adium@anon) has joined #jogamp
20130619 21:05:19 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20130619 22:25:17 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20130619 22:32:51 * h05h1 (~h05h1@anon) Quit (Ping timeout: 246 seconds)
20130619 23:09:31 <sgothel> http://docs.oracle.com/javase/7/docs/technotes/guides/jweb/no_redeploy.html -> JAR Manifest tags: Permissions, Codebase
20130619 23:09:57 <sgothel> + Sealed
20130619 23:10:44 <sgothel> idea: .. Using 'Sealed' on jogamp.* classes impl. / direct acessing the JNI code - so there will be no 'derivation'
20130619 23:11:27 <sgothel> Codebase: *.jogamp.org in JAR file would prevent somebody copying the signed stuff and survive a vulnerability deletion ..
20130619 23:12:02 <sgothel> i.e. if we delete it .. it's done - and if users like to use signed jars on their servers, they would sign it w/ their keys
20130619 23:13:03 <sgothel> yes, this is more like snake oil'ish .. not real security - more a policy how to share / locate and use / derive classes
20130619 23:13:36 <sgothel> the 'Sealed' cannot be used on jogl right now, since we 'mix/match' atomic package namespace .. well
20130619 23:14:02 <sgothel> just a little brainstorming .. will test this after 7u25 is working on build system
20130619 23:49:03 * void256 (~void@anon) Quit (Remote host closed the connection)
20130620 01:30:00 * juank_prada (~juank_pra@anon) has joined #jogamp
20130620 01:30:40 <juank_prada> Hello guys, Is there any issue right now with Jogl 2rc11 and Ati open source driver radeon?
20130620 01:31:04 <juank_prada> I am not able to render a single thing on the screen when running my application on a machine with that driver
20130620 01:31:13 <juank_prada> other linux graphic cards work well
20130620 01:31:14 <sgothel> pls test latest version .. see Wiki
20130620 01:31:34 <sgothel> not that I know of .. however, rc11 is quite old
20130620 01:32:03 <juank_prada> sgothel, is the one from git? the latest one?
20130620 01:32:32 <sgothel> yes .. and see aggregated builds in wiki ..
20130620 01:35:23 <juank_prada> gonna try it, thanks
20130620 01:35:35 <juank_prada> btw, what are the plans for the final release? is it comming soon?
20130620 01:36:02 <sgothel> yes .. before siggraph, but will do an RC12 to test new stuff / setup ..
20130620 01:36:43 <juank_prada> sgothel, sorry what is siggraph?
20130620 01:37:28 <sgothel> http://jogamp.org/doc/siggraph2012/
20130620 01:38:58 <sgothel> http://www.siggraph.org/
20130620 01:39:35 <juank_prada> oh, good. thanks
20130620 01:40:05 <juank_prada> Im looking forward to the final release. You guys are making a great work with Jogl! congrats and thanks!
20130620 01:54:36 <juank_prada> so... I tried one of the aggregated builds (the latest) and not only it didn't fixed my issue, but now I am not able to close the window. Instead when I click the x button i get this error: http://pastebin.com/h3s9E6nJ
20130620 01:55:17 <juank_prada> The line that triggers everything is: canvas.destroy(). If I remove it i dont get That error, but shouldnt I destroy the GLWindow before exiting?
20130620 01:55:35 <sgothel> not you in the listener .. nope
20130620 01:55:41 <sgothel> so besides this ..
20130620 01:55:51 <sgothel> whats the bug ?
20130620 01:56:08 <juank_prada> well that pretty much no rendering is done
20130620 01:56:11 <sgothel> i.e. GL related ? which Radeon and driver ?
20130620 01:56:35 <sgothel> I rarely tested the open source driver ..
20130620 01:56:43 <sgothel> when I did, I used latest mesa git
20130620 01:56:56 <sgothel> so .. you test mesa >= 9 ?
20130620 01:57:19 <sgothel> can you run some simple tests ?
20130620 01:57:26 <sgothel> i.e. our gearses2 .. etc ?
20130620 01:57:54 <sgothel> i.e. proven to be correct code :)
20130620 01:58:10 <juank_prada> well doing a glxinfo i get this about mesa: OpenGL version string: 3.0 Mesa 9.2.0-devel
20130620 01:58:25 <juank_prada> and OpenGL renderer string: Gallium 0.4 on AMD RS880
20130620 01:58:27 <sgothel> well .. w/ your application (see wiki/faq) .. you can enable debug propeties, i.e. jogl.debug.DebugGL and/or jogl.debug.TraceGL
20130620 01:58:41 <sgothel> looks like something up-to-date
20130620 01:58:51 <juank_prada> yup
20130620 01:59:07 <sgothel> again .. if our simple code / demos works .. maybe it's in your app -> GL debug properties
20130620 01:59:20 <juank_prada> right, gonna check one of the demos
20130620 01:59:36 <sgothel> scenarion: developed w/ NV and now testing w/ mesa/ati ?
20130620 01:59:40 <sgothel> *scenario
20130620 02:01:02 <juank_prada> sgothel, no, developed under intel hd something (cant remember exactly which intel chipset was) then it was ported to this very same ati card but in Windows which works fine.
20130620 02:01:15 <juank_prada> later i tried to run the app in the same machine but under linux
20130620 02:01:30 <sgothel> well .. GL driver ..
20130620 02:01:36 <juank_prada> yup
20130620 02:01:42 <juank_prada> im blaming it
20130620 02:02:01 <sgothel> so try to make sense out of the mentioned debug props .. if you are lucky, you will see a GL error exception :)
20130620 02:02:30 <juank_prada> yup, I will come back screaming if i find some weird messages there :)
20130620 02:02:36 <sgothel> if you use shaders and our shader-utils also try to enable GLSL debugging ..
20130620 02:02:55 <sgothel> -Djogl.debug.GLSLCode -Djogl.debug.DebugGL -Djogl.debug.TraceGL
20130620 02:03:07 <juank_prada> I do use shaders but not the shader-utils. Will the GLSL debugging still work?
20130620 02:03:12 <sgothel> jogl/make$ vi scripts/tests.sh <- many examples ..
20130620 02:03:20 <sgothel> yup
20130620 02:03:27 <sgothel> oh GLSL .. nope :)
20130620 02:03:33 <sgothel> but GL debugging does
20130620 02:03:40 <sgothel> and trace .. ofc
20130620 02:03:53 <juank_prada> oh ok, gonna check the shader-utils then :) thank you very much for your helo sgothel
20130620 02:03:56 <sgothel> most of the time .. it's a GL usage bug, tolerated here .. but not there ..
20130620 02:03:58 <juank_prada> *help
20130620 02:23:15 * [Mike] (~Mike]@anon) Quit ()
20130620 02:26:51 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=758
20130620 02:28:46 <juank_prada> sgothel, I think I can discard errors on my application. I just tried to run the demos (JNLP) and I dont get any rendering in any of them
20130620 02:29:15 <sgothel> so what are the DEBUG props saying ? anything useful ?
20130620 02:29:53 <juank_prada> heh havent checked that yet. But I will in a few minutes
20130620 02:31:04 <juank_prada> first I need to understand how to work with that :P
20130620 02:31:15 <sgothel> wiki / faq
20130620 02:31:46 <juank_prada> yup im into that
20130620 02:38:44 <juank_prada> so there is quite a lot of information i get from the debugging messages
20130620 02:39:00 <juank_prada> not sure I understand everything it says though
20130620 02:39:24 <juank_prada> sgothel, If you have the time can you help me with this? -> http://pastebin.com/HCNJUmUM
20130620 02:39:58 <juank_prada> there is one particular exception thrown there
20130620 02:40:08 <juank_prada> java.lang.Exception: Info: setGL (OpenGL null): main-Display-.x11_:0.0-1-EDT-1, <null> -> GL4bcImpl, jogamp.opengl.gl4.GL4bcImpl@56d637ab
20130620 02:40:49 <sgothel> Info
20130620 02:42:40 <juank_prada> sgothel, sorry?
20130620 02:43:02 <juank_prada> what do you mean?
20130620 02:43:36 <sgothel> it says 'Info' :)
20130620 02:44:04 <sgothel> I see, we allow 3.1 ctx on Radeon - we have disabled it for Intel (non compliant ..)
20130620 02:44:15 <sgothel> you can compile jogl ?
20130620 02:44:23 <juank_prada> yup
20130620 02:44:28 <sgothel> well then ..
20130620 02:44:45 <juank_prada> not sure how to disable 3.1 on radeon though :P
20130620 02:44:49 <sgothel> GLContextImpl.java line 1643
20130620 02:44:53 <juank_prada> oh ok
20130620 02:45:04 <sgothel> "if( glRenderer.contains("Intel(R)") && compatCtx && ( major > 3 || major == 3 && minor >= 1 ) ) " ->
20130620 02:45:30 <sgothel> "if( major > 3 || major == 3 && minor >= 1 )"
20130620 02:45:34 <sgothel> just to test ..
20130620 02:45:44 <juank_prada> ok let me try
20130620 02:46:04 <sgothel> then we would need to refine ..
20130620 02:46:23 <sgothel> however - pls file a bug report w/ that debug trace you just dumped!
20130620 02:46:31 <sgothel> 7z/zip it .. and attach it
20130620 02:47:17 <sgothel> say: No Rendering Results w/ Mesa VERSION and RENDERER (gpu name, galium/sw/?? name, version)
20130620 02:47:39 <juank_prada> sure thing
20130620 02:48:23 <sgothel> in above patch .. you might be able to keep the '&& compatCtx' alive .. maybe test w/ a core profile as well ..
20130620 02:48:46 <juank_prada> sgothel, so I should just remove the Intel part?
20130620 02:48:50 <sgothel> if this works out .. we need good detailed infos to edit the quirk
20130620 02:48:54 <sgothel> yup ..
20130620 02:48:57 <juank_prada> k
20130620 02:49:22 <sgothel> and then test GEarsES2 w/ core profile (Test*GearsES2*Newt has a -gl3 switch)
20130620 02:49:30 <sgothel> read source ..
20130620 02:49:38 <juank_prada> ok
20130620 02:49:44 <sgothel> it's sad .. but yeah, we need to maintain those quirks due to buggy drivers :(
20130620 02:49:52 <juank_prada> ust a minute. Dinner is ready :)
20130620 02:50:00 <sgothel> bin appetite
20130620 02:59:24 * [Mike] (~Mike]@anon) has joined #jogamp
20130620 03:51:04 <juank_prada> sgothel, so If I remove both glRenderer.contains("Intel(R)") && compatCtx things work on my application. If I leave compatCtx && ( major > 3 || major == 3 && minor >= 1 ) it will crash when creating the context
20130620 03:52:13 <sgothel> fascinating .. please note both in bug report .. maybe a [git] patch file .. will merge later
20130620 03:52:15 <sgothel> thank you!
20130620 03:52:43 <sgothel> so .. it's not just Intel .. but all Mesa 9* > 3.0 ctx .. :(
20130620 03:52:47 <juank_prada> well Im just starting with this git thing.... Im not sure how to create a patch file
20130620 03:53:14 <sgothel> a unified diff would be ok as well, or look up: git patch via email ..
20130620 03:53:31 <sgothel> so your name etc is preserved .. otherwise I will add yours from the bug report
20130620 03:54:30 <juank_prada> I cant really tell about all mesa. but I still would blame the radeon driver. It still sucks
20130620 03:55:10 <sgothel> only Intel and ATI is avail AFAIK .. have you tested software ? but .. there might be no software ..
20130620 03:55:22 <sgothel> so .. we may just add the Radeon renderer string next to Intel ..
20130620 03:55:27 <sgothel> pls try that ..
20130620 03:55:30 <sgothel> btw ..
20130620 03:55:45 <sgothel> you said 'If I leave compatCtx && ( major > 3 || major == 3 && minor >= 1 ) it will crash when creating the context" ?
20130620 03:55:53 <sgothel> w/ a core GL3 test ?
20130620 03:56:35 <juank_prada> i meant it will crash when creating the canvas (GLWindow)
20130620 03:56:42 <juank_prada> Im sleepy right now :P
20130620 03:56:44 <sgothel> what the proper value for Radeon, for "glRenderer.contains(RENDERER)" ?
20130620 03:56:58 <sgothel> oh .. so laters .. have a nap!
20130620 03:57:13 <sgothel> no rush .. tomorrow would be great! :)
20130620 03:57:14 <juank_prada> sorry. Im not following you about the glRenderer.contains
20130620 03:58:22 <sgothel> we may like to confine the quirk .. i.e. " ( ( glRenderer.contains("Intel(R)") & compatCtx ) || YOUR RADEON ID ? ) && ( major > 3 || major == 3 && minor >= 1 )
20130620 03:58:39 <sgothel> so we have glRenderer and glVendor AFAIK
20130620 04:00:21 <juank_prada> oh ok
20130620 04:01:08 <sgothel> if you can add the dump of etc/test_dbg.sh .. it's all in there AFAIK :)
20130620 04:03:50 <juank_prada> i think running that script is throwing some sort of exception
20130620 04:03:58 <juank_prada> I will take a look at that tomorrow if that is ok
20130620 04:04:06 <sgothel> thank you .. good night
20130620 04:04:33 <juank_prada> no, thank you. Now I got a working application on this crappy machine!
20130620 04:05:07 <juank_prada> I will do all I can to help you guys with this .... just... tomorrow :D
20130620 04:05:29 <juank_prada> sgothel, see you tomorrow then. and thanks again
20130620 04:05:30 <sgothel> 'thank you' <- I mean it ! relax :)
20130620 04:05:34 <sgothel> ay
20130620 04:05:51 * juank_prada (~juank_pra@anon) Quit (Quit: Leaving)
20130620 05:05:27 -CatOut- Continue @ http://jogamp.org/log/irc/jogamp_20130620050527.html