#jogamp @ irc.freenode.net - 20130710 09:21:02 (UTC)

20130710 09:21:02 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20130710021747.html
20130710 09:21:02 -NickServ- This nickname is registered. Please choose a different nickname, or identify via /msg NickServ identify <password>.
20130710 09:21:12 * jogamp (~jogamp@anon) has joined #jogamp
20130710 09:21:12 * Topic is 'http://jogamp.org | Hacking 3D Graphics, Multimedia and Processing across Devices'
20130710 09:21:12 * Set by rmk0 on 20130116 23:58:04
20130710 09:21:12 -NickServ- You are now identified for jogamp.
20130710 09:44:43 * xranby (~xranby@anon) has joined #jogamp
20130710 10:03:59 * xranby1 (~familjen@anon) has joined #jogamp
20130710 10:13:16 * xranby1 (~familjen@anon) Quit (Quit: Leaving.)
20130710 10:22:44 <xranby> olamedia: hi! i have an experimental branch that give you sound using the GLMediaPlayer on linux X11
20130710 10:23:36 <xranby> olamedia: the GLMediaPlayer is designed to use the system installed libffmpeg/libav on Linux/Windows/Mac systems.. thus you will have to install libffmpeg/libav on windows manually
20130710 10:24:34 <xranby> olamedia: in the future we may bundle legally ok licenced audio/video codecs with jogamp... such as webm hopefully
20130710 10:26:07 <xranby> olamedia: the GLMediaPlayer back-end for android is currently the most feature complete implementation, where we can use the android media framwork to decode both audio and video for many formats out of the box.
20130710 10:27:13 <olamedia> >_< my alert sound was too loud
20130710 10:29:19 <olamedia> what about .jar of dlls?
20130710 10:29:34 <olamedia> of ffmpeg
20130710 10:31:05 <olamedia> replaced sound of highlighted messages ^_^
20130710 10:31:27 <olamedia> so, back to sleep now
20130710 10:31:38 <olamedia> will check your branch
20130710 10:31:51 <xranby> olamedia: of course you may bundle your own build of ffmpeg with jogl
20130710 10:32:27 <xranby> the reason why jogamp currently do not bundle the libs is because of legal issues distributing decoders for some media formats
20130710 10:33:02 <xranby> usually mpegla require a license for products that can decode some formats in some countrys
20130710 10:33:16 <xranby> due to patents
20130710 10:34:42 <xranby> olamedia: you can merge in my github FFMPEGMediaPlayer branch on top of the current JOGL git master https://github.com/xranby/jogl/commit/6509c313dec8947723a932e7d1646a78dd6ac49a
20130710 10:34:55 <xranby> if you want to test sound under X11
20130710 10:35:23 <xranby> olamedia: we plan to hack a bit more on this branch later this week (hopefully) and add a small video frame cache
20130710 10:35:35 <xranby> that will improve audio/video sync a bit more
20130710 10:35:42 <xranby> especially under load
20130710 10:37:18 <olamedia> audio frame cache sounds better and smaller
20130710 10:51:26 <xranby> right now we sync video to the audio.. because us humans gets easily upset if audio stops for some reason.. and yes we use an audio cache
20130710 10:52:17 <olamedia> when a skip frames option for a video
20130710 10:52:20 <olamedia> *than
20130710 10:52:24 <olamedia> *then
20130710 10:52:37 <olamedia> oh... i'm still sleeping )
20130710 10:52:47 <xranby> we run into issues if the file is packed funny for example:
20130710 10:53:00 <xranby> a file that looks like this may get us into trouble:
20130710 10:53:26 <olamedia> "Compile FFmpeg without "--enable-gpl" and without "--enable-nonfree"." > and provide as a jar fallback
20130710 10:54:06 <xranby> V1 A1A2A3A4 V4 A5A6A7A8 V8 where the number is the desired presentation timestamp
20130710 10:54:39 <xranby> if the file is packed A1A2A3A4 V1 A5A6A7A8 V8 then everything is fine
20130710 10:55:07 <olamedia> Ax Vx - audio/video chunks?
20130710 10:55:13 <xranby> https://jogamp.org/bugzilla/show_bug.cgi?id=686
20130710 10:55:20 <xranby> yes A = audio frames V = video frames
20130710 10:55:35 <xranby> each frame got a decode timestamp and a presentation timestamt
20130710 10:56:04 <xranby> you want the frames to be heard or viewed at the desired presentation timestamp
20130710 10:56:15 <xranby> usually within 16ms
20130710 10:56:49 <xranby> without a video frame cache prevents us to buffer up audio
20130710 10:57:02 <olamedia> btw i need to get libav dlls or ffmeg one?
20130710 10:57:05 <xranby> if the video frame sits and blocks extraction of new audio frames
20130710 10:57:14 <olamedia> *ffmpeg
20130710 10:57:20 <xranby> olamedia: jogamp can use both libav or libffmpeg
20130710 10:57:28 <xranby> they both originate from the same project
20130710 10:58:10 <xranby> we dynamically lookup which functions are provided by the system installed libav/libffmpeg library
20130710 10:58:32 <olamedia> >_> but i saw an error than was trying to put ffmpeg dll into root of project... "codec: true format:false utils: true"
20130710 10:59:38 <olamedia> under windows
20130710 10:59:51 <xranby> olamedia: then we failed to lookup some functions in format
20130710 11:00:06 <olamedia> under linux all is working fine
20130710 11:00:08 <xranby> olamedia: can be caused by a too new libformat or too old version
20130710 11:00:47 <xranby> olamedia: some fedora user reported a similar issue with one version of ffmpeg
20130710 11:01:02 <xranby> i game him a suggested fix but he never returned and validated the fix
20130710 11:01:25 <olamedia> i have made a test app that is using remote (http) 3gp file to play on a "screen" shaped mesh
20130710 11:01:59 <olamedia> format-55 i have tested on windows
20130710 11:01:59 <xranby> olamedia: you can try and check if this fix makes your ffmpeg windows version more happy https://github.com/xranby/jogl/commits/avformat-F17
20130710 11:02:52 <xranby> 55 sounds like a new version
20130710 11:03:51 <olamedia> yes, latest
20130710 11:03:53 <xranby> olamedia: the debug output of jogl should list the symbol it fails to resolve
20130710 11:03:54 <xranby> http://forum.jogamp.org/How-to-use-GLMediaPlayer-FFMPEG-td4026567.html
20130710 11:04:27 <xranby> this forum thread explains how we investigate these kind of missing symbols
20130710 11:05:41 <xranby> like i said, unfortunally Danny02 did never test the fix hence its not merged
20130710 11:08:18 <olamedia> >_> or may be he tested it and was so happy to forgot post on forum ^_^
20130710 11:09:32 <olamedia> about openal... is it possible to feed it with some raw data from another decoder?
20130710 11:09:33 <xranby> well .. i doubt it.. if so he would be back when the rc11 release and rc12 release got out complainign its not working :)
20130710 11:10:24 <olamedia> or he's sticked to an old branch
20130710 11:10:24 <xranby> olamedia: raw data like in the compressed audio frames?
20130710 11:10:43 <xranby> olamedia: you can send raw unpacked data like 5.1 sound buffers
20130710 11:11:06 <xranby> and have it played back in all awesome 3d possition + doppler if you like
20130710 11:11:48 <olamedia> for example, something from java sound...
20130710 11:12:51 <olamedia> i.e. AudioSystem.getAudioInputStream(decodedFormat, inputStream); etc
20130710 11:12:54 <xranby> java sound uses simple unpacked audio buffers
20130710 11:13:08 <xranby> thus yes that work
20130710 11:13:31 <olamedia> what exactly i need to pass and to which class?
20130710 11:14:12 <xranby> olamedia: you can take a look at the ALut impementation
20130710 11:14:19 <xranby> it currently uses java sound to unpack wav files
20130710 11:14:31 <xranby> and feed thw wav into joal
20130710 11:15:21 <olamedia> should passing AudioInputStream to alutLoadWAVFile works?
20130710 11:15:32 <xranby> ok let me correct that.. it was using java sound
20130710 11:15:43 <xranby> now it is using gouessej's wav parser
20130710 11:16:02 <xranby> thus we can now use joal ALut on non j2se systems
20130710 11:17:13 <xranby> olamedia: here is the last java sound implementation http://jausoft.com/git/?p=joal.git;a=blob;f=src/java/com/jogamp/openal/util/WAVLoader.java;h=f997eb49faa9bfbe6100bbd71fc9ef4cb2e42a0b;hb=14186120ed0c26a053ca90969ac45d55dda53468
20130710 11:17:45 <xranby> that would prepare a WAVData http://jausoft.com/git/?p=joal.git;a=blob;f=src/java/com/jogamp/openal/util/WAVData.java;h=43c217da75c87aa30b299b9cb295ddd88e333c08;hb=14186120ed0c26a053ca90969ac45d55dda53468
20130710 11:18:17 <xranby> that in turn ALut use to prepare an OpenAL audio buffer here: http://jausoft.com/git/?p=joal.git;a=blob;f=src/java/com/jogamp/openal/util/ALut.java;h=2db47d0a1f0c2e9c6788f6dcd9c8f666b20e1e10;hb=14186120ed0c26a053ca90969ac45d55dda53468
20130710 11:18:50 <xranby> ALut.alutLoadWAVFile
20130710 11:19:45 <xranby> how you use the alutLoadWAVFile is explained by the JOAL tutorial!
20130710 11:20:49 <xranby> sgothel: rmk0: i am not sitting on a machine with the ssh keys to pull in the latest joal-demos www tutorial
20130710 11:21:04 <xranby> this link still points to the non updated joal-demo www content https://jogamp.org/joal-demos/www/devmaster/lesson1.html
20130710 11:21:58 <xranby> sorry for poking you if you are under stress
20130710 11:23:19 <xranby> olamedia: i have spent some hours cleaning up the tutorials yesterday
20130710 11:23:39 <xranby> the new version will be online in near time.. its found inside the joal-demos git
20130710 11:24:27 <xranby> olamedia: read this https://jogamp.org/git/?p=joal-demos.git;a=blob_plain;f=www/devmaster/lesson1.html;h=12c7c7f9309db5d8fd865f2b4ee1e8871c9b64f5;hb=06a3aaa5fd38097a4644921a269d6ca282fb31eb
20130710 11:27:44 <rmk0> xranby: how can i help?
20130710 11:28:34 <xranby> rmk0: i need someone with access to the jogamp.org www joal-demos folder to pull in the latest joal-demos from the jogamp git
20130710 11:28:55 <xranby> in order to get the up to date joal tutorials online
20130710 11:28:57 <sgothel> @Xerxes: pulled
20130710 11:29:07 <xranby> sgothel: thanks!
20130710 11:29:16 <xranby> rmk0: ok thank you, sgothel fixed it
20130710 11:29:25 * rmk0 merges back into shadows
20130710 11:29:46 <xranby> looks good https://jogamp.org/joal-demos/www/devmaster/lesson1.html
20130710 11:30:28 <xranby> i think this is the first time in joal 10 year history that you can actually follow the joal tutorial and get a working application :)
20130710 11:30:51 <xranby> only the demo code in the git have been working during the last decade
20130710 11:33:44 <sgothel> KUDOS Xerxes
20130710 11:44:30 * xranby (~xranby@anon) Quit (Ping timeout: 264 seconds)
20130710 11:59:31 * xranby (~xranby@anon) has joined #jogamp
20130710 12:36:21 <sgothel> for fun: http://www.flickr.com/photos/pedernorrby/sets/72157632277119513/with/8285997954/ (almost electronic art)
20130710 12:37:49 <xranby> sgothel: cool!
20130710 12:38:41 <olamedia> what they expecting from heightmap <_<
20130710 12:41:35 <xranby> it alctually looks quite real... compared to the recent UK railway landslide: http://www.southyorkshiretimes.co.uk/news/local/many-weeks-to-fix-moving-spoil-heap-1-5411053 http://www.manchestereveningnews.co.uk/news/uk-news/major-landslide-stainforth-south-yorkshire-1318795
20130710 12:41:52 <olamedia> http://shanghai.edushi.com/
20130710 12:42:44 <xranby> olamedia: that was actually quite cute!
20130710 12:43:00 <xranby> sim city IRL
20130710 12:43:32 <olamedia> a lot of major chinese cities
20130710 12:45:47 <olamedia> looks like a google map layer
20130710 12:46:21 <olamedia> hand-drawn layer
20130710 12:48:37 <xranby> olamedia: http://www.mapofthedead.com/map#12.00/34.0503/-118.2421 -- good survival map to have before the bof
20130710 12:52:10 <olamedia> japan is so big on this map
20130710 12:55:00 <xranby> speaking of maps.. this one is very interesting, especially if you plan to market applications for mobile devices: http://www.mapbox.com/labs/twitter-gnip/brands/#10/34.0211/-118.4117
20130710 12:56:02 <xranby> its quite interesting to see how ritch people tend to purcase expensive iphoens
20130710 12:56:09 <xranby> and normal people go for android
20130710 12:57:09 <olamedia> http://www.mapbox.com/labs/twitter-gnip/brands/#13/56.8211/60.6160
20130710 12:58:22 <olamedia> more randomly distributed
20130710 12:58:41 <xranby> well you can easily spot the airport and the main street
20130710 12:58:56 <xranby> all lit up by red lights
20130710 12:59:53 <xranby> spain is interesting
20130710 12:59:56 <xranby> all green
20130710 13:00:11 <olamedia> switch off the red.. it will be lit by green the same way
20130710 13:00:38 <xranby> ok,, might be a layer thing
20130710 13:00:50 <olamedia> tokyo http://www.mapbox.com/labs/twitter-gnip/brands/#10/35.6339/139.6912
20130710 13:02:00 <xranby> north korea: http://www.mapbox.com/labs/twitter-gnip/brands/#10/39.3985/126.3929
20130710 13:04:12 <sgothel> would be interesting to see the comparison w/ mobile web / email access
20130710 13:04:25 <olamedia> i though the map was not loaded yet
20130710 13:04:33 <olamedia> *t
20130710 13:04:38 <olamedia> XD
20130710 13:06:18 <xranby> yeah.. when you are online.. you only reach people living in the coloured light up cloud
20130710 13:07:30 <xranby> i assume mobile phones and cameras with gps images exist in all countrys with internet access
20130710 13:07:48 <xranby> these maps are based on twitter tweets with gps coordinates
20130710 13:07:59 <olamedia> oh.. marquee... http://www.kcna.kp/
20130710 13:08:42 <olamedia> i see marquee again!
20130710 13:37:54 <xranby> (15:05:23) sgothel: would be interesting to see the comparison w/ mobile web / email access <-- & SMS!
20130710 14:00:58 * xranby (~xranby@anon) Quit (Ping timeout: 247 seconds)
20130710 14:01:53 * xranby (~xranby@anon) has joined #jogamp
20130710 14:08:58 * xranby1 (~familjen@anon) has joined #jogamp
20130710 14:21:03 <xranby> Do someone know who "falexey" is on github? this person have ported the jogl-demos tennis to es2 https://github.com/falexey/jogl-demos/commits/master
20130710 14:21:41 <sgothel> yup .. he did the pon demo
20130710 14:21:43 <sgothel> png
20130710 14:21:46 <sgothel> sorry :)
20130710 14:21:49 <sgothel> PONG
20130710 14:22:00 <olamedia> pon.png
20130710 14:22:28 <sgothel> i.e. Tennis, I merged it
20130710 14:23:40 <olamedia> falexey pushed to master at falexey/jogl-demos
20130710 14:23:40 <olamedia> 0f85e3b ILoveGit
20130710 14:23:50 <olamedia> me too
20130710 14:25:38 <sgothel> @Xerxes: I moved his assets in it's own jar file (textures ..) - if one likes to resurrect this game .. pls do so (i.e. NApplet launcher on website etc)
20130710 14:27:28 <xranby> sgothel: i may look into it when time permists.. that is to create a demo launcher in order to more easily run the many 200+ demos
20130710 14:27:46 <sgothel> ahh .. right
20130710 14:27:50 <sgothel> would be neat ..
20130710 14:28:00 <xranby> i guess 190 demos uses swing
20130710 14:28:04 <sgothel> I currently try to digest the new header files
20130710 14:28:06 <xranby> while 10 uses newt
20130710 14:28:09 <sgothel> (GL and ES ..)
20130710 14:28:27 <sgothel> w/ some changes in our parser, we shall digest them mostly original
20130710 14:28:39 <xranby> right.. need some separatino in order to have the es compatible demos ready for consume on mobile
20130710 14:28:43 <sgothel> in the end .. we should have support for new GL/ES ..
20130710 14:28:52 <xranby> while keeping the gl swing demos ready for use on desktop
20130710 14:29:07 <xranby> using the imm mode sink?
20130710 14:29:07 <sgothel> some magic grep command maybe ?
20130710 14:29:33 <sgothel> well, if GL2ES* profile is being used and not awt/swing .. -> mobile
20130710 14:30:06 <sgothel> yes .. imm-mode-sink is one of the many ways used to compile the VBO for mobile
20130710 14:32:16 <xranby> https://gist.github.com/xranby/5966788 - this is my fisrt attempt to manually categorize the many jogl-demos
20130710 14:32:46 <xranby> its a mess :)
20130710 14:32:58 <xranby> i added the <3 for nice demos
20130710 14:33:35 <sgothel> if the demos would run as a simple GLEventListener .. well, then our NEWT applet launcher could start them ..
20130710 14:33:38 <xranby> all demos with UTIL after require the UTIL jars
20130710 14:33:46 <sgothel> i.e. if they do not use a surrounding 'main' class
20130710 14:34:31 <sgothel> http://jogamp.org/deployment/jogamp-current/jogl-applet-runner-newt-gears-normal-napplet.html <- like this one
20130710 14:34:49 <xranby> yes.. demos with GLNewtRun require the demos.GLNewtRun TOOL
20130710 14:35:22 <xranby> and the demos with label NEWT uses NEWT and a regular main class
20130710 14:35:29 <sgothel> I would use the more maintained com.jogamp.newt.awt.applet.JOGLNewtApplet1Run
20130710 14:35:44 <sgothel> the other one is .. err .. a bit outdated, I started w/ that
20130710 14:35:56 <xranby> it still work! :D
20130710 14:36:23 <xranby> i can use it to run the demos.es1.angeles.AngelesES from commandline
20130710 14:37:09 <xranby> ok time for me to travel some miles.. talk to you all later
20130710 14:37:16 <sgothel> ah .. maybe it can be merged / or use the applet mechanism .. etc, however - yes, thats a good starting point
20130710 14:37:20 <sgothel> laters
20130710 14:43:09 * xranby (~xranby@anon) Quit (Ping timeout: 246 seconds)
20130710 15:54:31 <olamedia> oh.. forgot.. how to rewind movie without reloading it?
20130710 15:55:01 <olamedia> player.seek(0) was trying something strange
20130710 15:55:20 <olamedia> can i rewind buffer?
20130710 16:30:20 -christel- [Global Notice] As you may have noticed we appear to have trotted over the grumpy troll's bridge -- the network is currently being attacked and as a result we're temporarily without services as we work to get a working copy of the db fired up elsewhere. Apologies for the inconvenience.
20130710 17:52:59 -mist- [Global Notice] Sorry for the continued lack of services, we're currently moving it to an alternate location. Thanks for your patience.
20130710 18:17:30 * olamedia (~olamedia@anon) Quit (Read error: Connection reset by peer)
20130710 19:12:33 * jk4 (~jk4@anon) Quit (Ping timeout: 240 seconds)
20130710 19:14:21 * jk4 (~jk4@anon) has joined #jogamp
20130710 19:47:22 <xranby1> (17:55:20) olamedia: can i rewind buffer? <-- there is some bug with the player.seek() using the X11 backend, i have fixed forward seek on my FFMPEGMediaPlayer branch using this commit https://github.com/xranby/jogl/commit/170826d2e202a9938595017b9621a25a6f36a50b and fixed backward seek & video sync using this commit https://github.com/xranby/jogl/commit/062f3e9e1ecfeb63ea61eb540d94d17a2de1412a#L0R488
20130710 19:48:41 <xranby1> olamedia: the issue was that the x11 backend forgot to update lastAudio & video PTS after the seek
20130710 19:49:22 <xranby1> preventing the sync code to show the old/waiting for new frames when seeking backward and forward
20130710 19:53:08 <xranby1> olamedia: thus with the seek code fixed seek(0) rewinds playback to the beginning
20130710 19:54:05 <xranby1> please test the FFMPEGMediaPleyer branch and check if seek work as intended for you as well!
20130710 19:55:01 <xranby1> olamedia: https://github.com/xranby/jogl/compare/FFMPEGMediaPlayer
20130710 19:59:43 -mist- [Global Notice] We're still working on getting services (nickserv, chanserv, alis, etc.) back up and running. Another global notice will be sent once we're happy that they are back properly.
20130710 20:00:04 * jk4 is now known as Guest47772
20130710 20:01:47 -mquin- [Global Notice] Services are now back but may be lagged for a little while as everyone identifies. Thank you for your patience and for flying freenode!
20130710 21:01:01 -mquin- [Global Notice] That should be the last bump, services are back now and should be working normally. Thank you all again for your patience and understanding.
20130710 21:17:32 * odin_ (~Odin@anon) Quit (*.net *.split)
20130710 21:39:16 * void256 (~void@anon) has joined #jogamp
20130710 21:39:23 * odin_ (~Odin@anon) has joined #jogamp
20130710 21:40:59 * odin_ is now known as Guest38561
20130710 22:07:09 * [Mike] (~Mike]@anon) has joined #jogamp
20130710 22:56:59 * olamedia (~olamedia@anon) has joined #jogamp
20130710 22:57:26 <olamedia> hmm
20130710 22:59:29 <olamedia> got libav working, ffmpeg dlls still failing
20130710 23:00:21 <sgothel> I had no troubles using the ffmpeg blobs from the site mentioned in the API doc, well .. tested a while ago.
20130710 23:00:45 <xranby1> olamedia: ok btw.. i posted some info in the backlog about seek()
20130710 23:00:54 <olamedia> latest -55, marked as "future" in sources
20130710 23:01:37 <olamedia> ?again?, where (at which stage) and how is better to check if i have got working GL context?
20130710 23:02:06 <xranby1> GLEventListener init is quite safe
20130710 23:02:21 <olamedia> my other machine with windows
20130710 23:02:44 <olamedia> app is launching without warnings and even going to init stage
20130710 23:03:07 <olamedia> GL2/GL2ES2 is shown as supported ("true") in log
20130710 23:03:13 <olamedia> :/
20130710 23:03:41 <olamedia> but GLMediaPlayer is failing because of "GL init error"
20130710 23:03:59 <xranby1> olamedia: are shaders supported?
20130710 23:04:30 <olamedia> glsl: flase
20130710 23:04:34 <olamedia> *false
20130710 23:04:41 <xranby1> olamedia: i have a machine at work running the windows GDI drivers ... that do not support any shaders
20130710 23:04:52 <xranby1> if the machine got an intel GPU card
20130710 23:05:04 <olamedia> u mean, gl is supported, but no shaders?
20130710 23:05:05 <xranby1> then get the latest OpenGL drivers from intels website
20130710 23:05:17 <xranby1> because the drivers shipped with windows do not support shaders
20130710 23:05:22 <olamedia> yes, internal intel card
20130710 23:05:47 <olamedia> i thought machine have no opengl at all
20130710 23:06:02 <xranby1> olamedia: microsoft is one of the companys in the world that pushes their own standard directX and try to break opengl
20130710 23:06:15 <xranby1> microsoft phones are for example the only mobile phone not supporting opengl
20130710 23:06:51 <olamedia> so, should i check for shaders at init() of GLEventListener?
20130710 23:06:59 <xranby1> yes
20130710 23:07:05 <xranby1> if glsl is false
20130710 23:07:17 <xranby1> then ther is no way we can get the current GLMediaPlayer working
20130710 23:07:37 <xranby1> because we use GLSL to convert YUV to RGB
20130710 23:07:40 <xranby1> using it
20130710 23:07:52 <olamedia> fastest way to check this? GLContext?
20130710 23:08:13 <xranby1> there is a function in context yes
20130710 23:08:42 <olamedia> it's probably possible to convert using libav...
20130710 23:09:08 <xranby1> no
20130710 23:09:13 <xranby1> thats outside the scope of libav
20130710 23:09:23 <xranby1> libav do not do rendering
20130710 23:09:56 <xranby1> it only soft the frame data and makes sure an av player app gets complete audio/video/text frames
20130710 23:10:02 <xranby1> sort
20130710 23:10:27 <xranby1> its the av apps task to render the frame / process teh audio buffer
20130710 23:10:35 <xranby1> with correct sync
20130710 23:11:19 <xranby1> many a/v apps uses simd instructions on the CPU to do this kind of processing
20130710 23:11:23 <xranby1> jogamp jogl uses the GPU
20130710 23:11:44 <sgothel> .. note: if one day we do subtitles, it will be also via OpenGL :)
20130710 23:12:10 <sgothel> ok .. cont. hacking ES3/GL4.3 .. all those enums and functions .. which have to be validated ..
20130710 23:12:55 <olamedia> 8.12 format
20130710 23:12:55 <olamedia> Convert the input video to one of the specified pixel formats. Libavfilter will try to pick one that is supported for the input to the next filter.
20130710 23:12:55 <olamedia> This filter accepts the following parameters:
20130710 23:12:55 <olamedia> �pix_fmts�
20130710 23:12:55 <olamedia> A �|�-separated list of pixel format names, for example "pix_fmts=yuv420p|monow|rgb24".
20130710 23:13:21 <olamedia> <_< but do converting
20130710 23:13:42 <sgothel> and hog the CPU - this is low low low prio - i.e. not before siggraph
20130710 23:14:18 <sgothel> (also incr. bandwidth ..)
20130710 23:14:38 <xranby1> olamedia: i give you a point for finding this
20130710 23:16:05 <xranby1> olamedia: try instal the intel drivers
20130710 23:16:12 * void256 (~void@anon) Quit (Remote host closed the connection)
20130710 23:17:22 <xranby1> http://www.intel.com/p/en_US/support/graphics
20130710 23:25:56 <olamedia> lol... intel detector installer crashed my os
20130710 23:26:26 <olamedia> into bsod..
20130710 23:29:41 <xranby1> well... i wonder if there exist an open bugtracker to report these kind of blue issues
20130711 00:09:09 <olamedia> lol.. glGenBuffers not available now XD (while building mesh)
20130711 00:09:29 <xranby1> sounds odd
20130711 00:10:34 <olamedia> well... i found the name of the card . it's sis 330, software opengl drivers
20130711 00:11:07 <olamedia> sounds odd that it tries to start opengl
20130711 00:19:45 <xranby1> sounds like a fixed func opengl 1.5 card
20130711 00:20:33 <sgothel> .. run GLWindow's main .. will dump what we find
20130711 00:25:23 <olamedia> GL.GL_VERSION 1.1.0
20130711 00:25:39 <olamedia> Microsoft GDI generic
20130711 00:26:33 <olamedia> what i did wrong that it allowed selecting GL2?
20130711 00:27:19 <sgothel> see Wiki/FAQ .. bug report .. i.e. how to produce debug logs ..
20130711 00:27:26 <sgothel> probably no other GL available
20130711 00:27:51 <sgothel> it dumps a list of GL mappings .. maybe you can post it on some web space and show us the link
20130711 00:30:15 <olamedia> INIT GL IS: gl4...impl
20130711 00:30:34 <sgothel> all desktop use that 'impl' !
20130711 00:30:50 <sgothel> GLContext .. version / string ..
20130711 00:31:01 <sgothel> etc .. best dump all somewhere and show us link
20130711 00:33:15 <olamedia> http://pastebin.com/Ey8wFhTD
20130711 00:34:56 <sgothel> this is not whole dump from GLWindow ..
20130711 00:35:10 <sgothel> however .. yes seems like GDI only
20130711 00:35:19 <sgothel> run GLWindow w/ all dbg flags ..
20130711 00:35:35 <sgothel> it has a main method .. !
20130711 00:38:52 <olamedia> http://pastebin.com/J6BhnFMf
20130711 00:42:54 <sgothel> .. and the debug version may give us a hint why ..
20130711 00:43:09 <sgothel> debug props .. -Djogl.debug=all
20130711 00:45:20 * xranby1 (~familjen@anon) Quit (Quit: Leaving.)
20130711 00:49:06 <olamedia> updated, same url
20130711 00:51:00 <sgothel> you didn't use pipe .. to file, lines are broken - hard to read
20130711 00:51:16 <olamedia> >_> pipe in cmd?
20130711 00:52:11 <sgothel> read mentioned wiki/faq/bugreport ..
20130711 00:52:17 <sgothel> links to etc/test_dbg.bat
20130711 00:56:26 <olamedia> <_< there's no .bat
20130711 00:57:03 <sgothel> jogl$ ls etc/test_dbg.bat (RTFM)
20130711 01:01:25 <olamedia> updated...
20130711 01:03:38 <sgothel> thx .. readable - yeah, simply no hw GL
20130711 01:04:30 <olamedia> yes, sw... i said this above
20130711 01:56:23 <olamedia> XD text drawing on newt works
20130711 01:57:15 <olamedia> all other draw calls failing at glGenBuffer
20130711 02:01:10 <olamedia> it's because ImmModeSink assumes that there're hardware buffers?
20130711 02:12:57 * [Mike] (~Mike]@anon) Quit ()
20130711 02:56:19 <sgothel> @Mark: http://forum.jogamp.org/Atomic-jars-in-Maven-Central-Gradle-build-system-tp4029555p4029560.html
20130711 02:56:26 <sgothel> .. for our Maven Man :)
20130711 03:31:50 <olamedia> XD got 5.8 fps (without video yet)
20130711 03:32:54 <olamedia> (added novbo support into immModeMesh.draw())
20130711 03:57:42 <olamedia> ok.. got 12 fps using display lists )
20130711 03:59:04 <olamedia> may be switching to fullscreen + making resolution lower will boost it more
20130711 05:06:45 * olamedia (~olamedia@anon) Quit (Read error: Connection reset by peer)
20130711 06:05:49 * olamedia (~olamedia@anon) has joined #jogamp
20130711 06:06:15 <olamedia> >_< and now my app made bsod trying to go fullscreen
20130711 06:38:51 * gabereiser (~GXReiser@anon) has joined #jogamp
20130711 06:39:14 <gabereiser> is anyone awake?
20130711 06:39:43 <olamedia> me awake.. with my dead windows
20130711 06:41:22 <gabereiser> dead windows?
20130711 06:41:59 <olamedia> yes... it died while i was trying to switch to fullscreen mode
20130711 06:42:09 <gabereiser> lol
20130711 06:42:21 <gabereiser> sorry
20130711 06:42:29 <olamedia> can't get out from bsod :/
20130711 06:42:39 <gabereiser> reboot
20130711 06:42:48 <olamedia> bsod after reboot
20130711 06:43:02 <gabereiser> reboot into safe mode?
20130711 06:43:04 <olamedia> < before welcome screen
20130711 06:43:20 <olamedia> and? i don't even know what happened
20130711 06:43:24 <gabereiser> install linux?
20130711 06:43:29 <olamedia> XD
20130711 06:43:43 <olamedia> i'm already on linux...
20130711 06:43:55 <olamedia> that was my old testing machine
20130711 06:44:07 <gabereiser> Virtualbox
20130711 06:44:16 <gabereiser> i don't touch real windows machines anymore
20130711 06:44:22 <olamedia> with opengl?
20130711 06:44:26 <gabereiser> yea
20130711 06:44:40 <gabereiser> OpenGL2
20130711 06:44:42 <olamedia> mesa?
20130711 06:44:55 <gabereiser> not sure
20130711 06:45:02 <gabereiser> Oracle's bastard drivers
20130711 06:45:09 <gabereiser> it works though
20130711 06:45:15 <gabereiser> for what I need it for
20130711 06:45:46 <gabereiser> I'm building a front-end for my raspi using jogamp so my requirements are pretty low
20130711 06:46:11 <gabereiser> you ever messed with a raspberry pi?
20130711 06:46:18 <olamedia> no
20130711 06:46:59 <gabereiser> you know what it is?
20130711 06:47:14 <olamedia> some mobile device
20130711 06:47:41 <gabereiser> more than just mobile… sub-mobile… it's a little $35 arm chip w/ 512mb ram
20130711 06:47:56 <gabereiser> it's pretty neat
20130711 06:48:13 <gabereiser> for $35, I just buy a handful of them and put them to use around the house
20130711 07:29:54 * hharrison (~chatzilla@anon) has joined #jogamp
20130711 07:30:59 <hharrison> gabereiser: if xranby shows up, he's been working with the RPi and might be of some help
20130711 07:33:05 <gabereiser> i don't have any questions or problems regarding it… I would like to know which GLMediaPlayer backend is used (I think omxplayer) cause I have a need to play video, but i'm not sure omxplayer supports my format (I could always add it since it's open source)
20130711 07:35:35 <gabereiser> my real issue is how to handle cookies and headers as it related to accessing a video file via network using GLMediaPlayer… I think I have an idea
20130711 07:38:10 <gabereiser> but i'm not there yet
20130711 07:38:15 <gabereiser> i'm still building UI elements
20130711 07:38:18 <gabereiser> :P
20130711 07:38:40 <gabereiser> using newt OpenGL directly
20130711 07:39:00 <gabereiser> and using textures/quads/TextRenderer
20130711 07:39:11 <gabereiser> which, by the way, needs some TLC
20130711 07:39:20 <gabereiser> it's good, but could be a hell of a lot better
20130711 07:39:32 <olamedia> mb u can test with instanceof?
20130711 07:39:47 <gabereiser> ?
20130711 07:40:34 <gabereiser> you mean for the video backend?
20130711 07:41:03 <gabereiser> i know on the RPi it's omxplayer… but there's got to be a way to detect that i guess
20130711 07:41:29 <gabereiser> or is that what you mean by instance of?
20130711 07:42:30 * hharrison (~chatzilla@anon) Quit (Quit: ChatZilla 0.9.90 [Firefox 22.0/20130621124006])
20130711 07:47:48 * Guest47772 is now known as jk4
20130711 07:48:00 * gabereiser (~GXReiser@anon) Quit (Quit: gabereiser)
20130711 07:48:13 * jk4 (~jk4@anon) Quit (Changing host)
20130711 07:48:13 * jk4 (~jk4@anon) has joined #jogamp
20130711 08:19:27 * olamedia (~olamedia@anon) Quit (Remote host closed the connection)
20130711 09:00:01 * xranby (~xranby@anon) has joined #jogamp
20130711 10:42:46 * xranby (~xranby@anon) Quit (Ping timeout: 240 seconds)
20130711 10:46:52 * xranby (~xranby@anon) has joined #jogamp
20130711 12:03:50 <rmk0> hm... quick question: there's a "gluegen-gl.jar" file in the atomic directory of jogamp-all-platforms.7z
20130711 12:04:04 <rmk0> which project does it belong to?
20130711 12:10:42 <rmk0> sgothel: should be able to deploy atomic jar files via the maven scripts now
20130711 12:10:46 <rmk0> doesn't require any new steps
20130711 12:10:50 <rmk0> would try it on jogamp.org first
20130711 12:26:56 <xranby> rmk0: nice
20130711 12:27:25 <xranby> rmk0: i assume there will be one atomic target for each os abi and arch combinatino jogamp support?
20130711 12:28:24 <rmk0> nope, but i'll explain why when i've finished documenting it
20130711 12:28:44 <rmk0> it does provide what you just said, but it sort of looks like it doesn't at first glance
20130711 12:31:09 <xranby> i look forward to see how the cheese is cut!
20130711 12:46:51 <rmk0> i've moved the Maven stuff to a separate page, as it's getting rather large
20130711 12:47:08 <rmk0> https://jogamp.org/wiki/index.php/Maven
20130711 12:48:22 <rmk0> so the atomic jar files are now provided as attachments to some new packages
20130711 12:48:33 <rmk0> the newt package contains all the NEWT atomic jar files, and so on
20130711 12:49:16 <rmk0> not really sure on the use cases for the atomic jars at the moment, but this is the simplest way to make sure that they're all usable from Maven and things like Gradle that use the same packages
20130711 12:53:40 <rmk0> dislike gradle as it seems to be a one way street
20130711 12:54:54 <rmk0> whole point of having something declarative like the POM xml is so that external tools can read and understand it without having to evaluate/execute anything
20130711 12:56:39 <rmk0> as an example... if i publish a project with a maven pom
20130711 12:56:47 <rmk0> that pom can be used by ivy, gradle, and friends
20130711 12:57:02 <rmk0> if i publish a project using gradle, i have to have gradle manually generate a pom file so that other build systems can use it
20130711 12:57:12 <rmk0> it's an optional step that i imagine nobody bothers with
20130711 12:57:27 <rmk0> hence the "one way street"
20130711 12:57:53 <rmk0> doesn't help that groovy is shit
20130711 12:58:19 <xranby> rmk0: did you push atomic maven files to jogamp.org? i only see 2.0.2-rc-20130710
20130711 12:58:38 <rmk0> not yet, no... i assumed they'd be pushed with the next dated rc
20130711 12:58:43 <xranby> ok!
20130711 13:12:58 * xranby (~xranby@anon) Quit (Quit: Leaving.)
20130711 15:43:12 <sgothel> (02:10:43 PM) rmk0: sgothel: should be able to deploy atomic jar files via the maven scripts now
20130711 15:43:12 <sgothel> (02:10:46 PM) rmk0: doesn't require any new steps
20130711 15:43:23 <sgothel> @Mark: KUDOS - thx
20130711 15:43:43 <rmk0> bit nervous about people using them to basically create non-portable projects
20130711 15:43:48 <rmk0> but i guess we can't exactly prevent that
20130711 15:45:04 <sgothel> why 'non partable' ?
20130711 15:45:07 <sgothel> portable ?
20130711 15:45:28 <rmk0> i mean if they have their project specifically depend on "linux" "windows" "x86" "x86_64"
20130711 15:45:38 <rmk0> as the build script in the forum did
20130711 15:45:50 <rmk0> harms portability for no good reason
20130711 15:45:53 <sgothel> nope .. not that atomic pls ..
20130711 15:46:05 <sgothel> i.e. atomic for our jar/atomic .. I thought
20130711 15:46:16 <sgothel> i.e. jogl-core.jar ..
20130711 15:46:24 <rmk0> it's exactly the atomic jars as given in jogamp-all-platforms.7z
20130711 15:46:32 <sgothel> great
20130711 15:58:17 * gabereiser (~GXReiser@anon) has joined #jogamp
20130711 15:59:10 <gabereiser> hi
20130711 15:59:31 <sgothel> Hey
20130711 16:06:06 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=661#c29
20130711 16:10:17 <sgothel> great - mostly noise and already resolved stuff @ forum, thats when you know we are getting stable :)
20130711 16:10:50 <sgothel> back to fiddling GL extensions into their sets ..
20130711 16:11:40 <gabereiser> lol
20130711 16:11:43 <sgothel> quite sad though .. the new khronos headers removed extension mapping :|, i.e. they went lazy and pushed most down the 4.[23] throat
20130711 16:12:03 <gabereiser> well… to be fair… it's for a good reason...
20130711 16:12:43 <gabereiser> drivers for olders cards aren't being updated anymore and khronos has long said it hates the fact that there's all this extension non-sense
20130711 16:12:59 <sgothel> yeah .. but for documentation issues .. it would be nice to have it both ways .. like early this year headers, explicitly stating they are part of extension and subsumed in X.y
20130711 16:13:00 <gabereiser> it's getting better
20130711 16:13:20 <gabereiser> or maybe a "compatibility" header
20130711 16:13:23 <sgothel> sure .. won't hurt - just our API doc will be less verbose then :)
20130711 16:13:26 <gabereiser> that includes that stuff
20130711 16:14:01 <gabereiser> honestly, I come from a long history of DirectX programming, I've been meaning to learn OpenGL for a long time now, just never did...
20130711 16:14:04 <sgothel> na .. will not maintain old header files - updating to latest, trying to keep diff most minimal for easy maintenance and sanity
20130711 16:14:05 <gabereiser> so it's all new to me
20130711 16:14:28 <sgothel> welcome to the API of the free platforms :)
20130711 16:14:50 <gabereiser> well… I was talking to my wife the other day when she saw me messing with jogamp
20130711 16:15:01 <gabereiser> She was like "Why aren't you messing with your own engine?"
20130711 16:15:11 <sgothel> then you maybe know .. whats the latest beef of Win8 mobile/RT and OpenGL ?
20130711 16:15:11 <gabereiser> I said "Cause it's directX and screw that"
20130711 16:15:26 <gabereiser> oh, yeah
20130711 16:15:32 <gabereiser> not gonna happen
20130711 16:15:45 <sgothel> (guess technically .. latest D3D and GL is equal)
20130711 16:15:50 <gabereiser> Microsoft wants people to use DirectX
20130711 16:16:01 <gabereiser> and pushes people away from "open standards" as much as possible
20130711 16:16:12 <gabereiser> that's one of the reasons I no longer work with DirectX
20130711 16:16:18 <sgothel> sure sure .. but I hope wrath of companies will tell 'em better .. like w/ Vista ..
20130711 16:16:34 <gabereiser> after writing two engines using DX9 and XNA, I'm done with em
20130711 16:16:47 <gabereiser> that, and the advent of mobile...
20130711 16:16:59 <gabereiser> where every device known to man runs OpenGL
20130711 16:17:07 <gabereiser> I'm kicking myself for not learning it sooner
20130711 16:17:22 <sgothel> yup .. _not_ exposing ES2/3 by winRT .. soo ridiculous
20130711 16:17:55 <gabereiser> but honestly, according to my companies analytics, winRT is like 0.01% of userbase
20130711 16:17:59 <gabereiser> so ignore them
20130711 16:18:22 <gabereiser> which brings me to why I'm using jog amp....
20130711 16:18:37 <gabereiser> I'm gonna be making a new set top box front-end with it
20130711 16:18:48 <sgothel> sure sure .. trying to give folks similar idea regarding iOS :) we will see .. if there is enough momentum / support to do a port.
20130711 16:18:51 <gabereiser> that runs on small devices like Raspberry PI or cubieboard
20130711 16:19:20 <gabereiser> I wouldn't worry about it until they allow some way to compile and run java on ios
20130711 16:19:42 <gabereiser> if they know OpenGL enough, they can just write their games/apps for iOS without too much trouble
20130711 16:20:01 <gabereiser> and look at the source for jogamp if they need some utility
20130711 16:20:14 <gabereiser> that's what burned me out the first time
20130711 16:20:24 <gabereiser> I write a game engine using DX7 VB
20130711 16:20:35 <sgothel> well .. it's our story to use JogAmp to minimize that dev roundtrip costs porting to other OS.. but we will see .. Avian, JamVM .. etc
20130711 16:20:36 <gabereiser> then when upgrading to DX8, I ported it to VB.Net
20130711 16:20:48 <sgothel> oh
20130711 16:20:57 <gabereiser> and found its runs slow as balls
20130711 16:21:01 <gabereiser> so we ported to C++
20130711 16:21:04 <sgothel> expensive .. wow
20130711 16:21:05 <gabereiser> MC++
20130711 16:21:14 <gabereiser> and then added DX9
20130711 16:21:19 <gabereiser> and it went well for years
20130711 16:21:22 <gabereiser> then...
20130711 16:21:24 <gabereiser> XNA
20130711 16:21:29 <gabereiser> you bastards
20130711 16:21:39 <sgothel> ?
20130711 16:22:05 <gabereiser> XNA framework is like "Managed directX" that runs on windows, windows phone, xbox360
20130711 16:22:09 <gabereiser> but only supports C#
20130711 16:22:12 <gabereiser> so...
20130711 16:22:19 <gabereiser> I broke off and rewrote the engine
20130711 16:22:23 <gabereiser> in C# for XNA
20130711 16:22:28 <sgothel> holy moly
20130711 16:22:30 <gabereiser> called Reactor3D
20130711 16:22:42 <gabereiser> released a game on xbox360 called Alchemist
20130711 16:22:50 <gabereiser> and the numbers just wasn't there
20130711 16:23:00 <gabereiser> so we abandoned the XNA port
20130711 16:23:08 <gabereiser> and I started to port it to OpenGL
20130711 16:23:12 <gabereiser> but don't know enough
20130711 16:23:14 <sgothel> well .. our story is not just to run on Java, but any JVM'ish thing .. so that we could deal w/ it if required.
20130711 16:23:33 <gabereiser> yeah, which is why I like it
20130711 16:23:42 <sgothel> would you like to write this down to 1-3 slides .. for our BOF ?
20130711 16:23:56 <sgothel> you coming ? you may present this story yourself ..
20130711 16:24:14 <gabereiser> when I was working on my engine, we had bindings for C/C++, C#, VB.Net, Pascal...
20130711 16:24:20 <gabereiser> no I'm not coming...
20130711 16:24:21 <sgothel> https://jogamp.org/wiki/index.php/JogAmp_BOF_SIGGRAPH_2013_PLANNING#Meet_JogAmp.27s_Ecosystem_Across_Platforms_and_Devices <- matches what we like to tell ..
20130711 16:24:23 <gabereiser> sorry..
20130711 16:24:33 <gabereiser> i got to work
20130711 16:24:38 <sgothel> but you could let us show your demo and slides .. !3min
20130711 16:24:42 <sgothel> ~3min.
20130711 16:24:59 <gabereiser> I could put something together...
20130711 16:25:04 <gabereiser> and have you guys show it
20130711 16:25:04 <sgothel> we record video .. lots of users watching it later on .. so would be neat
20130711 16:25:25 <sgothel> great - we have lots of such content from remote participators ..
20130711 16:26:01 <gabereiser> here's some stuff I did using my engine of 'old
20130711 16:26:02 <gabereiser> http://www.youtube.com/watch?v=dlISS1CIsm8
20130711 16:26:02 <sgothel> pls intro you and company, goal .. and how / why you went to Jogamp .. (see your writing above) .. would be awesome
20130711 16:26:14 <gabereiser> http://www.youtube.com/watch?v=g0Th2BlQsj4
20130711 16:26:27 <sgothel> plus a ~2min video in 720p .. and/or a demo we can show .. if avail.
20130711 16:26:28 <gabereiser> http://www.youtube.com/watch?v=XxK7ErKReyI
20130711 16:26:33 <sgothel> hoho
20130711 16:26:43 <gabereiser> http://www.youtube.com/watch?v=AmuY88uE6Rg
20130711 16:27:00 <sgothel> if possible send me/us the slides and infos by email, better https://jogamp.org/bugzilla/show_bug.cgi?id=661 .. etc
20130711 16:28:00 <sgothel> if you can make a compiled 720p video .. 2 minutes .. would be great
20130711 16:28:17 <sgothel> or 3 min.
20130711 16:28:33 <sgothel> shorter .. better, while reading your slides etc
20130711 16:28:37 <gabereiser> i just started with jogamp like 6 days ago… I don't really have anything to show yet...
20130711 16:28:42 <gabereiser> lol
20130711 16:29:09 <sgothel> doesn't matter .. i.e. your story is great, we can say .. this is your engine and what you have achieved .. now porting for those reasons
20130711 16:29:22 <sgothel> company (?) goals .. etc
20130711 16:29:36 <gabereiser> ok, i'll try to send something this weekend
20130711 16:29:42 <sgothel> open source .. great
20130711 16:30:17 <sgothel> http://jogamp.org/jogl/www/ <- pls send me a nice image 160x[120-160] .. and a little text
20130711 16:31:01 <gabereiser> k
20130711 16:31:17 <sgothel> awesome - thank you!
20130711 16:31:24 <sgothel> dinner now
20130711 16:41:57 <gabereiser> so my current reason for messing with jogamp is this: To make a low cost set top box for the company I work for. Brighthouse Networks (a cable company)
20130711 16:42:11 <gabereiser> a few months ago, Comcast announced X2 box… http://www.youtube.com/watch?v=snjNomkyoaE
20130711 16:42:21 <gabereiser> we have nothing to compete with that...
20130711 16:42:59 <gabereiser> so the idea (still in home-brew, can I do it, phase) is to take something like a Raspberry PI or cubieboard and build a platform using it for this purpose
20130711 16:43:10 <gabereiser> we already have apps for ios/android that do streaming
20130711 16:43:28 <gabereiser> so we have the infrastructure there… we just don't have a box for the living room that can do it
20130711 16:43:52 <gabereiser> and after looking for weeks for a framework that will help me, I found jogamp and it's glorious GLMediaPlayer class
20130711 16:44:09 <gabereiser> so I started to mess with it and got streaming working
20130711 16:44:51 <gabereiser> now to actually build a front-end for it, which is rather difficult due to Java's "pass by value" crap but not unsolvable.
20130711 16:45:35 <gabereiser> the biggest gripe I have so far is there's not an easy way to do text. Sure, the TextRenderer class allows me to render text but there's a lot missing for doing good text work
20130711 16:45:42 <gabereiser> like measuring the size of a string
20130711 16:45:47 <gabereiser> width/height
20130711 16:45:55 <gabereiser> wrapping to bounds
20130711 16:46:02 <gabereiser> weight
20130711 16:46:05 <gabereiser> etc...
20130711 16:46:24 <gabereiser> so a lot of this i'm having to code myself using the Font class and the TextRenderer
20130711 16:46:38 <gabereiser> but I guess this kinda goes hand in hand with building UI's
20130711 16:46:55 <gabereiser> it's just going slower than I would like
20130711 16:48:21 <gabereiser> i have a custom kernel img for the raspberry pi that I built from source that hides text output, boots with splash screen, inits the IR receiver service, detects the HDMI resolution and loads a stripped down x11 immediately to an app
20130711 16:48:30 <gabereiser> that app would be my jogamp app.
20130711 16:48:50 * rmk0 listens
20130711 16:49:08 <gabereiser> but for 6 days of work, I really shouldn't complain
20130711 16:49:20 <gabereiser> as that would have taken me months in straight C/C++
20130711 16:50:07 <sgothel> perfect .. pls put this into your slides (maybe downsize jogamp's gloriosity :)
20130711 16:50:19 <gabereiser> lol
20130711 16:50:31 <sgothel> but yeah .. thats the story .. awesome
20130711 16:51:02 <rmk0> type and memory safety is a net benefit to humanity
20130711 16:51:05 <sgothel> hope we have time to fix the media player stuff .. Xerxes and I will do it eventually .. also w/ using OpenMAX
20130711 16:51:33 <sgothel> yes Mark, how terrible ECMA is .. I learned w/ JPEG decoder porting
20130711 16:51:46 <rmk0> it's bad
20130711 16:51:53 <rmk0> it's not exactly in the minority either
20130711 16:52:21 <sgothel> and I was just tracking types through 1 file of 1000 loc ..
20130711 16:52:30 <gabereiser> eewww
20130711 16:52:54 <gabereiser> OpenMAX looks pretty neat...
20130711 16:53:15 <gabereiser> but ultimately, anything that can stream m3u8 is a benefit to me
20130711 16:53:25 <sgothel> at least for RaspPI it seems to be avail (thx 2 Xerxes who always finds out the details)
20130711 16:54:10 <sgothel> m3 ? unsigned 8 ? .. looks like one of these ole sox params
20130711 16:54:37 <gabereiser> Apple HLS format
20130711 16:54:47 <sgothel> haha
20130711 16:55:00 <gabereiser> it's basically a text file that defines HTTP streams at different bitrates
20130711 16:55:13 <sgothel> maybe you can help us Gabe ..
20130711 16:55:16 <gabereiser> ultimately letting one know the location of a TS file to decode
20130711 16:55:27 <gabereiser> i'd love to
20130711 16:55:43 <sgothel> i.e. we like us to deploy libav (w/o GPL .. w/o non-free codecs ..) to JOGL
20130711 16:55:53 <sgothel> so .. like JOAL's OpenAL-soft ..
20130711 16:56:20 <gabereiser> yeah cool… but still detect the presence of libavcodecs in the system
20130711 16:56:24 <sgothel> so one would need to build this beast on the platforms .. (cross .. etc)
20130711 16:56:44 <sgothel> in JOAL, I incl. OpenAL-soft branch .. etc .. and kick off cmake
20130711 16:56:57 <sgothel> works for all but Windows .. where we inject a binary
20130711 16:57:05 <sgothel> same should be done for libav ..
20130711 16:57:30 <gabereiser> windows has it's own libs for av though
20130711 16:57:36 <sgothel> this would be the last task before siggraph .. (next week leaving) .. and I dunno if I can make it until then
20130711 16:57:38 <gabereiser> i.e. not libav
20130711 16:58:05 <sgothel> well .. if you know them .. and can teach us or make a backend of our player .. even better!
20130711 16:58:10 <sgothel> same for OSX ..
20130711 16:58:20 <sgothel> ofc .. libav would be just like the fallback ..
20130711 16:58:55 <sgothel> Xerxes has the current stuff in his branch until I merge and jump in ..
20130711 16:58:57 <gabereiser> OSX has CoreVideo.framework which has inherent libav/libavcodecs support… XBMC uses it in lieu of ffmpeg
20130711 16:59:30 <sgothel> so you mean, using libav on OSX uses OSX hw accel ? great
20130711 16:59:39 <gabereiser> which, btw, xbmc codebase is a masterpiece of spaghetti code...
20130711 17:00:00 <gabereiser> yeah, it should as long as you include CoreVideo.framework when building the natives for OSX
20130711 17:00:01 <sgothel> I use it at home for browsing the files w/ a DB :)
20130711 17:00:38 <sgothel> if you can help here (Windows native backend, and libav provider ..) awesome .. well - however, thats what we have to do one day ..
20130711 17:00:44 <sgothel> besides OpenMAX ofc
20130711 17:01:09 <gabereiser> so GLMediaPlayer doesn't have a windows compatibility path?
20130711 17:01:26 <sgothel> GLMediaPlayer .. has no native Windows backend, right :)
20130711 17:01:32 <gabereiser> ok
20130711 17:01:35 <sgothel> I was testing it w/ libav blobs .. works
20130711 17:02:04 <sgothel> i.e. non perf. critical .. could use libav
20130711 17:02:18 <gabereiser> not hw/accel right?
20130711 17:02:39 <sgothel> right, but Win machines usually fast enough - so best 1st goal IMHO .. have libav compiled and deployed
20130711 17:02:47 <sgothel> then we jump to the native backends ..
20130711 17:03:33 <sgothel> Xerxes branch already has JOAL audio out support - we have to see how to do A52 and spdif/bitstream
20130711 17:03:56 <sgothel> then .. we are able to be used for a full featured TV box :)
20130711 17:04:35 <sgothel> should add this to a GLMediaPlayer roadmap .. hmm
20130711 17:04:51 <sgothel> have to go now .. back in errr .. ~4 hours or so
20130711 17:04:58 <sgothel> laters !
20130711 17:42:06 * gabereiser (~GXReiser@anon) Quit (Read error: Connection reset by peer)
20130711 17:42:26 * gabereiser (~GXReiser@anon) has joined #jogamp
20130711 19:10:00 * void256 (~void@anon) has joined #jogamp
20130711 19:46:37 * rmk0_ (~rmk0@anon) has joined #jogamp
20130711 19:52:40 * rmk0 (~rmk0@anon) Quit (Ping timeout: 246 seconds)
20130711 20:13:48 * xranby (~familjen@anon) has joined #jogamp
20130711 20:15:55 <xranby> gabereiser: hi, and welcome, let me read the back log
20130711 20:16:17 <gabereiser> the novel you mean?
20130711 20:16:55 <xranby> yup, looks like a good read
20130711 20:25:54 * void256 (~void@anon) Quit (Quit: ChatZilla 0.9.90 [Firefox 22.0/20130618035212])
20130711 20:31:40 <xranby> gabereiser: sgothel added an initial OpenMAX backend to the JOGL GLMediaPlayer back in 2008. It was setup to use OpenMAX on the first generation Nvidia tegra 1 devices
20130711 20:32:09 * gabereiser (~GXReiser@anon) Quit (Read error: Connection reset by peer)
20130711 20:32:29 * gabereiser (~GXReiser@anon) has joined #jogamp
20130711 20:34:44 <xranby> gabereiser: The only demo wehave on file of the OpenMAX GLMediaPlayer backend is this movie from 2008 http://www.youtube.com/watch?v=D6Lkw3eZK1w the backend used some nvidia only extensions that has to be replaced with generic code, sgothel and the jogl git got the details.
20130711 20:35:35 <xranby> gabereiser: so, in order to get the GLMediaPlayer fully hardware accelerated on the Raspberry Pi the first step is ofcourse to re-enable the openmax backend and polish it
20130711 20:38:18 <xranby> even with a basic libav backend on the pi we at least use the GPU for YUV -> RGB conversion using a GLSL fragment shader.. so thats the only hardware accelerated bit in use right now
20130711 20:38:39 <gabereiser> what about the omxplayer backend?
20130711 20:39:07 <xranby> like simply start the omxplayer commandline tool?
20130711 20:39:23 <gabereiser> i saw that in the code, that it's an impl for GLMediaPlayer
20130711 20:39:56 <xranby> yes thats the OpenMAX backend i talked about that is currently disabled due to the removed nvidia only code
20130711 20:40:14 <gabereiser> ok
20130711 20:40:46 <xranby> getting it working on the raspberry pi would make sense
20130711 20:41:04 <xranby> since the pi got good omx support
20130711 20:43:44 <gabereiser> so I found that the GLMediaPlayerFactory creates the android player and the ffmpeg player but no omxplayer
20130711 20:43:50 <gabereiser> so that's a start I guess
20130711 20:43:56 <xranby> that is correct. re-enable it by adding https://github.com/xranby/jogl/commit/39ccda2107284ad7c6c9e2fae99c05e7a9a9d4f1
20130711 20:44:59 <gabereiser> yeah, but how would one detect the presence of OpenGL max supported hardware ...
20130711 20:45:15 <gabereiser> cause I don't want to just create a omxplayer without knowing it's supported
20130711 20:45:47 <gabereiser> platform checks on the rpi say it's just linux arm
20130711 20:46:00 <gabereiser> so maybe detect linux arm and use omxplayer...
20130711 20:46:15 <gabereiser> but then, do all arm processors support OpenGL MAX?
20130711 20:46:50 <xranby> the backend should return true for isAvailable if the OMXGLMediaPlayer is able to initialize openmax
20130711 20:47:09 <xranby> we use dynamic lookup of library symbols
20130711 20:47:20 <xranby> thanks to helper classes of gluegen
20130711 20:47:58 <xranby> so that we can check if A) the system library can be loaded and B) the library implements the needed functions
20130711 20:47:58 <gabereiser> ok
20130711 20:48:29 <xranby> for the libav/ffmpeg backend we use this dynamic lookup to stay compatible with N-versions of libav/ffmpeg
20130711 20:48:37 <gabereiser> I'll like to add support for Platform.OS_TYPE
20130711 20:48:42 <gabereiser> to check for raspberry pi
20130711 20:48:48 <gabereiser> and other ARM based systems
20130711 20:48:51 <gabereiser> like cubieboard
20130711 20:49:15 <xranby> openmax is not supported on all arm systems
20130711 20:51:15 <xranby> TI omap chips got openmax drivers
20130711 20:51:47 <xranby> the broadcom video core iv supports openmax as well
20130711 20:51:53 <gabereiser> yeah
20130711 20:52:17 <gabereiser> but on the rpi a simple uname -a doesn't list anything that would say it's a raspberry pi :(
20130711 20:53:12 <gabereiser> it just says armv6
20130711 20:53:18 <gabereiser> armhf
20130711 20:53:54 <xranby> yes its an issue.. we currently guess it is a raspberry pi if the /opt/vc/ video core drivers are found.
20130711 20:54:04 <gabereiser> that works
20130711 20:54:38 <xranby> hence jogamp jogl then use the video core dispman x api to setup the rendering surface
20130711 20:54:50 <xranby> for opengl es use
20130711 20:57:09 <gabereiser> which brings up another question...
20130711 20:57:24 <gabereiser> jogl can draw straight to the fb using dispman x right?
20130711 20:57:27 <gabereiser> no x11 required?
20130711 20:57:43 <xranby> correct, no X11 required!
20130711 20:57:53 <gabereiser> so how is keyboard input handled?
20130711 20:58:02 <gabereiser> no change?
20130711 20:58:10 <xranby> we parse the /dev/event/ input files directly
20130711 20:58:11 <gabereiser> because of jinput?
20130711 20:58:15 <gabereiser> sweet
20130711 20:58:17 <xranby> since we cant use the X11 api
20130711 20:58:55 <gabereiser> cause my IR remote sends commands via irw to the /dev/event/kb out
20130711 20:59:11 <gabereiser> so i could be able to get it working relatively easily
20130711 21:00:01 <xranby> as long as the linux kernel generates events then it would be possible to extend the LinuxEventDeviceTracker in jogl to handle it
20130711 21:00:21 <xranby> the linux kernel sends events for a lot of things like headphone insertion
20130711 21:00:24 <gabereiser> it's just a matter of listening for key events like from a keyboard
20130711 21:00:30 <xranby> power on/off button
20130711 21:00:34 <xranby> etc. yes
20130711 21:00:46 <xranby> we currently only decode the keyboard events
20130711 21:00:53 <xranby> for use on the raspberry pi
20130711 21:00:59 <gabereiser> i already wrote a mapper for IR codes to keyboard keys
20130711 21:01:06 <gabereiser> using irw
20130711 21:01:18 <gabereiser> so i'm good there
20130711 21:02:17 <gabereiser> so I'll write some additional stuff for Platform.class to detect raspberry pi and use that for GLMediaPlayerFactory to instantiate the omxplayerimpl
20130711 21:02:35 <gabereiser> i'll probably do this on github and send a merge request
20130711 21:03:04 <xranby> that would be grat.. i belive you may ask gluegen if it is on a raspberry pi
20130711 21:03:13 <xranby> since it already got the autodetection code
20130711 21:03:19 <gabereiser> ok
20130711 21:03:27 <gabereiser> ill look into that
20130711 21:04:20 <xranby> let me look up the git commit for the autodetection
20130711 21:10:23 <sgothel> back .. glanced over backlog .. (great stuff)
20130711 21:10:34 <xranby> gabereiser: sorry, the autodetection is located in jogl.. http://jogamp.org/git/?p=jogl.git;a=commit;h=8a1e50dc33e14e9423e2f4bc066b32a835bb33fb guessBroadcomVCIV() in NativeWindowFactory
20130711 21:10:35 <sgothel> @Gabe: I was using the windows blob ..
20130711 21:11:08 <sgothel> as described in FFMPEGMediaPlayer, i.e. http://ffmpeg.zeranoe.com/builds
20130711 21:11:24 <sgothel> but would be great to use an official libav version/git-repo
20130711 21:11:57 <sgothel> @Gabe/Xerxes: Best advancement would be now to have libav being build like we do openal-soft.
20130711 21:12:12 <sgothel> today/tomorrow I finalize ES3/GL4.3 .. a 1st draft
20130711 21:12:23 <xranby> sgothel: KUDOS!
20130711 21:12:32 <sgothel> then I need to check Bug 776
20130711 21:12:42 <sgothel> then .. I finally join you in A/V stuff
20130711 21:13:11 <sgothel> (lets say .. I will have 4 full days for it)
20130711 21:13:19 <sgothel> should be enough ..
20130711 21:13:27 <sgothel> Xerxes .. you leave on 18th .. right ?
20130711 21:13:44 <xranby> yes early on the 18'th
20130711 21:14:22 <sgothel> so if you can assist me on Monday/Tuesday .. would be great I guess. Then we have 2.0.2 out .. as it is
20130711 21:14:41 <xranby> OK sounds good
20130711 21:14:59 <sgothel> Dunno about slides and stuff .. hmm, guess we will merge them last minute ?
20130711 21:15:01 <xranby> evenings or mornings?
20130711 21:15:06 <sgothel> Well, we have a few .. :)
20130711 21:15:10 <sgothel> oh .. right :)
20130711 21:15:32 <sgothel> I will be in your time space continuum :) .. which is ?
20130711 21:15:36 <sgothel> you tell me
20130711 21:16:11 <xranby> i am usually free at evenings 22pm -> 3am like right now
20130711 21:16:34 <sgothel> wow .. (and busy in the morning) .. ay ay
20130711 21:17:13 <sgothel> We can meet in afternoon in Anaheim w/ Rami, etc
20130711 21:17:14 <xranby> yup.. i am teaching my daughters cousin electronics
20130711 21:17:18 <sgothel> Sunday w/ Harvey
20130711 21:17:21 <xranby> we light up leds by touching things
20130711 21:17:30 <sgothel> hoho
20130711 21:18:03 <xranby> darlington transistors is the best education tool!
20130711 21:18:43 <sgothel> I failed getting my daughter interested in technical stuff .. oh well :)
20130711 21:18:47 <xranby> they got usually 30000 x amplification thus it can detect signals passing through several humans
20130711 21:19:14 <sgothel> feels .. intriguing
20130711 21:20:22 <xranby> shake hands and light a led,, that is asesome instantly
20130711 21:20:28 <sgothel> ok .. back to the GL stuff .. laters - Gabe/Xerxes: libav build :) well, we may inject manual builds if time is too short for a proper setup (like openal-soft for windows, see JOAL ant-files)
20130711 21:20:55 <sgothel> (yes OMX is not forgotten)
20130711 21:31:41 * Markt_ (32c4d781@anon) has joined #jogamp
20130711 21:33:17 <Markt_> Hi, Has anyone had any trouble locking a drawble surface on linux? I keep freezing randomly while attempting to dsi = ds.GetDrawingSurfaceInfo(); in X11JAWTWindow.java
20130711 21:33:45 <sgothel> Hi 'Markt'
20130711 21:34:08 <Markt_> Hello
20130711 21:34:18 <sgothel> Well .. no troubles when using it 'properly' within surface lock/unlock, otherwise out unit tests would fail
20130711 21:34:22 <sgothel> *our*
20130711 21:34:41 <Markt_> All i'm doing is calling display() on my GLCanvas
20130711 21:34:51 <sgothel> i.e. this code path is pretty mature for years
20130711 21:34:56 <Markt_> it doesn't seem to freeze ever on windows
20130711 21:35:06 <Markt_> I see. perhaps its my driver
20130711 21:35:19 <sgothel> hmm .. in such case, please produce a full bugreport (Wiki -> JOGL FAQ -> Bugreport)
20130711 21:35:22 <Markt_> X windows seems to be stuck over on an IOCTL
20130711 21:35:35 <Markt_> on GetWindowAttributes
20130711 21:35:41 <sgothel> when having all details of your environment etc .. and debug logs .. then we can discuss it.
20130711 21:36:26 <sgothel> pls attach the test_dbg.log to your new bugzilla entry (7z if too large)
20130711 21:36:52 <sgothel> in short: nope - you are the 1st reporting this issue
20130711 21:37:31 <sgothel> can you do this ?
20130711 21:41:01 <Markt_> In my particular industry, i'm not particularly allowed to show the entire environment.
20130711 21:41:26 <Markt_> i'm going to attempt to comment out: //dsi = ds.GetDrawingSurfaceInfo(); //if (dsi == null) { // unlockSurfaceImpl(); // return LOCK_SURFACE_NOT_READY; //} //updateBounds(dsi.getBounds());
20130711 21:41:27 <sgothel> well .. you can run the test as described (etc/test_dbg.sh)
20130711 21:41:33 <sgothel> not much we dump there !
20130711 21:41:51 <sgothel> pls review .. use a non critical test machine .. etc
20130711 21:42:06 <sgothel> I will not discuss such bugs w/o required data
20130711 21:42:21 <sgothel> b/c those bugs require us to know a bit more about the used modules ..
20130711 21:42:34 <Markt_> understood thanks
20130711 21:43:45 <Markt_> https://jogamp.org/bugzilla/ I can try to put the non-critial data in there
20130711 21:43:46 <sgothel> you are welcome. if you are able to disclose non-critical details to 'your industry' .. then we can help you. you can also hire us - see wiki maintainer etc.
20130711 21:44:03 <sgothel> yes pls
20130711 21:45:07 <sgothel> this is our process of tracking issues, everything else (IRC and forum and emails) is considered transient ..
20130711 21:45:48 <sgothel> our industry is pretty anal about trackable issues and git-commit association as well :)
20130711 21:52:04 <Markt_> :)
20130711 21:52:47 <Markt_> I have written up most of it, what other info might be useful? I have the OS, driver version, stack traces from X, and JVM, and from Java itself.
20130711 21:53:26 <xranby> Markt_: include JOGL version
20130711 21:53:57 <Markt_> Its a few months out, how do i get the version number?
20130711 21:53:58 <sgothel> the test_dbg.log. If you have a specific _small_ test case which reproduces the issue (i.e. test_dbg.sh does not) .. then please add this test case as well, plus it's debug logs
20130711 21:54:14 <sgothel> test_dbg.log will tell us
20130711 21:54:25 <sgothel> in case you use official JAR files
20130711 21:54:40 <sgothel> we have version info incl. in JAR manifest and read it out @ runtime
20130711 21:55:06 <sgothel> if you cannot reproduce the issue w/ latest JOGL - it is a no issue for us!
20130711 21:55:19 <sgothel> so pls test w/ latest RC12 or latest aggregated build
20130711 21:56:10 <Markt_> let me check the jar, I just built the latest yesterday, but I can't remmeber if i reverted before or after i sent the binaries to the machine
20130711 21:57:36 <Markt_> Implementation-Version: 2.0.2-rc-20130710
20130711 21:57:41 <Markt_> Implementation-Build: 2.0-bmanual-20130710
20130711 21:57:51 <sgothel> the whole thing to the bugreport pls
20130711 21:58:08 <sgothel> git commit sha-1 is most important ofc
20130711 21:58:44 <sgothel> I will check this offline .. maybe in the next week, no time right now anyways
20130711 21:59:00 <sgothel> but if reproducible .. sure will fix it
20130711 21:59:04 <Markt_> got it
20130711 21:59:25 <sgothel> thank you for helping make our stuff stable!
20130711 21:59:27 <Markt_> so right now, it happens every few hours, and seemly.. it recovers on its own after seemingly about 20 minutes
20130711 22:00:07 <Markt_> Its been pretty stable, this seemingly just started so i don't even know if jogl is the source, or if its mesa, or X, or the way i'm rendering now or what
20130711 22:00:11 <sgothel> all to -> bug report please, will discuss it w/ you
20130711 22:00:27 <sgothel> could be .. anything
20130711 22:00:53 <sgothel> hence naming all details is important (distri, OS, X11 modules/version .. blabla)
20130711 22:01:22 <sgothel> ie: how to reproduce .. can I install some distri on a virtual machine .. etc
20130711 22:01:48 <sgothel> we sometimes have troubles w/ ATI fglrx on linux ..
20130711 22:02:12 <Markt_> its intel
20130711 22:02:35 <sgothel> tested w/ latest debian and ubuntu here: works fine
20130711 22:03:06 <sgothel> .. lets cont. this in the bug report .. me cont. ES3/GL4.3 hacking :)
20130711 22:03:43 <Markt_> ya, i'll have to just trail and error it most likely
20130711 22:03:51 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=781 : pls add the stuff s detailed as I requested above
20130711 22:04:03 <Markt_> Fun =p
20130711 22:04:44 <sgothel> well .. you know, janitor work .. especially for the type of bugs .. we need as much data as possible
20130711 22:06:24 <xranby> Markt_: since you use AWT please include your JRE/JDK version
20130711 22:07:02 <sgothel> XReply/xcb .. oh dear, not another threading bug in their libraries
20130711 22:07:26 <sgothel> again: the test case and jogl debug log
20130711 22:07:50 <sgothel> Fedora Core 14 x64 .. not testing on those machines here, but may do after siggraph
20130711 22:08:39 <sgothel> recommendation: if you can use an opaque GL 'canvas' in your AWT/Swing UI .. well, or don't need AWT/Swing really, use NEWT / NewtCanvasAWT
20130711 22:09:09 <sgothel> you will experience much better threading behavior and input .. and less AWT issues
20130711 22:09:32 <Markt_> How do i change to the other type?
20130711 22:09:50 <Markt_> i really dont need awt or swing
20130711 22:09:53 <sgothel> look at the many unit tests ..
20130711 22:10:00 <sgothel> then you could use plain NEWT
20130711 22:10:30 <sgothel> http://jogamp.org/jogl/doc/NEWT-Overview.html
20130711 22:11:19 <sgothel> this way .. and w/ GL2ES2 .. etc .. your application will also be 'ready' for mobile
20130711 22:12:22 <sgothel> http://jogamp.org/wiki/index.php/Maintainer_and_Contacts#Contacts <- for consultancy services for your industry :)
20130711 22:12:45 <sgothel> but we are pretty packed until after siggraph .. well
20130711 22:14:12 <Markt_> =)
20130711 22:14:27 <xranby> Markt_: with NEWT you start by creating a window using GLWindow.create( example here is the conversion from AWT -> NEWT inside the niftygui project https://github.com/xranby/nifty-gui/commit/ca9e68c096ff6a5074095b9603bc590a6850c2ba#L1L66
20130711 22:14:27 <sgothel> pls try w/ fedora 19 as well, seems to be the latest version
20130711 22:14:39 <Markt_> Well now that i see i can by pass awt... i will prolly try that next
20130711 22:15:05 <sgothel> will reduce bug severity .. until reproducible w/ latest fedora release
20130711 22:15:22 <Markt_> I am not actually able to change my OS at this time
20130711 22:15:50 <Markt_> Thats fine. ya i suspect its a combination of weird circumstances
20130711 22:17:19 <sgothel> try java7, java6 is EOL
20130711 22:17:25 <sgothel> (as well .. )
20130711 22:18:02 <Markt_> Had lots of issues with compatibility in 7. Type mismatches all over the place =(
20130711 22:18:05 <sgothel> but probably an XCB issue, we had tons of it
20130711 22:18:32 <sgothel> but maybe you favor the security fixes .. and 'cleanup' your code ..
20130711 22:19:07 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=758
20130711 22:19:42 <Markt_> Oo i see
20130711 22:19:43 <xranby> sgothel: btw, redhat now work hard to backport security fixes to 6: this got released yesterday: http://blog.fuseyism.com/index.php/2013/07/10/security-icedtea-1-11-12-1-12-6-for-openjdk-6-released/
20130711 22:19:46 <Markt_> that could world
20130711 22:19:47 <Markt_> work
20130711 22:20:04 <xranby> or two days ago...
20130711 22:20:10 <sgothel> yeah right ..
20130711 22:20:38 <xranby> its the 7u25 security fixes backported
20130711 22:21:14 <sgothel> well, surely no Orcale/RedHat/OpenJDK can be considered secure now anyways, but the idea of backporting to an EOL java6 seems sort of inducing complications and giving a solution for no problem.
20130711 22:22:25 <xranby> its the solution used for the long time support "enterprise" distributions.
20130711 22:22:31 <sgothel> best advise if you run on a unix like machine: run a JVM in a process of a diff user w/ most little access to other systems/files if possible!
20130711 22:22:51 <xranby> if someone decide to run linux 1.0 forever redhat would surely supply patches as long as someone pays for it
20130711 22:23:02 <sgothel> opinions differ here .. but thats ok
20130711 22:23:11 <sgothel> lol
20130711 22:24:00 <sgothel> and Gothel Software e.K. and other are happy to provide great releases as well - but no backports (yet).. Hmm .. Harvey mentioned something though.
20130711 22:24:12 <sgothel> for a huge fee maybe :)
20130711 22:24:52 <sgothel> good thing .. WW2 uses JOGL2 by now, b/c we are stubborn :)
20130711 22:24:59 <sgothel> etc
20130711 22:25:48 <xranby> nice to see the saturn spacecraft planner updated to java3d using jogl 2 as well!
20130711 22:26:13 <sgothel> yup, trust me the former almost begged me to do a backport :)
20130711 22:26:59 <sgothel> so when 'enterprise' gets over their lazy ass .. they will see it's stupid to use something not naturally maintained
20130711 22:27:09 <sgothel> which is .. the rolling release
20130711 22:28:06 <sgothel> but security .. illusion, if not separated by proper mechanisms .. JVM simply cannot itself
20130711 22:29:08 <sgothel> had a discussion about the many hoops webgl does .. horrible costs for the illusion IMHO
20130711 22:29:32 <sgothel> a GPU simply has no ring philosophy .. and process / memory separation
20130711 22:36:24 <gabereiser> ok
20130711 22:36:29 <gabereiser> pull requests sent
20130711 22:39:16 <xranby> gabereiser: https://github.com/sgothel/jogl/pull/68 i think you have one more file to check in
20130711 22:40:07 <xranby> with the change that adds the RASPBERRYPI enum
20130711 22:41:44 <xranby> ah!
20130711 22:41:44 <xranby> https://github.com/sgothel/gluegen/pull/15
20130711 22:41:59 <xranby> gabereiser: i missed the gluegen change
20130711 22:43:06 <xranby> gabereiser: well i guess i simplty have to compile this and check the status on the Pi!
20130711 22:43:45 <xranby> to see what needs to be added to the OMX backend
20130711 22:45:26 <xranby> gabereiser: kudos, i will experiment with this, perhaps suggest we wait with the final merge into jogl when the omx backend work better compared to the ffmpeg backend
20130711 22:45:34 <xranby> on the pi
20130711 22:46:43 <gabereiser> yeah
20130711 22:46:47 <gabereiser> thats fine
20130711 22:47:18 <gabereiser> i need to test it out myself
20130711 22:47:27 <gabereiser> but this will at least let me instantiate one
20130711 22:48:22 <gabereiser> https://github.com/huceke/omxplayer
20130711 22:48:33 <gabereiser> this is the omxplayer that comes preinstalled on raspbian
20130711 22:52:11 <xranby> good reference
20130711 22:52:18 <gabereiser> sorry: https://github.com/popcornmix/omxplayer this is the new location for updates
20130711 22:52:26 * Markt_ (32c4d781@anon) Quit (Quit: Page closed)
20130711 23:27:14 <xranby> good night, happy to see more activity improving the a/v backends
20130711 23:28:03 * xranby (~familjen@anon) Quit (Quit: Leaving.)
20130711 23:49:08 * gabereiser (~GXReiser@anon) Quit (Quit: gabereiser)
20130712 01:04:41 * gabereiser (~GXReiser@anon) has joined #jogamp
20130712 01:53:56 * gabereiser (~GXReiser@anon) Quit (Read error: Connection reset by peer)
20130712 01:54:19 * gabereiser (~GXReiser@anon) has joined #jogamp
20130712 05:11:22 * gabereiser (~GXReiser@anon) Quit (Ping timeout: 276 seconds)
20130712 11:09:14 * monsieur_max (~maxime@anon) has joined #jogamp
20130712 11:09:25 * monsieur_max (~maxime@anon) has left #jogamp
20130712 11:51:48 <rmk0_> hm, jesus zuzueta's problem in the forum...
20130712 11:51:50 * rmk0_ is now known as rmk0
20130712 11:51:58 * rmk0 (~rmk0@anon) Quit (Changing host)
20130712 11:51:58 * rmk0 (~rmk0@anon) has joined #jogamp
20130712 11:53:07 <rmk0> it seems like a slightly better algorithm might be to construct the jar file natives name as is being done, but to then iterate through the class path and see if any classpath entry ends with that name, and then attempt to open that one
20130712 11:53:11 <rmk0> i think that'd fix his problem
20130712 11:53:28 <rmk0> right now, it obviously just constructs the name and then tries to open a file with that name in the same directory
20130712 11:53:44 <sgothel> always: pls do not forget every lookup is a network lookup!
20130712 11:53:51 <sgothel> _and_: Good Morning! :)
20130712 11:53:55 <rmk0> lo
20130712 11:54:05 <rmk0> yes, but i think this would still only be one lookup, as it already is
20130712 11:54:50 <sgothel> but you ask to try to iterate through some pattern
20130712 11:55:04 <sgothel> each a lookup
20130712 11:55:22 <sgothel> one guys said .. that ivy can do what maven can plus more ..
20130712 11:55:33 <rmk0> unless i've grossly misunderstood... doing string comparisons on the classpath doesn't look up anything on the network
20130712 11:55:42 <rmk0> it only does that when you actually try to open a file
20130712 11:55:46 <sgothel> forget 'classpath' ..
20130712 11:56:00 <sgothel> we assemble one URI .. and look it up
20130712 11:56:23 <sgothel> we get the base URI by the defining class's
20130712 11:56:32 <rmk0> hm, right
20130712 11:56:45 <sgothel> native jars are not in classpath - and should not be
20130712 11:56:46 <rmk0> then i suppose it's up to him
20130712 11:57:36 <sgothel> since all in classpath would get parsed and interrogated at general class lookup, that would mean JVM would lookup up all native jar files
20130712 11:57:37 <rmk0> i'm not sure this is solvable with gradle
20130712 11:57:43 <sgothel> ah gradle ..
20130712 11:57:55 <sgothel> well, as I said before .. he could add a plugin
20130712 11:58:17 <sgothel> i.e. a NJ.. (forgot my fancy abbreviation) .. :)
20130712 11:58:32 <sgothel> where the jar naming pattern is being plugged in
20130712 11:58:43 <rmk0> right
20130712 11:59:05 <sgothel> dunno why nobody reads my mails :)
20130712 11:59:21 <rmk0> i rarely see any of them
20130712 11:59:27 <sgothel> hehe
20130712 11:59:27 <rmk0> don't have time to watch the forum these days
20130712 11:59:34 <rmk0> tend to respond when people point me at things!
20130712 11:59:37 <sgothel> that is .. err .. great :)
20130712 11:59:41 <rmk0> hehe
20130712 11:59:59 <rmk0> it seems like gradle uses an internal repository format that's going to be extremely hard to use here
20130712 12:00:00 <sgothel> i.e. Wade did one of these plugins for his Eclipse usage
20130712 12:00:21 <rmk0> if it organizes files by what looks like SHA checksums... not going to be able to predict those in order to be able to locate other files
20130712 12:00:24 <sgothel> so I mentioned this thing a fw post above ..
20130712 12:00:47 <sgothel> well .. so it would be all possible, he would just do an implementation :)
20130712 12:01:04 <rmk0> what i mean is
20130712 12:01:29 <rmk0> gradle's going to download all of the jars from maven central, take the SHA checksums of them, and then store them individually in directories named after those checksums
20130712 12:01:50 <rmk0> not sure how you're supposed to locate them again afterwards from the gluegen code... you'd need to know the checksums to locate them
20130712 12:02:05 <sgothel> they must have a mapping
20130712 12:02:11 <rmk0> hopefully!
20130712 12:02:16 <sgothel> i.e. java does same for network cached files
20130712 12:02:38 <sgothel> i.e. all our URL lookups uses the internal java jar cache
20130712 12:02:50 <rmk0> whether it's accessible through some reliable mechanism is another question
20130712 12:02:57 <sgothel> so they can combine all of this
20130712 12:02:59 <rmk0> neither gradle or groovy strike me as carefully thought out
20130712 12:03:32 <sgothel> as long as we can have a clean plugin mechanism and not query the world, I am fine w/ a solution.
20130712 12:04:51 <sgothel> tbh .. I don't know either right now .. and have to go back to booring task of checking GL extensions .. yaaaawn
20130712 12:05:44 <rmk0> not exactly sure what to tell jz
20130712 12:06:13 <rmk0> guess i'll just repeat what you just told me
20130712 12:06:52 <sgothel> emphasize my reply - plugin for naming resolution - and enhancing our existing URL resolver plugin for that cache thingy if required - should do the trick!
20130712 12:07:44 <sgothel> important that they get actively involved and provide a patch .. to discuss!
20130712 12:13:29 <rmk0> replied
20130712 12:14:20 <rmk0> sgothel: on another note entirely... we never did talk about SMF
20130712 12:14:42 <rmk0> i'm mildly concerned that nabble will fall apart in the inevitable SIGGRAPH traffic spike
20130712 12:14:46 <sgothel> you are trying my memory ? .. scared ..
20130712 12:14:57 <sgothel> ah .. forum :)
20130712 12:15:02 <rmk0> yesh
20130712 12:15:22 <rmk0> i think i might be able to migrate user accounts, but i've been looking at how nabble store posts in the database and... it's a nightmare
20130712 12:15:26 <sgothel> my last words were like .. import the data ?
20130712 12:15:31 <rmk0> urhur
20130712 12:15:43 <sgothel> yup .. and lots of work
20130712 12:15:44 <rmk0> we can get some of it, but migrating the posts seems like an enormous job
20130712 12:16:10 <rmk0> they actually store mail messages in the database as plain text along with ordinary forum posts... so we'd first be needing to parse email headers
20130712 12:16:13 <sgothel> so I cannot possibly ask anybody of doing this .. the huge amount of work
20130712 12:16:29 <sgothel> oh .. I have one sneaky idea :)
20130712 12:16:36 <rmk0> shoot!
20130712 12:16:51 <sgothel> since I have almost most (of not all) of entires as emails ..
20130712 12:17:04 <sgothel> maybe we can simply inject the emails !
20130712 12:17:15 <rmk0> it might work!
20130712 12:17:30 <rmk0> not sure if preserving the thread structure would be tricky
20130712 12:17:42 <rmk0> they don't even store them as threads in a strict sense in the database
20130712 12:17:50 <rmk0> every post is a "node", and a node may or may not have a parent
20130712 12:18:02 <rmk0> you can't look at a post in isolation and have any idea which "thread" it's in
20130712 12:18:23 <rmk0> is pretty hairy
20130712 12:18:56 <rmk0> i'd have to configure SMF on jogamp.org to do the mail stuff, as my mail setup on io7m.com isn't sufficient (read: barely exists)
20130712 12:19:15 <rmk0> so... we may find that it's complete crap after setting it up, because i've not been able to test that side yet
20130712 12:19:52 <sgothel> I assume all this will take up days to setup properly .. naturally
20130712 12:20:04 <rmk0> the initial setup is pretty quick, actually
20130712 12:20:12 <sgothel> hence doing this before siggraph .. I don't know
20130712 12:20:14 <rmk0> took about an hour, and i didn't really know what i was doing the first time
20130712 12:20:23 <rmk0> migration could take ages
20130712 12:20:37 <sgothel> sure .. one (you) can go ahead and test it .. hmm but yes, proper migration will take time
20130712 12:21:13 <sgothel> at least .. you can have all necessary access to machine to 'play' w/ it .. err. evaluate
20130712 12:21:21 <rmk0> hehe
20130712 12:21:29 <rmk0> is there a database running?
20130712 12:21:31 <sgothel> if you are up to the task .. KUDOS
20130712 12:21:35 <sgothel> mysql
20130712 12:21:39 <rmk0> right
20130712 12:21:56 <sgothel> sendmail, procmail, dovecot
20130712 12:23:19 <rmk0> any idea what sort of traffic caused nabble to fall over last time?
20130712 12:23:40 <sgothel> dunno
20130712 12:23:42 <rmk0> i'm making the wild assumption that you usually see traffic spikes after things like SIGGRAPH... that may not be the case, of course
20130712 12:24:00 <rmk0> would be pretty unfortuate if all these new people flooded in and couldn't access the forum
20130712 12:24:27 <sgothel> well, we do not have that much readers .. see the web forum stats on each post .. just a few thousand hits
20130712 12:24:46 <sgothel> well .. I mean, awesome for technical stuff though :)
20130712 12:24:58 <rmk0> yeah
20130712 12:25:11 <rmk0> imagine quite a few are one-time visitors too
20130712 12:25:18 <rmk0> something breaks, they turn up, get it fixed, never come back
20130712 12:25:18 <sgothel> other 'content' would surely get more attentions, but I guess .. we are diff. industry .. lol
20130712 12:25:31 <sgothel> sure
20130712 12:26:02 <rmk0> hm... maybe we should wait then
20130712 12:26:10 <sgothel> JogAmp website has about \textbf{20k} unique visitors per month, \textbf{1.5k} per day,
20130712 12:26:10 <sgothel> looking at \textbf{400k} pages per month causing \textbf{200GB} monthly bandwidth utilization.
20130712 12:26:24 <rmk0> nice
20130712 12:26:28 <sgothel> you know our stats page right ?
20130712 12:26:35 <rmk0> i've forgotten it
20130712 12:26:55 <sgothel> http://jogamp.org/log/stats/
20130712 12:26:59 <rmk0> right
20130712 12:27:03 <sgothel> http://jogamp.org/log/stats/jogamp.org.201307.html
20130712 12:27:22 <sgothel> ha - it's the hollidays :)
20130712 12:28:08 <rmk0> i'm quite surprised that so many people use the source archive directly
20130712 12:28:19 <rmk0> vastly more hits on that than the other distfiles
20130712 12:28:39 <sgothel> yes .. very good
20130712 12:28:58 <sgothel> sure the blobs are being cached
20130712 12:29:35 <rmk0> anyway... i'll delay SMF until after siggraph
20130712 12:29:43 <sgothel> .. relieve .. :)
20130712 12:29:46 <rmk0> will try and migrate everything after that calms down
20130712 12:29:56 <rmk0> wonder what sort of grave i'm digging for myself :)
20130712 12:29:57 <sgothel> yup .. and will assist ofc
20130712 12:30:03 <sgothel> hehe
20130712 12:30:29 <sgothel> so if it's possible to use the email threading for SMF graph threading (nodes ..) .. we are fine
20130712 12:30:49 <sgothel> sure .. pictures etc get lost .. but who cares ..
20130712 12:31:02 <rmk0> think it can handle attachments
20130712 12:31:15 <sgothel> but why didn't nobody else did that yet .. hmm
20130712 12:31:22 <rmk0> no idea
20130712 12:31:27 <sgothel> maybe start a thread at SMF dev-forum .. and see ..
20130712 12:31:55 <rmk0> expect they want all of that sort of thing written in php
20130712 12:32:02 <rmk0> so that people can run it on their awful shared hosting
20130712 12:32:09 <rmk0> they can't make me!
20130712 12:32:25 <sgothel> well .. this is an importer tool, must not work 'online'
20130712 12:33:27 <sgothel> on the other hand .. if we would use something like plain mailman .. we would have little work - plus some mailman web archive like they use for LKML
20130712 12:33:32 <sgothel> sure .. no real web forum
20130712 12:34:25 <sgothel> ok .. me shut up .. you are the man, will pick this up later .. surely support you/us :)
20130712 12:34:29 <rmk0> i mildly prefer it, but i think we're in the minority there, unfortunately
20130712 12:34:35 <rmk0> the mailing list, i mean
20130712 12:34:40 <rmk0> got no time to read web forums
20130712 12:35:03 <sgothel> yeah, emails I prefer as well .. better process of dealing w/ things
20130712 17:02:32 * gabereiser (~GXReiser@anon) has joined #jogamp
20130712 18:41:09 * gabereiser (~GXReiser@anon) Quit (Ping timeout: 248 seconds)
20130712 18:42:35 * gabereiser (~GXReiser@anon) has joined #jogamp
20130712 21:18:41 * MarkT (32c4d781@anon) has joined #jogamp
20130712 21:19:06 <MarkT> Hello, I'm trying to switch to NEWT from AWT, but addKeyListener seems to ignore Fkeys?
20130712 21:22:48 * MarkT (32c4d781@anon) Quit (Client Quit)
20130712 21:30:08 <gabereiser> let me test that real quick
20130712 21:31:25 <gabereiser> works for me
20130712 21:32:31 <gabereiser> http://pastebin.com/jSZQSmEb
20130712 23:05:18 * gabereiser (~GXReiser@anon) Quit (Quit: gabereiser)
20130712 23:08:33 * gabereiser (~GXReiser@anon) has joined #jogamp
20130712 23:12:06 * gabereiser (~GXReiser@anon) Quit (Client Quit)
20130712 23:21:21 * gabereiser (~GXReiser@anon) has joined #jogamp
20130713 00:41:37 * gabereiser (~GXReiser@anon) Quit (Quit: gabereiser)
20130713 05:06:02 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20130713050602.html