#jogamp @ irc.freenode.net - 20150408 05:05:30 (UTC)

20150408 05:05:30 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20150407050530.html
20150408 05:05:30 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20150408050530.html
20150408 05:27:01 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150408 05:46:13 * badshah400 (~badshah40@anon) has joined #jogamp
20150408 06:15:17 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150408 06:24:17 * kles4enko (~kles4enko@anon) has joined #jogamp
20150408 06:24:41 * zzuegg_ (~zzuegg___@anon) has joined #jogamp
20150408 06:27:05 * zzuegg__ (~zzuegg___@anon) Quit (Ping timeout: 246 seconds)
20150408 06:46:56 * monsieur_max (~maxime@anon) has joined #jogamp
20150408 07:06:44 * kles4enko_ (~kles4enko@anon) has joined #jogamp
20150408 07:07:01 * kles4enko (~kles4enko@anon) Quit (Ping timeout: 264 seconds)
20150408 07:27:45 * eclesia (~husky@anon) has joined #jogamp
20150408 07:28:02 <eclesia> good morning
20150408 07:38:55 * crys|s (~crys|s@anon) has joined #jogamp
20150408 07:55:46 <sgothel> goody
20150408 07:56:09 * crys|s (~crys|s@anon) Quit (Quit: Leaving)
20150408 07:56:41 <monsieur_max> hi there
20150408 08:02:47 <monsieur_max> http://stackoverflow.com/research/developer-survey-2015
20150408 08:11:05 <monsieur_max> nice stats there
20150408 08:11:27 <zubzub> idd
20150408 08:11:30 <zubzub> most dreaded
20150408 08:11:43 <zubzub> I'm happy that sharepoint is in there
20150408 08:17:42 <monsieur_max> :) i'm currently working on a office add-in for word, not quite my dream project ...
20150408 08:18:44 <zubzub> I'd probably have killed myself many times over if I had to do that
20150408 08:19:04 <bruce-> aye
20150408 08:20:35 <monsieur_max> zubzub: well, at least, it's a bit different of Java, new thing to discover...ah, who am i fooling ...
20150408 08:37:30 <sgothel> median .. they always tend to ignore the median, just publishing avrg.
20150408 08:43:14 * kles4enko_ (~kles4enko@anon) Quit (Ping timeout: 252 seconds)
20150408 08:51:19 <zubzub> and even avg is useless without deviation
20150408 09:21:42 * xranby_GTX580 (~familjen@anon) has joined #jogamp
20150408 09:23:19 <xranby_GTX580> Pixar releases a classic library of textures (w/ bump/normal maps too) under CC 850mb of texture data: https://community.renderman.pixar.com/article/114/library-pixar-one-twenty-eight.html license: http://creativecommons.org/licenses/by/4.0/
20150408 09:24:29 <xranby_GTX580> thats very usefull assets!
20150408 09:26:01 <xranby_GTX580> i have recived an ubuntu phone. writing qml apps was easy, still trying to figure out how to deploy java applications, the phone uses a new packageformat
20150408 09:27:16 <monsieur_max> xranby_GTX580: oh nice :)
20150408 09:27:29 <xranby_GTX580> i had hoped it was as easy as running apt-get install openjdk but for some reason the apt respository on the phone is locked and instead they have this new click package installer
20150408 09:27:43 <xranby_GTX580> so it is ubuntu + some new deployment layer
20150408 09:29:02 <monsieur_max> is it really a ubuntu distro running ?
20150408 09:29:40 <xranby_GTX580> tes
20150408 09:29:42 <xranby_GTX580> yes
20150408 09:29:51 <zubzub> does it use mir?
20150408 09:30:00 <xranby_GTX580> yes
20150408 09:30:11 <xranby_GTX580> it have a terminal
20150408 09:30:13 <xranby_GTX580> as well
20150408 09:30:22 <zubzub> running java ui's wont work then
20150408 09:30:22 <xranby_GTX580> so you can run vi etc
20150408 09:30:36 <zubzub> newt afaik does not support mir
20150408 09:30:47 <xranby_GTX580> zubzub: there is some opengl es test applications that i can launch from the terminal
20150408 09:30:56 <zubzub> written in java?
20150408 09:30:57 <xranby_GTX580> so i am optimistic that by using opengl
20150408 09:31:02 <xranby_GTX580> i can get a window
20150408 09:31:03 <zubzub> yes you can use opengl
20150408 09:31:10 <xranby_GTX580> so jogamp will work
20150408 09:31:15 <zubzub> yup
20150408 09:31:19 <xranby_GTX580> awt not so much
20150408 09:31:20 <zubzub> jogamp works on android right?
20150408 09:31:28 <zubzub> then it will work on ubuntu phone
20150408 09:31:29 <xranby_GTX580> yes jogamp work on android
20150408 09:31:38 <zubzub> the thing is ubuntu phone has mir display server
20150408 09:31:48 <zubzub> and newt needs to have support for that display server
20150408 09:32:04 <zubzub> or it will not be able to do opengl shizzle
20150408 09:32:22 <zubzub> it is possible?
20150408 09:32:22 <zubzub> yes
20150408 09:32:25 <xranby_GTX580> i will investigate how much customization is needed
20150408 09:32:35 <xranby_GTX580> to newt
20150408 09:32:39 <xranby_GTX580> to support this os
20150408 09:32:42 <zubzub> but you will need to add some client platform code to newt
20150408 09:33:14 <zubzub> while you're add it, can you add wayland support as well?
20150408 09:33:16 <zubzub> :p
20150408 09:33:19 <xranby_GTX580> sgothel helped out to add the raspberry pi newt client code
20150408 09:33:32 <monsieur_max> zubzub: :)
20150408 09:33:33 <zubzub> you can use my java wayland bindings if you want
20150408 09:33:41 <zubzub> simply replace all the jna code with jni
20150408 09:33:50 <zubzub> and just use the client side code
20150408 09:34:05 <zubzub> I used jna direct mapping so the jna functions map 1to1 to native function
20150408 09:34:07 <xranby_GTX580> this mir is hopefully only require some heuristics on what newt client backend to activate
20150408 09:34:18 <xranby_GTX580> maybe the android code work as is
20150408 09:34:36 <xranby_GTX580> the phone have the android drivers
20150408 09:35:01 <zubzub> yeah and it uses libhybris iirc
20150408 09:35:14 <zubzub> to translate android egl to common sense egl
20150408 09:35:23 <xranby_GTX580> sven have a sailfish os device
20150408 09:35:33 <xranby_GTX580> so maybe we can bring that up as well
20150408 09:35:36 <zubzub> cool
20150408 09:35:45 <zubzub> my java wayland bindings also has a java test client
20150408 09:35:56 <zubzub> you should be able to run it as-is
20150408 09:36:02 <zubzub> on the sailfish phone
20150408 09:36:20 <xranby_GTX580> we need to compile a jre for sailfish os
20150408 09:36:33 <xranby_GTX580> or look into other deployment options
20150408 09:36:37 <zubzub> shouldn't be too hard
20150408 09:36:39 <xranby_GTX580> like ahead of time compilation
20150408 09:36:52 <zubzub> I assume it's arm hf
20150408 09:37:05 <zubzub> you can use the oracle jdk for testing purposes
20150408 09:37:20 <zubzub> sailfish is also pretty standard gnu/linux I assume?
20150408 09:37:26 <xranby_GTX580> glibc library versioning + libffi can drive you nuts sometimes
20150408 09:37:54 <xranby_GTX580> have prevented me from migrating jre's into some embedded distributions
20150408 09:38:06 <zubzub> yeah those things can be a PITA
20150408 09:38:09 <xranby_GTX580> instead i had to recompile the jre
20150408 09:38:13 <xranby_GTX580> for the target
20150408 09:40:20 <zubzub> it also sucks that there is no good jit for arm for java
20150408 09:40:58 <zubzub> unless you use oracle's hotspot
20150408 09:42:37 <xranby_GTX580> zubzub: i owuld not be that hard on the free implementations
20150408 09:42:47 <xranby_GTX580> i think they are actually pretty good
20150408 09:42:59 <zubzub> they're good enough :)
20150408 09:43:06 <xranby_GTX580> people should use them in real world cenarios
20150408 09:43:08 <zubzub> so yeah, they're not bad :)
20150408 09:44:07 <xranby_GTX580> distributions are unfortunally slow on updating the free jits after they have been improved
20150408 09:44:33 <xranby_GTX580> hence it would be awesome to have a buildfarm providing builds that people can use
20150408 09:44:41 <xranby_GTX580> of the latest versions
20150408 09:44:45 <xranby_GTX580> for example
20150408 09:45:21 <xranby_GTX580> cacao vm .. distributions still ship the "stable" version based on the now old 2009 c codebase
20150408 09:45:35 <xranby_GTX580> the jvm has been rewritten in c++
20150408 09:45:45 <xranby_GTX580> so if you compile the tip is a remastered jit
20150408 09:46:38 <xranby_GTX580> having a build farm that provides build using this jit + some headless classes rock on single core arm systems
20150408 09:47:17 <xranby_GTX580> the thumb2 jit found in icedtea have recently also been updated to provide a jit on all 32bit arm platforms
20150408 09:47:47 <xranby_GTX580> thus performance is much improved on old embedded devices such as the old raspberry pi
20150408 09:48:01 <xranby_GTX580> the new raspberry pi that got thumb2 have a fast jit out of the box
20150408 09:48:25 <xranby_GTX580> jamvm have been updated to support openjdk 9
20150408 09:48:51 <xranby_GTX580> then we have robovm
20150408 09:48:54 <xranby_GTX580> and avian jvm
20150408 09:49:15 <xranby_GTX580> the latter two provide some interesting deployment options
20150408 09:49:56 <xranby_GTX580> gcj static compilation is also pretty awesome if you java app can be compiled into native code
20150408 09:50:26 <zubzub> last thing I headr (long time ago) was that gcj was pretty bad/slow
20150408 09:50:36 <xranby_GTX580> the interpreter is quite slow
20150408 09:50:50 <xranby_GTX580> but the compiler is producing good quality code
20150408 09:52:06 <xranby_GTX580> i created this chart some years ago on the old raspberry pi 1 https://docs.google.com/a/gudinna.com/spreadsheet/oimg?key=0ArR4FJuzvXEsdDZkek5QcXd3c2ROMm1IajE1WmFwYkE&oid=3&zx=owyi7culzhlp
20150408 09:52:20 <xranby_GTX580> GCJ outperformed the oracle jvm
20150408 09:52:50 <xranby_GTX580> at least on this numberchrunch benchmark
20150408 09:53:13 <xranby_GTX580> GCJ static ahead of time compilation is the purple bars
20150408 09:54:02 <xranby_GTX580> avian and cacao also gave impressive results
20150408 09:54:28 <xranby_GTX580> slightly below the oracle jit
20150408 09:55:39 <xranby_GTX580> so you can see gcj interpreter is one of the slowest while gcj compiled is the fastest
20150408 10:04:35 <zubzub> ic
20150408 10:04:37 <zubzub> interesting
20150408 10:10:08 <zubzub> how does it use ordinary bytecode java?
20150408 10:13:55 * jvanek (jvanek@anon) has joined #jogamp
20150408 10:14:20 <xranby_GTX580> zubzub: gcj treat bytecode as some sort of sourcefile
20150408 10:15:03 <xranby_GTX580> it tries to resolve all functions at compiletime thus all bytecode must be known at compiletime
20150408 10:15:03 <zubzub> so it compiles it on the fly?
20150408 10:15:17 <zubzub> ah
20150408 10:15:34 <zubzub> so you can not run a program with a previously unknown plugin say?
20150408 10:15:38 <xranby_GTX580> the end result is a native executable
20150408 10:15:48 <xranby_GTX580> just like c and c++
20150408 10:16:07 <zubzub> I wonder how it treats things like jna
20150408 10:17:50 <xranby_GTX580> there is a mechanism that is kind of experimental -findirect-dispatch
20150408 10:18:20 <xranby_GTX580> example
20150408 10:18:23 <xranby_GTX580> gcj -findirect-dispatch craftbukkit.jar -o craftbukkit --main=org.bukkit.craftbukkit.Main
20150408 10:18:35 <xranby_GTX580> will try copmile the minecraft bukkit server
20150408 10:19:07 <xranby_GTX580> and allow the compilation to succeed despite the jar containing references to external jars, like ant and lucene
20150408 10:19:29 <zubzub> gcj seems to support just 1.5?
20150408 10:19:33 <zubzub> https://gcc.gnu.org/java/
20150408 10:19:35 <xranby_GTX580> yes
20150408 10:19:44 <xranby_GTX580> thats one of the more serve limitations
20150408 10:19:49 <xranby_GTX580> so it is fast
20150408 10:20:05 <rmk0> i feel like it'd be a bad idea to depend on gcj for anything... have bad experiences with "yeah i promise we're really maintaining this" bits of gcc
20150408 10:20:05 <xranby_GTX580> but noone updates it to stay in sync with openjdk
20150408 10:20:20 <xranby_GTX580> gcj is mostly used to compile the bootstrap compiler for openjdk
20150408 10:20:43 <xranby_GTX580> which is awesome when someone wants to port openjdk to a new platform
20150408 10:21:19 <zubzub> yeah
20150408 10:21:29 <zubzub> would be really nice to have a bytecode to native compiler
20150408 10:21:39 <xranby_GTX580> zubzub: you can try robovm
20150408 10:21:53 <xranby_GTX580> it uses llvm to produce native code
20150408 10:21:58 <zubzub> oh nice
20150408 10:22:01 <xranby_GTX580> and uses more up to date classes
20150408 10:22:09 <zubzub> but first I'm going to get some food
20150408 10:22:13 <xranby_GTX580> robovm is used to get a lot of java applications into ios
20150408 10:22:55 <xranby_GTX580> https://github.com/robovm/robovm
20150408 10:23:22 <xranby_GTX580> you can compile it yourself
20150408 10:25:05 <bruce-> does that work with jogl?
20150408 10:25:39 <xranby_GTX580> bruce-: all jogl bugs i know of is fixed by now
20150408 10:25:48 <bruce-> interesting
20150408 10:27:42 <xranby_GTX580> bruce-: you have to tell robovm what classes uses reflection inside jogamp like this: https://github.com/robovm/robovm/issues/534#issuecomment-58342768
20150408 10:28:05 <xranby_GTX580> -forcelinkclasses jogamp.**:com.jogamp.**:javax.media.**
20150408 10:28:36 <xranby_GTX580> else robovm will forget to compile the jogl classes
20150408 10:28:45 <bruce-> and does that work on IOS devices too?
20150408 10:29:06 <xranby_GTX580> i dont think so because we have not added ios client support
20150408 10:29:09 <xranby_GTX580> inot newt
20150408 10:29:13 <xranby_GTX580> into
20150408 10:29:17 <bruce-> aah
20150408 10:29:56 <xranby_GTX580> some code is needed in newt to initialize opengl and input etc
20150408 10:30:49 <xranby_GTX580> bruce-: i ran a series of experiments to use robovm on embedded gnu linux systems
20150408 10:30:59 <xranby_GTX580> only tested i386 sucessfully
20150408 10:31:10 <xranby_GTX580> i need to recompile llvm to support arm and x86_64
20150408 10:31:49 <xranby_GTX580> (the llvm provided by robovm only support ios (different calling convention compared to gnu/linux arm) and i386
20150408 10:32:22 * gouessej (5ee4b442@anon) has joined #jogamp
20150408 10:32:26 <gouessej> Hi
20150408 10:32:30 <xranby_GTX580> gouessej: hello!
20150408 10:32:33 <rmk0> lo
20150408 10:32:58 <gouessej> xranby_GTX580: I updated LibGDX a few days ago
20150408 10:33:04 <xranby_GTX580> gouessej: !!!
20150408 10:33:21 <gouessej> xranby_GTX580: Let me know which of your changes should be kept
20150408 10:33:51 <gouessej> JoglGL30.java is a bit incomplete
20150408 10:33:52 <xranby_GTX580> thank you, i think we should "prune" the history old binarys before we submit the backend to be merged
20150408 10:33:59 <gouessej> JoglGL20 should work
20150408 10:34:16 <xranby_GTX580> git keeps deleted files in the history
20150408 10:34:30 <xranby_GTX580> so we may need to greate a git branch with only the latest tip
20150408 10:34:36 <xranby_GTX580> and remove some of our history
20150408 10:34:39 <gouessej> I know that our backend will be refused for sure
20150408 10:34:45 <xranby_GTX580> to favour a clean merge
20150408 10:34:54 <gouessej> but we should keep it as clean as possible
20150408 10:35:19 <xranby_GTX580> yes, in our case libgdx maintainers will be unable to merge if they cant understand all commits
20150408 10:35:42 <xranby_GTX580> so a clean history may be important
20150408 10:35:55 <gouessej> No, they won't merge even though they understand our stuff
20150408 10:36:10 <gouessej> They don't want to maintain one more desktop backend
20150408 10:36:21 <gouessej> They refused to merge a second iOS backend
20150408 10:36:31 <xranby_GTX580> we should pitch it as an embedded backend
20150408 10:36:45 <xranby_GTX580> we add support for platforms that their current backend do not support
20150408 10:37:01 <gouessej> We should put it into our Maven repository. I agree with you
20150408 10:38:02 <gouessej> In my humble opinion, we have to be conformant and isofunctional (first)
20150408 10:38:11 <xranby_GTX580> for best impact we can make some games easy to run on the new raspberry pi
20150408 10:38:33 <gouessej> We can use their demos instead of writing another game from scratch
20150408 10:38:34 <xranby_GTX580> having a working demonstration makes a clear case
20150408 10:38:42 <xranby_GTX580> yes we pacakge their demo games
20150408 10:38:54 <gouessej> They moved them into separate subprojects
20150408 10:38:56 <xranby_GTX580> and foss libgdx games
20150408 10:39:24 <xranby_GTX580> maybe publish some of the games in the "raspberry pi store"
20150408 10:39:28 <xranby_GTX580> for example
20150408 10:39:31 <gouessej> :)
20150408 10:39:34 <gouessej> Excellent idea
20150408 10:40:00 <xranby_GTX580> the new raspberrypi 2 with a quad core cpu is really something
20150408 10:40:12 <xranby_GTX580> its a fast little single board computer
20150408 10:40:38 <gouessej> I plan to buy one to make a small server
20150408 10:40:52 <xranby_GTX580> yes get the new one with the armv7
20150408 10:40:58 <xranby_GTX580> instructionset
20150408 10:41:21 <gouessej> What is missing in our backend to be isofunctional? the natives?
20150408 10:41:49 <xranby_GTX580> the "small details"
20150408 10:42:09 <xranby_GTX580> like some interesting graphics effects occour if you run the mouse outside the screen
20150408 10:42:31 <xranby_GTX580> maybe we should pick input from Xorg if X is running
20150408 10:43:04 <xranby_GTX580> maybe overlap the opengl screen to track an empty xorg window
20150408 10:43:36 <xranby_GTX580> thus things can get better
20150408 10:43:44 <gouessej> Ok, it will require Newt Input API and I have still some work to do on the HiDpi in JMonkeyEngine and JogAmp's Ardor3D Continuation
20150408 10:44:14 <xranby_GTX580> when you run a jogmap application from console then we have some newt imput from the linux kernel directly
20150408 10:44:29 <xranby_GTX580> the raspberry pi do not use xorg for opengl
20150408 10:44:40 <gouessej> I remember that there is a Linux tracker
20150408 10:44:47 <xranby_GTX580> yes its good for console
20150408 10:44:56 <xranby_GTX580> can be confusing if you run xorg at the same time
20150408 10:45:03 <xranby_GTX580> (dual mousepointers)
20150408 10:45:22 <gouessej> What other small details are missing?
20150408 10:45:34 <xranby_GTX580> awt integration
20150408 10:45:46 <xranby_GTX580> (thats a large detail)
20150408 10:45:52 <xranby_GTX580> focus on newt
20150408 10:46:02 <xranby_GTX580> openal is working last time i checked
20150408 10:46:08 <xranby_GTX580> opengl es is working
20150408 10:46:16 <xranby_GTX580> basically things work
20150408 10:46:19 <xranby_GTX580> games run
20150408 10:46:24 <xranby_GTX580> and the experience is good
20150408 10:46:28 <xranby_GTX580> at least from console
20150408 10:46:51 <xranby_GTX580> ... the keyboard input from console do not support any internationalization
20150408 10:47:29 <xranby_GTX580> the question is how to support something like that without reimplementing xorg
20150408 10:47:30 <gouessej> ok
20150408 10:48:04 <xranby_GTX580> the keyboard do not know what is printed on the keys
20150408 10:48:17 <gouessej> I assume that we just need to provide a JAR with the native libraries matching with LibGDX convention
20150408 10:48:31 <xranby_GTX580> its only the os that can provide keymaps for all zillions of keymaps
20150408 10:48:46 <xranby_GTX580> yes
20150408 10:48:58 <gouessej> it's trivial
20150408 10:49:06 <gouessej> I can do it
20150408 10:49:22 <xranby_GTX580> we are set to deploy some games :)
20150408 10:49:46 <gouessej> Let me know what is buggy and I'll do my best.
20150408 10:50:03 <gouessej> What is the package type on an Ubuntu Phone?
20150408 10:50:07 <xranby_GTX580> click
20150408 10:50:22 <xranby_GTX580> https://developer.ubuntu.com/en/publish/packaging-click-apps/
20150408 10:51:32 <xranby_GTX580> the click apps contains a manifest with a declared security policy for each app https://developer.ubuntu.com/en/publish/security-policy-groups/
20150408 10:52:40 <xranby_GTX580> AppArmour enforces the policys
20150408 10:52:47 <bruce-> TextureIO.newTextureData does not seem to support basic auth for urls like http://user:pass@host/image.png
20150408 10:54:29 <gouessej> Is it possible to install an application not from the Software Centre, for example by downloading a package on a website?
20150408 10:55:45 <xranby_GTX580> gouessej: yes, but you have to enable "developer mode"
20150408 10:56:03 <gouessej> It's unfair
20150408 10:56:11 <gouessej> Yes another centralized crap
20150408 10:56:13 <xranby_GTX580> about the phone -> enable develoepr mode
20150408 10:56:36 <gouessej> Does it still support DEB packages?
20150408 10:57:31 <xranby_GTX580> technically yes, all the core os applications are installed using debs
20150408 10:57:49 <gouessej> Ok I'll stick to DEB then
20150408 10:58:01 <xranby_GTX580> but the end user do not have an easy way to install an deb :/
20150408 10:58:08 <xranby_GTX580> the os update process uses debs
20150408 10:58:13 <xranby_GTX580> you can go to the terminal
20150408 10:58:18 <xranby_GTX580> but that is kind of clunky
20150408 10:58:47 <xranby_GTX580> i guess people will create rouge app stores
20150408 10:59:35 <xranby_GTX580> so there is the official way creat click app -> submit it for review crap
20150408 11:00:05 <xranby_GTX580> developers have an easy way to test apps
20150408 11:00:13 <xranby_GTX580> plug in an usb cable
20150408 11:00:19 <xranby_GTX580> click run from the sdk
20150408 11:00:24 <xranby_GTX580> and it runs on the phone
20150408 11:03:30 <gouessej> I don't want to use a store
20150408 11:03:48 <gouessej> direct relationship from creators to end users
20150408 11:03:54 <gouessej> no censorship
20150408 11:04:31 <gouessej> click packages can't be very big anyway
20150408 11:04:42 <gouessej> it's not interesting for me
20150408 11:04:54 <xranby_GTX580> Be less than 3GB in size
20150408 11:05:03 <xranby_GTX580> do you need more than 3GB?
20150408 11:05:50 <xranby_GTX580> thats the only size limitation i have seen for click packags
20150408 11:05:51 <gouessej> Not now but in the future yes
20150408 11:06:10 <xranby_GTX580> nods
20150408 11:06:20 <gouessej> Does OpenJDK work well on it?
20150408 11:07:57 <xranby_GTX580> ringt now not at all, becaus I have issues using apt to install openjdk
20150408 11:08:04 <xranby_GTX580> the apt cache is locked
20150408 11:08:13 <gouessej> :s
20150408 11:08:15 <xranby_GTX580> read only system
20150408 11:08:37 <xranby_GTX580> i have root pn the phone so .. i guess it can be solved somehow
20150408 11:08:50 <xranby_GTX580> but i have not found nay easy way to install openjdk
20150408 11:09:53 <xranby_GTX580> mount / -o remount,rw
20150408 11:10:31 <gouessej> ok I hope it's possible. If Ubuntu Phone isn't open enough, I won't try to support it
20150408 11:10:40 <xranby_GTX580> ok installing openjdk now
20150408 11:10:53 <xranby_GTX580> making the root rw made apt happy
20150408 11:11:07 <xranby_GTX580> running apt-get install openjdk-7-jre now
20150408 11:12:05 <xranby_GTX580> i guess the easiest thing is to bundle openjdk with the click app for now
20150408 11:12:06 <gouessej> :)
20150408 11:12:28 <gouessej> I bundle OpenJDK with the DEB but nothing for ARM right now
20150408 11:14:03 <xranby_GTX580> \o/ java -version work ;)
20150408 11:14:27 <xranby_GTX580> uses the armhf ubuntu openjdk
20150408 11:16:18 <zubzub> try running swing ;)
20150408 11:16:22 <zubzub> maybe it uses xmir :p
20150408 11:18:27 <xranby_GTX580> i think this is a "first" https://twitter.com/xranby/status/585763496396267520/photo/1
20150408 11:19:27 <gouessej> :)
20150408 11:19:47 <gouessej> Good job
20150408 11:20:17 <xranby_GTX580> good to put some pressure on them :)
20150408 11:20:23 <xranby_GTX580> demonstrating what developers want
20150408 11:20:52 <xranby_GTX580> i think canonical has to listen to these early developer reactions
20150408 11:21:05 <xranby_GTX580> there is only ~2000 phones out there
20150408 11:21:15 <xranby_GTX580> thus one developers voice is large
20150408 11:21:19 <zubzub> yup
20150408 11:21:31 <xranby_GTX580> because only 20 or so are speaking about it
20150408 11:22:34 <zubzub> if they want to pay me, I can probably make some mir java bindings :p
20150408 11:22:50 <zubzub> and have javafx run on ubuphone
20150408 11:23:13 <xranby_GTX580> i will not do that port
20150408 11:23:17 <gouessej> Have you tried to run JOGL on it?
20150408 11:23:28 <xranby_GTX580> gouessej: .. soon i have wget
20150408 11:23:51 <xranby_GTX580> do doy have some nice prepacakged game?
20150408 11:24:02 <xranby_GTX580> that i can wget untar and java run?
20150408 11:24:09 <xranby_GTX580> do you
20150408 11:24:14 <gouessej> The ARM OPenJDK is missing from the DEB package of TUER
20150408 11:24:27 <gouessej> I can fix it very soon
20150408 11:24:28 <xranby_GTX580> i already have openjdk on the phone :)
20150408 11:24:48 <xranby_GTX580> so i only need the jogl native armv7 jar
20150408 11:24:58 <xranby_GTX580> lets see.. sing
20150408 11:25:00 <xranby_GTX580> swing
20150408 11:25:02 <xranby_GTX580> hmm
20150408 11:25:08 <zubzub> wont work unless xmir works
20150408 11:25:25 <gouessej> JogAmp's Ardor3D Continuation doesn't work very well with OpenGL ES yet because of GLU
20150408 11:26:05 <xranby_GTX580> i need a bluetooth keyboard or something
20150408 11:26:14 <xranby_GTX580> typing using a touchscreen is quite something
20150408 11:26:23 <zubzub> use ssh
20150408 11:28:49 <gouessej> This bug is in my way :( https://jogamp.org/bugzilla/show_bug.cgi?id=1045
20150408 11:31:23 <xranby_GTX580> is GLU fixed function only?
20150408 11:32:34 <gouessej> It's written to work with GL2
20150408 11:38:25 <gouessej> Currently, if you pass a GL ES 2 instance, it creates a very limited GLU object
20150408 11:38:33 <gouessej> unable to create any mipmap
20150408 11:42:37 * badshah400 (~badshah40@anon) has joined #jogamp
20150408 11:45:14 * badshah400 (~badshah40@anon) Quit (Client Quit)
20150408 11:47:14 <xranby_GTX580> interesting
20150408 11:47:21 <xranby_GTX580> it appers the phone uses this repository http://derived.archive.canonical.com/ubuntu-rtm/
20150408 11:47:29 <xranby_GTX580> its like only the core ubuntu developement tools
20150408 11:47:53 <xranby_GTX580> so.. that the jdk was included was a good surprise
20150408 11:48:55 <xranby_GTX580> i can probably make my phone a mess by adding the "normal" ubuntu repository
20150408 11:53:11 <zubzub> probably
20150408 11:53:19 <zubzub> normal repos assume an X server
20150408 11:53:30 <zubzub> and non phone libs and whatnot
20150408 11:59:59 * badshah400 (~badshah40@anon) has joined #jogamp
20150408 12:03:29 <monsieur_max> xranby_GTX580: nice, really interesting phone suddenly ;)
20150408 12:04:55 <xranby_GTX580> zubzub: i have now added the normal ubuntu repository with a lower apt pinning
20150408 12:05:14 <xranby_GTX580> lets see... apt-get install eclips
20150408 12:05:26 <monsieur_max> haha :)
20150408 12:06:29 <zubzub> I hope you have a way to reset the phone :)
20150408 12:06:44 <xranby_GTX580> zubzub: we will see
20150408 12:06:54 <monsieur_max> "yolo"
20150408 12:07:05 <zubzub> yopo
20150408 12:07:06 <xranby_GTX580> fetching 169mb of archives...
20150408 12:07:11 <zubzub> eclipse wont work
20150408 12:07:16 <zubzub> swt doesnt support mir afaik
20150408 12:07:17 <xranby_GTX580> ssh -X may work
20150408 12:07:21 <zubzub> it does have wayland support
20150408 12:07:22 <zubzub> yeah ok
20150408 12:07:25 <zubzub> that's cheating :p
20150408 12:07:29 <xranby_GTX580> hehe
20150408 12:07:56 <xranby_GTX580> how about netbeans or hmm robotwars
20150408 12:08:00 <zubzub> nope
20150408 12:08:03 <zubzub> netbeans is swing
20150408 12:08:08 <xranby_GTX580> your stuff then?
20150408 12:08:13 <zubzub> nope
20150408 12:08:18 <zubzub> my stuff is wayland :)
20150408 12:08:28 <zubzub> what you need are java mir client side bindings
20150408 12:08:35 <zubzub> or get xmir working
20150408 12:11:03 <zubzub> what qt version does it use?
20150408 12:11:06 <zubzub> I assume 5.x?
20150408 12:11:06 <xranby_GTX580> phablet@ubuntu-phablet:~$ apt-cache search xmir
20150408 12:11:07 <xranby_GTX580> ubuntu-desktop-mir - System compositor for Ubuntu Desktop
20150408 12:11:07 <xranby_GTX580> xserver-xorg-xmir - Xorg - the X.Org X server (module for running nested in Mir)
20150408 12:11:14 <zubzub> yup
20150408 12:11:17 <zubzub> try that
20150408 12:11:40 <xranby_GTX580> 5.3
20150408 12:12:04 <zubzub> if you can get xmir started on your phone, alle basic X programs should work
20150408 12:12:26 <zubzub> but given the clusterf... eerrr complexity of X, your milage may vary
20150408 12:12:31 <xranby_GTX580> installing
20150408 12:12:47 <xranby_GTX580> xmir pulled in libfontenc1 libpciaccess0 libxaw7 libxfont1 libxkbfile1 libxmu6 libxpm4
20150408 12:12:47 <xranby_GTX580> libxt6 x11-xkb-utils xserver-common xserver-xorg-core xserver-xorg-xmir
20150408 12:12:58 <zubzub> look sok
20150408 12:13:57 <xranby_GTX580> had to disable ipv6 on the phone
20150408 12:14:05 <xranby_GTX580> else it tried to fetch packages using it
20150408 12:14:25 <xranby_GTX580> sad that ipv6 dont work everywhere
20150408 12:16:43 <xranby_GTX580> unpacking... eclipse ant and friends
20150408 12:17:04 <zubzub> you can try a less heavy X client as well
20150408 12:17:31 <xranby_GTX580> ok... f*** (no space left on device... this was a cheap phone hmm
20150408 12:17:45 <zubzub> try a less heavy X client
20150408 12:17:47 <zubzub> xev or something
20150408 12:17:54 <zubzub> if that works then X works
20150408 12:18:23 <xranby_GTX580> all apt packages unpacked into the small 2gb root partition
20150408 12:18:29 <xranby_GTX580> hmm
20150408 12:18:31 <zubzub> :)
20150408 12:22:34 <xranby_GTX580> cleaning up /var/apt/cache always help
20150408 12:23:46 * badshah400 (~badshah40@anon) Quit (Quit: badshah400)
20150408 12:24:11 <xranby_GTX580> 6.4mb left on / eclipse is installed \o/
20150408 12:24:48 <monsieur_max> xranby_GTX580: which device did you buy ?
20150408 12:25:20 <xranby_GTX580> monsieur_max: the bq aquaris e4.5 ubuntu edition
20150408 12:25:42 <xranby_GTX580> http://www.bq.com/gb/ubuntu.html
20150408 12:26:40 <xranby_GTX580> its the only ubuntu phone that you can buy preinstalled atm
20150408 12:26:42 <monsieur_max> ah you still have a µSD port
20150408 12:27:08 <xranby_GTX580> yes
20150408 12:28:17 <xranby_GTX580> org.eclipse.swt.SWTError: No more handles [gtk_init_check() failed]
20150408 12:28:17 <xranby_GTX580> at org.eclipse.swt.SWT.error(SWT.java:4387)
20150408 12:28:18 <xranby_GTX580> hmm
20150408 12:29:34 <xranby_GTX580> no idea how to start the xmir thing
20150408 12:29:42 <xranby_GTX580> will try reboot
20150408 12:31:41 <zubzub> you can probably start it manually
20150408 12:31:44 <zubzub> as a stand alone app
20150408 12:32:04 <zubzub> just like you would run X on windows
20150408 12:32:06 <zubzub> or X on wayland
20150408 12:32:08 <zubzub> :)
20150408 12:32:22 <zubzub> (or X on mac)
20150408 12:32:28 <zubzub> simply
20150408 12:32:28 <zubzub> X
20150408 12:32:31 <zubzub> might work
20150408 12:32:52 <zubzub> unles mir has an X socket open and starts & redirects the communication
20150408 12:32:57 <zubzub> (like weston the wayland ref compositor does)
20150408 12:33:11 <zubzub> but I doubt the ubuntu guys are that advanced
20150408 12:39:47 <xranby_GTX580> cant get X11 to work on the phone
20150408 12:39:59 <xranby_GTX580> will stop experimenting with that now
20150408 12:47:04 <gouessej> :/
20150408 12:47:16 <gouessej> I would have loved to see JOGL work on your phone
20150408 12:47:54 <zubzub> only if xmir supports gl
20150408 12:48:06 <zubzub> that or newt has to implement mir client side bindings
20150408 12:48:52 <zubzub> which reminds me I probably should have to have a look at newt and wayland again
20150408 12:49:20 <xranby_GTX580> gouessej: i will run the standard jogl test script
20150408 12:49:31 <xranby_GTX580> to probe for opengl capabilities
20150408 12:49:43 <gouessej> xranby_GTX580: Thanks, good idea :)
20150408 12:50:18 <xranby_GTX580> installing p7zip so that i can unpack it
20150408 12:59:49 <xranby_GTX580> https://gist.githubusercontent.com/xranby/09e82bd1161d8652696a/raw/test_dbg.log
20150408 13:01:16 <xranby_GTX580> a new GLDrawableFactory.static - Native OS Factory for this device is required
20150408 13:01:30 <xranby_GTX580> the X11 probing did not work..
20150408 13:01:40 <xranby_GTX580> at least good to see the opengl functions resolving
20150408 13:02:34 <zubzub> :)
20150408 13:03:58 <zubzub> as the device is running on android drivers, egl should work just fine
20150408 13:04:22 <zubzub> it's just that getting an egl surface and and egl display, you need mir client side java bindings
20150408 13:05:25 <xranby_GTX580> yes EGL initialized
20150408 13:05:51 <zubzub> or you can try killing mir and run on pure egl implementation
20150408 13:06:05 <zubzub> although I don't know how you'd initialize that
20150408 13:06:18 <zubzub> as I don't think you'll have a generic kms routine
20150408 13:06:24 <zubzub> (being all androidy and all)
20150408 13:06:44 <xranby_GTX580> i think the heuristics in jogamp needs to be fixed
20150408 13:06:54 <xranby_GTX580> it thinks its on a x11 system
20150408 13:07:32 <zubzub> it should try all systems that might be supported
20150408 13:07:56 <zubzub> the time that linux display graphics were only X are long gone
20150408 13:08:24 <xranby_GTX580> there is a way to overrride it hmm
20150408 13:08:46 <zubzub> it should try kms, <supprted newt egl implementations>, X, directfb, mir, wayland, ...
20150408 13:10:00 <xranby_GTX580> later
20150408 13:42:43 <gouessej> Good luck. I agree with you about the heuristics
20150408 13:43:07 * gouessej (5ee4b442@anon) Quit (Quit: Page closed)
20150408 14:10:31 <bruce-> would it be possible to place the ffmpeg dll files (for windows) as resources in a jar, copy them to a temp file and use loadLibrary to load them and have that work with glmediaplayer?
20150408 14:13:09 <monsieur_max> bruce-: mmmh i'm curious about the loadLibrary part , what do you mean ?
20150408 14:14:00 <bruce-> sorry I meant load() https://docs.oracle.com/javase/7/docs/api/java/lang/System.html#load(java.lang.String)
20150408 14:14:14 * xranby (~xranby@anon) has joined #jogamp
20150408 14:14:48 <xranby> bruce-: yes bundling ffmpeg or libav librarys should work
20150408 14:15:42 <monsieur_max> bruce-: did not know about this one, nice, does it work well ?
20150408 14:15:45 <xranby> jogamp itself cant do it unless the legal situation gets cleared up
20150408 14:16:03 <xranby> https://jogamp.org/bugzilla/show_bug.cgi?id=686
20150408 14:16:13 <bruce-> xranby: I realize that, I also imagine that to be a completely fuzzy situation
20150408 14:17:08 <bruce-> since the ffmpeg people don't really seem to be able to tell which parts infringe patents
20150408 14:17:19 <xranby> bruce-: the hardest part to comply with is patents
20150408 14:17:46 <xranby> its always up to the patents holders discretion to sue
20150408 14:17:51 <xranby> everyone is a moving target
20150408 14:17:52 <bruce-> monsieur_max: that generally works well for the libraries I wrap
20150408 14:18:15 <xranby> everyone should have a lawyer
20150408 14:18:17 <monsieur_max> bruce-: great, I'll replace my current code with this then
20150408 14:18:22 <xranby> for personal protection
20150408 14:20:21 <xranby> many companys file their own paternts just to do the review process, at leasat they know the new novel stuff then do not infringe on someone elses patent
20150408 14:21:01 <xranby> some later turn troll
20150408 14:21:08 <bruce-> I wonder how even that can be done with high accuracy these days
20150408 15:02:53 * xranby tries to read the javadoc to find the -D debug plags to use to force com.jogamp.nativewindow.egl
20150408 15:02:56 <xranby> flags
20150408 15:04:37 * jvanek (jvanek@anon) Quit (Ping timeout: 255 seconds)
20150408 15:16:57 <bruce-> hm
20150408 15:17:34 <bruce-> strange, glmediaplayer does not work when I load avcodec-55.dll myself
20150408 15:18:22 * jvanek (jvanek@anon) has joined #jogamp
20150408 15:21:10 <xranby> bruce-: did you load avformat and friends as well?
20150408 15:21:57 <xranby> bruce-: jogamp can be turned verbose and it will tell why it did not load
20150408 15:22:10 <bruce-> how do I do that?
20150408 15:22:21 <bruce-> I tried loading all the libraries too, I am wondering if order matters?
20150408 15:25:05 <xranby> bruce-: if you have the jogamp sourcecode
20150408 15:25:37 <xranby> you can look for Debug.debug( inside the sources and you will locate the keywords used to enable debugging of various subsystems
20150408 15:26:55 <xranby> try adding -Djogl.debug.GLMediaPlayer -Djogl.debug.GLMediaPlayer.Native
20150408 15:27:14 <xranby> bruce-: -Djogl.debug.GLMediaPlayer -Djogl.debug.GLMediaPlayer.Native
20150408 15:27:59 <xranby> you can also enable super verbose using -Djogl.debug=all
20150408 15:34:26 * xranby (~xranby@anon) Quit (Remote host closed the connection)
20150408 15:36:50 <bruce-> rah
20150408 15:37:00 <bruce-> I just made a stupid, calling loadLibrary instead of load
20150408 15:52:41 <monsieur_max> ah again !! ;)
20150408 15:52:48 <bruce-> fml
20150408 15:53:01 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150408 16:02:35 * eclesia (~husky@anon) has left #jogamp
20150408 16:26:15 * monsieur_max (~maxime@anon) has joined #jogamp
20150408 16:34:00 * kles4enko_ (~kles4enko@anon) has joined #jogamp
20150408 16:53:55 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20150408 17:46:30 * xranby_GTX580 (~familjen@anon) Quit (Remote host closed the connection)
20150408 19:20:10 * zzuegg_ (~zzuegg___@anon) Quit (Ping timeout: 250 seconds)
20150408 20:53:20 * monsieur_max (~maxime@anon) Quit (Ping timeout: 246 seconds)
20150408 21:00:53 * monsieur_max (~maxime@anon) has joined #jogamp
20150408 21:20:03 * monsieur_max (~maxime@anon) Quit (Ping timeout: 276 seconds)
20150408 22:42:25 * kles4enko_ (~kles4enko@anon) Quit (Quit: Konversation terminated!)
20150408 23:51:44 * badshah400 (~badshah40@anon) has joined #jogamp
20150409 05:05:30 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20150409050530.html