#jogamp @ irc.freenode.net - 20130909 05:06:11 (UTC)


20130909 05:06:11 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20130908050611.html
20130909 05:06:11 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130909050611.html
20130909 06:47:47 * monsieur_max (~maxime@anon) has joined #jogamp
20130909 07:36:56 <xranby> sgothel: 7u40 is still not released, now to be released in september,http://openjdk.java.net/projects/jdk7u/releases/7u40.html https://jdk7.java.net/download.html <- JDK 7 Update 40 Early Access Release
20130909 07:37:09 <xranby> sgothel: chich build did you test?
20130909 07:37:12 <xranby> which
20130909 09:12:39 * xranby (~xranby@anon) has left #jogamp
20130909 09:35:11 * sgothel (~sven@anon) has joined #jogamp
20130909 09:35:11 * ChanServ sets mode +v sgothel
20130909 15:04:41 <sgothel> https://jogamp.org/wiki/index.php/SW_Tracking_Report_Objectives_for_the_release_2.1.0#Multimedia <- cleaned up
20130909 15:33:27 * hharrison (~chatzilla@anon) Quit (Remote host closed the connection)
20130909 15:48:36 * hharrison (~chatzilla@anon) has joined #jogamp
20130909 15:56:53 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20130909 17:26:51 <sgothel> Hi Harvey ..
20130909 17:27:03 <sgothel> re OSX 7u40 .. CALayer .. etc
20130909 17:31:45 <hharrison> yes?
20130909 17:32:01 <sgothel> how can I verify this ? .. will add a bug report .. hmm
20130909 17:32:25 <sgothel> i.e. I was testing w/ 7u40 fc .. and I haven't seen any artifacts .. i.e. all good
20130909 17:32:34 <hharrison> I've reached out to some OSX testers for results with the u40 release
20130909 17:32:49 <sgothel> jdk-7u40-fcs-bin-b40-macosx-x86_64-16_aug_2013.dmg .. I was using
20130909 17:33:14 <sgothel> http://forum.jogamp.org/Mac-OS-X-10-7-Oracle-JRE-7-Swing-integration-issues-td4027780.html#a4029292 <- forum post
20130909 17:33:25 <sgothel> whats a good test case ?
20130909 17:33:27 <hharrison> perhaps we should treat this as 'wait for Harvey to get test results form people, file bugs if needed'
20130909 17:33:33 <sgothel> :)
20130909 17:33:39 <hharrison> VZome was a good easy to run test case
20130909 17:33:42 <sgothel> i.e. a good unit test pure JOGL ?
20130909 17:34:06 <sgothel> at least I saw the misplacement in applets .. sometimes (even w/ our workaround) .. now not
20130909 17:34:18 <hharrison> I'll have to think about that, it's very possible there's other problems that may be Java3D's fault
20130909 17:34:19 <sgothel> I may remove workaround .. and see if it looks ok
20130909 17:34:38 <hharrison> Removing the workaround would likely be the best baseline to work from...I dunno
20130909 17:34:58 <sgothel> yup .. my thinking too .. plus, this would properly document the workaround :)
20130909 17:35:03 <sgothel> (i.e. the code sections)
20130909 17:35:39 <hharrison> I'd suggest in the long run, it may be best _if_ we can get by without the workaround
20130909 17:35:49 <sgothel> ay
20130909 17:36:03 <hharrison> Not hard to restore it if required, but just say "requires u40 or better for Java7"
20130909 17:36:07 <hharrison> ...on OSX
20130909 17:36:45 <hharrison> August seemed to have a good test case for this, I'll ping him and ask what his tests involved
20130909 17:37:07 <sgothel> a good summary line ? "OSX AWT/CALayer components get positioned erratically .." ?
20130909 17:38:09 <sgothel> .. or "JAWT native CALayer not positioned over Canvas" -> http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=2229714 :)
20130909 17:38:19 <sgothel> guess thats best :)
20130909 17:39:23 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=816 <- magic of bugzilla :)
20130909 17:39:49 <hharrison> The test case attached to 816 may be best here
20130909 17:40:21 <hharrison> If that works for you, I'l treat any other problems that come up as 'likely Java3D's fault'
20130909 17:40:55 <hharrison> Or, Java3D's fault, until proven otherwise ;-)
20130909 17:41:20 <hharrison> But I do have testers for this, so shouldn't take long to track down I think
20130909 17:41:36 <sgothel> goody
20130909 17:42:55 <sgothel> oh .. for next siggraph, since I guess you know the folks who organize things .. (?) .. can you try to get a schedule so we don't compete w/ Khronos ? :)
20130909 17:44:32 <sgothel> .. and since we have video input now, maybe your video encoder stuff might come in handy .. dunno :)
20130909 17:44:50 <sgothel> trying to brainstorm here ..
20130909 17:45:17 <sgothel> (at least API wise .. like GLMediaRecorder .. dunno)
20130909 17:45:30 <sgothel> then .. why 'GL' :)
20130909 17:46:24 <hharrison> I know one person, will make the request
20130909 17:46:36 <sgothel> awesome - thx! Have you ever tested stereo GL ?
20130909 17:47:21 <sgothel> me personally a bit sick of the lack of standardization and craziness (gpus, types, fake/real, ..)
20130909 17:47:36 <sgothel> last time was .. err .. around 2001 I guess :)
20130909 17:53:29 <hharrison> I think it did work at one point, but I think the changes to GLCapabilities (the Java3d version) may have omitted all the stereo handling bits
20130909 17:53:47 <hharrison> Would have to compare the old OpenGL code and the JOGL backend
20130909 17:54:10 <hharrison> So...likely broken because we just don;t turn it on currently, but the code is probably all in place
20130909 17:54:49 <sgothel> so this stereo should work .. err w/ those 120hz lcd's and those shutter glasses plus a pro-gpu-card I assume ?
20130909 17:54:56 <hharrison> I need to look through that code anyway and stomp out the latent PBuffer stuff, rely on the JOGL offscreen interface
20130909 17:55:13 <hharrison> I'd assume so
20130909 17:55:17 <sgothel> that would be nice ..
20130909 17:56:00 <sgothel> especially b/c this will expose FBObject more .. and hence get audited/stabilized more ..
20130909 17:56:42 * hrenovka (hrenovka@anon) has joined #jogamp
20130909 17:56:52 <sgothel> some code path right now .. like GLJPanel are everything but 'straight forward' :) .. more like surreal :)
20130909 17:57:03 <hharrison> Java3d was built for cave environements, and really tries hard to keep all the needed state for stereo viewing
20130909 17:57:34 <sgothel> which is ? 4 buffers ?
20130909 17:57:38 * hrenovka is now known as Dawn
20130909 17:58:13 * Dawn is now known as The
20130909 17:58:34 * The is now known as The_Dawn
20130909 17:58:36 <hharrison> I believe it keeps the concept of stereo and double/single buffering orthogonal
20130909 17:59:32 <hharrison> So you either have mode CYCLOPIAN (no joke) or it exposees the separate EYE_LEFT, EYE_RIGHT rendering pass and allocates the extra buffers if needed
20130909 17:59:47 <sgothel> hmm .. thought you just need to render to back-left, back-right .. w/ 2 viewports (eye distance) .. swap, done
20130909 17:59:55 <hharrison> I'll admit, I have not really even looked at what happens in the backend pipeline, just descirbing the Java3d rendering model
20130909 18:00:03 <sgothel> ah
20130909 18:00:09 <sgothel> hehe cyclops :)
20130909 18:00:50 <sgothel> the graph .. yes .. must deal w/ the 2 cameras as well ..
20130909 18:46:15 <hharrison> Let me know if you puch out a build with the workarounds removed, I'll make sure to point testers at that one as well as plain 2.0.2
20130909 20:39:36 * chakie_ is now known as chakie
20130909 20:44:20 * void256 (~void@anon) has joined #jogamp
20130909 21:26:16 * The_Dawn (hrenovka@anon) Quit ()
20130909 22:37:51 * void256 (~void@anon) Quit (Remote host closed the connection)
20130910 03:13:38 * juank_prada (~juank_pra@anon) has joined #jogamp
20130910 04:57:11 * juank_prada (~juank_pra@anon) Quit (Read error: Connection reset by peer)
20130910 05:06:11 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20130910050611.html