#jogamp @ irc.freenode.net - 20130621 05:05:27 (UTC)


20130621 05:05:27 -CatOut- Previous @ http://jogamp.org/log/irc/jogamp_20130620050527.html
20130621 05:05:27 -CatOut- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130621050527.html
20130621 06:30:28 * monsieur_max (~maxime@anon) has joined #jogamp
20130621 06:31:51 * [Mike] (~Mike]@anon) Quit (Read error: Connection reset by peer)
20130621 07:38:08 * ebrayet (~Adium@anon) has joined #jogamp
20130621 08:43:52 * h05h1 (~h05h1@anon) has joined #jogamp
20130621 09:13:16 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20130621 09:14:02 * ebrayet (~Adium@anon) has joined #jogamp
20130621 10:11:11 <sgothel> @ALL: Latest security changes tested manually Linux/Window and Android - all good.
20130621 10:11:29 <sgothel> Besides .. toolchain is now working w/ java7
20130621 10:12:11 <sgothel> now .. finishing the jenkins nodes to acually do builds again - and to cont. w/ ES3 + GL4.3 :)
20130621 10:14:06 <sgothel> @ALL: Whoever uses JogAmp w/ SecurityManager (JNLP / Applet / ..) _please_ update - and remove old binaries !! (we will do so soon!)
20130621 10:16:25 <sgothel> having a nap now .. laters
20130621 10:20:03 <sgothel> ohh .. @Xerxes and @All: SIGGRAPH SLIDES please :/
20130621 10:38:03 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20130621 10:43:58 * xranby1 (~familjen@anon) has joined #jogamp
20130621 10:49:15 <xranby1> good day,
20130621 10:49:56 <monsieur_max> sgothel: well done :)
20130621 10:50:00 <xranby1> sgothel: a little direction, do we want the slides to cover shore to shore of all activity of jogamp or do we want some specific focus for this years bof?
20130621 10:50:42 <xranby1> i need to check if the rewritten security code have improved jogamp + icedtea-web use
20130621 10:54:50 <xranby1> sgothel: do any of the jenkins nodes test the AMD fglrx driver using older gpu cards? i started to see issues on my ATI RV770 [Radeon HD 4850] system after switching from the opensource mesa drivers to the fglrx driver, a crash at shutdown inside the driver. :/ i will create a bugreport when i have a reliable way to reproduce it
20130621 10:57:04 <xranby1> i will try reproduce this issue with http://jogamp.org/deployment/archive/master/gluegen_686-joal_448-jogl_1011-jocl_810-signed/archive/
20130621 10:57:53 * ebrayet (~Adium@anon) has joined #jogamp
20130621 11:27:07 <xranby1> sgothel: i am able to reproduce the above by performing an unclean shutdown triggered by catching an ALException inside the GLEventListener display http://paste.ubuntu.com/5786534/
20130621 11:28:29 <xranby1> "reproducer" http://paste.ubuntu.com/5786542/
20130621 11:28:46 * ebrayet (~Adium@anon) Quit (Quit: Leaving.)
20130621 11:34:58 <xranby1> the crash can be more or less serve: http://paste.ubuntu.com/5786553/ any sugegstions on how to debug this further is welcome
20130621 11:37:50 <xranby1> http://paste.ubuntu.com/5786557/ - the hotspot crash log
20130621 11:47:06 <xranby1> crash variant 2 http://paste.ubuntu.com/5786578/
20130621 11:47:48 <xranby1> crash log 2 http://paste.ubuntu.com/5786579/
20130621 12:01:24 <xranby1> looks like a possible race condition calling jogamp.newt.driver.x11.DisplayDriver.dispatchMessagesNative() after closeNativeImpl() ... javaObjectAtom and
20130621 12:01:24 <xranby1> windowDeleteAtom may be 0 ?!
20130621 12:11:49 * h05h1 (~h05h1@anon) Quit (Quit: Leaving)
20130621 12:14:35 <xranby1> nope.. installing X11 debugsymbols
20130621 12:15:30 <xranby1> crash in xcb_in.c:533
20130621 12:17:12 <xranby1> poll_for_next_event (c=0x0, queued=0) at ../../src/xcb_in.c:533
20130621 12:21:46 <sgothel> yup .. the 'dirty exit' problem .. if possible, pls create a _short_ manual test
20130621 12:22:06 <sgothel> IMHO this only happens sometimes w/ fglrx due to race conditions - I suspect the driver
20130621 12:22:25 <sgothel> or .. display handle invalid
20130621 12:23:41 <sgothel> yes a post shutdown .. dispatchMessagesNative() is still kicking .. :/
20130621 12:24:16 <sgothel> we live under the assumption that this is not possible :)
20130621 12:25:18 <sgothel> http://paste.ubuntu.com/5786557/ ..
20130621 12:26:21 <sgothel> so the X11Util shutdown-hook must have the ability to register a callback to ensure that 'client' is dead before X11Util cont.
20130621 12:26:23 * xranby1 (~familjen@anon) Quit (Ping timeout: 260 seconds)
20130621 12:26:33 <sgothel> killing display handles ..
20130621 12:26:56 <sgothel> a nice ( .. ohh xerxes gone )
20130621 12:27:34 <sgothel> .. (log) .. > test w/ disabling closing remaining display handles @ X11Util shutdown hook
20130621 12:30:54 <sgothel> : closePendingDisplayConnections() -> test w/o it .. - enforcing closing the display handles exists purely for aesthetic reasons, trying to tame a buggy driver to bring back some order :]
20130621 12:37:04 * xranby1 (~familjen@anon) has joined #jogamp
20130621 12:37:17 <xranby1> back, i had to reboot, my GPu locked up
20130621 12:37:20 <xranby1> https://jogamp.org/bugzilla/show_bug.cgi?id=761
20130621 12:37:45 <xranby1> that was the best backtrace i was able to reproduce
20130621 12:40:38 <xranby1> sgothel: you want me to try disable the quirk?
20130621 12:40:52 <xranby1> Quirks [DontCloseX11Display]
20130621 12:46:13 <xranby1> commented out // closePendingDisplayConnections()
20130621 12:46:19 <xranby1> lets see
20130621 12:50:19 <xranby1> sgothel: it looks to behave well after commenting out that line yes
20130621 13:08:47 <rmk0> anyone in here Noel Grandin?
20130621 13:18:35 * xranby1 (~familjen@anon) Quit (Read error: No route to host)
20130621 13:48:06 * petrs (~petrs@anon) Quit (Read error: Connection reset by peer)
20130621 13:52:50 <rmk0> guess not!
20130621 14:18:53 * monsieur_max (~maxime@anon) has left #jogamp
20130621 15:03:10 <hharrison> Anybody have any experience with Win8 and jogl?
20130621 15:04:07 <hharrison> good/bad/ugly?
20130621 15:23:51 * [Mike] (~Mike]@anon) has joined #jogamp
20130621 15:27:38 <rmk0> i've had someone running windows 8 run my unit tests
20130621 15:28:01 <rmk0> quite a few failed, but they failed upon trying to actually open a display with a specific format
20130621 15:28:17 <rmk0> so there may be a lack of flexibility there... haven't had the chance to examine it further
20130621 15:28:32 <rmk0> that was on an intel HD 4000
20130621 16:00:31 <hharrison> Thanks for that, we'll likely start having some Win8 machines in the mix over here soon
20130621 16:01:05 <hharrison> There's also a report this morning in the Java3d forum with problems on a HD4000 and Win8
20130621 16:01:36 <hharrison> Do you have any specifics on formats that passed/failed?
20130621 16:02:42 <rmk0> hm, i thought it was a specific format, but it may just have been an offscreen buffer
20130621 16:03:00 <rmk0> http://fossil.io7m.com/repo.cgi/io7m-jcanephora/artifact/4b30a4ef16b084f2d96508c5508feda1d5a1f645
20130621 16:03:30 <rmk0> is those functions
20130621 16:05:00 <rmk0> are there any windows 8 jenkins nodes currently?
20130621 16:14:08 <hharrison> Also, Win8 is not shipping Intel drivers by default anymore, you need to explicitly install the opengl driver from Intel to have any hope
20130621 16:14:24 <rmk0> ugh!
20130621 16:14:54 <hharrison> Well, Intel opengl drivers anyways, you get the shitty OpenGL 1.1 basic driver otherwise
20130621 16:15:17 * monsieur_max (~maxime@anon) has joined #jogamp
20130621 18:22:52 <hharrison> I still get a kick out of each Java3d-project that crawls out of the woodwork
20130621 18:57:03 <monsieur_max> hharrison: any link to provide about a recent java3D stuff ?
20130621 18:57:34 <hharrison> Just the note that came across today in the java3d forum
20130621 18:58:00 <hharrison> But recently, I've had almost zero time for java3d stuff
20130621 18:59:08 <hharrison> almost managed to get java3d-utils and java3d-core properly segregated so utils depends only core instead of needing to co-build them
20130621 18:59:46 <hharrison> Two functions away from finishing that, then I can get java3d packaged for fedora (they don't like the utils license and won't include it)
20130621 19:00:48 <hharrison> That is essentially what I am going to release as 1.6.0
20130621 20:05:44 <monsieur_max> not really related to jogamp but still quite interesting : http://games.greggman.com/game/android-vs-ios-game-myths/
20130621 21:40:15 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20130622 00:29:59 * hharrison (~chatzilla@anon) Quit (Quit: ChatZilla 0.9.90 [Firefox 22.0/20130617145905])
20130622 03:01:02 <sgothel> @Max: Thank you - was telling for ~2years that the iOS thing won't matter anymore :)
20130622 03:01:44 <sgothel> other official numbers match this of the game, i.e. iOS / Android deployment
20130622 03:02:09 <sgothel> ofc .. the difficulty to generate revenue is something orthogonal :)
20130622 03:02:27 <sgothel> Win8 ? They stopped doing GL drivers already ? Wow.
20130622 03:02:40 <sgothel> I just played w/ Win8 Beta .. it worked on the old API
20130622 03:03:23 <sgothel> We consider[ed] getting a Win8 test node .. hmm, maybe even support the new API (multi-touch, ..) - But w/o GL drivers ?
20130622 03:09:42 <sgothel> Re: iOS vs Android: What hasn't taken into account is direct deployment and other app stores - just look at China, they usually don't use Google'e App Store but their own ..
20130622 03:10:26 <sgothel> So my advise to deploy an Android game actually would be to also use a website w/ signed APKs :)
20130622 03:11:17 <sgothel> Since Google doesn't allow you to generate revenue other than w/ their store (IANAL ..) .. you have to rethink that maybe ..
20130622 05:05:27 -CatOut- Continue @ http://jogamp.org/log/irc/jogamp_20130622050527.html