#jogamp @ irc.freenode.net - 20130905 05:06:10 (UTC)


20130905 05:06:10 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20130904050610.html
20130905 05:06:10 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130905050610.html
20130905 06:30:39 * hharrison (~chatzilla@anon) Quit (Remote host closed the connection)
20130905 09:29:21 <xranby> i wonder what microsoft will do with the OpenGL ES enabled java me "Asha" phones that did bring some income.
20130905 09:39:37 <xranby> btw.. http://www.nextmicrosoftceo.com/ i voted for Richard Stallman
20130905 10:49:51 <sgothel> haha
20130905 10:51:14 <sgothel> wonder whether MS will be sued for the 'steal' approach .. i.e. long planned course of action w/o telling shareholders .. fascinating
20130905 10:54:45 <xranby> truly fashinating that microsofts hope to get a marketshare in india is to sell a phone that is not based on windows
20130905 10:55:38 <xranby> http://articles.timesofindia.indiatimes.com/2013-09-04/strategy/41764408_1_microsoft-nokia-asha-phones-lumia
20130905 10:55:49 <xranby> "'Asha' is the new hope for Microsoft"
20130905 10:56:27 <sgothel> Nokia .. did some blackmail against India, ala .. if we need to pay tax here .. we can also go to China :)
20130905 10:57:05 <xranby> the asha 501 is a meego UI copy based on the Nokia S40 OS
20130905 10:57:16 <xranby> running java me applications
20130905 10:57:23 <xranby> and OpenGL ES
20130905 10:57:33 <sgothel> hmm .. never heard of 'em .. where to buy ?
20130905 10:58:05 <xranby> in india!
20130905 10:58:18 <xranby> and pakistan
20130905 10:58:24 <sgothel> :)
20130905 10:58:32 <xranby> http://www.nokia.com/global/products/phone/asha501/
20130905 10:59:05 <sgothel> interesting .. so they had an alternative to Win phones ..
20130905 10:59:35 <sgothel> should make those 'observers' and judges in such a MS/Nokia case vs shareholders .. curious at least
20130905 11:00:23 <sgothel> so this is like 'jolla' .. Nokia tech .. and good software ontop
20130905 11:01:16 <sgothel> QVGA (320 x 240) .. well
20130905 11:02:12 <xranby> the 501 is probably designed to brings internet to the indian people
20130905 11:02:50 <sgothel> these .. or Jolla's .. would be nice to see Ji Gong on it
20130905 11:03:08 <sgothel> .. and the one you have .. the only one .. the one :)
20130905 11:03:58 <xranby> i have signed up for an nokia dev account... so that i may eventually deploy packages
20130905 11:05:04 <xranby> sgothel: btw... switching subject to jogl.. i prepared a 64bit dev machine yesterday capable of running the 9.0.x mesa drivers
20130905 11:05:30 <xranby> and was able to reproduce the "blue screen" VBO errors
20130905 11:05:38 <xranby> that some users are reporting in the forum
20130905 11:06:15 <sgothel> me unaware .. bug report ? will take a look later .. guess related to some gl3 core thing
20130905 11:07:15 <xranby> when you use mesa in combination with VBO you can reuse VBO handles after glDeleteBuffers unless you explicitly generate new using glDeleteBuffers
20130905 11:07:45 <xranby> all other drivers allows you to reuse the handles after a glDeleteBuffer without fuzz
20130905 11:07:55 <sgothel> "when you use mesa .." GL spec ?
20130905 11:08:07 <xranby> GL 3 core
20130905 11:08:08 <sgothel> or mesa special under the hood .. feature ? :)
20130905 11:08:38 <sgothel> so .. glDeleteBuffers is more like 'earmark for deletion' ?
20130905 11:08:43 <sgothel> in GL3 core ..
20130905 11:09:20 <sgothel> well .. 'nice' trait .. but .. why relating on it ?
20130905 11:09:29 <sgothel> *depending*
20130905 11:09:37 <xranby> can you find any good spec that clearly describes the recommended VBO lifecycle
20130905 11:09:38 <xranby> ?
20130905 11:09:49 <sgothel> the obvious .. ofc ..
20130905 11:09:54 <sgothel> create - use - delete
20130905 11:10:11 <sgothel> use: bind - draw - unbind
20130905 11:10:23 <xranby> i will update my toy gl2es2demo code and fix this.. to support the mesa GL 3 core drivers
20130905 11:10:50 <sgothel> still don't get it ..
20130905 11:11:13 <sgothel> so you are not using VBO as described above w/ obvious lifecycle ?
20130905 11:11:25 <xranby> nope :)
20130905 11:11:31 <sgothel> welll ..
20130905 11:11:58 <sgothel> then pls also use GLContext GLSL version string .. to settle this issue as well :)
20130905 11:12:12 <xranby> sgothel: yup. thank you for pointing that out as well!
20130905 11:12:38 <sgothel> me back to printing - generic tile rendring is done so far
20130905 11:13:17 <xranby> we have a user who gets a blue screen using WorldWind + JME and that is most likely also caused by wrong VBO lifecycle
20130905 11:13:19 <xranby> IMHO
20130905 11:14:03 <sgothel> hmm
20130905 11:14:28 <sgothel> our GLDataArray .. VBO handler do bind/unbind when necessary
20130905 11:14:46 <sgothel> but in our user code, we always bind-draw-unbind .. it's easier
20130905 11:15:09 <sgothel> and they all work everywhere .. so rock solid
20130905 11:15:26 <xranby> sgothel: what is confusing about VBo IMHO is that you have three calls to setup the VBO glBindBuffer + glBufferData + glBindBuffer
20130905 11:15:37 <sgothel> buffer data only ONCE!
20130905 11:15:42 <xranby> and only one call to destroy .glDeleteBuffers
20130905 11:15:53 <sgothel> why confusing ?
20130905 11:16:03 <xranby> so the spec is not clear what gets deleted, is it the data the handle or both
20130905 11:16:07 <sgothel> you create a buffer, you fill it w/ data .. and you use it
20130905 11:16:14 <sgothel> both!
20130905 11:16:28 <sgothel> the buffer is a handle for data
20130905 11:16:34 <sgothel> which gets set by the BufferData ..
20130905 11:16:36 <sgothel> like texture ..
20130905 11:16:51 <xranby> .. glGenBuffers + glBufferData + glBindBuffer
20130905 11:16:52 <sgothel> lazy memory allocation .. similar to textures (same thing) .. and OpenAL :)
20130905 11:17:17 <xranby> is there any call that only removes the buffered data while leaving the handle intact?
20130905 11:17:31 <sgothel> GLVBOArrayHandler
20130905 11:17:42 <sgothel> BufferData needs it to be bound as well
20130905 11:18:01 <sgothel> why ?
20130905 11:18:10 <sgothel> you could do another BufferData w/ less size :)
20130905 11:18:18 <sgothel> i.e. reuse it
20130905 11:18:25 <sgothel> if the spec allows it ..
20130905 11:18:32 <xranby> nope that would lead to memory leaks using the tegra 2 drivers
20130905 11:18:49 <sgothel> forget tegra2 .. what a PIA :)
20130905 11:18:56 <sgothel> but yeah .. just delete buffer
20130905 11:18:59 <xranby> if i call glBufferData using the same handle the tegra will run out of GPU memory
20130905 11:19:16 <sgothel> never saw that .. but hey, we only BufferData once
20130905 11:19:18 <xranby> and lock up the system
20130905 11:19:29 <sgothel> w/ GLVBOArrayHandler
20130905 11:19:45 <sgothel> all buffer handling (VBO, FBO, ..) should be very fast!
20130905 11:19:58 <sgothel> b/c this is the new way of doing things .. all buffers and draw
20130905 11:20:28 <sgothel> OpenGL is now more like a tool for GPU memory management ..
20130905 11:20:41 <xranby> ok i will try write a unit test and a quirk for the tegra driver
20130905 11:21:03 <xranby> looks like i need to patch this GLVBOArrayHandler
20130905 11:21:25 <sgothel> it doesn't work for you ? don't think so
20130905 11:21:25 <xranby> sine this class will try reuse the handle
20130905 11:21:34 <sgothel> hmm
20130905 11:22:06 <sgothel> but quirk for tegra2 is great .. i.e. having it detected
20130905 11:22:30 <sgothel> I need it also for the GLSL discard .. command, i.e. to skip it - crashes on tegra2 only :)
20130905 11:23:08 <sgothel> clearly a very odd bug .. as many .. in old tegra2 driver
20130905 11:23:18 <sgothel> good that nobody uses them anymore :/
20130905 11:24:31 <xranby> hehe.. well have to test if it is fixed using the tegra 3
20130905 11:24:56 <xranby> and that forces me to compile and test on android a well
20130905 11:25:00 <xranby> using my tegra 3 tablet
20130905 11:30:56 <sgothel> tegra2 devices here as well to test ..
20130905 11:31:42 <sgothel> highly probable that this is not tegra in general but related to driver version! your ac100 ?
20130905 11:32:01 <sgothel> so these quirks must be carefully applied
20130905 11:40:41 <xranby> sgothel: i have to look it up. running ubuntu 12.10 armhf https://wiki.ubuntu.com/ARM/TEGRA/AC100#Installing_Ubuntu_12.04_or_Lubuntu_12.10_on_the_AC100 + the drivers provided by ubuntu.
20130905 11:41:03 <xranby> so its at least version R16.1
20130905 11:41:35 <xranby> according to nvidia there exist now R16.2 and R16.3 (that may be incompatible with the ac100) https://developer.nvidia.com/linux-tegra
20130905 11:42:19 <xranby> the ac100 was based on the "harmony" design that is no longer supported
20130905 11:42:54 <xranby> the "Ventana" design is in use by the TrimSlice tegra 2 products
20130905 11:43:28 <xranby> so what i can do is to test the R16.3 drivers on my timslice and check if the problem still exist
20130905 11:44:02 <xranby> the future for the ac100 is now in the hands of the reverse engineering driver efforts
20130905 11:44:58 <xranby> apparently nvidia have now dropped support for "Ventana" as well
20130905 11:45:07 <xranby> since the R16.3 was the last driver update provided
20130905 11:45:13 <xranby> for the tegra 2
20130905 11:45:17 <xranby> Tegra 2 == EOL
20130905 11:45:40 <sgothel> sure ..
20130905 11:45:46 <sgothel> and actually Tegra* == EOL
20130905 11:46:00 <sgothel> next devices will be desktop-ish GeForce designs
20130905 11:47:06 <xranby> http://www.phoronix.com/scan.php?page=news_item&px=MTQ0NzI <- Nouveau Driver Ported To Work On ARM
20130905 11:52:19 <xranby> there is also the tegra-re effort https://github.com/kusma/tegra-re
20130905 11:52:49 <xranby> https://github.com/grate-driver/grate
20130905 11:53:08 <xranby> Grate: Open source Tegra 2D/3D user-space
20130905 11:53:54 <xranby> etc...
20130905 11:56:59 <xranby> https://www.eff.org/deeplinks/2013/09/hundreds-pages-nsa-spying-documents-be-released-result-eff-lawsuit - "Hundreds of Pages of NSA Spying Documents to be Released As Result of EFF Lawsuit"
20130905 11:59:20 <sgothel> .. and Manning and Co.. are in jail .. oh dear
20130905 12:01:44 <sgothel> Now even the German secret service claims those gas attacks are real .. w/o proof .. simply blubbering what their masters told them. Hope the 'leaders' will obey the peoples will .. and not do a thing - especially not w/o a proof. And actually .. well .. bombs can only make peace in an Orwell'ean 1984 works .. which seems to be kind of close.
20130905 12:22:52 <xranby> and we have a nobel peace winner who wants to start a war
20130905 12:23:35 <sgothel> right .. well, pleasing the monetary contributors :/
20130905 12:27:29 <xranby> and the syrian government claims to be in a fight with al-quaida,,, so if USA jumps into this makes US the Al-Quiada airforce?
20130905 12:27:54 <sgothel> haha
20130905 12:28:33 <sgothel> AL Quiada (sic) .. was always a product of US, Saudi-Arabia and Israel -> Watch Carlie Wilsons war :)
20130905 12:29:23 <sgothel> Hilarious how Saudi-Arabia (SA) tried to blackmail Putin .. good he has a backbone. Hope China stands tight as well.
20130905 12:30:33 <sgothel> if the whole middle east trouble would not be .. and oil traded for gold .. and pipelines could get build via pakistan/india .. etc .. well ..
20130905 12:31:07 <sgothel> .. that is not gonna happen .. thx to Al-Quiada (sic), i.e. US, SA, Israel .. and Europe ofc
20130905 12:32:21 <xranby> http://www.madmagazine.com/blog/2013/08/08/mads-war-on-privacy-cover <-- i think i will switch to this news agency for a while
20130905 12:34:40 <sgothel> hehe
20130905 12:40:40 <sgothel> http://jausoft.com/Files/misc/SyriaRevolObamaSingleDictator.jpg :)
20130905 12:43:53 <xranby> http://www.nytimes.com/2013/09/05/world/europe/obama-arrives-in-sweden-before-g-20-summit.html?pagewanted=2 - "Still, the leader of a small political party said Mr. Obama should stop tapping her e-mail account and give back his Nobel Peace Prize." http://youtu.be/CKQ3sKuirkw <- Leader of the Swedish Pirate Party Welcomes Obama to Sweden
20130905 12:44:18 <xranby> sgothel: nice poster, who made it?
20130905 12:44:32 <sgothel> dunno .. found it on 4chan /pol/ :)
20130905 12:45:43 <xranby> you follow 4chan?! :D
20130905 12:47:54 <sgothel> /pol/ is sometimes a good relaxation mixing up 'not so weird' ideas .. as it seems :)
20130905 12:48:18 <sgothel> 2013, the year where all conspiracy theories were proven true :)
20130905 12:48:56 <sgothel> what a change .. few years ago .. all newspaper forum post would be like 'why didn't we start the war already'
20130905 12:49:07 <sgothel> now .. it's like .. haha, we see what you did ..
20130905 12:49:25 <sgothel> nobody believes the BS anymore
20130905 13:04:42 <xranby> sgothel: is all JOGL commits pushed. i got a missing symbol : PMVMatrixCallback
20130905 13:04:46 <xranby> ?
20130905 13:05:15 <xranby> [javac] location: class com.jogamp.opengl.util.RandomTileRenderer
20130905 13:05:26 <xranby> [javac] public void setupPrint(final int printWidth, final int printHeight, RandomTileRenderer.PMVMatrixCallback pmvMatrixCB)
20130905 13:06:12 <sgothel> hmm
20130905 13:06:20 <sgothel> doesn't exist anymore
20130905 13:07:39 <xranby> jogl$ grep -r PMVMatrixCallback src
20130905 13:07:40 <xranby> src/jogl/classes/com/jogamp/opengl/util/TileRendererBase.java: * {@link #attachAutoDrawable(GLAutoDrawable, int, PMVMatrixCallback)}
20130905 13:07:40 <xranby> src/jogl/classes/javax/media/opengl/awt/GLCanvas.java: public void setupPrint(final int printWidth, final int printHeight, RandomTileRenderer.PMVMatrixCallback pmvMatrixCB) {
20130905 13:07:55 <xranby> git pull
20130905 13:07:55 <xranby> Already up-to-date.
20130905 13:08:27 <sgothel> oh dear .. I have pushed GLCanvas ?
20130905 13:08:46 <xranby> i am using jogl 3f2e0d3a9a6cab80936851c44f61107a2993ddfe
20130905 13:08:59 <sgothel> oops
20130905 13:10:40 <sgothel> pushed .. thx
20130905 13:13:54 <xranby> sgothel: thx! BUILD SUCCESSFUL
20130905 13:15:31 <sgothel> fighting w/ AWT .. high-dpi printing .. looks like it's forced to 72 dpi .. ? hmm
20130905 13:15:37 <sgothel> you know about it ?
20130905 13:15:55 <sgothel> thats what the tiled renderer is for .. high dpi slices ..
20130905 13:16:08 <sgothel> (sure .. can be used for anything)
20130905 13:16:12 <xranby> using openjdk or oracles jdk?
20130905 13:16:39 <sgothel> currently oracle ? is it different API ?
20130905 13:17:52 <xranby> yes.. i did find that the output was different based on implementation: http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=801
20130905 13:18:24 <xranby> unfortunally noone did a follow up on this bug
20130905 13:18:24 <sgothel> so .. how to force i.e. 300 dpi ?
20130905 13:18:39 <sgothel> transormation matrix ?
20130905 13:18:49 <sgothel> i.e. 72dpi gears .. well :)
20130905 13:20:49 <xranby> sgothel: hope this help: https://forums.oracle.com/thread/1290877?tstart=1846 - java.awt.print: Setting the DPI
20130905 13:22:59 <sgothel> will try .. thx!
20130905 13:24:05 <xranby> sgothel: cool so you are using opengl for high-dpi prints!
20130905 13:24:33 <sgothel> the GLCanvas / GLJPanel [ / NewtCanvasAWT ] .. yes
20130905 13:24:48 <sgothel> using AWT API print ..
20130905 13:24:59 <sgothel> still fiddling w/ it ..
20130905 13:31:31 <sgothel> Note: unit tests already show how to create a high-res PNG image .. for example
20130905 13:34:06 <sgothel> works .. hmm .. so gears appears smaller .. :)
20130905 13:34:25 <sgothel> so now .. how-to scale it 'big' .. etc etc
20130905 13:38:55 * xranby (~xranby@anon) Quit (Ping timeout: 260 seconds)
20130905 13:44:59 * xranby (~xranby@anon) has joined #jogamp
20130905 13:46:16 <xranby> http://labb.zafena.se/jogamp/test/tile-nvidia/
20130905 13:47:01 <sgothel> excellent :)
20130905 13:47:09 <sgothel> but it uses GL2 ..
20130905 13:52:54 <xranby> sgothel: do you want the output log from the two runs?
20130905 13:52:56 <sgothel> TestRandomTiledRendering2GL2NEWT, TestTiledRendering2GL2NEWT <- add GL2ES2 .. ish maybe - have to think about rowLength .. equires GL2ES3
20130905 13:53:07 <sgothel> nope .. looks good
20130905 14:43:05 <xranby> sgothel: i will send you a scan of the printout .. after pushing the 300dpi button using openjdk 7 IcedTea 2.3.10
20130905 14:45:20 <xranby> sgothel: it is scanned from the printed a4
20130905 15:22:37 <sgothel> it's wrong .. ofc
20130905 15:22:47 <sgothel> i.e. buggy not done yet ..
20130905 15:23:10 <sgothel> also .. I test w/ 'output to file' .. postscript .. saving some trees :)
20130905 16:46:09 * hharrison (~chatzilla@anon) has joined #jogamp
20130905 18:56:00 * dfj is now known as MaoistLama
20130905 18:58:05 * MaoistLama is now known as TaoistLlama
20130905 18:59:21 * TaoistLlama is now known as dfj
20130905 20:58:17 * void256 (~chatzilla@anon) has joined #jogamp
20130905 21:14:35 * void256 (~chatzilla@anon) Quit (Quit: ChatZilla 0.9.90.1 [Firefox 23.0.1/20130814063812])
20130905 22:59:57 * void256 (~void@anon) has joined #jogamp
20130906 00:11:01 * void256 (~void@anon) Quit (Remote host closed the connection)
20130906 02:14:11 * xranby (~xranby@anon) Quit (Ping timeout: 245 seconds)
20130906 02:22:55 * xranby (~xranby@anon) has joined #jogamp
20130906 04:39:00 * hharrison_ (~chatzilla@anon) has joined #jogamp
20130906 05:06:10 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20130906050610.html