JogAmp IRC Chat Logs

#jogamp @ irc.freenode.net - 20140518 05:05:15 (UTC)

20140518 05:05:15 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20140517050515.html
20140518 05:05:15 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20140518050515.html
20140518 08:03:52 * monsieur_max (~maxime@anon) has joined #jogamp
20140518 09:24:11 * hija (~hija@anon) has joined #jogamp
20140518 11:34:55 * monsieur_max (~maxime@anon) Quit (Remote host closed the connection)
20140518 11:35:28 * monsieur_max (~maxime@anon) has joined #jogamp
20140518 12:03:47 * guardianp (~gp@anon) has joined #jogamp
20140518 15:00:04 <hharrison> Has anyone seen problems with NewtCanvasAWT on OSX Mavericks machines since 2.1.5?
20140518 15:01:15 * rmk0 (~rmk0@anon) Quit (Remote host closed the connection)
20140518 15:01:46 <hharrison> Buying a Mac this weekend to try and track it down, just curious if anyone else is seeing this
20140518 15:02:09 <hharrison> No GL errors, just no contents make it to the canvas
20140518 15:47:08 <sgothel> A similar bug has been fixed - SecurityManager + JAWTWindow -> NPE ..
20140518 15:48:04 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=1004
20140518 15:49:32 <hharrison> Hmm, no stack traces in this case, but I'll pull in that fix and see if it makes a difference
20140518 15:50:04 <hharrison> Until I have hardware (Tuesday), I only have external testers
20140518 15:50:29 <sgothel> com.jogamp.opengl.test.junit.jogl.demos.es2.awt.TestGearsES2GLJPanelsAWT <- reproducible w/ this test ?
20140518 15:50:42 <sgothel> com.jogamp.opengl.test.junit.jogl.demos.es2.newt.TestGearsES2NewtCanvasAWT <- this one I mean :)
20140518 15:50:48 <hharrison> I will let you know Tuesday ;-)
20140518 15:50:59 <sgothel> ok .. I will check later today
20140518 15:51:32 <hharrison> But thereafter I will have an OSX machine at my desk to directly test stuff, this remote debug _sucks_
20140518 15:55:11 <hharrison> got a few minutes to chat re:unit tests?
20140518 15:55:59 <hharrison> I've been playing with importing a common math helpers package to sweep up a lot of the helpers currently in-tree
20140518 15:56:05 <hharrison> com.jogamp.math
20140518 15:56:43 <hharrison> But, when integrating the unit tests, I'd really prefer to write the test classes in the same package, just at a different path in the repo
20140518 15:56:56 <sgothel> hmm
20140518 15:57:16 <hharrison> src/com/jogamp/math/{implementation classes}
20140518 15:57:16 <sgothel> you mean our test package prefix thingy used in jogl ?
20140518 15:57:32 <hharrison> src/com/jogamp/math/{testclasses}
20140518 15:57:53 <hharrison> but the test classes are also in the package com.jogamp.math
20140518 15:57:57 <sgothel> could be test/src/com/jogamp/math/*
20140518 15:58:02 <sgothel> for better separation
20140518 15:58:13 <sgothel> (producing jar files)
20140518 15:58:13 <hharrison> In jaamsim I stole the maven guidelines
20140518 15:58:26 <hharrison> src/main/java/...implementation
20140518 15:58:39 <hharrison> src/test/java/...unit tests
20140518 15:58:39 <sgothel> well .. we implicit disable white-box tests while using a diff package name
20140518 15:59:04 <sgothel> but we also have same package names for some white-box access
20140518 15:59:08 <hharrison> So I have com.jaamsim.math.Vec2d
20140518 15:59:17 <hharrison> com.jaamsim.math.TestVec2d
20140518 15:59:29 <sgothel> understood ..
20140518 15:59:30 <hharrison> and can use all the package-private stuff in tests and implementation
20140518 15:59:38 <sgothel> the issue is white-box or black-box
20140518 15:59:56 <hharrison> sorry....white-box?
20140518 15:59:56 <sgothel> i.e. I personal don't like tests using non-accessible stuff
20140518 16:00:14 <sgothel> yes .. non-accessible stuff .. looking into the implementation
20140518 16:00:22 <sgothel> where black-box simply only sees the API
20140518 16:00:47 <sgothel> just a matter of taste and target
20140518 16:00:52 <hharrison> so you'd prefer com.jaamsim.math.test.{testclasses}
20140518 16:01:03 <hharrison> in the above axample
20140518 16:01:41 <sgothel> for black-box tests only using the API.. if possible .. otherwise - if white-box tests using magic stuff is required .. as you wish
20140518 16:02:11 <sgothel> in JOGL tests .. I have added some whote-box magic accessor classes like .. one-sec ..
20140518 16:02:39 <sgothel> jogamp.newt.WindowImplAccess
20140518 16:02:39 <hharrison> Also, further trivia for gluegen:
20140518 16:02:41 <hharrison> https://github.com/hharrison/gluegen.git master
20140518 16:02:55 <sgothel> yup .. gluegen is mixed up :)
20140518 16:03:31 <sgothel> ^^ you merge some of your math to our math package ? great!
20140518 16:04:18 <hharrison> That's what I'm working through, it's just that I chose the white-box style of tests
20140518 16:04:37 <hharrison> And I'm trying to figure out the best way to add the code in
20140518 16:05:06 <sgothel> so you keep using our current float/float[] access style w/o the class footprint (Vec2D, ..) ?
20140518 16:05:34 <hharrison> as a start the array-based stuff can just be consolidted to helpers
20140518 16:05:36 <sgothel> if you need white-box access .. then simply do so .. i.e. in the required new package name
20140518 16:05:57 <sgothel> thats what our static math methods are all about .. yup
20140518 16:06:05 <hharrison> Going forward we can discuss actual Vec classes/interfaces as/when they prove their worth
20140518 16:06:18 <hharrison> two discussions really
20140518 16:06:43 <sgothel> as I mentioned .. footprint especially for temp. objects (GC) and math etc .. can be huge ..
20140518 16:06:50 <hharrison> yep
20140518 16:07:24 <hharrison> My class-based lib is essentially garbage free, with good marchal-to-array helpers
20140518 16:07:30 <hharrison> sigh...marshal
20140518 16:08:09 <hharrison> But then again, there's an implied way to use that effectively that doesn't fit everyone's uses
20140518 16:08:59 <sgothel> best discussed when it's there I guess, for now ofc I would like to see things being added to FloatUtil etc as-is .. but hey, I am flexible .. so lets see :) thank you!
20140518 16:09:13 <hharrison> Yep, will say it with code
20140518 16:09:49 <hharrison> until then, enjoy the little findbugs-fixes at the gluegen link above
20140518 16:10:34 <sgothel> ah .. will merge those later .. thx .. also playing w/ high-dpi and graph-ui ... enough fun stuff :)
20140518 16:12:12 <hharrison> also, I'm still trying to get the code generators to emit proper @override for classes like GL4bcimpl
20140518 16:12:30 <sgothel> nice!
20140518 16:12:37 <hharrison> It really makes findbugs bitchy with ~2000 missing @override in those guys
20140518 16:12:43 <sgothel> hehe
20140518 16:14:22 <hharrison> Just trying to make the real stuff easier to see
20140518 16:14:23 <sgothel> since the GlueGen CStruct annotation processor works now, another test to be done would be using GL UniformObject/Buffers .. (sic)
20140518 16:14:46 <sgothel> will play w/ this for GraphUI w/ it's many uniforms .. hmm
20140518 16:15:17 <hharrison> The previous cleanups I did added @Override for the debug/tracer, the impl ones are proving harder
20140518 16:15:41 <sgothel> plus a few hardcoded impl.
20140518 16:15:43 <hharrison> As it misses @Override for anything provided by one of the interfaces
20140518 16:16:11 <hharrison> And I can't figure out how to test in the code emitter when the method is provided by an interface
20140518 16:16:30 * hija (~hija@anon) Quit (Quit: hija)
20140518 16:16:59 <hharrison> anyways, don't waste any time on it, I keep poking as I have time/energy
20140518 16:17:05 <sgothel> all impl. methods generated are @Overrides, no ?
20140518 16:17:52 <sgothel> just not the buffer/long (pointer) variations
20140518 16:18:29 <sgothel> (they have a postfix *Impl AFAIK)
20140518 16:26:34 <hharrison> No, most of the methods n the Impl classes don't get generated with @Override
20140518 16:27:15 * Eclesia (~eclesia@anon) has joined #jogamp
20140518 16:27:27 <Eclesia> hi
20140518 16:28:13 <hharrison> afk for a few minutes
20140518 16:33:08 <Eclesia> question on JOAL : there are constants for MONO and STEREO 8/16 bits PCM. is this a limitation of OpenAL or is it the binding ? since some formats (flac) have channels for from,rear and other speakers, how can I declare and use those ?
20140518 16:33:19 <Eclesia> from=front*
20140518 16:42:28 <sgothel> sadly a limitation
20140518 16:42:50 <sgothel> I would like to see a bitstream extension (aka SPDIF etc)
20140518 16:44:02 <sgothel> bitstream (PCM in this case .. forgot the other format names using PCM) .. sure we can't otherwise encoded stuff
20140518 16:45:36 <sgothel> .. on the other hand there are 5.1 channels configs .. hmm
20140518 16:46:20 <Eclesia> 5.1 is available ?
20140518 16:46:42 <Eclesia> with openal ? I didn't find anything like that in the code
20140518 16:46:56 <sgothel> or me confused by libav/ffmpeg .. have to check one sec ..
20140518 16:49:01 <sgothel> ALHelpers.getDefaultALChannelLayout(..)
20140518 16:49:58 <sgothel> ALAudioSink.isSupported(AudioFormat af) <- we check here ..
20140518 16:50:50 <sgothel> FFMPEGMediaPlayer.isAudioFormatSupported(..)
20140518 16:51:00 <sgothel> in short: yes, supported using soft buffers
20140518 16:51:28 <sgothel> ALHelpers.getALFormat(..)
20140518 16:51:49 <Eclesia> what are soft buffers ? (sorry for my ignorance, still a newbie with audio)
20140518 16:52:23 <sgothel> np - it's an extension - see last method ^^
20140518 16:54:24 * Eclesia think more hours of audio format reading will come soon ...
20140518 16:55:21 <Eclesia> why can't they make simple stuffs...
20140518 16:56:31 <Eclesia> sgothel: thanks for the help as always :)
20140518 16:56:39 <sgothel> it is actually 'simple' .. you may follow our FFMPegPlayer / ALAudioSink code .. it has all the pieces in place .. (feel free to copy, I wrote all of it)
20140518 16:56:54 <sgothel> (it's just .. annoying :)
20140518 16:57:00 <sgothel> afk .. laters
20140518 17:02:09 * guardianp (~gp@anon) Quit (Quit: liteIRC for Android)
20140518 17:43:32 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20140518 18:54:25 * guardianp (~gp@anon) has joined #jogamp
20140518 19:22:38 * kermyt (~kermyt@anon) Quit (Quit: ZNC - http://znc.in)
20140518 20:00:02 * guardianp is now known as ngtech
20140518 20:00:43 * rmk0 (~rmk0@anon) has joined #jogamp
20140518 20:00:43 * rmk0 (~rmk0@anon) Quit (Changing host)
20140518 20:00:43 * rmk0 (~rmk0@anon) has joined #jogamp
20140518 20:23:30 * ngtech (~gp@anon) Quit (Ping timeout: 240 seconds)
20140518 20:25:57 * hija (~hija@anon) has joined #jogamp
20140518 20:27:44 * kermyt (~kermyt@anon) has joined #jogamp
20140518 21:19:52 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140518 22:13:44 * kermyt (~kermyt@anon) Quit (*.net *.split)
20140518 22:13:44 * hija (~hija@anon) Quit (*.net *.split)
20140518 22:13:45 * Schrostfutz (~who@anon) Quit (*.net *.split)
20140518 22:13:46 * odin_ (~Odin@anon) Quit (*.net *.split)
20140518 22:13:46 * sgothel (~sgothel@anon) Quit (*.net *.split)
20140518 22:13:47 * hharrison (~chatzilla@anon) Quit (*.net *.split)
20140518 22:13:47 * bbbruce (~bx@anon) Quit (*.net *.split)
20140518 22:13:47 * xranby (~xranby@anon) Quit (*.net *.split)
20140518 22:13:48 * magaio (~magaio@anon) Quit (*.net *.split)
20140518 22:19:59 * kermyt (~kermyt@anon) has joined #jogamp
20140518 22:19:59 * hija (~hija@anon) has joined #jogamp
20140518 22:19:59 * hharrison (~chatzilla@anon) has joined #jogamp
20140518 22:19:59 * odin_ (~Odin@anon) has joined #jogamp
20140518 22:19:59 * bbbruce (~bx@anon) has joined #jogamp
20140518 22:19:59 * magaio (~magaio@anon) has joined #jogamp
20140518 22:19:59 * sgothel (~sgothel@anon) has joined #jogamp
20140518 22:19:59 * Schrostfutz (~who@anon) has joined #jogamp
20140518 22:19:59 * xranby (~xranby@anon) has joined #jogamp
20140518 23:11:24 * kermyt (~kermyt@anon) Quit (Ping timeout: 265 seconds)
20140518 23:16:29 * kermyt (~kermyt@anon) has joined #jogamp
20140518 23:16:30 * kermyt (~kermyt@anon) Quit (Changing host)
20140518 23:16:30 * kermyt (~kermyt@anon) has joined #jogamp
20140519 01:44:11 * guardianp (~gp@anon) has joined #jogamp
20140519 01:48:22 * guardianp is now known as ngtech
20140519 01:55:50 * ngtech (~gp@anon) Quit (Quit: liteIRC for Android)
20140519 02:15:59 * guardianp (~gp@anon) has joined #jogamp
20140519 02:42:14 * guardianp (~gp@anon) Quit (Ping timeout: 240 seconds)
20140519 05:05:15 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20140519050515.html