#jogamp @ irc.freenode.net - 20150810 05:05:28 (UTC)


20150810 05:05:28 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20150809050528.html
20150810 05:05:28 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20150810050528.html
20150810 06:40:06 * elect_ (~elect@anon) has joined #jogamp
20150810 06:40:46 <elect_> hi
20150810 07:29:14 * eclesia (~husky@anon) has joined #jogamp
20150810 07:29:20 <eclesia> hi
20150810 08:02:59 * monsieur_max (~maxime@anon) has joined #jogamp
20150810 08:28:25 * jvanek (jvanek@anon) has joined #jogamp
20150810 09:52:04 * gouessej (5ee4b442@anon) has joined #jogamp
20150810 09:52:15 <gouessej> Hi
20150810 09:52:28 <gouessej> elect_: Are you still working on DDS?
20150810 09:52:38 <elect_> yes
20150810 09:53:03 <xranby> good morning
20150810 09:53:13 <elect_> I guess I did one of the most complete list on dx9 and dx10 formats
20150810 09:53:20 <gouessej> elect_: You should look at this: https://jogamp.org/bugzilla/show_bug.cgi?id=1042#c10
20150810 09:53:22 <elect_> but we will support only the most common
20150810 09:53:48 <xranby> gouessej: hi! have you started to test your engines on raspberry pi?
20150810 09:53:51 <gouessej> elect_: I don't want to break anything in JOGL
20150810 09:54:26 <xranby> gouessej: i have been thinking about the idea how to inform the user if no opengl is available
20150810 09:54:35 <elect_> me neither
20150810 09:54:46 <gouessej> xranby: I have to buy an HDMI cable. As far as I know, I have a few things to fix. Maybe I can make JogAmp's Ardor3D Continuation work with ES 1 with a few efforts.
20150810 09:55:10 <gouessej> elect_: I have a doubt on the reading of the magic number
20150810 09:55:33 <gouessej> elect_: it should be 'D' 'D' 'S'
20150810 09:56:32 <gouessej> xranby: What are your suggestions about a mean to inform the end user of the absence of any OpenGL support?
20150810 09:56:35 <xranby> gouessej: jogl may inform the user by using a regular newt window and drawing the messages to the user using some simple api that uses the cpu.. i looked at the posts from lwjgl where they used third party tools to display a dialogue box bug i think we have most of the parts required inside newt to do it
20150810 09:57:03 <elect_> this is the magic
20150810 09:57:04 <elect_> 0x20534444
20150810 09:57:08 <gouessej> elect_: Please can you try my few changes on a DDS image file?
20150810 09:57:30 <xranby> we can initialize a regular window on all platforms that we support
20150810 09:57:45 <xranby> thus the "hardest" part is to print some text in it using the cpu
20150810 09:58:03 <elect_> let me check
20150810 09:58:03 <xranby> maybe displaying an image is the "easy" solution
20150810 09:58:03 <gouessej> xranby: This is what the source code I quoted does
20150810 09:58:26 <gouessej> xranby: It displays a message box, it supports Unicode
20150810 09:59:00 <gouessej> xranby: We just have to improve it to make it work under Android and to check that it works on the Raspberry Pi
20150810 09:59:27 <xranby> i only saw the link to the lwjgl forum http://jogamp.org/log/irc/jogamp_20150806050528.html#l74
20150810 09:59:34 <xranby> where is the code that you posted?
20150810 10:00:09 <elect_> yep, right
20150810 10:00:13 <elect_> look
20150810 10:00:19 <gouessej> xranby: the code is in the post
20150810 10:00:21 <xranby> the code by kappa uses third party tools to bring up the dialogue boxes
20150810 10:00:32 <elect_> https://github.com/elect86/joglDX10/blob/master/joglDX10/src/texture/spi/DDSresources.java#L564-L566
20150810 10:00:39 <elect_> to translate words to int
20150810 10:00:40 <gouessej> xranby: it's better than nothing
20150810 10:01:16 <elect_> pass makeFourCC('D', 'D', 'S', ' ') and it will return 542327876, the decimal of 0x20534444
20150810 10:06:32 <xranby> gouessej: personally i would rather simply add the minimal amount of classes required to render some text using the cpu into the native window to display the error message when no opengl is available.. i am sure there exist an public domain implementation in eclesia's unlicense project
20150810 10:06:44 <xranby> compared to rely on third party tools
20150810 10:07:31 <gouessej> xranby: Good idea, it would be more reliable on the long term
20150810 10:08:49 <elect_> gouessej btw, I am not yet done with dds, I am still working on it, I can try your changes once I am done
20150810 10:09:49 <eclesia> which formats are you doing ?
20150810 10:10:02 <gouessej> elect_: Ok. I'll write a unit test
20150810 10:10:41 <xranby> gouessej: i need to have one simple way to clear the native surface using the cpu on the raspberry pi to fix corner case 6 of the enhancement bug 1178 https://jogamp.org/bugzilla/show_bug.cgi?id=1178#c3
20150810 10:11:06 <elect_> eclesia, essentially all of them https://msdn.microsoft.com/en-us/library/windows/desktop/bb943991%28v=vs.85%29.aspx
20150810 10:11:24 <xranby> thus i will start look into something minimal to clear the native surface of a newt window
20150810 10:11:28 <elect_> minus those who don't have a GL equivalent and plus the new BC6 and BC7
20150810 10:14:19 <elect_> http://www.reedbeta.com/blog/2012/02/12/understanding-bcn-texture-compression-formats/
20150810 10:15:26 <eclesia> made BC1 to 5 already if you need them
20150810 10:16:58 <elect_> the problem is that your structure is completely different from the one in jogl
20150810 10:17:36 <elect_> moreover you just have to calculate the 16x16 size and pass the right GL enum and you should be pretty done
20150810 10:17:49 <eclesia> it's very unique yes :D but it covers all imaging scopes, drawing, 3d, weather, science, medical etc...
20150810 10:18:26 <xranby> in case there is technical merits of doing it the eclesia way then we should consider using it in jogl
20150810 10:21:24 <elect_> eclesia, are you taking in account the dxt1 with and without alpha?
20150810 10:21:35 <eclesia> I do
20150810 10:21:38 <gouessej> elect_: I'll use http://en.webhex.net to check I'm right
20150810 10:22:07 <xranby> gouessej: you mention in many bugreports that using angle in combination with jogl may cause crashes.. do we have a bugreport to track these crashes?
20150810 10:23:17 <elect_> you check the DDS_RGB and DDS_RGBA to know if you have alpha?
20150810 10:24:15 <eclesia> elect_: don't remember. perhaps. you should check the source code
20150810 10:24:35 <gouessej> xranby: I got one crash under Windows XP, Sven talked about a crash when using the ANGLE native library provided with Chrome or Chromium, I'm not sure that we have a bug report for this but maybe look at those about ANGLE. We had to add an option to skip it
20150810 10:26:04 <gouessej> Does anybody have a DDS file somewhere online?
20150810 10:26:58 <eclesia> some are available in my test files for dxt1-5 and at1-2
20150810 10:30:15 <gouessej> Where are your tests? https://bitbucket.org/Eclesia/ ?
20150810 10:31:07 <eclesia> https://bitbucket.org/Eclesia/unlicense/src/425ee5f199bfdd9cf2ff2845035f9ffe45b6ada9/storage/image/image-dxt/src/test/resources/un/storage/imagery/dxt/?at=default
20150810 10:31:28 <eclesia> hm only 3 files. I thinkd I had more sorry :/
20150810 10:34:54 <gouessej> eclesia: Thanks. I've just used it for my check
20150810 10:35:05 <gouessej> elect_: My code seems to be correct
20150810 10:35:32 <gouessej> elect_: I see "44 44 53 20" in the hex viewer
20150810 10:37:09 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150810 10:39:06 <gouessej> sgothel: I'll improve my source code and make a pull request soon.
20150810 10:49:54 <elect_> cant you search inside the repo on bitbucket?
20150810 10:50:49 <elect_> where are you assigning the GL enum, eclesia?
20150810 10:53:44 * monsieur_max (~maxime@anon) has joined #jogamp
20150810 10:54:04 <xranby> elect_: you can search quickly if you checkout the hg to your own machine
20150810 10:54:31 <xranby> its only 201Mb
20150810 10:54:54 <elect_> it is not an option for me
20150810 10:55:01 <xranby> why not?
20150810 10:55:10 <elect_> I cant handle tens of different projects
20150810 10:55:23 <elect_> did you checkout it?
20150810 10:55:43 <xranby> yes
20150810 10:55:53 <xranby> its a cool project
20150810 10:55:54 <elect_> could you just tell me where you find this
20150810 10:56:02 <xranby> no
20150810 10:56:09 <xranby> i will not be your search tool
20150810 10:56:27 <xranby> its easier and more polite if you check it out yourelf
20150810 10:56:34 <elect_> ok
20150810 11:06:42 <gouessej> elect_: Is it here? https://bitbucket.org/Eclesia/unlicense/src/425ee5f199bfdd9cf2ff2845035f9ffe45b6ada9/storage/image/image-dxt/src/main/java/un/storage/imagery/dxt/?at=default
20150810 11:07:27 <elect_> no
20150810 11:07:37 <elect_> I already went file by file there
20150810 11:15:49 <gouessej> elect_: un.impl.opengl.resource.Texture?
20150810 11:17:21 <gouessej> elect_: maybe in TextureUtils
20150810 11:17:24 <elect_> where?
20150810 11:17:58 <gouessej> elec_: https://bitbucket.org/Eclesia/unlicense/src/425ee5f199bfdd9cf2ff2845035f9ffe45b6ada9/engine/engine-opengl-util/src/main/java/un/impl/opengl/resource/TextureUtils.java?at=default#TextureUtils.java-181
20150810 11:18:50 <gouessej> elect_: https://bitbucket.org/Eclesia/unlicense/src/425ee5f199bfdd9cf2ff2845035f9ffe45b6ada9/engine/engine-opengl-util/src/main/java/un/impl/opengl/resource/TextureUtils.java?at=default#TextureUtils.java-181
20150810 11:19:55 <elect_> unfortunately neither
20150810 11:19:59 <elect_> anyway never mind
20150810 11:20:01 <gouessej> elect_: https://bitbucket.org/Eclesia/unlicense/src/425ee5f199bfdd9cf2ff2845035f9ffe45b6ada9/engine/engine-opengl-util/src/main/java/un/impl/opengl/resource/TextureUtils.java?at=default#TextureUtils.java-266 ???
20150810 11:21:01 <gouessej> eclesia, you create some indirect NIO buffers when creating textures, is it really intended?
20150810 11:23:41 * gouessej (5ee4b442@anon) Quit (Quit: Page closed)
20150810 11:25:07 <eclesia> elect_: TextureUtils.loadTexture
20150810 11:25:30 <eclesia> my image API is completely separate from OpenGL
20150810 11:26:15 <eclesia> so the loadTexture look at the image models and try to use GL types when possible
20150810 11:26:30 <eclesia> otherwise it converts it to plain RGBA
20150810 11:26:57 <eclesia> this way I can always load a texture
20150810 11:33:01 <elect_> maybe I didnt make myself clear, but I meant where you convert the dxt to GL enum
20150810 11:35:25 <eclesia> then I don't understand your question :X
20150810 11:35:47 <eclesia> gl enum = internalformat/imageinternalformat ?
20150810 11:35:54 <eclesia> imageformat*
20150810 11:39:35 <elect_> where you convert, for example, D3DFMT_DXT1 to COMPRESSED_RGB_S3TC_DXT1_EXT
20150810 11:40:16 <eclesia> nowhere, the compression is lost in my model. images are uncompressed at reading (for now).
20150810 11:42:23 <elect_> why are you doing that?
20150810 11:43:56 <eclesia> because 3D isn't the only objective of my library :) .
20150810 11:44:32 <eclesia> but I will have compressed buffers one day.
20150810 11:44:46 <elect_> ok ^^
20150810 11:47:25 <eclesia> elect_: you can help if you want :p
20150810 11:47:53 <elect_> I have too many things in parallel
20150810 11:48:03 <eclesia> so do I :D
20150810 11:48:05 <elect_> we shouldnt split out efforts
20150810 11:48:30 <elect_> you could use the dds jogl once I am done
20150810 11:48:34 <elect_> ^^
20150810 11:48:50 <elect_> you could use the textureData as common layer
20150810 11:49:14 <elect_> I did myself a culture
20150810 11:49:14 <eclesia> I appreciate the offer but I won't. my objective is to have everything in public domain.
20150810 11:49:26 <elect_> with dds
20150810 11:49:34 <elect_> I can give you some hints
20150810 11:49:54 <eclesia> I'll ask if I have a problem ;)
20150810 11:49:58 <elect_> I saw at the moment you are supporting only bc1-5 compressed formats
20150810 11:50:01 <elect_> am I right?
20150810 11:50:08 <elect_> no plain ones
20150810 11:50:22 * gouessej (5ee4b442@anon) has joined #jogamp
20150810 11:50:23 <eclesia> just because I didn't do the other, not because I can't
20150810 11:52:19 <gouessej> eclesia: http://jogamp.org/log/irc/jogamp_20150810050528.html#l106
20150810 11:53:00 <eclesia> 20150810 11:21:01 <gouessej> eclesia, you create some indirect NIO buffers when creating textures, is it really intended?
20150810 11:54:01 <eclesia> gouessej: should I use direct buffers .
20150810 11:54:03 <eclesia> ?
20150810 11:54:33 <gouessej> eclesia: If you don't, JOGL will have to create a direct NIO buffer anyway :s
20150810 11:54:42 <gouessej> under the hood
20150810 11:55:05 <eclesia> k will change that
20150810 11:55:09 <elect_> if you will do, remember to follow MS hints, since you cannot tell if dwFourCC is a dx9 or dx10, except for few
20150810 11:55:10 <gouessej> eclesia: but direct NIO buffers use native memory
20150810 11:55:36 <elect_> this means checking flags, rgb bit count and rgb bit masks
20150810 11:55:46 <elect_> something like this https://github.com/elect86/joglDX10/blob/master/joglDX10/src/texture/spi/DDSutil.java
20150810 11:56:10 <gouessej> eclesia: It means that it's up to you to manage them
20150810 11:56:40 <gouessej> eclesia: Ensure that you use the cleaner when you destroy the texture
20150810 11:57:09 <gouessej> eclesia: There is no public API to do it yet (it will be different in Java 1.9)
20150810 11:57:22 <eclesia> cleaner ?
20150810 11:58:13 <gouessej> eclesia: You have to use the reflection API to access some private fields and methods. The cleaner is the object that releases the native memory in this private API
20150810 11:58:44 <gouessej> eclesia: It's not needed in Apache Harmony / DVM / ART, it works with Oracle Java / OpenJDK
20150810 11:59:16 <gouessej> eclesia: Apache Harmony has a method called ByteBuffer.free() unlike OpenJDK
20150810 11:59:28 <gouessej> "named"
20150810 12:01:13 <eclesia> I will check ... later. working on C to Java transcoder for now
20150810 12:03:54 <elect_> btw eclesia, uncompressing texture means you get rid of any advantage and get only the contro
20150810 12:04:33 <eclesia> that's right.
20150810 12:04:44 <eclesia> but it works whatever the format ^^
20150810 12:05:10 <elect_> ..you support
20150810 12:05:35 <elect_> hw compress/decompress on the fly basically for free
20150810 12:05:54 * gouessej (5ee4b442@anon) Quit (Ping timeout: 246 seconds)
20150810 12:05:55 <eclesia> elect_: IF the graphic card support it
20150810 12:06:11 <elect_> dont look at the graphic card
20150810 12:06:17 <elect_> look at the gl version first
20150810 12:06:46 <eclesia> sure sure, just saying it won't work on in all cases
20150810 12:07:21 <elect_> sure, but we have 4.5 support for cards from 2010
20150810 12:07:38 <xranby> mesa3d do not yet support opengl 4
20150810 12:08:46 <elect_> what support?
20150810 12:09:05 <xranby> http://mesamatrix.net/ <- all green for desktop opengl 3.3
20150810 12:09:19 <elect_> then you get up to bc5
20150810 12:09:20 <elect_> http://renderingpipeline.com/2012/07/texture-compression/
20150810 12:10:01 <elect_> if we look at the first revision of s3tc, it is 2000
20150810 12:10:09 <elect_> we are speaking about 15 years ago..
20150810 12:10:20 <elect_> an ethernity in IT world
20150810 12:10:27 <elect_> https://www.opengl.org/registry/specs/EXT/texture_compression_s3tc.txt
20150810 12:12:31 <eclesia> you forget server who doesn't have graphic cards
20150810 12:14:08 <elect_> what has to do texture compression in this case?
20150810 12:14:14 <elect_> I dont get the scenario
20150810 12:15:48 <eclesia> I make Map desktop and server applications. even if the server doesn't have a gpu, it must still be able to process compressed image such as dxt. so cpu decompression will still be needed.
20150810 12:17:18 <elect_> I dont get, if he doesnt even have gpu, why does it have to process them instead keeping them as dds?
20150810 12:17:29 <elect_> dont have to answer now if dont have time
20150810 12:23:39 <eclesia> server do side work, collect datas, process them and output them for various display types, could be raw data (often tiff), or web maps (OGC WMS) or 3d views (Cesium mesh). so it needs to convert from anything to anything. datas aren't store in dds, often they go in a database in WKB Raster format.
20150810 12:26:26 <elect_> ah, got it
20150810 12:48:55 <eclesia> \o/
20150810 13:02:34 <elect_> ^^
20150810 13:38:07 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20150810 13:42:17 <sgothel> ^^ mesa3d does GL 4.1 .. lately
20150810 13:42:29 <sgothel> (on supported drivers .. )
20150810 13:43:51 <xranby> http://www.phoronix.com/scan.php?page=news_item&px=NVIDIA-OpenGL-2015-Linux NVIDIA Releases "OpenGL 2015" + OpenGL ES 3.2 Linux Driver
20150810 13:44:48 <xranby> http://www.phoronix.com/scan.php?page=article&item=sig-gles32-glu&num=1 Khronos Debuts OpenGL ES 3.2 & New GL Extensions, But No Vulkan This Week
20150810 13:45:04 <sgothel> Vulkan .. end of this year maybe .. hmm
20150810 13:47:17 <xranby> at least they only plan one vulkan (no vulkan es or vulkan desktop)
20150810 13:47:39 <xranby> and google wants to have vulkan added to android
20150810 14:01:47 <xranby> sgothel: shall i add https://jogamp.org/bugzilla/show_bug.cgi?id=1189 to the 2.3.2 list?
20150810 14:01:58 <xranby> - Add OpenGL ES 3.2 and new GL 4.5 Extensions support
20150810 14:04:57 <sgothel> hmm .. yes
20150810 14:04:58 <xranby> added it to the features list
20150810 14:05:09 <sgothel> so 2.3.2 will take a bit longer :)
20150810 14:06:49 * gouessej (5ee4b442@anon) has joined #jogamp
20150810 14:07:10 <gouessej> xranby: Maybe the bug 1189 can wait for the next minor release, can't it?
20150810 14:07:43 <sgothel> let me walk over it next week ..
20150810 14:07:57 <gouessej> xranby: lol it's already planned for JOGL 2.3.2
20150810 14:08:05 <sgothel> this week, I am 'off' WED - SUN .. so 2.3.2 will not happen before my return anyways
20150810 14:08:30 <xranby> gouessej: yes we are moving fast now :)
20150810 14:08:44 <sgothel> however, I will build a new test version for NEWT changes: Bug 1188, Bug 1186 and Bug 637
20150810 14:09:15 <sgothel> (@eclesia: the whole ICCM X11 stuff is working already .. checking in now ..)
20150810 14:09:17 <xranby> great 1188 - minimize ! \o/
20150810 14:09:32 <eclesia> sgothel: iccm ?
20150810 14:09:44 <sgothel> http://standards.freedesktop.org/wm-spec/wm-spec-latest.html
20150810 14:10:22 <eclesia> ho fame icon not displayed in toolbar !
20150810 14:10:25 <eclesia> frame*
20150810 14:10:52 <eclesia> and also icon frame by frame ?
20150810 14:11:01 <eclesia> sgothel: you did both ?
20150810 14:11:06 <sgothel> that not yet .. and I dunno .., actually brought back the iconized on X11 if !visible :)
20150810 14:11:37 <sgothel> (taskbar/switcher ..)
20150810 14:11:40 <gouessej> eclesia: anyway it has no chance to work under OS X
20150810 14:12:17 <eclesia> gouessej: the first or second ?
20150810 14:12:48 <gouessej> eclesia: "icon frame by frame"
20150810 14:13:00 <eclesia> k, that's less important
20150810 14:15:30 <eclesia> sgothel: I will test this evening
20150810 14:17:20 <sgothel> http://jogamp.org/git/?p=jogl.git;a=commit;h=2d837a7a7130702ad36b694875613fae77c7ef06
20150810 14:18:07 <sgothel> TODO: probably adding a 'supported' bitfield as well, so user can query upfront what is supposed to work
20150810 14:19:41 <gouessej> sgothel: good job :)
20150810 14:19:48 <sgothel> NEWT X11 impl. also queries the Atoms only once at window creation .. reducing the server roundtrip
20150810 14:21:28 <gouessej> Is it safe?
20150810 14:23:05 <xranby> i will try this new implementation and check if the reduced roundtrips workaround the https://jogamp.org/bugzilla/show_bug.cgi?id=1187
20150810 14:23:21 <xranby> X11 glamour vc4: X11 IOError Resource temporarily unavailable
20150810 14:25:36 <gouessej> sgothel: What is the interest of getStateMask()?
20150810 14:25:54 <sgothel> well .. as the API doc states ..
20150810 14:26:08 <sgothel> you can see all states via its bit mask at once
20150810 14:26:22 <sgothel> instead of querying it sequentially is<FEATURE>()
20150810 14:26:40 <sgothel> yes it is safe, since an Atom is unique per server
20150810 14:27:23 <gouessej> sgothel: Thank you for the explanations
20150810 14:27:51 <sgothel> the bit-mask is also now [almost] taken 1:1, i.e. shared .. w/ the native implementation
20150810 14:28:00 <sgothel> reducing complexity
20150810 14:31:31 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=1187#c3 <- close it as invalid, iff 'likely inside the X11 experimental glamour driver'
20150810 14:31:51 <sgothel> well .. yeah .. maybe we can do a workaround or something .. you do it Xerxes?
20150810 14:32:35 <sgothel> @ALL, Xerxes, Julien: Do I have something to merge ? (git)
20150810 14:32:43 <xranby> sgothel: i am looking at it to check iff it is on our turf... the workaround may be that the X11 server is slow?
20150810 14:32:45 <sgothel> fetching all repos now .. to see :)
20150810 14:33:05 <xranby> sgothel: i only have one fix comments patch
20150810 14:33:18 <sgothel> unavail -> no window/dpy .. not just slow :)
20150810 14:33:23 <xranby> https://jogamp.org/bugzilla/attachment.cgi?id=716
20150810 14:33:32 <sgothel> yup .. don't say 'only'
20150810 14:33:37 <sgothel> can I have it via git ?
20150810 14:33:50 <sgothel> or is it a git-email patch ? .. lets see
20150810 14:34:04 <sgothel> very good - thx!
20150810 14:34:11 <xranby> youre welcome
20150810 14:38:04 <sgothel> hmm .. not formated well .. for 'git am'
20150810 14:38:33 <xranby> hmm... odd ... ok let me push it to my git then
20150810 14:39:00 <sgothel> when you do that .. format as 'email patch' .. so 'git am' works properly
20150810 14:43:27 <xranby> ok, i will try that way of exporting it next time... i tried to use gitk -> export git commit
20150810 14:43:47 <gouessej> sgothel: I have to implement your suggestion, the rest is working
20150810 14:44:08 <xranby> sgothel: https://github.com/xranby/jogl/commit/217d8b78a3d70d9be59d4537c7565118dfe1e277
20150810 14:44:40 <gouessej> sgothel: look at the very last comments: https://jogamp.org/bugzilla/show_bug.cgi?id=1042
20150810 14:50:07 * gouessej (5ee4b442@anon) Quit (Quit: Page closed)
20150810 14:52:25 <xranby> sgothel: https://github.com/xranby/jogl/commit/93753f2aa7e6638ac29b5405f541d356a66ef988 ShaderCode: Fix [ -> ] typo in javadoc and comments.
20150810 14:56:31 <sgothel> @Julien: Great thx
20150810 14:56:43 <sgothel> @Xerxes: ditto :)
20150810 14:59:13 <sgothel> Xerxes: 93753f2aa7e6638ac29b5405f541d356a66ef988 <- is wrong
20150810 14:59:44 <xranby> it is?
20150810 14:59:47 <sgothel> '[1.30 .. 1.50[' -> glslVersion.compareTo(GLContext.Version1_30) >= 0 && glslVersion.compareTo(GLContext.Version1_50) < 0
20150810 15:00:01 <sgothel> so the open-bracket means 'excluding 1.50' :)
20150810 15:00:20 <sgothel> its a math style ..
20150810 15:00:50 <xranby> oh.. i feel more math clever now
20150810 15:01:17 <xranby> thank you. skip that patch then
20150810 15:02:35 <xranby> i have removed the commit from my master
20150810 15:06:01 <sgothel> ditto -- merged Xerxes/Julien
20150810 15:06:14 <xranby> kudos!
20150810 15:11:11 <sgothel> @Xerxes: You may verify whether I broke anything related to Bug 1178 (X11 underlay) .. ahem :-)
20150810 15:11:36 <sgothel> but AFAIK .. you don't use X11 WindowDriver native details .. so it should work
20150810 15:11:58 <xranby> sgothel: no problem!
20150810 15:12:03 <xranby> sgothel: i will check it
20150810 15:12:25 <sgothel> Bug 1189 - good description w/ references thx! looks doable ..
20150810 15:12:29 <sgothel> (in short time)
20150810 15:19:38 * gouessej (5ee4b442@anon) has joined #jogamp
20150810 15:20:43 <gouessej> xranby: Maybe we should fill a bug report about the message box even though we can't target the version 2.3.2
20150810 15:21:48 <xranby> gouessej: yes, always a good idea to file bugreports
20150810 15:23:45 <xranby> sgothel: i think there is one regression for the underlay tracker when X11 fails to initialize i now see a RuntimeException instead of an ExceptionInInitializerError
20150810 15:24:48 <xranby> and this makes the driver fail to switch to the fallback to read teh /dev/ files directly https://github.com/sgothel/jogl/commit/f84af439535640e1072b6cd670aa44cbe91ef052#diff-58483f2d267c239a5a18b6157b1cd92cR76
20150810 15:25:00 <gouessej> xranby: What about that? http://stackoverflow.com/a/1394677
20150810 15:25:32 <gouessej> xranby: it requires only Xm, Xt and X11
20150810 15:26:59 <xranby> sgothel: let me check if changing that line 76 to catch a RuntimeException fixes it OR shall i look into why it no longer sends an ExceptionInInitializerError ?
20150810 15:27:55 <sgothel> @Xerxes: yes please .. check changed behavior first .. hmm, RuntimeEx -> could mean we determine it doesn't work, where ExInit means java could not even run it?
20150810 15:28:27 <sgothel> @message-box: well .. I surely have a split feeling about that :)
20150810 15:28:43 <sgothel> iff a message box, then only for 'GL n/a' !
20150810 15:28:56 <sgothel> otherwise .. all other message boxes can be created using NEWT and GL
20150810 15:29:20 <sgothel> and this message-box should be invoked explicitly .. by the user app ..
20150810 15:29:24 <xranby> how about mesasge boxes created using NEWT + CPU?
20150810 15:29:42 <sgothel> we already give a n/a exception, or GL profile set .. to query at runtime
20150810 15:30:06 <sgothel> @Xerxes: that would be more acceptable .. yes
20150810 15:30:59 <gouessej> Is it already possible to use NEWT without OpenGL support?
20150810 15:31:06 <sgothel> ofc
20150810 15:31:20 <sgothel> just use a Window .. not the GLWindow :)
20150810 15:31:55 <sgothel> NEWT can even be build w/o JOGL .. hence the subpackage opengl for GLWindow
20150810 15:32:38 <gouessej> excellent :)
20150810 15:33:04 <xranby> Caused by: com.jogamp.nativewindow.NativeWindowException: X11Util.Display: Unable to create a display(nil) connection. Thread main-Display-.bcm.vc.iv_nil-1-EDT-1
20150810 15:33:04 <xranby> at jogamp.nativewindow.x11.X11Util.openDisplay(X11Util.java:453)
20150810 15:33:09 <xranby> Caused by: java.lang.RuntimeException: com.jogamp.nativewindow.NativeWindowException: X11Util.Display: Unable to create a display(nil) connection. Thread main-Display-.bcm.vc.iv_nil-1-EDT-1
20150810 15:33:09 <xranby> at jogamp.newt.ScreenImpl.create(ScreenImpl.java:143)
20150810 15:33:19 <xranby> lets see
20150810 15:33:54 <sgothel> dunno whether I changed anything here .. hmm, but that is a 'good exception', i.e. useful
20150810 15:34:03 <sgothel> better than .. Ex..Init..
20150810 15:34:16 <sgothel> tbh .. just catch both .. :)
20150810 15:34:24 <gouessej> yes, it was throwing an exception in a static initializer
20150810 15:34:31 <sgothel> ok .. I have to go now .. back in later tonight
20150810 15:34:36 * elect_ (~elect@anon) Quit (Ping timeout: 264 seconds)
20150810 15:34:36 <gouessej> me too. Bye
20150810 15:34:42 * gouessej (5ee4b442@anon) Quit (Quit: Page closed)
20150810 15:34:43 <xranby> sgothel: gouessej: bye!
20150810 15:34:45 <sgothel> a few regressions to fix later .. it seems :)
20150810 15:55:38 * elect (~elect@anon) has joined #jogamp
20150810 15:57:25 <xranby> regression caused by the fix for 1178 cc7: https://jogamp.org/bugzilla/show_bug.cgi?id=1178#c6
20150810 15:57:30 <xranby> problem understood
20150810 15:57:55 <xranby> assigned this new corner case as cc10
20150810 15:58:16 <xranby> Since we no longer try to make a window visible inside the static constructor of the X11UnderlayTracker.java generates a new side effect that we no longer throw an ExceptionInInitializerError that can be caught on line 76 of bcm/vc/iv/WindowDriver
20150810 16:00:51 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20150810 16:04:07 * eclesia (~husky@anon) has left #jogamp
20150810 16:07:33 * elect (~elect@anon) Quit (Ping timeout: 256 seconds)
20150810 16:33:18 <xranby> sgothel: https://github.com/xranby/jogl/commit/36e9671177c690ebd71d93021b8f797487785dda Bug 1178: Fix cc10 regression caused by the fix for cc7.
20150810 16:38:39 <xranby> sgothel: regression resolved with the above commit ^
20150810 16:41:10 <xranby> the bitfield stuff appears to work
20150810 16:41:18 <xranby> maximize etc
20150810 17:08:27 * elect (~elect@anon) has joined #jogamp
20150810 18:18:35 * elect (~elect@anon) Quit (Ping timeout: 252 seconds)
20150810 19:43:28 * elect (~elect@anon) has joined #jogamp
20150810 19:53:58 <sgothel> @Xerxes: great
20150810 20:58:05 * elect (~elect@anon) Quit (Ping timeout: 252 seconds)
20150811 03:15:39 * bob (7046558f@anon) has joined #jogamp
20150811 03:16:03 * bob is now known as Guest91456
20150811 03:17:21 * Guest91456 (7046558f@anon) has left #jogamp
20150811 03:24:22 * aaaaa (7046558f@anon) has joined #jogamp
20150811 03:26:12 * aaaaa (7046558f@anon) has left #jogamp
20150811 05:05:28 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20150811050528.html