#jogamp @ irc.freenode.net - 20160519 05:06:23 (UTC)
20160519 05:06:23 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20160518050623.html
20160519 05:06:23 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20160519050623.html
20160519 06:28:16 * elect (~GBarbieri@anon) has joined #jogamp
20160519 06:31:33 <elect> hi
20160519 06:45:11 <elect> what is called before and after display()?
20160519 06:45:28 <elect> is jogl doing a glcontext.makeCurrent() and release()?
20160519 06:48:35 <elect> I am looking at the GLDrawableHelper.invokeGLImpl()
20160519 06:48:39 <elect> what does this mean?
20160519 06:48:40 <elect> // Exclusive Cases:
20160519 06:48:40 <elect> // 1: lock - unlock : default
20160519 06:48:40 <elect> // 2: lock - - : exclusive, not locked yet
20160519 06:48:40 <elect> // 3: - - - : exclusive, already locked
20160519 06:48:40 <elect> // 4: - - unlock : ex-exclusive, already locked
20160519 07:08:11 <elect> and if I call glWindow.display() from different threads, shall I have the security that calls will not overlap each others?
20160519 07:10:35 * monsieur_max (~maxime@anon) has joined #jogamp
20160519 07:11:58 * jvanek (jvanek@anon) has joined #jogamp
20160519 07:21:51 * Eclesia (~husky@anon) has joined #jogamp
20160519 07:25:21 <Eclesia> hi
20160519 07:25:37 <monsieur_max> hello
20160519 08:12:33 <monsieur_max> intel not suporting its hd3000 chipset on win10, such a shame
20160519 08:12:42 <elect> at all?
20160519 08:29:06 <monsieur_max> nope
20160519 08:29:16 <monsieur_max> support starts only with hd4000
20160519 08:29:50 <monsieur_max> the hd3000 drivers for win10 is provided by MS, and only supports DX
20160519 08:29:58 <monsieur_max> ( how surprising ! )
20160519 08:30:02 <elect> great..
20160519 08:30:45 <monsieur_max> so, players with hd3000 upgrading from win7 to win10 are actually losing functionalities
20160519 08:32:24 <monsieur_max> explain that to a player ... your game is a 2D pixel game that cannot run on a modern GPU with the latest version of the Os that is used the most
20160519 08:33:59 <zubzub> blame intel
20160519 08:34:11 <monsieur_max> of course i blame Intel
20160519 08:34:28 <monsieur_max> but that won't change this fact
20160519 08:34:49 <zubzub> indeed it won't
20160519 08:35:27 <monsieur_max> how come there is no open source driver for such situation ...
20160519 08:35:55 <monsieur_max> what are people doing ? I want answers !
20160519 08:35:57 <elect> how much share does the hd3000 have?
20160519 08:36:05 <elect> & win10
20160519 08:36:36 <monsieur_max> well, elect, I don't have the numbers, but I'd say "significant"
20160519 08:36:44 <bruce-> hd3000 is about 5 years old now
20160519 08:36:54 <zubzub> it's widely used as an integrated mobo gfx card no?
20160519 08:37:05 <elect> what about using the win7 driver?
20160519 08:37:15 <elect> I know sometimes you can use the previous driver
20160519 08:37:54 <monsieur_max> elect: that was suggested on a forum, but it seems that it's not working
20160519 08:41:10 <elect> how many users reported that?
20160519 08:42:43 <monsieur_max> elect: well, some. The minimal requirement for the game is intelHD4400, so it may have kept people away
20160519 08:43:19 <elect> because if they are few it might still work with on other pcs
20160519 08:43:30 <monsieur_max> it's more about how intel deprecates its own hardware and users don't have a clue
20160519 08:44:18 <monsieur_max> elect: it's not, it's failing at opengl init, only software profile is available and it fails on glActiveTexture
20160519 08:44:45 <elect> what's the error?
20160519 08:46:02 <monsieur_max> elect: that intel does not provide opengl drivers for hd3000 on win10
20160519 08:46:30 <monsieur_max> no point in investigating, it's a damn fact, nothing can fix that
20160519 08:48:30 <elect> there was one guy here working on a pure java software opengl implementation
20160519 08:48:36 <elect> I dont remember the nixk
20160519 08:48:57 <elect> it may be perfect for your case
20160519 08:49:06 <elect> hd3000 & win10 fallback
20160519 08:49:20 <bruce-> I don't expect that to perform well enough
20160519 08:49:43 <elect> I know, but we are talking about a 2d game however
20160519 08:49:48 <elect> it may be enough
20160519 08:49:51 <elect> I dont know
20160519 08:50:02 <elect> guillaum1, maybe?
20160519 09:21:21 <xranby> monsieur_max: microsoft's "solution" is that you should ship your game with microsoft opengl -> directx driver implementation
20160519 09:21:34 <xranby> it only support opengl es
20160519 09:22:07 <monsieur_max> so much power
20160519 09:22:46 <xranby> https://github.com/MSOpenTech/angle/wiki
20160519 09:23:08 <xranby> the idea seems to drive people to publish games for windows store.. at least that is the only easy way to use this technology
20160519 09:23:40 <xranby> in theory this git repository should compile
20160519 09:23:45 <xranby> into something usefull
20160519 09:24:24 <monsieur_max> interesting
20160519 09:29:19 <xranby> so intel hd graphics 3000 is found on all core i5 and i7 Sandy Bridge systems
20160519 09:30:05 <monsieur_max> that's why its so common
20160519 09:30:35 <xranby> if you plug the monitor to the on board dvi/vga connectoor on the motherboard you use it
20160519 09:33:18 <monsieur_max> xranby: the more common case is laptop with no gpu
20160519 09:40:57 <xranby> monsieur_max: do you have one such system to test on?
20160519 09:41:15 <monsieur_max> unfortunately, no
20160519 09:45:07 <guillaum1> elect, monsieur_max: Angle will give far better result than my software implementation, sure :-)
20160519 09:45:10 <xranby> so the only known working solution is then to install linux
20160519 09:45:27 <xranby> where the driver have mesa 3d support out of the box
20160519 09:45:31 <xranby> the GPU
20160519 09:45:51 <monsieur_max> xranby: yeah, well, I'd hardly call it a "solution" for the users :)
20160519 09:46:18 <monsieur_max> http://steamcommunity.com/app/440880/discussions/0/371918937265098747/
20160519 09:48:15 <guillaum1> sad :-|
20160519 09:51:19 <monsieur_max> yeah
20160519 09:51:43 <guillaum1> xranby, monsieur_max: why not compile osmesa (Mesa software renderer) for windows and use it as a backend for JOGL ?
20160519 09:51:54 <guillaum1> http://www.mesa3d.org/osmesa.html
20160519 09:51:59 <monsieur_max> you overestimate me :)
20160519 09:52:07 <monsieur_max> what does "compile" mean ?
20160519 09:52:14 <guillaum1> :-D
20160519 09:52:19 <monsieur_max> software ?
20160519 09:53:53 <guillaum1> monsieur_max: osmesa rendering is done by the CPU so no hardware problem anymore (but it's slower obviously)
20160519 09:54:34 <monsieur_max> yeah, gotta be sure which ogl version is supported. For this game, I only need a GL2 profile
20160519 09:55:08 <monsieur_max> nothing fancy is done, except using multiple color attachment on FBO , shaders are 1.30
20160519 09:55:15 <monsieur_max> gotta keep things simple
20160519 09:58:25 <guillaum1> https://wiki.qt.io/Cross_compiling_Mesa_for_Windows
20160519 09:59:08 <guillaum1> seems osmesa binaries are available here
20160519 10:00:10 <guillaum1> So someone has to hack a bit JOGL to load opengl32sw.dll
20160519 10:03:10 <guillaum1> xranby: do you think it's hard (big changes) to do ?
20160519 10:04:23 <xranby> you only need to add that name to the list of dll's jogl use when it scans for opengl drivers
20160519 10:05:52 <xranby> http://jogamp.org/git/?p=jogl.git;a=blob;f=src/jogl/classes/jogamp/opengl/windows/wgl/WindowsWGLDynamicLibraryBundleInfo.java;h=d13e4ddada6cfb4ae2b098fecbbdcdd1e4d95e3b;hb=HEAD#l45
20160519 10:06:00 <xranby> there on line 45
20160519 10:06:33 <xranby> libsGL.add("opengl32sw");
20160519 10:07:30 <guillaum1> xranby: ok thanks. I'll try this afternoon
20160519 10:08:02 <monsieur_max> xranby: any idea when the next release will be ?
20160519 10:09:28 <xranby> the general idea is that the next release 2.4 is out when the list work items have been fixed http://jogamp.org/wiki/index.php/SW_Tracking_Report_Objectives_for_the_release_2.4.0
20160519 10:10:53 <monsieur_max> ok thanks
20160519 10:15:27 <elect> xranby, how can I contribute?
20160519 10:15:38 <elect> I have two (very easy) bug fixes
20160519 10:16:01 <elect> https://github.com/sgothel/jogl/pulls
20160519 10:20:14 <xranby> that look like a good way to contribute.. you have both filed bugs and added solutions
20160519 10:20:39 <xranby> what is needed is more people reviewing
20160519 10:32:48 <xranby> monsieur_max: here you have some official statements from intel that HD graphics 3000 and 2000 will not be supported with windows 10 http://www.intel.com/content/www/us/en/support/graphics-drivers/000005526.html
20160519 10:32:57 <xranby> https://communities.intel.com/message/330854#330854
20160519 10:33:46 <monsieur_max> xranby: great link, thanks
20160519 10:49:12 <guillaum1> Exception in thread "main" com.jogamp.opengl.GLException: Profile GL2 is not available on null, but: []
20160519 10:49:12 <guillaum1> at com.jogamp.opengl.GLProfile.get(GLProfile.java:991)
20160519 10:49:12 <guillaum1> at com.jogamp.opengl.GLProfile.get(GLProfile.java:1004)
20160519 10:49:12 <guillaum1> at Line.main(Line.java:37)
20160519 10:49:24 <guillaum1> doesn't work out of the box...
20160519 10:51:59 <xranby> guillaum1: can you create a runtime version check by executing the etc/test_dbg.bat file?
20160519 10:52:28 <xranby> https://jogamp.org/wiki/index.php/Jogamp_Versioning_and_Releases#Runtime_Debug_Logs
20160519 10:52:34 <xranby> sh etc/test_dbg.sh (Unixoid), or etc\test_dbg.bat (Windows)
20160519 10:59:58 <guillaum1> xranby: https://gist.github.com/glegris/f12543afbb2db2f1ac3101e4cf0105f3
20160519 11:11:29 <guillaum1> seems the java.library.path parameter I've set makes things wrong
20160519 11:14:16 <elect> you may want to add those link in the steam thread, monsieur_max
20160519 11:31:12 <guillaum1> http://forum.jogamp.org/JOGL-and-mesa-opengl32-dll-on-Windows-td4031340.html
20160519 11:35:37 <xranby> guillaum1: are you still using your old jogl jar? It still only looks for OpenGL32.dll https://gist.github.com/glegris/f12543afbb2db2f1ac3101e4cf0105f3#file-test_dbg-log-L357-L381
20160519 11:38:13 <guillaum1> xranby: I just downloaded jogamp-all-platforms.7z from http://jogamp.org/deployment/jogamp-current/archive/
20160519 11:39:25 <xranby> ok you can try rename the mesa3d dll and place it at one of the locations jogl used to search for the dll
20160519 11:40:05 <xranby> or you can add libsGL.add("opengl32sw"); to WindowsWGLDynamicLibraryBundleInfo.java
20160519 11:41:11 <monsieur_max> elect : added in the "known issues" thread, where i already talk about intel chipsets
20160519 11:44:18 <guillaum1> I downloaded another version of the dll from here: https://cristianadam.eu/20150514/grim-fandango-remastered/
20160519 11:44:51 <guillaum1> xranby: I got better result (and I added libsGL.add("opengl32sw") )
20160519 11:45:31 <guillaum1> Natives
20160519 11:45:31 <guillaum1> GL4bc false
20160519 11:45:31 <guillaum1> GL4 false
20160519 11:45:31 <guillaum1> GLES3 false
20160519 11:45:31 <guillaum1> GL3bc true [3.3 (Compat profile, arb, compat[ES2, ES3], FBO, hardware)]
20160519 11:45:31 <guillaum1> GL3 true [3.3 (Core profile, arb, compat[ES2, ES3], FBO, hardware)]
20160519 11:45:31 <guillaum1> GL2 true [3.3 (Compat profile, arb, compat[ES2, ES3], FBO, hardware)]
20160519 11:45:32 <guillaum1> GLES2 false
20160519 11:45:32 <guillaum1> GLES1 false
20160519 11:45:33 <guillaum1> Count 3 / 8
20160519 11:45:46 <xranby> nice
20160519 11:46:50 <guillaum1> xranby : alas the AWT windows starts but nothing is drawn
20160519 11:46:51 <xranby> is it fast?
20160519 11:46:53 <xranby> oh
20160519 11:46:59 <guillaum1> don't know yet
20160519 12:48:02 * dwbrite (4474ac1d@anon) has joined #jogamp
20160519 12:48:09 <dwbrite> Good morning
20160519 12:50:34 <elect> hi
20160519 12:54:55 <xranby> greetings
20160519 12:55:36 <monsieur_max> guillaum1: and with a newt window ? :)
20160519 12:55:45 <elect> xranby, btw what have been working on lately? I am looking forward for this new graph ui
20160519 13:01:21 <monsieur_max> warning very graphic content : http://i.imgur.com/JRBv5pf.webm
20160519 13:01:41 <xranby> elect: do you have a usecase that needs to render a lot of curves / plot cad drawings fast on mobile gpus? the implementation of graph have today been driven by the needs to render curves and cad drawings fast for C3D's products
20160519 13:03:12 <elect> not really honestly
20160519 13:03:21 <elect> I need more a gui for newt
20160519 13:03:54 <xranby> elect: in that case Eclesia have a gui for you!
20160519 13:04:02 <elect> yeah, I tried
20160519 13:04:32 <xranby> nifry gui also work with newt
20160519 13:04:50 <elect> but perfs weren't so great
20160519 13:05:02 <elect> I know it take s alot of work
20160519 13:05:03 <xranby> and libgdx Scene2D
20160519 13:05:12 <elect> I dont wanna blame
20160519 13:07:18 <xranby> IMHO: the best way to speed up developmenet of new features is to secure income for steady development
20160519 13:08:20 <xranby> providing engineers is also great
20160519 13:09:13 * Eclesia will have to work on perfs to seduce elect
20160519 13:09:25 <elect> \o/
20160519 13:09:36 <elect> I'd also like a separate project
20160519 13:09:45 <elect> kind of a stand-alone
20160519 13:10:01 <Eclesia> for just the ui ?
20160519 13:10:12 <elect> yep
20160519 13:10:29 <Eclesia> it's already in it's own module unrelated to opengl
20160519 13:10:58 <elect> last wish, github
20160519 13:11:12 <elect> dont look at me that way
20160519 13:11:18 <elect> I know you doing
20160519 13:11:23 <elect> :D
20160519 13:11:58 <Eclesia> if github has mercurial repos maybe
20160519 13:12:39 <elect> http://hg-git.github.io/
20160519 13:13:05 <Eclesia> http://git-hg.github.io/
20160519 13:13:11 <Eclesia> XD
20160519 13:13:17 <elect> >.>
20160519 13:13:21 <elect> :D
20160519 13:13:40 <Eclesia> you don't like mercurial ?
20160519 13:13:50 <elect> how can I put it?
20160519 13:13:54 <elect> ehm.... no
20160519 13:14:31 <elect> let's say, I don't find its interface particularly
20160519 13:14:50 <elect> I just find the github's one muuuuuch better
20160519 13:15:11 <elect> try it
20160519 13:15:13 <elect> let's do a test
20160519 13:16:24 <elect> and git seems to be universally accepted as the state-of-the-art source control tool
20160519 13:16:28 <Eclesia> I use both already
20160519 13:16:30 <elect> in b4 shitstorm
20160519 13:16:53 <elect> you see, you have been already infected
20160519 13:17:15 <Eclesia> you have a windows right, same problem
20160519 13:17:37 * Eclesia troll time
20160519 13:17:53 <elect> ^^
20160519 13:19:30 <elect> monsieur_max, about your very graphic content, http://www.overclock-and-game.com/news/pc-gaming/46-gtx-1080-what-s-not-being-discussed
20160519 13:20:41 <monsieur_max> nice :)
20160519 13:25:04 <guillaum1> monsieur_max: doesn't work with newt. The Redsquare performance test goes crazy too: https://gist.github.com/glegris/fdb682b09007d072b315dceccf01d602
20160519 13:25:09 <guillaum1> 2500 fps !!!
20160519 13:26:54 <monsieur_max> more !
20160519 13:27:04 <elect> OT, youtube 2016 most disliked video, https://www.youtube.com/watch?v=w3ugHP-yZXw
20160519 13:27:13 <elect> comments are delivering some discrete lulz
20160519 13:34:10 <guillaum1> xranby: is there a way to grab pixels from a native surface when using NEWT ?
20160519 13:35:46 <zubzub> glreadpixel?
20160519 13:38:33 <guillaum1> zubzub: sure, but I don't want to get pixels with the OpenGL API. I need to debug the native surface itself
20160519 13:39:35 <zubzub> I don't think that's possible
20160519 13:39:53 <zubzub> as pushing pixels van de gpu memory to the screen is done by the driver
20160519 13:39:53 <xranby> guillaum1: one moment.. there is an api for it
20160519 13:39:58 <zubzub> not by the native display driver
20160519 13:40:19 <zubzub> unless you take a detour
20160519 14:08:35 <guillaum1> monsieur_max, zubzub, xranby: the Mesa rendering works (so glReadPixels already returns a conform result). So I just need to understand why the NEWT system hangs
20160519 14:09:01 <zubzub> 15:39 < zubzub> not by the native display driver
20160519 14:09:05 <zubzub> *native display server
20160519 14:11:53 <xranby> guillaum1: i was thinking about this API .. maybe that is not what you are looking for https://jogamp.org/deployment/jogamp-next/javadoc/jogl/javadoc/com/jogamp/opengl/util/GLReadBufferUtil.html
20160519 14:14:18 <xranby> guillaum1: do non jogl applications work with this driver on your system?
20160519 14:15:36 <guillaum1> xranby: don't know
20160519 14:16:14 <xranby> i was thinking if this driver uses a non-windows native window then windows will have issues rendering it
20160519 14:27:55 * guillaum1 debugs NEWT sources
20160519 15:18:17 <zubzub> man this google trial reads like a horror movie
20160519 15:18:35 <zubzub> you know who's the bad guy
20160519 15:18:48 <zubzub> and you know the jury is completely clueless
20160519 15:28:58 <Eclesia> if oracle wins, it's the end of Java RT
20160519 15:29:35 <guillaum1> Hopefully, there is the Eclesia Runtime to replace it
20160519 15:30:20 <zubzub> if oracle wins, I'm going to find another language to dive into
20160519 15:30:50 <zubzub> I already know C so I might pickup C++/objc and maybe swift
20160519 15:31:16 <Eclesia> if it comes to that, I'll switch to Ruby or D
20160519 15:31:26 * guillaum1 looks at COBOL 60
20160519 15:46:59 <rmk0> suspect my exit strategy would have to be rust. dynamically typed languages are not usable for software that has any interest in being correct, period. that leaves whatever there is out there that's statically typed (with algebraic types) and is memory safe. that leaves about three options and of those options, only one of them has good concurrency support
20160519 15:47:42 <Eclesia> rmk0: message is cut at ' has good concurrency support'
20160519 15:47:49 <rmk0> that was the end of it
20160519 15:48:06 <Eclesia> and the 3 options are ?
20160519 15:48:30 <rmk0> ocaml, haskell, rust. i could maybe include F# in there, but .net is ... not in a good way, last i heard
20160519 15:48:48 <rmk0> poor ecosystem
20160519 15:48:55 <rmk0> windows-centric
20160519 15:49:38 <rmk0> i've written close to a million lines of code in java since late 2011. i don't enjoy the thought of having to rewrite
20160519 15:50:59 <monsieur_max> zubzub: c# is nice too :), i plan to move to it soon enough
20160519 15:51:39 <monsieur_max> with microsoft opensourcing things, it may be something to look at
20160519 15:51:43 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20160519 15:52:14 <rmk0> think it will take decades for the .net ecosystem to catch up, and they don't have the same culture of open source, so it may not
20160519 15:52:23 <Eclesia> rmk0: same feeling, if it really happens I would make a program to convert the code
20160519 15:52:50 <rmk0> i don't think that's feasible
20160519 15:52:58 <guillaum1> Eclesia: https://github.com/mono/sharpen
20160519 15:53:01 <guillaum1> works fine
20160519 15:53:15 <Eclesia> well 90% should be possible
20160519 15:53:49 <bruce-> does that produce code fits in well with other C# code?
20160519 15:53:50 <rmk0> you can do a syntactic conversion, but it won't capture the subtleties of either language because inevitably nobody understands either of them
20160519 15:53:55 <bruce-> or will it feel alien?
20160519 15:53:57 <rmk0> semantics matter, syntax doesn't
20160519 15:54:17 <rmk0> i wouldn't trust any code that was transformed like that as far as i could throw it
20160519 15:54:32 <bruce-> don't worry, you also convert your unit tests with the same tool
20160519 15:54:49 <rmk0> i'd rewrite from scratch
20160519 15:56:02 <guillaum1> Sharpen does a pretty good job IMHO. I use it to convert my JVM to C# ahead of time. It works out of the box
20160519 15:56:14 <Eclesia> I don't have enough willpower to restart from scratch again
20160519 15:56:22 <rmk0> as labour intensive as a rewrite would be, it would be less labour intensive in the long run that trying to fix code generated by a "we promise to break your code very subtle ways" tool
20160519 15:56:30 <rmk0> program transformation is a HARD problem
20160519 15:57:17 <rmk0> .. *in
20160519 15:57:54 <guillaum1> Look at: https://github.com/xamarin/XobotOS/tree/master/android/generated
20160519 15:58:10 <guillaum1> Android converted to C# with sharpen
20160519 15:58:58 <rmk0> and where are the soundness proofs that demonstrate that the generated code is semantically equivalent to the original code
20160519 15:59:12 <rmk0> ("what's a soundness proof? i just put words in and words come out. that's good enough, right?!"
20160519 15:59:15 <rmk0> )
20160519 16:00:32 <guillaum1> rmk0: unitary tests ?
20160519 16:00:54 <rmk0> not good enough for this, unfortunately
20160519 16:01:51 <Eclesia> then let's make our own VM and language (in the public domain) :-)
20160519 16:01:58 <rmk0> anyway... what are the actual practical effects that oracle winning this case would have?
20160519 16:02:09 <rmk0> i've heard a lot of hyperbole and that's about it
20160519 16:02:37 <bruce-> we can no longer have nice things
20160519 16:02:43 <rmk0> Eclesia: would be nice
20160519 16:02:48 <Eclesia> if API are subject to copyrights then it will become a reaaaaal mess
20160519 16:03:29 <rmk0> and how would switching languages help that? if they are copyrightable, then they are copyrightable and switching languages won't make you exempt
20160519 16:04:45 <Eclesia> going home, back later ++
20160519 16:04:58 * Eclesia (~husky@anon) has left #jogamp
20160519 16:05:30 <rmk0> my question still stands!
20160519 16:05:34 <rmk0> running away won't save you!
20160519 16:05:42 <rmk0> ticktockticktockticktockticktockticktock
20160519 16:07:28 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20160519 16:08:23 <elect> why all so little pussies
20160519 16:08:28 <elect> dont be so negative
20160519 16:08:48 <elect> java is the language n1
20160519 16:09:07 <elect> they wont screw it up
20160519 16:09:55 <elect> http://arstechnica.com/tech-policy/2016/05/oracle-java-architect-conscripts-harry-potter-in-making-the-case-against-google/
20160519 16:10:39 <rmk0> i still don't have an answer about why this ruling would screw up java specifically
20160519 16:22:54 * jvanek (jvanek@anon) has joined #jogamp
20160519 16:23:41 * elect (~GBarbieri@anon) Quit (Ping timeout: 276 seconds)
20160519 16:30:14 * monsieur_max (~maxime@anon) has joined #jogamp
20160519 16:31:05 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20160519 16:56:22 <guillaum1> monsieur_max, xranby: got it !
20160519 16:56:55 <guillaum1> I got the Mesa SW renderer work
20160519 16:56:59 <xranby> guillaum1: news!
20160519 16:57:57 <guillaum1> 1) Add libsGL.add("OpenGL32sw"); to WindowsWGLDynamicLibraryBundleInfo
20160519 16:58:04 <monsieur_max> guillaum1: what a hero :) is it something that could be used without compiling a new version of jogamp ?
20160519 16:58:34 <xranby> monsieur_max: new version of jogamp will be required
20160519 16:58:47 <monsieur_max> one can always dream :)
20160519 16:58:52 <guillaum1> 2) Use GLCapabilites.setHardwareAccelerated(false); in your code
20160519 16:59:59 <guillaum1> Maybe the second step can be done automatically on driver detection
20160519 17:00:45 <xranby> we usually prefer hardware accelerated drivers before using software emulated drivers
20160519 17:01:04 <guillaum1> monsieur_max: I can send you a jar + dll for testing
20160519 17:01:36 <monsieur_max> guillaum1: that'd be great :) I can provide a place to upload them if you need it
20160519 17:01:44 <guillaum1> xranby: sure, just for help on some weird cases
20160519 17:02:15 <monsieur_max> guillaum1: won't it work hw rather than sw ?
20160519 17:03:03 <guillaum1> as it's harcoded yet, it will only only work in SW mode
20160519 17:04:00 <monsieur_max> ah, yes, of course, but I was curious if you tried with HW :)
20160519 17:09:05 <guillaum1> monsieur_max: do you use GLCanvas or NEWT in yor game ?
20160519 17:09:35 <monsieur_max> newt and glwindow
20160519 17:09:48 <monsieur_max> like a true viking
20160519 17:12:15 <guillaum1> monsieur_max: can you add GLCapabilites.setHardwareAccelerated(false) in your code or you can't and I force it in the patched JOGL ?
20160519 17:12:42 <monsieur_max> guillaum1: of course, i can do anything :)
20160519 17:22:21 <guillaum1> monsieur_max: you can grab the zip here: https://framadrop.org/r/gRAodOSKdU#bRMwearWpJgL8c8ZYT79PgEZdX/bFOzojvU+rl0Rdjg=
20160519 17:23:05 <guillaum1> 1) Unzip
20160519 17:23:05 <guillaum1> 2) Add jogl-sw.jar _before_ the other jar
20160519 17:23:37 <monsieur_max> ah a frama user ;) saw they have a mattermost instance recently, how nice of them :)
20160519 17:25:51 <guillaum1> 3) Add GLCapabilites.setHardwareAccelerated(false); in your code
20160519 17:26:22 <monsieur_max> mmmh no windows available here right now, this'll have to wait tomorrow :(
20160519 17:26:58 <monsieur_max> ah wait, maybe it'll work inside a virtualbox VM :)
20160519 17:33:48 <guillaum1> monsieur_max: yes it's pure software, so it should :-)
20160519 17:33:57 <monsieur_max> nice
20160519 17:34:17 <monsieur_max> ok, gonna try this right away, i just finished a task on another game
20160519 18:25:56 * guillaum1 looks for a demo with a complex scene in jogl-demo to make a small benchmark
20160519 19:13:54 <guillaum1> xranby: what do you think of trying to load opengl32sw if loading opengl32 fails. And force GLCapabilites.setHardwareAccelerated(false) in order to not change user code (transparent detection and settings)
20160519 19:16:17 <guillaum1> or just use a -Drenderer.software=true flag
20160519 19:18:05 <guillaum1> if (softwareRendering) libsGL.add("OpenGL32sw"); else libsGL.add("OpenGL32");
20160519 19:47:07 * Eclesia (~eclesia@anon) has joined #jogamp
20160519 20:01:35 <Eclesia> rmk0: answering the question : it was a problem 4years ago but not anymore. java 1.5 syntax is fine with me, if only there was some other good vm to use as replacement.
20160519 20:03:06 <rmk0> i forget what the question was
20160519 20:04:33 <Eclesia> nevermind :)
20160519 20:06:23 <guillaum1> software renderer tested on Windows XP 32 bits and Windows 10
20160519 20:37:49 <rmk0> ... Eclesia: i'm still trying to work out what i asked that you could have answered with that /o\
20160519 20:54:05 <Eclesia> 20160519 16:03:29 <rmk0> and how would switching languages help that? if they are copyrightable, then they are copyrightable and switching languages won't make you exempt
20160519 20:56:34 <rmk0> i suppose by "they" i should have made it clear that i meant APIs and not languages
20160519 20:58:27 <Eclesia> 4years ago it would have been a problem, but today with Unlicense-lib I don't care what they decide since I don't need the jvm api anymore.
20160519 21:00:34 <rmk0> it's not clear why anyone other than those that are producing jvm and/or jre implementations would care at all
20160519 21:01:46 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20160519 21:03:43 <Eclesia> the way oracle accused google is the problem, they claim API are copyrightable and that google should have payed a license. so in clear : 'to implement a jvm api you have to pay'
20160519 21:05:17 <rmk0> so... it's not clear why anyone other than those that are producing jvm and/or jre implementations would care at all
20160519 21:06:49 <Eclesia> well the next step would be for the Java RT apis, collection, comparator. a hanging damocles sword :/
20160519 21:07:33 <Eclesia> who knows how far oracle will go ?
20160519 21:08:09 <rmk0> well, again, if the ruling says that APIs are copyrightable, then it won't matter which language you use
20160519 21:09:35 <Eclesia> sure, as i said I don't use those apis anymore. just the vm. if there was a nice one which support reflection and jni that would be a way to break free of it.
20160519 21:24:46 <Eclesia> has anyone build jogamp with avian (without the openjdk) ?
20160519 22:48:21 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20160520 05:06:23 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20160520050623.html