#jogamp @ irc.freenode.net - 20150813 05:05:29 (UTC)
20150813 05:05:29 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20150812050529.html
20150813 05:05:29 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20150813050529.html
20150813 06:22:39 * jvanek (jvanek@anon) has joined #jogamp
20150813 06:23:57 * elect_ (~elect@anon) has joined #jogamp
20150813 06:26:39 <elect_> hi
20150813 06:48:49 * monsieur_max (~maxime@anon) has joined #jogamp
20150813 07:23:53 <zubzub> https://www.youtube.com/user/java/videos
20150813 07:23:58 <zubzub> cool talks
20150813 07:30:53 * eclesia (~husky@anon) has joined #jogamp
20150813 07:30:59 <eclesia> good morning
20150813 07:36:22 <zubzub> good morning
20150813 07:36:38 <zubzub> I'll repost my link again for you :p
20150813 07:36:40 <zubzub> https://www.youtube.com/user/java/videos
20150813 07:41:23 * elect_ (~elect@anon) Quit (*.net *.split)
20150813 07:42:44 * 7GHAAX6X6 (~elect@anon) has joined #jogamp
20150813 07:42:49 * elect_ (~elect@anon) has joined #jogamp
20150813 07:46:46 * 7GHAAX6X6 (~elect@anon) Quit (Ping timeout: 256 seconds)
20150813 08:11:26 * packet0 (~packet0@anon) Quit ()
20150813 09:09:18 <xranby_> funny reminder.. oak (java) was initially designed by the FirstPerson team in 1992 to run on mobile embedded devices such as the *7 ("StarSeven") device http://tech-insider.org/java/research/1998/05.html
20150813 09:10:41 <monsieur_max> so sexy : http://tech-insider.org/java/research/1998/images/gosling.jpg
20150813 09:16:12 <xranby_> great content to put on slides
20150813 09:21:06 <xranby_> flashback from 2008 https://www.youtube.com/watch?v=DeupVAMnvFA - At the JavaOne Conference in San Francisco, Ken Russell and Sven Gothel of Sun Microsystems explain how the Nvidia APX2500 chip allows developers to write Java apps on a desktop and run them directly to cell phones. Users will be able to play games and navigate cities in 3D using GPS.
20150813 09:21:39 <monsieur_max> that's the future !
20150813 09:22:08 <xranby_> this was back then when javafx and NEWT was the stuff we now provide using JogAmp!
20150813 09:22:45 <elect_> im wet
20150813 09:22:47 <xranby_> for example back in 2005 Package com.sun.javafx.newt -> https://www.cs.duke.edu/courses/fall09/cps124/resources/jogldoc/com/sun/javafx/newt/package-summary.html
20150813 09:23:22 <zubzub> we did oracle massacre javafx anyway?
20150813 09:24:02 <xranby_> zubzub: well... they did it before publishing "javafx 2"
20150813 09:24:16 <xranby_> javafx 1 was using jogl and was working on many mobiule devices
20150813 09:24:25 <xranby_> using the new language
20150813 09:24:44 <zubzub> javafx 1 looked to me as qml/qt inspired well desgined widget toolkit
20150813 09:24:49 <xranby_> yes
20150813 09:25:06 <zubzub> java fx 2+ looks like a wpf inspired disfunctional disaster
20150813 09:25:13 <monsieur_max> haha :)
20150813 09:25:25 * zubzub uses javafx @work
20150813 09:25:55 * elect_ is sorry for zubzub
20150813 09:26:12 <zubzub> the problem is, is that there is no alternative
20150813 09:26:26 <elect_> sven looks like a gigant compared to Ken :D
20150813 09:26:39 <xranby_> elect_: he is :)
20150813 09:27:33 <elect_> why Sven left Sun?
20150813 09:27:42 <zubzub> because Oracle? :p
20150813 09:27:43 <zubzub> dunno
20150813 09:28:12 <xranby_> elect_: in order to stay a free man
20150813 09:28:44 <elect_> has he ever written anything about?
20150813 09:28:44 <xranby_> Johan Vos have some good sumamry of javafx history here https://www.youtube.com/watch?v=qukopcMA-9M
20150813 09:28:53 <xranby_> elect_: yes on his blog
20150813 09:29:19 <elect_> http://jausoft.com/blog/?
20150813 09:29:38 <xranby_> https://blogs.oracle.com/kbr/entry/java_on_the_nvidia_apx <- more images from 2008
20150813 09:29:57 <xranby_> so javafx 1 released in 2009 using jogl worked on mobile
20150813 09:30:06 <xranby_> javafx 2 in 2001 dropped mobile
20150813 09:30:09 <xranby_> 2011
20150813 09:31:11 <elect_> I can't find anything aboout, xranby_
20150813 09:31:39 <elect_> this is the oldest post http://jausoft.com/blog/page/5/
20150813 09:32:48 <xranby_> http://jausoft.com/blog/2009/11/09/jogl-is-dead-long-live-jogl/
20150813 09:33:53 <elect_> I honestly thought something more elaborated
20150813 09:33:56 <elect_> but never mind
20150813 09:38:10 <zubzub> is johan vos on this channel?
20150813 09:38:18 * gouessej (5ee4b442@anon) has joined #jogamp
20150813 09:38:34 <gouessej> Hi
20150813 09:39:22 <xranby_> elect_: http://www.java-gaming.org/topics/jogl2-newt-future/21417/msg/175749/view.html#msg175749 "However, yes, JOGL/NEWT/etc could not have been continued by myself
20150813 09:39:22 <xranby_> in it's free spirit as long I was employed by Sun, due to policy changes.
20150813 09:39:22 <xranby_> This problem is now solved."
20150813 09:39:35 <gouessej> "navigate cities in 3D using GPS". It's partially possible in TUER, there is already a level made of a map, I can run in 3D close to Paris :)
20150813 09:39:50 <xranby_> gouessej: fun!
20150813 09:40:39 <xranby_> so you plan to run around with argumented reality doing a virtual revolution?
20150813 09:42:39 <gouessej> I don't have a lot of spare time and the levels made with OSM require a lot of work to become interesting
20150813 09:42:43 <elect_> interesting link, xranby_
20150813 09:43:18 <elect_> so basically most of people switched from jogl to lwjgl at that time?
20150813 09:43:25 <elect_> when things looks uncertain?
20150813 09:43:46 <xranby_> elect_: maybe, many projnects never switched
20150813 09:43:46 <gouessej> Lol, a very few developers dropped JOGL
20150813 09:44:15 <xranby_> especially if you targeted mobile
20150813 09:44:24 <elect_> jogl is older than lwjgl?
20150813 09:44:35 <gouessej> difficult question, I think no
20150813 09:45:22 <gouessej> L**** already existed when some APIs had been merged into JOGL
20150813 09:45:35 <elect_> I mean, jogl was developed by sun, who had economical interests, but what was the point for making lwjgl then?
20150813 09:46:03 <gouessej> We don't need Sun/Oracle/whatever corporation
20150813 09:46:18 <elect_> no, I mean the reason at the birth
20150813 09:46:50 <xranby_> sgothel had been workin on http://jausoft.com/gl4java/ since 1999 before he joined sun and started working on jogl
20150813 09:46:52 <xranby_> http://jogamp.org/jogl/doc/Overview-OpenGL-Evolution-And-JOGL.html
20150813 09:47:05 <gouessej> JOGL took the best of several similar projects, Magician, Jungle, GL4Java, ...
20150813 09:50:29 <gouessej> Ken Russell was behind Jungle with Christopher Kline
20150813 09:56:28 <xranby_> Jogl supports the "composable pipeline" paradigm introduced by the Magician
20150813 09:57:56 <elect_> what's that?
20150813 09:58:07 <elect_> btw, the recorded stream of yesterday at siggraph
20150813 09:58:08 <elect_> http://www.ustream.tv/recorded/70729406
20150813 09:58:28 <xranby_> "composable pipeline" is the DebugGL and TraceGL.. for example The DebugGL pipeline calls glGetError after each OpenGL call, reporting any errors found. It can greatly speed up development time because of its fine-grained error checking as opposed to the manual error checking usually required in OpenGL programs written in C.
20150813 09:59:06 <elect_> I didnt know that
20150813 09:59:23 <xranby_> it is also used for the fixed function emulation pilelines that you can use to emulate fixed function on shader only hardware
20150813 10:00:12 <xranby_> elect_: you can enable these at runtime by simply passing -Djogl.debug.DebugGL -Djogl.debug.TraceGL
20150813 10:00:17 <xranby_> How_to_use_the_Debug_and_Trace_GL_Pipeline_and_Debug_GL_Context
20150813 10:00:20 <elect_> yeah, I knew them
20150813 10:00:24 <xranby_> http://jogamp.org/wiki/index.php/How_to_use_the_Debug_and_Trace_GL_Pipeline_and_Debug_GL_Context
20150813 10:00:32 <elect_> I didnt know the debug calls glGetError after each call
20150813 10:01:31 <xranby_> thank you for the stream link
20150813 10:02:13 <elect_> yop
20150813 10:03:33 * gouessej (5ee4b442@anon) Quit (Quit: Page closed)
20150813 10:15:32 <xranby_> https://youtu.be/qukopcMA-9M?t=10m12s <- look there is a javafx 1 phone showcased in this video (the cool stuff that uses jogl)
20150813 10:16:40 <xranby_> maybe showcased is an overstatement
20150813 10:16:45 <xranby_> but its there at least
20150813 10:43:24 * gouessej (5ee4b442@anon) has joined #jogamp
20150813 10:44:10 <gouessej> xranby_: How do you know that it uses JOGL?
20150813 10:44:57 <gouessej> xranby_: I remember that JOGL was used under the hood in JavaFX 1
20150813 10:45:11 <gouessej> xranby_: I thought this demo was showcasing JavaFX 2
20150813 10:45:52 * i-make-robots (~i-make-ro@anon) Quit ()
20150813 10:46:08 <xranby_> he show one of the old javafx 1 phones as a reminder that javafx was wokring on mobile in 2009 and he started the jacafxports project in frustration that javafx2 never added mobile support
20150813 10:49:11 <zubzub> I feel his pain
20150813 10:49:40 <xranby_> http://steveonjava.com/2009/07/ <- the phone is a HTC Diamond
20150813 10:51:40 <zubzub> If he was situated 20km closer to where I live, I'd be applying for a job :p
20150813 10:55:40 <gouessej> xranby_: Have you looked at how the manifest of the fat JAR is made?
20150813 10:56:02 <xranby_> gouessej: the fat jar is made by the jogamp-scripting project
20150813 10:56:16 <xranby_> it basically unpacks many of the ordinary jars into one fat-jar
20150813 10:56:31 <xranby_> thus i guess the manifest is overwritten
20150813 10:56:38 <xranby_> the manifest is actually deleted
20150813 10:56:49 <xranby_> to fix the corner case 2 we need to combine the manifest files
20150813 10:57:38 <gouessej> xranby_: There is a risk of collision, the same attributes have different values in those JARs
20150813 10:57:51 <xranby_> http://jogamp.org/git/?p=jogamp-scripting.git;a=blob_plain;f=jenkins-builds/funcs_promotion.sh;hb=HEAD
20150813 10:58:05 <xranby_> function prom_make_fatjar
20150813 10:59:45 <xranby_> hmm.. to properly test these scripts i really need to setup my own jenkins server
20150813 11:00:17 <xranby_> because thats the only place we use these scripts during the jenkins build promotion
20150813 11:00:31 <gouessej> why do the script of this subproject still use bash instead of Ant?
20150813 11:00:40 <xranby_> its called by jenkins
20150813 11:00:43 <gouessej> "scripts"
20150813 11:00:47 <gouessej> ok
20150813 11:01:34 <xranby_> all maven deployment is handled here as well
20150813 11:01:49 <xranby_> like marks work that upload new releases to maven central
20150813 11:02:25 <gouessej> I assume that your change on the location of native libraries may affect those scripts
20150813 11:02:46 * rmk0 lurks
20150813 11:02:53 <xranby_> the script will not care.. it simply tries to move natives from / to the new /natives/os-arch/
20150813 11:03:13 <xranby_> and because the changes i did already made /natives/os-arch populated
20150813 11:03:17 <xranby_> should not break this script
20150813 11:03:27 <gouessej> "find . -maxdepth 1"??
20150813 11:03:33 <xranby_> rmk0: i have one question
20150813 11:03:50 <gouessej> Xerxes, find will have to look deeper
20150813 11:03:51 <xranby_> rmk0: you mention in one of your blog posts that the reason why gradle work
20150813 11:04:04 <xranby_> gouessej: it do not need to look deeper
20150813 11:04:16 <xranby_> because the natives are already at the right location
20150813 11:04:24 <rmk0> oh, right, yes
20150813 11:04:43 <xranby_> rmk0: the reason is that we look for the natives on the "classpath"
20150813 11:05:27 <rmk0> the way i remember it is that it used to work by accident
20150813 11:05:30 <xranby_> rmk0: i am wokring on making the maven deployment plugin work and the eclipse extract all jars into one jar work
20150813 11:05:43 <xranby_> but this change may break gradle
20150813 11:05:45 <xranby_> hmm
20150813 11:05:49 <xranby_> maybe
20150813 11:05:54 <rmk0> it just so happened that due to the way maven worked internally, everything ended up in the right place
20150813 11:05:56 <xranby_> i reviewed the native loader
20150813 11:06:02 <rmk0> then gradle changed how it worked, and that broke
20150813 11:06:09 <xranby_> and the native loader actually try ask the classloader where the native is
20150813 11:06:15 <xranby_> if the native is found inside one jar
20150813 11:06:15 <rmk0> so we switched to looking up natives on the classpath and fixed both permanently
20150813 11:06:37 <xranby_> i guess we will have to run some manual tests using gradle
20150813 11:06:45 <xranby_> to check that it still work with the new fat-jar
20150813 11:07:05 <rmk0> this new fat-jar is news to me... i've not been paying much attention here as i've been busy
20150813 11:07:16 <xranby_> the changes i do makes the many smaller native jars put the natives inside a new location inside the jars
20150813 11:07:24 <xranby_> so that if you extract all the jars
20150813 11:07:29 <xranby_> no native gets overwritten
20150813 11:07:54 <gouessej> Oops, I have forgotten to keep the dylib for OpenAL in my own scripts :s
20150813 11:09:12 <xranby_> rmk0: https://jogamp.org/bugzilla/show_bug.cgi?id=1145 <- the fat jar
20150813 11:09:17 * rmk0 eyes it
20150813 11:09:55 <rmk0> ... i'd be inclined not to deploy this to maven
20150813 11:10:27 <rmk0> the goal is to give something for people to play around with, but if they're using something like maven, they should be using the correct dependencies
20150813 11:10:48 <xranby_> rmk0: ok.. let me show the next change that adds support for the maven deployment plugin
20150813 11:11:26 <xranby_> rmk0: https://jogamp.org/bugzilla/show_bug.cgi?id=1172#c5
20150813 11:11:40 <xranby_> the fix https://github.com/xranby/gluegen/commit/3f73bbbd44721cc666e4d3505fcf163490636ba8
20150813 11:11:40 <rmk0> i think if someone is already using something like maven, then using a fat jar is no easier and it means the fat jar then has to work with maven and so on... feels unnecessary
20150813 11:12:07 <xranby_> rmk0: bug 1172 deals with an error currently happening when using maven in combination with the correct dependencies
20150813 11:12:48 <xranby_> the maven maven-assembly-plugin single goal jar deployment extracts all jars into one jar
20150813 11:12:55 <xranby_> causing overwrites of the natives
20150813 11:13:22 <rmk0> blech
20150813 11:13:35 <xranby_> i can solve 1172 by using a new in-jar folder structure
20150813 11:13:49 <xranby_> so that all the natives do not get overwritten by the plugin
20150813 11:13:56 <zubzub> yeah
20150813 11:14:03 <zubzub> you need a special descriptor to make it work
20150813 11:14:07 <xranby_> no
20150813 11:14:14 <zubzub> that's how I solved it
20150813 11:14:17 <rmk0> i think it's safe to say i am not a fan of fat jars in any form
20150813 11:14:36 <rmk0> they cause more problems than they solve, and fix a non-problem
20150813 11:14:43 <zubzub> it's not what you want, it's waht the customer wants :)
20150813 11:15:05 <xranby_> its nice that we have the tag classes in all jars with natives
20150813 11:15:07 <rmk0> if the customer can't deal with typing in a single maven dependency...
20150813 11:15:21 <xranby_> makes itpossible to locate the classloader to be used to find the native inside the jars
20150813 11:16:10 <xranby_> zubzub: i do not follow why you need a new descriptor
20150813 11:16:20 <xranby_> and what bug you see if you do not
20150813 11:16:33 <zubzub> what I main is you need a special maven assembly configuration
20150813 11:16:41 <xranby_> aha
20150813 11:16:54 <xranby_> what do this assembly configuration do?
20150813 11:17:08 <zubzub> depens how you configure it :p
20150813 11:17:14 <gouessej> I agree with rmk0 about not deploying the fat JAR on Maven
20150813 11:17:27 <zubzub> xranby_: let me check in my git history
20150813 11:17:28 <xranby_> i also agree that we do not need to deploy the fat jar to maven
20150813 11:17:42 <xranby_> but i think we want users who use the maven maven-assembly-plugin single goal jar deployment
20150813 11:17:46 <xranby_> to have working deplopyments
20150813 11:18:22 <gouessej> Is it possible to make maven-assembly-plugin work without our fat JAR?
20150813 11:18:31 <xranby_> gouessej: yes
20150813 11:18:55 <xranby_> gouessej: my patch https://github.com/xranby/gluegen/commit/3f73bbbd44721cc666e4d3505fcf163490636ba8 should be enough to fix it
20150813 11:19:18 <gouessej> xranby_: your patch is ok for me, I understand it now
20150813 11:19:23 <zubzub> xranby_: https://github.com/Zubnix/westmalle/blob/71413ec845c4b263672f50bfd37b378a6046097e/bootstrap/src/main/assembly/static.xml
20150813 11:19:41 <gouessej> xranby_: it makes a legitimate case work
20150813 11:20:10 <xranby_> it also makes the eclipse "one-jar" -> export -> java runnable jar file -> * "Extract required libraries into generated JAR" deployment option work
20150813 11:20:24 <xranby_> that is the default option when using eclipse one-jar
20150813 11:20:53 <xranby_> eclipse can use a maven project with correct dependencies
20150813 11:21:04 <xranby_> and use this deployment option to create a single jar with all dependencies
20150813 11:21:06 <gouessej> xranby_: It's a good piece of news but I don't see what would break Gradle
20150813 11:21:23 <xranby_> good, it probably will not
20150813 11:21:34 <xranby_> but as i recall there was something special with gradle
20150813 11:21:51 <xranby_> and kettle
20150813 11:22:13 <xranby_> or is it called kotlin?
20150813 11:22:29 <gouessej> It reminds me that I have forgotten to update my Gradle scripts when I switched to the version 2.3.2 RC of JOGL :s
20150813 11:22:46 <gouessej> kotlin?
20150813 11:22:58 <gouessej> kettle, no
20150813 11:23:33 * xranby_ is having a hard time to remember
20150813 11:23:38 <xranby_> it was a new language
20150813 11:23:42 <xranby_> using a teapot symbol
20150813 11:23:48 <xranby_> something
20150813 11:23:57 <xranby_> i think rmk0 tested it out
20150813 11:24:05 <xranby_> and ended up in a 2 day debug session with the developers
20150813 11:24:09 <rmk0> ceylon
20150813 11:24:15 <xranby_> ceylon! ah! thank you
20150813 11:24:18 <gouessej> "The critical internal APIs proposed to remain accessible in JDK 9 are: sun.misc.Cleaner" I'm reassured: http://openjdk.java.net/jeps/260
20150813 11:24:44 <rmk0> we decided there isn't a good way to support ceylon directly... would need to unpack the jar files and produce native ceylon packages with everything arranged the way they need it
20150813 11:25:10 <xranby_> what was the issue with ceylon?
20150813 11:25:40 <xranby_> it didnt allow that you loaded natives using one classloder to bu used by gl code running on a differnt classloader?
20150813 11:26:01 <elect_> gouessej, why there are no more jogl thread on jgo?
20150813 11:26:01 <rmk0> yep
20150813 11:26:11 <xranby_> elect_: politics
20150813 11:26:31 <gouessej> rmk0: "It runs on both Java and JavaScript ". Have you used Ceylon a lot?
20150813 11:26:44 <rmk0> gouessej: nope!
20150813 11:26:46 <elect_> xranby_, can you elaborate?
20150813 11:26:52 <xranby_> elect_: better have the jogl quiestions here at the jogamp forum
20150813 11:26:54 <gouessej> elect_: It was a nightmare
20150813 11:27:25 <elect_> but this is actually counterproductive
20150813 11:27:26 <gouessej> elect_: It was my choice to encourage the developers to ask questions on our official forum
20150813 11:27:36 <gouessej> elect_: Please be less peremptory
20150813 11:27:49 <elect_> what do you mean?
20150813 11:27:56 <xranby_> elect_: jgo is a forum for game developers, the forum turned into a mess developers trying to convince each other of the gl binding
20150813 11:28:23 <xranby_> gl binding support quiestions should not be part of such a forum
20150813 11:28:26 <gouessej> elect_: It was unberable, some (a few??) supporters of the other library were promoting their stuff very often when I tried to help JOGL users
20150813 11:28:54 <gouessej> elect_: xranby_ sums it up better than me
20150813 11:29:33 <elect_> but you do realize that in this way you left the door opened for lwjgl?
20150813 11:30:08 <gouessej> elect_: Do you realize that it's not so simple?
20150813 11:30:11 <xranby_> elect_: there is many frameworks build on top of jogl and lwjgl that is discussed in the forum
20150813 11:30:12 <elect_> you are right xranby_, but this may discourage people to come to jogl forum and ask
20150813 11:30:50 <gouessej> elect_: The flamewars were discouraging people
20150813 11:31:10 <xranby_> now we have a better situation, happy forum posts in our own forum
20150813 11:31:18 <gouessej> exactly
20150813 11:31:18 <xranby_> the search engines finds our forum
20150813 11:31:32 <elect_> but much lower activity
20150813 11:31:44 <elect_> have you ever done some comparison?
20150813 11:31:46 <xranby_> we are not the jgo team
20150813 11:32:07 <xranby_> thus we work on moving visitors to the jogamp forum
20150813 11:32:16 <gouessej> The JGO team isn't in our team......
20150813 11:32:17 <xranby_> where we can support them better
20150813 11:32:26 <gouessej> we can support them peacefully
20150813 11:32:33 <gouessej> without being bothered
20150813 11:33:09 <xranby_> i am maintaining the raspberry pi forum
20150813 11:33:20 <xranby_> where we have a lot of entusiastic users
20150813 11:33:23 <xranby_> because jogl work
20150813 11:33:29 <xranby_> on their devices
20150813 11:33:48 <gouessej> xranby is known to be of help on embedded devices with Java in general
20150813 11:33:57 <xranby_> and gouessej do lot of work on the stack overflow
20150813 11:34:37 <gouessej> It's really difficult to do the same on JGO
20150813 11:35:24 <gouessej> elect_: I am on JGO
20150813 11:35:34 <elect_> i know
20150813 11:35:37 <xranby_> JGO is good to showcase games
20150813 11:35:44 <gouessej> I subscribed in 2007
20150813 11:36:18 <xranby_> for example deployments of games using libgdx + jogl may be sowcased fine on JGO
20150813 11:36:23 <gouessej> A few supporters do the dirty job anyway
20150813 11:36:46 <elect_> but jgo catalyze a lot of indie devs, and if they see lwjgl everywhere, it is very likely they are going to use that
20150813 11:36:52 <gouessej> I have to update LibGDX soon
20150813 11:37:45 <xranby_> elect_: please write more blog posts :)
20150813 11:37:56 <elect_> ?
20150813 11:38:15 <xranby_> describing your work and what technology you find working
20150813 11:38:16 <gouessej> elect_: Minecraft crashed under GNU Linux with some drivers
20150813 11:38:34 <elect_> on my blog?
20150813 11:38:39 <xranby_> yes why not
20150813 11:38:44 <elect_> I dont have time for that
20150813 11:38:56 <xranby_> why not?
20150813 11:39:13 <elect_> I have a couple of projects with higher priority
20150813 11:39:17 <gouessej> elect_: Don't worry, it isn't the end of the world. They aren't married with a particular API.
20150813 11:39:17 <xranby_> thats the best way to grow awareness of technology that you like
20150813 11:40:16 <elect_> but it might be the end of jogl in the long run
20150813 11:40:22 <gouessej> elect_: No
20150813 11:40:27 <xranby_> elect_: no way :)
20150813 11:40:42 <xranby_> elect_: because, there is no architecture out there as capable
20150813 11:40:43 <elect_> btw, I dont see what writing more blog posts has to do it with jgo and jogl
20150813 11:40:55 <xranby_> elect_: do not write about jgo
20150813 11:41:01 <xranby_> elect_: write about your projects
20150813 11:41:08 <monsieur_max> elect_: its easy : see http://shscatequesis.blogspot.fr/2012/12/catequesis-tech-inside-1.html
20150813 11:41:16 <gouessej> elect_: It's just a mean of helping people to discover our stuff
20150813 11:42:24 <elect_> I guess the best way to help people embracing jogl is a better wiki, a complete list of simple test scenario for the most common questiona and a better forum
20150813 11:42:28 <elect_> but you know that ^^
20150813 11:42:48 <elect_> and this is what I am trying to do
20150813 11:43:07 <gouessej> and I'm grateful to you about that
20150813 11:43:16 <elect_> writing post is relatively useful only if you have a public, which I dont
20150813 11:43:25 <xranby_> monsieur_max: your last game looks fantastic https://curvedcatgames.wordpress.com/tales-of-pocoro/
20150813 11:43:38 <xranby_> i envy people who have good artists :)
20150813 11:43:55 <gouessej> elect_: it was just a suggestion. Your examples are already very useful
20150813 11:43:59 <monsieur_max> xranby_: well, to be honest, no artists, only free resources
20150813 11:44:08 <xranby_> amazing
20150813 11:44:23 <monsieur_max> xranby_: but my current project have a real artist, and it's really really great :)
20150813 11:45:07 <elect_> I wondered why no1 considered an hello triangle before
20150813 11:45:20 <xranby_> elect_: hello triangles are the best :D
20150813 11:45:25 <elect_> if you go on lwjgl website, they have one right on the download page
20150813 11:45:40 <monsieur_max> it helps keeping the motivation... people only sees the graphics, placeholders are not rewarding even if the engine is great
20150813 11:45:41 <elect_> moreover, I'd like to remember again a general cleaning
20150813 11:45:58 <gouessej> monsieur_max: I like your reply about ios
20150813 11:46:02 <elect_> a lot of stuff are outdated, but new people dont know that
20150813 11:46:20 <gouessej> "absolutely zero chance of it being ported to IPhone" :)
20150813 11:46:40 <gouessej> elect_: there is the same problem with Java3D
20150813 11:46:54 <monsieur_max> ;) yeah, i ported it to desktop with jogl, but stopped there
20150813 11:46:58 <gouessej> elect_: it is less problematic than with JOGL
20150813 11:47:23 <gouessej> monsieur_max: I assume that you don't use JOGL under Android
20150813 11:47:29 <monsieur_max> xranby_: i still remember my first jogl triangle !
20150813 11:47:55 <monsieur_max> gouessej: nope, at the time, i used the android gl bindings
20150813 11:48:06 <monsieur_max> very similar to jogl
20150813 11:48:44 <monsieur_max> that's why I chose jogl, i like the simplicity, and almost use only the bindings.
20150813 11:49:02 <gouessej> monsieur_max: We are a source of inspiration for Google, Kenneth is probably behind ;)
20150813 11:49:06 <monsieur_max> the only feature i use , beside bindings, is the mediaplayer
20150813 11:49:32 <monsieur_max> gouessej: yes he is, i learned it from Romain Guy
20150813 11:50:15 <gouessej> monsieur_max: We have a tactical advantage on our main competitor :)
20150813 11:50:37 <xranby_> time is our main competitor :/
20150813 11:50:41 <xranby_> hehe
20150813 11:51:04 <xranby_> the 2.3.2-rc looks fantastic
20150813 11:51:53 <xranby_> all the small details that makes the whole shine
20150813 11:52:15 <monsieur_max> xranby_: you're looking at the changelog ?
20150813 11:52:22 <xranby_> yes a lot
20150813 11:52:25 <gouessej> rmk0: Is there a mean of converting Java code to Ceylon in order to run it in a web browser?
20150813 11:52:27 <xranby_> i am looking at the
20150813 11:52:40 <xranby_> http://jogamp.org/wiki/index.php/SW_Tracking_Report_Objectives_for_the_release_2.3.2
20150813 11:52:53 <xranby_> as the main work list
20150813 11:53:50 <rmk0> gouessej: i don't think they do bytecode translation for java dependencies, no...
20150813 11:55:16 <xranby_> rmk0: to fix Ceylon properly we need to deploy using one fat-jar -> one classloader OR make sure that we re-load natives for each classloader?
20150813 11:55:47 <rmk0> xranby_: i think the main problem was that gluegen and jogl would live in different classloaders
20150813 11:55:59 <xranby_> ok then the fat-jar should solve ceylon
20150813 11:56:02 <rmk0> the solution was going to be to pack everything into one ceylon package and then it'd magically work
20150813 11:56:13 <rmk0> they don't really consume jar files, they have their own thing
20150813 11:56:32 <rmk0> what they have is closer to osgi bundles
20150813 11:57:18 <rmk0> to be clear... i'm not a user of the language
20150813 11:57:30 <rmk0> i had a look at it when it reached 1.0, because it has an interesting type system
20150813 12:07:42 <xranby_> pointer0: it is happening, the japanese java community finds out about jogamp thanks to the API change in 2.3.X http://acochan.hatenablog.com/entry/2015/08/08/164517
20150813 12:09:05 <gouessej> xranby_: It was a good idea :)
20150813 12:09:22 <xranby_> the javadoc also looks fantastic now
20150813 12:09:38 <xranby_> making the modules fit better togeather
20150813 12:18:29 <xranby_> https://www.youtube.com/watch?v=D6Lkw3eZK1w - JOGL on APX 2500 (Tegra1) Dec 2008
20150813 12:45:53 * gouessej (5ee4b442@anon) has left #jogamp
20150813 13:19:29 * packet0 (~packet0@anon) has joined #jogamp
20150813 13:19:55 <xranby_> packet0: it is happening, the japanese java community finds out about jogamp thanks to the API change in 2.3.X http://acochan.hatenablog.com/entry/2015/08/08/164517
20150813 13:20:18 <elect_> I cant see any method in FloatUtils able to convert a 16 float array (mat4) into a 9 (mat3)
20150813 13:23:12 <packet0> xranby_: great to see that happening. took sort of long though...
20150813 13:24:45 <elect_> packet0, where you from exactly?
20150813 13:25:04 <packet0> japan
20150813 13:25:19 <elect_> I supposed, city?
20150813 13:25:37 <xranby_> elect_: you go first
20150813 13:25:52 <elect_> ?
20150813 13:25:57 <elect_> mine?
20150813 13:26:14 <xranby_> elect_: yes, i actually do not understand why knowing someones location is that important
20150813 13:26:21 <elect_> just curiosity
20150813 13:26:44 <elect_> sometimes I read wiki's about
20150813 13:27:13 <elect_> I have a friend close to tokyo
20150813 13:27:29 <elect_> and one of my wife's relatives is there right now
20150813 13:28:02 <elect_> I also know one of my town person, who married a japanese and moved there
20150813 13:28:09 <elect_> it is an interesting land
20150813 13:28:21 <elect_> interesting culture
20150813 13:28:37 <packet0> the ip lookup geolocation is basically close enough to where i am...
20150813 13:28:53 <elect_> dont worry, if you dont wanna share you dont have to
20150813 13:29:27 <elect_> this relative sent us some photo about the Shinkanzen Nozomi
20150813 13:29:33 <elect_> they are awesome
20150813 13:31:17 <packet0> its literary really close, like 20min on a train...
20150813 13:32:00 <xranby_> elect_: thank you for clarifying
20150813 13:32:15 <elect_> np
20150813 13:32:25 <elect_> btw I come from turin if you are interested
20150813 13:33:48 <monsieur_max> never forgetti mom's spaghetti !
20150813 13:34:36 <xranby_> sounds like a good place to enjoy gelato and chocolate
20150813 13:35:22 <bruce-> I once had plans to move to Milano
20150813 13:35:31 <elect_> yep
20150813 13:36:02 <elect_> if you ever go there
20150813 13:36:18 <elect_> I can give you a detailed map about gastonomy
20150813 13:36:38 <elect_> ^^
20150813 13:36:39 <xranby_> you may need to update the map for a vegan tour
20150813 13:36:46 <bruce-> my (ex?)girlfriend moved there
20150813 13:36:56 <elect_> are you vegan?
20150813 13:37:10 <elect_> bruce, I am glad you didnt go to milan ^^
20150813 13:37:12 <xranby_> of course
20150813 13:37:35 <elect_> icecream is not a problem, for the pizza you can easily avoid meat
20150813 13:37:37 <bruce-> elect_: is this Turino/Milano rivalry or are there real reasons? :)
20150813 13:37:39 * monsieur_max cuts all his spaghettis 20 times with a knife in front of elect_, staring at him right in the eyes
20150813 13:37:58 <bruce-> monsieur_max: m…monster
20150813 13:38:17 <elect_> a bit rivarly, a bit a pity, coz we have such a much nicer places than milan, but people seem knowing always just milan..
20150813 13:38:20 <xranby_> monsieur_max: is an expert on scary topics
20150813 13:38:30 <elect_> YOU DONT DARE
20150813 13:38:39 <monsieur_max> i'm testing elect_ resistance
20150813 13:38:45 <bruce-> elect_: the funny thing is, while looking for place where I would like to work I only found places in Turino
20150813 13:38:48 <elect_> I'll show you the holy tecnique
20150813 13:39:00 <elect_> !
20150813 13:39:04 <elect_> which field?
20150813 13:39:23 <packet0> my swiss friend use to cut his spaghetti into pieces, is that where that custom is from?
20150813 13:39:45 <bruce-> interaction design … but the part that does not do websites and apps
20150813 13:40:07 <elect_> pussies cut them, real men eat them as they are ^^
20150813 13:40:41 <elect_> bruce-, and you where you from? You know some italian?
20150813 13:41:28 <monsieur_max> elect_: joking, I never cut my pastas ;) I'm not a barbarian
20150813 13:41:31 <elect_> if you latin it shouldnt be that hard
20150813 13:41:41 <elect_> it is better for you, monsieur_max
20150813 13:42:14 <bruce-> elect_: I am from the netherlands. My Italian is close to none eventhough I lived with an Italian GF for 3 years.
20150813 13:42:18 <monsieur_max> elect_: same for the pizza, i don't cut it, rolling it and eating it
20150813 13:42:30 <elect_> oh god
20150813 13:42:41 <elect_> I have a nice friend of mine from nl
20150813 13:42:50 <elect_> a lot of ppl from unreal tournament
20150813 13:43:16 <monsieur_max> name from the past ... reviving old memories
20150813 13:43:26 <elect_> My sister is in Rotterdam, when we went there to visit there, we bought a tandem on the way
20150813 13:43:38 <monsieur_max> ancient lan parties ...
20150813 13:43:57 <elect_> near Gouda
20150813 13:44:06 <elect_> btw, excellent kase
20150813 13:44:12 <elect_> ^^
20150813 13:44:38 <bruce-> feh, no where close near parmigiano
20150813 13:44:40 <monsieur_max> ach so ! sehr gut käse !
20150813 13:44:54 <bruce-> (kaas btw)
20150813 13:45:07 <elect_> and btw, I didnt understood why you are used to close so often the left lane on the autobahn
20150813 13:45:31 <elect_> althought they were no reconstruction or similar
20150813 13:45:36 <elect_> totally empty and clean
20150813 13:45:56 <elect_> and I guess I took some fine on the way
20150813 13:46:27 <elect_> I had the app for the blitzer, but I discovered only later what the small icon "SC" meant
20150813 13:46:51 <elect_> it stays for Speed Control, you do calculate the average speed as well as we do in some part in italy..
20150813 13:47:09 <monsieur_max> *speedi controli"
20150813 13:48:08 <elect_> btw I got a fine also close to grenoble
20150813 13:48:42 <monsieur_max> good good, give us your money !
20150813 13:48:42 <elect_> but because the police guy said I cross the line, you know where you have 2 lanes when there is a slope in order to let people pass slower tir, monsieur_max ?
20150813 13:48:55 <elect_> I think i didnt, but I had no chances
20150813 13:49:22 <elect_> he was "holding the knife from the handle" we say
20150813 13:49:33 <elect_> you mean, we poor italians!
20150813 13:49:35 <monsieur_max> mmmh
20150813 13:50:13 <monsieur_max> "holding the knife from the handle" what does it mean ... i mean, it's the right way to use a knife here in France :P
20150813 13:51:23 <elect_> it is a way to say I had nothing to support my part
20150813 13:51:34 <elect_> you dont say something similar in francais?
20150813 13:51:56 <monsieur_max> no, as you may know it, we always surrender, so ...
20150813 13:52:00 <elect_> :D
20150813 13:52:07 <monsieur_max> no need for such a saying ;)
20150813 14:10:59 * gouessej (5ee4b442@anon) has joined #jogamp
20150813 14:13:24 <gouessej> xranby: You should try the Gentle Gourmet in Paris, I find this vegan restaurant very good.
20150813 14:15:58 * packet0 (~packet0@anon) has left #jogamp
20150813 14:16:56 <gouessej> monsieur_max: "holding the knife from the handle" means "avoir le couteau sous la gorge"
20150813 14:17:46 * packet0 (~packet0@anon) has joined #jogamp
20150813 14:27:26 <monsieur_max> elect_: "having the knife on your throat"
20150813 14:27:34 <monsieur_max> gouessej: thanks :)
20150813 14:37:26 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20150813 15:20:03 <xranby_> gouessej: thanks
20150813 15:22:02 * elect_ (~elect@anon) Quit (Ping timeout: 260 seconds)
20150813 15:26:14 * xranby_ is trying to find the implementatino for dispatch_eglGetConfigAttrib0
20150813 15:30:18 <xranby_> found it build/jogl/gensrc/native/jogl/egl/EGL_JNI.c
20150813 15:32:55 * elect (~elect@anon) has joined #jogamp
20150813 15:33:46 * elect (~elect@anon) Quit (Client Quit)
20150813 15:43:23 * gouessej (5ee4b442@anon) Quit (Quit: Page closed)
20150813 15:47:46 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150813 16:05:07 * eclesia (~husky@anon) has left #jogamp
20150813 16:59:23 * packet0 (~packet0@anon) Quit ()
20150813 17:00:06 * monsieur_max (~maxime@anon) has joined #jogamp
20150813 18:16:18 <xranby_> Erik De Rijcke has been working on tarvis CI integration and a jogl newt linux framebuffer gbm/EGL NEWT driver! -> https://github.com/Zubnix/jogl/commits/master
20150813 18:16:58 <xranby_> also this adds some logic for a NEWT wayland driver
20150813 18:23:30 <xranby_> https://jogamp.org/bugzilla/show_bug.cgi?id=1156 - Newt EGL/GBM (KMS) driver
20150813 18:24:26 <xranby_> zubzub: do you think your diver is in good shape for merging?
20150813 18:28:04 * Eclesia (~eclesia@anon) has joined #jogamp
20150813 18:33:10 * a11mad11 (c0e2cd65@anon) has joined #jogamp
20150813 18:33:52 <xranby_> a11mad11: greetings
20150813 18:34:04 <a11mad11> hi
20150813 18:34:32 <zubzub> xranby_: no not at all
20150813 18:34:49 <zubzub> I actually stopped working on it a bit
20150813 18:35:19 <zubzub> because the whole egl innitialization stack of newt needs to change
20150813 18:35:26 <zubzub> for it to properly work
20150813 18:36:23 <zubzub> so it uses eglGetPlatformDisplay
20150813 18:36:35 <zubzub> but for that the x11 and others need to be adapted as well
20150813 18:36:46 <zubzub> so the scope of my project suddenly became a whole lot bigger
20150813 18:37:15 <zubzub> and as I was/still am not really familiar with newt (and sgothel was absent at the time) I stopped working on it
20150813 18:39:17 <xranby_> zubzub: yes always details
20150813 18:40:03 <xranby_> i think the detection can be done in steps because a kms driver may exist on all linux systems
20150813 18:40:20 <xranby_> thus instead to probe for /dev/*/card* files
20150813 18:41:18 <xranby_> one idea is to 1st try to open a wayland window, then try to open an xorg window and if that fail then try open an bcm/EGL window
20150813 18:43:05 <xranby_> i am sure sgothel will be around more when his baby has grown a little
20150813 18:43:38 <xranby_> in any case. this year i will have quite a lot of time to work on newt
20150813 18:44:12 <xranby_> because using a raspberry pi with an attached tft looks like a viable prototype for the things i do at my work
20150813 18:45:04 <xranby_> zubzub: thus if you want a "partner in crime" to hack up the newt bcm/egl support i am all ears
20150813 18:45:06 <zubzub> xranby_: it's not the gradual detection stuff
20150813 18:45:10 <zubzub> it's the egl init stuff
20150813 18:45:28 <xranby_> zubzub: i have to look deeper at your bugreport in order to gasp it
20150813 18:45:38 <zubzub> currently newt has a generic egl init system
20150813 18:45:49 <zubzub> but that one needs to change a bit
20150813 18:46:37 <zubzub> normall you'd do eglGetDisplay(nativedisplay) and you're done.
20150813 18:46:41 <zubzub> (or something)
20150813 18:47:16 <zubzub> problem is that the the underlying implementation has to guess what the nativedisplay is if you have multiple egl displays available
20150813 18:47:30 <zubzub> (x11, wayland and gbm for example)
20150813 18:47:57 <zubzub> so there eglGetPlatformDisplay(...) where you can specify what 'type' of native display your passing in
20150813 18:48:05 <zubzub> *you're
20150813 18:48:35 <xranby_> i read you
20150813 18:48:49 <xranby_> you need to pass a specific EGLNativeWindowType, EGLNativePixmapType and EGLNativeDisplayType depending on what the driver expects
20150813 18:49:01 <zubzub> exactly
20150813 18:49:17 <zubzub> currently the newt egl stack does not have those things in place
20150813 18:49:27 <xranby_> afaik, if you pass the wrong type, egl will not crash, it will simply error
20150813 18:49:35 <zubzub> probably
20150813 18:49:36 <xranby_> telling you that you passed the wrong type
20150813 18:49:46 <zubzub> you can also query what types egl supports
20150813 18:50:00 <zubzub> on your system
20150813 18:50:36 <xranby_> yes we currently guess based on some heuristics
20150813 18:50:52 <zubzub> https://www.khronos.org/registry/egl/extensions/EXT/EGL_EXT_platform_base.txt
20150813 18:51:00 <xranby_> that if you are on a linux system then your EGL probably wants an unix x11 Window
20150813 18:51:13 <xranby_> nice
20150813 18:51:50 <xranby_> do we know if the drivers we target support this EGL extension?
20150813 18:52:22 <zubzub> yeah if you query the egl extensions
20150813 18:55:03 <xranby_> last time i looked at mesa... i think mesa supported that you enabled many egl types at the same time
20150813 18:55:46 <xranby_> thus i think we may face linux systems that have support for x11, wayland and gbm at the same time
20150813 18:55:55 <zubzub> oh yest definitly
20150813 18:56:17 <xranby_> thus we still need to have some heuristics to detect what the user use right now
20150813 18:56:36 <zubzub> the user could be using some things at the same time too
20150813 18:56:38 <xranby_> so that we use x11 if the user is inside an xorg session
20150813 18:56:53 <zubzub> well a user could be running x11
20150813 18:56:56 <xranby_> wayland supports both x11 and wayland
20150813 18:56:56 <zubzub> on wayland :)
20150813 18:56:58 <zubzub> yes
20150813 18:57:11 <xranby_> therefore i suggest to query for wayland first
20150813 18:57:22 <zubzub> that makes sense
20150813 18:57:26 <zubzub> if the user is not running wayland on x11 :p
20150813 18:57:40 <zubzub> but the latter is more for testing purposes anyway
20150813 18:57:44 <xranby_> noone wants to do that for performance reasons
20150813 18:57:59 <zubzub> so the order should probably be wayland->x11->gbm
20150813 18:58:05 <xranby_> thus if someone run wayland on x11 they probably want an x11 newt window
20150813 18:58:13 <xranby_> yes
20150813 18:59:07 <xranby_> same thing when we look how to aquire input
20150813 18:59:41 <xranby_> first try to get input from wayland then x11 and last using the LinuxEventTracker reading the /dev/ files directly
20150813 18:59:49 <zubzub> or use libinput :)
20150813 19:00:20 <xranby_> we already have a "working" implementation
20150813 19:00:26 <xranby_> that we use on the raspberry pi
20150813 19:00:35 <xranby_> if the user runs opengl es content from console
20150813 19:00:48 <xranby_> we then read the /dev/ files directly
20150813 19:00:48 <zubzub> http://wayland.freedesktop.org/libinput/doc/latest/
20150813 19:01:08 <xranby_> ah
20150813 19:01:08 <xranby_> ok
20150813 19:01:24 <xranby_> guess libinput is the way to handle input especially if you have wayland installed
20150813 19:01:29 <zubzub> no
20150813 19:01:39 <zubzub> libinput is a stand alone user input driver
20150813 19:01:51 <zubzub> but it's used by wayland compositors/display servers
20150813 19:01:58 <zubzub> or by the x11 libinput driver
20150813 19:02:05 <zubzub> or by a stand alone applicatoin (gbm)
20150813 19:02:09 <xranby_> does it handle localization?
20150813 19:02:22 <zubzub> I would assume so vey much
20150813 19:02:33 <zubzub> quite a few intel devs are working really hard on it
20150813 19:02:44 <xranby_> the kernel only give raw pressed key id's
20150813 19:02:49 <xranby_> without localization
20150813 19:03:00 <xranby_> thus you have to apply an keymap at some level
20150813 19:03:14 <xranby_> if you want the stickers on the keyboard to map reality
20150813 19:03:28 <zubzub> a keymap :)
20150813 19:04:29 <xranby_> indeed, always the question how to configure such an keymap
20150813 19:04:38 <xranby_> noone wants to edit textfiles these days
20150813 19:05:01 <zubzub> normally you use xkbcommon
20150813 19:05:22 <zubzub> http://xkbcommon.org/
20150813 19:05:55 <xranby_> is that suitable to use to piggyback if you use gbm/EGL?
20150813 19:06:16 <xranby_> users who use console they use a different way to setup the keymap for the console
20150813 19:06:48 <zubzub> http://xkbcommon.org/doc/current/md_doc_quick-guide.html
20150813 19:07:14 <xranby_> http://www.linuxfromscratch.org/lfs/view/6.2/chapter07/console.html
20150813 19:07:54 <zubzub> so basically you need to get your keymap definition
20150813 19:08:07 <zubzub> and pass that to the xkbcommon lib and it will do the rest
20150813 19:08:35 <zubzub> (for gbm/console)
20150813 19:09:43 <xranby_> but we cant use the XCB common lib because we will be running without x11
20150813 19:10:14 <zubzub> *xkbcommon
20150813 19:10:16 <zubzub> not xcb :)
20150813 19:10:24 <zubzub> the name is misleading
20150813 19:10:32 <zubzub> it has nothing to do anymore with x11 or wayland or whatever
20150813 19:10:41 <zubzub> it's a stand alone library
20150813 19:10:55 <zubzub> however it *can* work with x11 or wayland or raw input
20150813 19:11:45 <zubzub> it's basically a keyboard mapping and state keeping library
20150813 19:12:10 <zubzub> so you don't have to ;)
20150813 19:12:44 <xranby_> good, because i dont want to do yet a new keymap mapper :D
20150813 19:13:14 <zubzub> and libinput is mostly a generic make-sense-of-funky-input-events library (think touch, relative touch, relative mouse movement, mouse with different dpi etc)
20150813 19:13:27 <xranby_> we will need to do a xkb eents -> newt events translator
20150813 19:13:31 <xranby_> events
20150813 19:13:53 <zubzub> yup
20150813 19:15:27 <zubzub> well the input events come from the display platform (wayland, x11, libinput), and xkb common can interprete them (and keep keyboard state) for you iirc
20150813 19:15:39 <zubzub> it's that state that you then have to translate to newt input
20150813 19:15:41 <xranby_> can kms/egl render content into a window on top of xorg ?
20150813 19:15:51 <xranby_> a screen on top of the other screen
20150813 19:16:13 <zubzub> good quetion, I assume you can do it if xorg doesn't take ownershipt of the /dev/../card*
20150813 19:16:30 <zubzub> but I don't know how it will look/will work ok
20150813 19:17:14 <zubzub> kms/gbm is basically the successor of the good ol' framebuffer api
20150813 19:17:27 <xranby_> yes.. i have had some fun testing it
20150813 19:18:13 <zubzub> the cool stuff is that you can draw with it it using egl/ogl :)
20150813 19:18:39 <zubzub> and you have a pure hw agnostic api
20150813 19:18:50 <xranby_> https://twitter.com/xranby/status/507099354113777664
20150813 19:18:54 <zubzub> so no more driver specific display structs
20150813 19:19:27 <xranby_> i have enjoyed seeing opengl content rendered by the main gpu ant put on an external usb display using kms
20150813 19:19:49 <xranby_> #JogAmp JOGL renders #OpenGL 3 out of the box using Mesa nouveau Gallium driver on external DisplayLink USB monitors
20150813 19:20:42 <xranby_> but i have not tested using mesa to render directly on the kms devices from console yet
20150813 19:20:49 <xranby_> like you do
20150813 19:20:56 <zubzub> I have tested it :p
20150813 19:20:57 <zubzub> yeah
20150813 19:21:01 <zubzub> didnt work
20150813 19:21:09 <zubzub> because newt uses eglGetDisplay
20150813 19:21:22 <zubzub> and it expected an xcb display struct
20150813 19:21:35 <zubzub> while I was passing in a gbm struct
20150813 19:21:38 <zubzub> segfault
20150813 19:21:45 <xranby_> bad mesa :P
20150813 19:21:56 <zubzub> well normally you need to cast to the correct type in C
20150813 19:21:57 <xranby_> it should have returned an error
20150813 19:22:09 <zubzub> but since we use java and a long iirc that's not really an option :p
20150813 19:22:48 <xranby_> but at the linker lever there must be two functions you can call
20150813 19:23:02 <xranby_> two "text" blocks in the shared library
20150813 19:23:20 <xranby_> one taking a xcb and one taking a gbm struct
20150813 19:24:00 <xranby_> thus we should be able to let gluegen dynamically link the two versions
20150813 19:24:44 <xranby_> since jogamp got gluegen we dont call the function directly from java
20150813 19:24:58 <xranby_> we call it from the gluegen generated jni part
20150813 19:25:06 <zubzub> I believe you :p
20150813 19:25:26 <xranby_> if c/c++ can do it then we can do it (tm)
20150813 19:29:13 <xranby_> zubzub: ok.. to get started i basically need a kernel with kms drivers (check) and mesa compiled with gbm/EGL & x11/EGL support and start testing
20150813 19:29:33 * elect (~elect@anon) has joined #jogamp
20150813 19:30:05 <elect> salut a11mad11
20150813 19:30:16 <zubzub> xranby_: correct
20150813 19:30:18 <elect> finally we met °ö°
20150813 19:30:21 <a11mad11> hi elect
20150813 19:30:33 <zubzub> you might also google a bit around to init a gbm stack
20150813 19:30:48 <a11mad11> yes i have some trouble to join this channel
20150813 19:31:04 <elect> dont worry, the worst has still to come ^^
20150813 19:31:08 <elect> tell me everything
20150813 19:31:08 <zubzub> xranby_: https://github.com/robclark/kmscube
20150813 19:31:27 <elect> at which point are you, right now?
20150813 19:31:51 <a11mad11> for no i try to learn about the pipline
20150813 19:32:01 <elect> good
20150813 19:32:19 <zubzub> xranby_: I believe the whole page flipping thing in gbm is also a bit more tricky
20150813 19:32:22 <elect> if you want to invest some time
20150813 19:33:14 <a11mad11> is it an obligation with shader to hanle the pipeline
20150813 19:33:22 <elect> yes
20150813 19:33:25 <elect> but dont worry
20150813 19:33:33 <elect> shaders are quite easy
20150813 19:33:44 <elect> you just have to learn the basic concept beyond
20150813 19:34:04 <elect> and then if you dont need anything special, your shaders could be pretty easy
20150813 19:34:12 <elect> anyway this is one of the best tutorial on opengl you can find
20150813 19:34:17 <elect> http://web.archive.org/web/20140209181347/http://www.arcsynthesis.org/gltut/
20150813 19:34:31 <elect> what you see it is an archive, since the domain expired
20150813 19:34:39 <elect> but everything is luckily still there
20150813 19:34:56 <a11mad11> thanks i will read it
20150813 19:36:38 <elect> your welcome
20150813 19:36:52 <elect> if you find those boring, you could see the video tutorial https://www.youtube.com/watch?v=JNahswHfXfw&list=PLgRQpWV
20150813 19:37:14 <elect> or if you want to learn fast, by samples without much theory, open.gl
20150813 19:37:20 * elect (~elect@anon) has left #jogamp
20150813 19:37:27 * elect (~elect@anon) has joined #jogamp
20150813 19:37:30 <a11mad11> just to know, my english is how?
20150813 19:37:43 <elect> seems ok
20150813 19:38:06 <zubzub> xranby_: the lesser cool part is if your x server uses a kms enabled driver
20150813 19:38:17 <elect> anyway, if you want to ask something in french, we have many francais here
20150813 19:38:32 <zubzub> as only a single process can claim ownership of your card device
20150813 19:38:52 <zubzub> if you want to test without having to exit X each time
20150813 19:39:00 <a11mad11> elect: I prefer reading
20150813 19:39:01 <zubzub> (and eg debug using your ide)
20150813 19:39:08 <elect> :)
20150813 19:39:21 <zubzub> xranby_: luckly there's still qemu with a special qemu kms driver that you can use
20150813 19:39:31 <zubzub> which should work in theory
20150813 19:39:35 <xranby_> zubzub: i uses raspberry pi's to test on and develop on a remote machine
20150813 19:39:48 <zubzub> rpi 1 doesnt have kms iirc, 2 does(?)
20150813 19:39:49 <xranby_> thus i do not need to run x11 on the same device
20150813 19:39:54 <xranby_> it does!
20150813 19:40:01 <xranby_> using the new driver
20150813 19:40:03 <zubzub> rpi 1?
20150813 19:40:05 <zubzub> ah
20150813 19:40:12 <zubzub> ok that explains :p
20150813 19:40:32 <xranby_> eric anholts work work with both the pi 1 and pi 2
20150813 19:40:41 <xranby_> you actually get an opengl 2 driver
20150813 19:40:57 <xranby_> thus you can run more application compared to using the broadcom binary blob
20150813 19:44:44 * elect (~elect@anon) Quit (Remote host closed the connection)
20150813 19:45:02 <xranby_> Getting Eric Anholt's Mesa3D VC4 driver running on a Raspberry Pi is easily done thanks to the work by gohai.
20150813 19:45:02 <xranby_> gohai started out roughly following Eric's notes here: http://dri.freedesktop.org/wiki/VC4/
20150813 19:45:02 <xranby_> And then put together a buildbot in Python, to produce system images for use on the Pi or Pi2.
20150813 19:45:10 <xranby_> Kernel, Mesa, XServer packages & dependencies are from git. System image produced using gohai's buildbot
20150813 19:45:10 <xranby_> https://github.com/gohai/vc4-buildbot
20150813 19:45:17 <xranby_> gohai publish daily builds of raspbian patched to the Mesa3D vc4 driver using his bot at:
20150813 19:45:17 <xranby_> http://sukzessiv.net/~gohai/vc4-buildbot/build/
20150813 19:45:22 <xranby_> using gohai's rasbian image i can start run quite a lot of OpenGL 2 content
20150813 19:45:22 <xranby_> https://www.raspberrypi.org/forums/viewtopic.php?f=81&t=115350
20150813 19:45:22 <xranby_> https://www.raspberrypi.org/forums/viewtopic.php?p=787240#p787240
20150813 19:46:22 <xranby_> gohai does all this as part of his google summer of code assignment to get processing 3 running on the raspberry pi
20150813 19:46:32 <xranby_> and other arm devices
20150813 19:49:56 <zubzub> nice
20150813 19:51:40 <xranby_> replicating his build bot is easy since he uses a regular pi as the bot
20150813 19:52:43 <zubzub> once I have touch & kms implemented for my java wayland compositor, I'll try to run it on my rpi with touch display :p
20150813 19:52:52 <xranby_> nice
20150813 19:53:38 <zubzub> so far I only have it running on X with mouse & keyboard input (no modifier key support yet)
20150813 20:17:04 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20150813 20:46:11 <xranby_> https://www.youtube.com/watch?v=cQoQE_HDG8g - The real story behind Wayland and X
20150813 20:47:25 <zubzub> daniel stone is one of the main authors of xkbcommon and libinput
20150813 20:47:35 <zubzub> also a large contributor of libwayland
20150813 20:49:20 <xranby_> i have to see this video to understand the ideology behind the changes and goals to be archived
20150813 20:49:26 <zubzub> yup
20150813 20:49:32 <zubzub> it's a very good presentation
20150813 21:01:56 * Eclesia (~eclesia@anon) has joined #jogamp
20150813 21:05:18 * a11mad11 (c0e2cd65@anon) Quit (Ping timeout: 246 seconds)
20150813 21:11:14 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20150813 21:19:07 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150813 22:15:22 <xranby_> Christian Thalinger (who previously has been working on the cacao jvm) now reveals the plans that java will get an AOT compiler https://www.youtube.com/watch?feature=player_embedded&v=Xybzyv8qbOc - JVMLS 2015 - Java Goes AOT
20150813 22:34:46 <xranby_> hmm so it is an AOT compiler for static functions.. only... ?
20150813 22:46:25 <xranby_> the news is the announcement
20150813 22:46:36 <xranby_> have to test it myself to measure bootup times
20150814 02:26:50 * badshah400 (~badshah40@anon) has joined #jogamp
20150814 05:05:29 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20150814050529.html