JogAmp IRC Chat Logs

#jogamp @ irc.freenode.net - 20151008 05:05:16 (UTC)

20151008 05:05:16 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20151007050515.html
20151008 05:05:16 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20151008050516.html
20151008 05:23:35 <xranby_f22> odd.. i would have expeccted this driver to have gles 3.2 GLES3 false <---> GL_VERSION 4.5.0 NVIDIA 355.00.05
20151008 05:24:09 <xranby_f22> sgothel: have you tested GLES3 support with the new nvidia drivers?
20151008 05:24:21 <xranby_f22> testing using the latest git ofc
20151008 05:25:52 <xranby_f22> Natives
20151008 05:25:52 <xranby_f22> GL4bc true [4.5 (Compat profile, arb, compat[ES2, ES3, ES31, ES32], FBO, hardware)]
20151008 05:25:52 <xranby_f22> GL4 true [4.5 (Core profile, arb, compat[ES2, ES3, ES31, ES32], FBO, hardware)]
20151008 05:25:52 <xranby_f22> GLES3 false
20151008 05:25:52 <xranby_f22> GL3bc true [4.5 (Compat profile, arb, compat[ES2, ES3, ES31, ES32], FBO, hardware)]
20151008 05:25:53 <xranby_f22> GL3 true [4.5 (Core profile, arb, compat[ES2, ES3, ES31, ES32], FBO, hardware)]
20151008 05:25:54 <xranby_f22> GL2 true [4.5 (Compat profile, arb, compat[ES2, ES3, ES31, ES32], FBO, hardware)]
20151008 05:25:54 <xranby_f22> GLES2 true [2.0 (ES profile, arb, compat[ES2], FBO, software)]
20151008 05:25:54 <xranby_f22> GLES1 true [1.1 (ES profile, arb, compat[FP32], software)]
20151008 05:25:55 <xranby_f22> Count 7 / 8
20151008 05:32:10 <xranby_f22> one moment.. will reopen https://jogamp.org/bugzilla/show_bug.cgi?id=1189
20151008 05:49:11 <xranby_f22> reopened. debug log attached
20151008 05:50:39 <xranby_f22> tested with gluegen 444af35253943cfa5bad88ff752981c311689e49 jogl 56eb5025694064ad3e25cf2fb7ba0cabbda8ee67 (the latest git)
20151008 05:50:58 <sgothel> well .. we just try w/ ES 3.2 no?
20151008 05:51:15 <sgothel> good that ES32 compat is detected at least
20151008 05:51:40 <sgothel> still figuring out w/ the gnome issue .. horrible
20151008 05:52:47 <xranby_f22> sgothel: i will go through my notes.. checking if i have more bugs to report... when you run ant junit.run you get the false impression everything is ok because junit.run ends with the message "everything is OK" :/
20151008 05:53:07 <xranby_f22> even in situations when tests have failed
20151008 05:53:37 <xranby_f22> on my machine i see that some tests fails to terminate properly.. we catch the ThreadDeath
20151008 05:53:59 <xranby_f22> which if caught should be rethrown to terminate the thread
20151008 05:54:21 <xranby_f22> will look into that and file a new bugreport.. likely a new regression
20151008 05:56:00 * xranby_f22 is currently only looking for regressions
20151008 06:01:59 <xranby_f22> sgothel: happy to hear that the gnome issue triggered on your machine as well, sorry to cause release delays hoever, good that regressions is found!
20151008 06:02:04 <xranby_f22> however
20151008 06:05:42 <xranby_f22> it is odd that with the nvidia driver 355.00.05 we pick software for GLES2 and GLES1... hmm
20151008 06:05:42 <xranby_f22> GLES2 true [2.0 (ES profile, arb, compat[ES2], FBO, software)] GLES1 true [1.1 (ES profile, arb, compat[FP32], software)]
20151008 06:05:42 <xranby_f22> have we used mesa3d here instead of the nvidia driver
20151008 06:05:42 <xranby_f22> ?
20151008 06:11:56 <sgothel> use one of the profile tests
20151008 06:12:02 <sgothel> i.e. dump proper info
20151008 06:12:09 <sgothel> me cont w/ show/hide bug
20151008 06:12:18 * elect (~GBarbieri@anon) has joined #jogamp
20151008 06:12:52 <sgothel> NVIDIA 355.11
20151008 06:13:19 * netlike (~netlike@anon) has joined #jogamp
20151008 06:13:28 <sgothel> so I also have to install the patched driver .. duh
20151008 06:13:33 <elect> hi
20151008 06:13:37 <sgothel> thought they would finally have it in ..
20151008 06:13:40 <sgothel> hi
20151008 06:15:22 <xranby_f22> 355.11 is the stable driver 355.00.05 is the new features driver
20151008 06:15:35 <xranby_f22> nvidia versioning is fun
20151008 06:16:13 <xranby_f22> nvidias windows new features driver is 355.97 https://developer.nvidia.com/opengl-driver
20151008 06:16:34 <xranby_f22> NVIDIA's Windows driver version 355.97 and Linux drivers version 355.00.09 provide beta support for OpenGL 2015 ARB extensions and OpenGL ES 3.2 on capable hardware.
20151008 06:17:24 <xranby_f22> odd why is thed river reporting .05 of my machine... instead of .09?
20151008 06:18:03 <elect> I wondered which were the OpenGL 2015 ext, but I didnt find a list
20151008 06:18:09 <xranby_f22> possibly because they have updted it >D
20151008 06:18:25 <xranby_f22> i will install the .09 beta driver and retest
20151008 06:20:02 * xranby_f22 (~familjen@anon) Quit (Remote host closed the connection)
20151008 06:20:16 <sgothel> gnome is _not_ clearing WM_STATE _NET_WM_STATE_HIDDEN .. 'awesome'
20151008 06:20:23 <sgothel> (on visible -> true
20151008 06:21:03 <sgothel> neither IconicState -> NormalState, nor MapWindow ..
20151008 06:26:35 <sgothel> so we would need to query whether user uses gnome .. duh
20151008 06:28:22 * xranby_f22 (~familjen@anon) has joined #jogamp
20151008 06:28:30 <xranby_f22> good news, upgrading from the 355.00.05 beta driver to the 355.00.09 beta driver fixed the issue!
20151008 06:28:37 <sgothel> ah
20151008 06:28:41 <xranby_f22> GLES3 true [3.2 (ES profile, arb, compat[ES2, ES3, ES31, ES32], FBO, hardware)]
20151008 06:28:45 <sgothel> then pls make it VALIDATED :)
20151008 06:29:18 <sgothel> (and I don't need to do that .. test)
20151008 06:29:19 <xranby_f22> the VALIDATED option os not found in bugzilla hmm
20151008 06:29:30 <sgothel> hehe .. 1) -> FIXED
20151008 06:29:35 <sgothel> then 2) VALIDATED
20151008 06:29:47 <xranby_f22> VALIDATED
20151008 06:30:33 <sgothel> IconicState / _NET_WM_STATE_HIDDEN is a nice 'hidden' window mode, where its still visible in task bar
20151008 06:30:44 <sgothel> but .. seems like at least gnome/* doesn't work
20151008 06:30:48 <sgothel> works well w/ KDE
20151008 06:30:54 <sgothel> so what can we do?
20151008 06:31:07 <sgothel> before - we simply did Unmap/Map
20151008 06:31:18 <sgothel> but then .. no task bar icon
20151008 06:32:21 <sgothel> Gnome Bug: if state is _NET_WM_STATE_HIDDEN, it cannot be cleared neither w/ NormalState nor XMapRaised (if having a child window)
20151008 07:05:06 * monsieur_max (~maxime@anon) has joined #jogamp
20151008 07:21:43 <xranby_f22> sgothel: do _NET_WM_STATE_HIDDEN differ from what happens when a user presses "minimize" ?
20151008 07:23:05 <xranby_f22> from a users perspective.. is it expected to see the window in the taskbar when you use visible(false) ?
20151008 07:25:01 <xranby_f22> i as a user and developer woul expect minimize -> visible in the taskbar and visible false -> make it dissapear, the user should not be able to interact with the window
20151008 07:27:31 <xranby_f22> hence I consider using _NET_WM_STATE_HIDDEN the wrong thing to use ... not sure when this is a good mode to use
20151008 07:27:59 <zubzub> xranby_f22: check ewmh
20151008 07:28:10 <zubzub> http://standards.freedesktop.org/wm-spec/wm-spec-latest.html
20151008 07:28:43 <zubzub> Windows which are actually iconified or minimized should have the _NET_WM_STATE_HIDDEN property set, to communicate to pagers that the window should not be represented as "onscreen."
20151008 07:31:14 <xranby_f22> zubzub: ok so we should use this for minimized and iconified windows
20151008 07:31:30 <xranby_f22> zubzub: however we should not use it for non visible windows, do we agree?
20151008 07:31:36 * jvanek (jvanek@anon) has joined #jogamp
20151008 07:32:29 <xranby_f22> thus if you use GLWindow.setVisible(false) we should not use _NET_WM_STATE_HIDDEN
20151008 07:33:30 * eclesia (~husky@anon) has joined #jogamp
20151008 07:33:31 <xranby_f22> "_NET_WM_STATE_HIDDEN should be set by the Window Manager to indicate that a window would not be visible on the screen if its desktop/viewport were active and its coordinates were within the screen bounds. The canonical example is that minimized windows should be in the _NET_WM_STATE_HIDDEN state. Pagers and similar applications should use _NET_WM_STATE_HIDDEN instead of WM_STATE to decide whether to display a window in minia
20151008 07:33:58 <eclesia> hi
20151008 07:33:58 <sgothel> jaja ..
20151008 07:34:25 <sgothel> no, thats not reallty dfjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjbn
20151008 07:34:26 * jvanek_ (jvanek@anon) has joined #jogamp
20151008 07:34:35 <sgothel> oos d dffffff
20151008 07:34:36 <xranby_f22> hello all
20151008 07:34:49 * xranby_f22 reads the freedesktop spec for the first time
20151008 07:34:51 <sgothel> hehe .. son is typib
20151008 07:34:58 <xranby_f22> hi sven and sven!
20151008 07:35:02 <sgothel> http://standards.freedesktop.org/wm-spec/wm-spec-latest.html#idm140200472615568
20151008 07:35:03 <elect> junior sven
20151008 07:35:25 <elect> he wants to learn
20151008 07:35:51 <sgothel> hewlklo
20151008 07:36:23 <xranby_f22> :-D
20151008 07:36:31 <xranby_f22> O:-)
20151008 07:36:44 <xranby_f22> hello!
20151008 07:37:37 <sgothel> you don't set that state directly .. but via IconicState/NormalState
20151008 07:37:51 <sgothel> KDE works fine .. as usual, gnome won't clear the hidden state
20151008 07:38:49 <sgothel> hideous :)
20151008 07:39:49 <xranby_f22> so we must clear it implicit by doing unrelated thing A we get the result we expect B ?
20151008 07:40:11 * jvanek_ (jvanek@anon) Quit (Quit: Leaving)
20151008 07:49:55 <xranby_f22> yoga lifehacks> press your feets down to stand taller & press the bones in your butt down into the chair to sit with a straight back
20151008 07:51:33 <zubzub> instructions unclear, desk too low, computer screen badly visible, arms painful
20151008 07:53:25 <elect> I code by standing up
20151008 07:53:30 <elect> it's nice
20151008 07:54:21 <zubzub> I code with a painful back
20151008 07:54:29 <zubzub> by lack of proper desk equipment
20151008 07:55:20 <zubzub> don't worry, I already got a plan
20151008 07:55:24 <zubzub> ask for 30% raise
20151008 07:55:33 <zubzub> if they refuse I'm out
20151008 07:55:40 <zubzub> the last full time senior dev of the team
20151008 08:03:21 <zubzub> mense die readmes in html formaat steke moete ze de kop inslaan -_-
20151008 08:03:32 <zubzub> wrong channel :p
20151008 08:16:42 <elect> it seems a good plan
20151008 08:17:24 <elect> zubzub, you need an holiday?
20151008 08:17:29 <elect> time to ask also more days ^^
20151008 08:17:44 <elect> take two pigeon with a bait
20151008 08:18:45 <elect> btw, the martian book is really great, I strongly suggest you to read it as well, guys
20151008 08:18:55 <elect> it's addicting
20151008 08:19:05 <elect> I think I am gonna see also the movie
20151008 08:19:23 <elect> althought they say who read the book would be a kind of disappointed
20151008 08:19:25 <elect> nothing new
20151008 08:19:44 <elect> book->movie is usually be that way
20151008 08:20:39 <xranby_f22> the movie is good, the book more realistic
20151008 08:20:49 <elect> have you seen it?
20151008 08:20:54 <elect> *watched
20151008 08:21:13 <xranby_f22> I have watched the movie
20151008 08:21:24 <elect> 3d or 2d?
20151008 08:21:29 <xranby_f22> 3d
20151008 08:21:37 <elect> worth over the 2d?
20151008 08:21:54 <xranby_f22> its 3d
20151008 08:22:06 <elect> there is also in 2d
20151008 08:22:13 <elect> as far as I saw
20151008 08:22:31 <xranby_f22> i have not seen the 2d version, thus i cant compare
20151008 08:25:30 <xranby_f22> the downside with 3d is that you have to watch the movie wearing polaroid glasses
20151008 08:25:52 <elect> I admit I always rejected the idea of watching a 3d movie
20151008 08:26:03 <elect> but now I am taking it in consideration for the first time
20151008 08:26:15 <xranby_f22> the polaroid glasse adds a metallic look that is unnatural
20151008 08:26:31 <xranby_f22> the polaroid glasses also makes my nose hurt
20151008 08:26:48 <xranby_f22> thus 2s is likely more enjoyable
20151008 08:27:24 <xranby_f22> if you enjoy seeing the real world wearing raybans.. then go for it
20151008 08:27:31 <xranby_f22> the 3d will look the same
20151008 08:29:22 <xranby_f22> the movie is good that you can leave the cinema and feel awe that the movie have no obvious logical flaws
20151008 08:29:22 <xranby_f22> except that the nasa got the budget to send people to mars
20151008 08:29:38 <elect> :D
20151008 08:29:48 <xranby_f22> but that flaw is in the book as well :)
20151008 08:32:49 <bruce-> is there a good alternative to glutesselate?
20151008 08:33:01 <bruce-> preferably a pure java implementation that is fast
20151008 08:40:08 <monsieur_max> bruce-: maybe eclesia has something in his great magic bag :)
20151008 08:40:26 <monsieur_max> bruce-: never used something else than glutesselate though
20151008 08:46:23 <eclesia> hm ? what does glutesselate do ?
20151008 08:46:53 <bruce-> chops up concave polygons with holes into triangles
20151008 08:47:38 <eclesia> triangulation, yes I have
20151008 08:48:05 <eclesia> convex,concave,holes... I use it for font glyphs
20151008 08:50:21 <eclesia> final EarClipping clipper = new EarClipping();
20151008 08:50:22 <eclesia> final Sequence triangles = clipper.triangulate(geom);
20151008 08:50:42 <eclesia> demo in : demo.geometry.TriangulationEC
20151008 08:51:01 <bruce-> why can no one here talk in links to code or documentation? :)
20151008 08:51:25 <bruce-> like, clickable ones
20151008 08:52:02 <eclesia> I'm at work, will give you the link at lunch :p
20151008 09:30:39 <monsieur_max> "if you need it, Eclesia has it"
20151008 09:31:13 <monsieur_max> bruce-: eclesia has probably a library that generates the links to code or documentation
20151008 09:32:15 <eclesia> nope, websemantic, speech interpretor and deduction API aren't started yet
20151008 09:38:55 <xranby> someone seen or used Vert.x ? http://vertx.io/ Vert.x is a tool-kit for building reactive applications on the JVM. Vert.x is event driven and non blocking. This means your app can handle a lot of concurrency using a small number of kernel threads. Vert.x lets your app scale with minimal hardware. Vert.x is open source and dual licensed under the Eclipse Public License 1.0 and Apache License 2.0.
20151008 09:39:24 <zubzub> xranby: it's nice
20151008 09:39:30 <zubzub> until you look at the code
20151008 09:40:04 <zubzub> so I gues it's ok to use it
20151008 09:40:09 <zubzub> but don't try to fix it :p
20151008 09:40:22 <xranby> thank you for the review
20151008 09:43:14 <monsieur_max> zubzub: I've always been told that beauty is inside :(
20151008 09:43:15 <xranby> http://vertx.io/vertx2/embedding_manual.html - "Many of the Vert.x core classes are not thread-safe. When running Vert.x in the Vert.x platform you don't have to worry about that as Vert.x guarantees that your verticle code is never executed by more than one thread concurrently. When running core embedded you have to be more careful as there is no container to make such guarantees. Therefore it's up to you the developer to guard against concurrent acces
20151008 09:43:40 <xranby> so.. vertx tries to make it beautiful by making your application singlethreaded?
20151008 09:43:47 <xranby> meh
20151008 09:44:10 <bruce-> but it is async!
20151008 09:44:37 <xranby> :)
20151008 09:44:49 <xranby> i will look elsewhere
20151008 09:49:44 <zubzub> xranby: are you searching for async message based framework?
20151008 09:55:03 <xranby> zubzub: yes, i have some legacy multithreaded event driven applications that i need to migrate to the new world
20151008 09:56:39 <xranby> ok this has changed from 2 to 3 http://vertx.io/docs/vertx-core/java/#_thread_safety "Most Vert.x objects are safe to access from different threads. However performance is optimised when they are accessed from the same context they were created from."
20151008 09:58:14 <xranby> still .. "most" may mean it is not
20151008 09:59:13 <zubzub> there this async frameowork
20151008 09:59:15 <zubzub> forgot the name
20151008 09:59:18 <zubzub> will look it up
20151008 09:59:21 <zubzub> aysnc threading
20151008 09:59:43 * netlike (~netlike@anon) Quit (Quit: Leaving)
20151008 11:02:25 <eclesia> bruce-: https://bitbucket.org/Eclesia/unlicense/src/tip/demo/src/main/java/demo/geometry/TriangulationEC.java?at=default&fileviewer=file-view-default
20151008 11:03:20 <eclesia> bruce-: https://bitbucket.org/Eclesia/unlicense/src/tip/api/api-geometry/src/main/java/un/impl/geometry/triangulate/EarClipping.java?at=default&fileviewer=file-view-default
20151008 11:03:22 <eclesia> :p
20151008 11:30:29 <xranby> https://twitter.com/xranby/status/652082418468392960 - This is the ultimate & compact, inverted, Raspberry Pi + 7" screen mount! Hidden cable with access to SD Card & GPIO!
20151008 11:31:53 <xranby> JogAmp runs on it of course :)
20151008 11:37:20 <bruce-> eclesia thanks :)
20151008 11:38:13 <eclesia> hand made, no guarantee :p
20151008 11:56:56 * netlike (~epyon@anon) has joined #jogamp
20151008 12:02:46 <xranby> https://sites.google.com/site/itstheshappening/ first known SHA-1 collision found
20151008 12:36:53 <rmk0> oooooooh!
20151008 12:49:53 <monsieur_max> gotta love research
20151008 12:54:53 <xranby> if someone really pays that much to rent the EC2 cloud for a couple of months to break one SHA-1.. how on earth can that go undetected? :D
20151008 12:56:09 <zubzub> 14:02 < xranby> https://sites.google.com/site/itstheshappening/ first known SHA-1 collision found
20151008 12:56:15 <zubzub> *first publicly known
20151008 12:56:15 <zubzub> ;)
20151008 12:56:20 <xranby> "we estimate the SHA-1 collision cost today (i.e., Fall 2015) between 75K$ and 120K$ renting Amazon EC2 cloud computing over a few months." "within the resources of a criminal syndicate"
20151008 12:56:24 <xranby> zubzub: good point!
20151008 12:56:56 <zubzub> I suspect NSA, FSB, GHCQ and friend already have a nice table
20151008 12:57:39 <xranby> yes.. they got a budget to store all internet traffic.. what about if they used it to build a rainbowtable
20151008 12:58:10 <xranby> they must use the unused storage for something.. right
20151008 12:58:26 <xranby> ?
20151008 13:38:17 <monsieur_max> do you mean ... they are storing the jogamp channel logs ? :( oh my
20151008 13:59:12 <eclesia> monsieur_max: nahh we are good guys, we would Never say anything bad about them. putin bad boy, ogm is good, islam evil. We live eating bigmacs, drinking cola and voting bush. they don't spy on honest citizen. ;)
20151008 14:00:01 <monsieur_max> ahhh yes ... of course !! who would live differently ? that'd be insane !
20151008 14:00:23 <elect> they do it for our own safety
20151008 14:17:19 <elect> http://marc.info/?l=linux-kernel&m=137392506516022&w=2
20151008 14:42:42 <elect> https://www.youtube.com/watch?v=-ZRvHbHxr-k
20151008 14:54:21 <zubzub> elect: your mail is from 2013
20151008 14:54:25 <zubzub> still relevant though
20151008 14:54:52 <elect> wut
20151008 14:54:58 <elect> ah
20151008 15:10:48 * odin_ (~Odin@anon) Quit (Quit: Leaving)
20151008 15:11:32 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20151008 15:13:30 * xranby_f22 (~familjen@anon) Quit (Ping timeout: 246 seconds)
20151008 15:14:29 * xranby_f22 (~familjen@anon) has joined #jogamp
20151008 15:48:14 * doev (~doev@anon) has joined #jogamp
20151008 15:52:45 * elect (~GBarbieri@anon) Quit (Ping timeout: 255 seconds)
20151008 15:55:17 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20151008 16:02:55 * eclesia (~husky@anon) has left #jogamp
20151008 16:42:52 <sgothel> hehe red-pill
20151008 16:50:05 <sgothel> https://stackoverflow.com/questions/30192347/how-to-restore-a-window-with-xlib
20151008 16:50:26 <sgothel> _NET_ACTIVE_WINDOW reverses IconicState .. duh!
20151008 16:53:13 * monsieur_max (~maxime@anon) has joined #jogamp
20151008 16:57:52 * Eclesia (~eclesia@anon) has joined #jogamp
20151008 17:28:54 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20151008 17:41:52 * odin_ (~Odin@anon) has joined #jogamp
20151008 17:53:57 * monsieur_max (~maxime@anon) has joined #jogamp
20151008 18:13:51 <sgothel> @Xerxes: Bug 1249 <http://jogamp.org/git/?p=jogl.git;a=commit;h=dca5d36370ec5eb44998bae593880e3b10cc9a4e> PLEASE TEST
20151008 18:53:30 * netlike (~epyon@anon) Quit (Quit: Leaving)
20151008 20:32:04 <xranby_f22> sgothel: hi, will test, starting now
20151008 20:47:04 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20151008 20:59:12 <xranby_f22> sgothel: parenting tests now passes, i have viewed the computer perform the test and the visuals are good
20151008 20:59:44 <sgothel> still not reliable 1/10 fails ..
20151008 20:59:59 <xranby_f22> ther is 3 or 4 parenting tests that occasionally fail
20151008 21:00:03 <xranby_f22> here as well
20151008 21:00:13 <sgothel> grrr :)
20151008 21:00:16 <xranby_f22> i am testing with monitor configurations
20151008 21:00:26 <sgothel> well .. further analysis ..
20151008 21:00:27 <xranby_f22> to check if there is some logic behind
20151008 21:00:33 <sgothel> I even use TestGLWindows00
20151008 21:00:34 <xranby_f22> yes.. much improved!
20151008 21:00:49 <xranby_f22> le me test TestGLWindows00
20151008 21:00:54 <sgothel> I have 3 virt monitors attached
20151008 21:01:12 <sgothel> https://jogamp.org/chuck/view/fwd/job/jogl/1465/label=linux-x86_64-amd/testReport/com.jogamp.opengl.test.junit.newt.parenting/TestParenting01NEWT/test03bReparentWin2TopReparentNative/ <- also
20151008 21:03:41 <sgothel> _NET_WM_STATE_DEMANDS_ATTENTION <- last chance
20151008 21:06:22 <xranby_f22> test03bReparentWin2TopReparentNative occasionally fails here as well.. for me it usually fail on case 3 line 656 your buildbot fail on case 4 line 665
20151008 21:06:41 <sgothel> some setVisible(false) I assume?
20151008 21:06:58 <xranby_f22> yes on setVisible(false)
20151008 21:07:17 <sgothel> ok .. so any setVisible then
20151008 21:07:24 <xranby_f22> you buildbot fails on setVisible(true)
20151008 21:07:46 <xranby_f22> indeed
20151008 21:07:48 <sgothel> cruel world :)
20151008 21:08:08 <sgothel> now .. checking whether action is performed ..
20151008 21:08:30 <sgothel> so we may need to fake visibility signal .. if its not coming .. I hate it
20151008 21:14:13 * Eclesia emit a new theory : Unlike natural bugs, computer bugs lay eggs at the time of their death, therefore it may be more efficient to cage it and document it rather then kill it.
20151008 21:15:14 <xranby_f22> do you mean add a statemask.put(STATE_BIT_VISIBLE,visible) in waitForVisible #or actually issue a signal?
20151008 21:15:26 <xranby_f22> in case we timeout in waitForVisible
20151008 21:15:56 <xranby_f22> Eclesia: you are like my daughter, she say... why fix it when it is beautiful
20151008 21:15:58 <sgothel> we do receive ConfigureNotifty
20151008 21:16:29 <sgothel> if that does not satisfy .. we could fake it IFF the action is performed (obersvation)
20151008 21:16:49 <sgothel> that would mean .. the WM is not updating states .. but performs actions
20151008 21:17:12 <sgothel> IFF actions are not perfomed .. well, we have to not use that 'fancy hidden' thing
20151008 21:17:51 <sgothel> dunno beautiful here thought :-/
20151008 21:18:07 <sgothel> *though*
20151008 21:34:38 <sgothel> http://tronche.com/gui/x/xlib/ICC/client-to-window-manager/wm-hints.html <- another 'trick' urgency bit .. hmm
20151008 22:00:20 * i-make-robots (~i-make-ro@anon) has joined #jogamp
20151008 22:00:27 <i-make-robots> hello, again.
20151008 22:02:00 <sgothel> neither urgency nor 'demands attention' makes a difference ..
20151008 22:02:22 <sgothel> no seeing whether the action is performed, despite the state .. by deadlocking in WindowImpl wait for ..
20151008 22:02:28 <sgothel> hi robot maker
20151008 22:05:45 * i-make-robots salutes
20151008 22:06:43 <xranby_f22> hi!
20151008 22:11:47 <xranby_f22> testing reveals a list of junit test subdirectorys with always passing tests, such a list is good to give to distributions to have them validate that jogl work after compilation, all passing tests should keep passing
20151008 22:12:29 <xranby_f22> especially good to have if you try to port jogl to a new platform
20151008 22:12:56 <sgothel> failures per machine < ~6
20151008 22:26:04 <sgothel> ha! .. action is taken - at FocusOut probing WM_STATE shows its hidden ..
20151008 22:32:48 <xranby_f22> after i updated to this new nvidia beta driver .. newt.mm testing have started to display bad visual output. like a corrupt mousepointer after multimonitor mode switch
20151008 22:33:00 <xranby_f22> i need to reboot my machine and retest
20151008 22:33:28 <sgothel> until NV produces a 'stable normal driver' .. I would not digg too far into it
20151008 22:33:45 <sgothel> i.e. happy that ES 3.2 works .. done w/ that beta one
20151008 22:35:26 <Eclesia> good night ++
20151008 22:35:29 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20151008 22:35:33 <xranby_f22> ther are a lot of tests that pass flawlessly using the new driver like, graph, acore.anim, acore.ect acore.slels, awt.text, caps, es1.newt, es2.newt, es2.awt, gl2.awt, gl3.newt, gl4.awt, gl4.newt, glsl, math, jogl.newt, jogl.offscreen, jogl.perf, jogl.tile, jogl.util, jogl.util.texture
20151008 22:36:27 <xranby_f22> this is my list of allways passing tests :-*
20151008 22:36:29 <sgothel> sure .. just mean .. that after we detect a few issues w/ a beta driver .. maybe we shall relax before getting all crazy :)
20151008 22:36:51 <sgothel> Xerxes good test coverage list
20151008 22:38:34 <xranby_f22> yes, we should chill
20151008 22:39:19 <sgothel> we should get contacts at the vendors .. one client of us has it .. hmm
20151008 22:39:36 <sgothel> and they are interested somewhat or willing to run tests
20151008 22:39:59 <sgothel> now .. would be a good time to get a handle there as well, directly
20151008 22:40:15 <sgothel> so the 'one click unit test' feature shall soon be there
20151008 22:40:38 <sgothel> since we are a high profile project :)
20151008 22:41:41 <xranby_f22> indeed, also would be good to have contacts with the window manager teams
20151008 22:41:45 <sgothel> w/ query at FocusOut .. survived 100 visibility toggles :)
20151008 22:41:59 <xranby_f22> ohh sweet!
20151008 22:42:14 <sgothel> I will prepare that unit test package .. so they can handle it ..
20151008 22:42:36 <sgothel> yeah, that is a bug ..i.e. state become visible after ConfigureNotify, at FocusOut
20151008 22:42:49 <sgothel> while sending the event as a configure one ..
20151008 22:42:54 <sgothel> well
20151008 22:45:12 <xranby_f22> did you have large test teams at sun doing this kind of work?
20151008 22:46:01 <xranby_f22> all in all, this opens up good ways to collaborate across free software teams
20151008 22:46:39 <xranby_f22> ways to test gnu drivers, ways to test windowmanger implementations, ways to test distibutions
20151008 22:46:58 <xranby_f22> ways to test GPU drivers
20151008 22:51:09 <sgothel> nope, we were just a bunch of people .. some upstream projects were testing, yes - but .. TBH, it was not professional, i.e. deterministic at all, not auto-tests
20151008 22:51:37 <sgothel> so when I started adding jenkins and stuff - quality was seriously enhanced
20151008 22:52:04 <sgothel> junit tests .. since .. jogamp
20151008 22:53:16 <sgothel> at ATI the team did test ofc for the proprietary driver - synthetic stuff from games and apps, even comparing the resulting frames
20151008 22:54:10 <sgothel> hmm .. so this fixes the missed invisible at FocusOut ..
20151008 22:54:26 <sgothel> but .. we have also one missed visible -> true ..
20151008 22:57:28 * xranby_f22 (~familjen@anon) Quit (Quit: Leaving.)
20151008 22:59:31 * xranby_f22 (~familjen@anon) has joined #jogamp
20151008 23:17:39 <xranby_f22> multi monitor tests ran without visual issues after rebooting my machine
20151008 23:19:42 <xranby_f22> sgothel: all in all your current visibility fix implementation have fixed 8 failures on my machine compared to 4 days ago KUDOS!
20151008 23:20:36 <xranby_f22> and i look forward to test your current work in progress
20151008 23:35:54 <sgothel> almost there ..
20151008 23:36:11 <sgothel> great you have found this issue
20151008 23:55:08 <sgothel> http://jogamp.org/git/?p=jogl.git;a=commit;h=e418a665756af52fe2ca691ca220644e9b27c22c
20151008 23:55:20 <sgothel> pushed .. continuing w/ a bit more testing
20151008 23:55:58 <sgothel> src/newt/native/X11Common.h: enable the VERBOSE_ON <- if you encounter an issue
20151008 23:56:22 <sgothel> touch all natives or delete newt obj folder .. -> recompile
20151008 23:59:47 <sgothel> -Dnewt.debug.Window.visibility.failure.freeze <- notice this one :)
20151009 00:07:23 <sgothel> also use followup commit ..
20151009 00:07:31 <sgothel> hmm .. gnome seems to be fine now
20151009 00:07:39 <sgothel> KDE setVisible(true) issue ..
20151009 00:07:58 <sgothel> guess I have to put back the XMap* .. for visibility here ..
20151009 00:24:36 <sgothel> http://jogamp.org/git/?p=jogl.git;a=commit;h=ecdaa59aa48c16a558ec7a0b50df9419a9ccedc1 <- last one I guess :)
20151009 00:34:02 <sgothel> goody .. no failures here anymore
20151009 00:39:45 <sgothel> now closing 10th of browser tabs .. w/ API docs :)
20151009 01:27:17 * doev (~doev@anon) Quit (Ping timeout: 246 seconds)
20151009 01:27:28 * doev (~doev@anon) has joined #jogamp
20151009 01:36:58 <sgothel> http://www.zerohedge.com/news/2015-10-08/nato-talks-tough-troop-deployment-kremlin-calls-wests-bluff ..
20151009 01:37:28 <sgothel> lets see whether Turkey will be the new 'patsy' eh? gosh
20151009 01:38:17 <sgothel> (while they (Turkey) asked for EU/NATO's blessing to rage war against Kurds/PKK .. just great ..)
20151009 01:38:55 <sgothel> (blessing? no, as part of a deal to keep the many million refugees ..)
20151009 01:39:10 <sgothel> blackmail .. some would it call .. I guess
20151009 01:40:05 <sgothel> (besides asking to finally becoming part of EU, what a EU ..)
20151009 01:40:26 <sgothel> https://jogamp.org/chuck/view/fwd/job/jogl/1467/ <- I want to see some blue ..
20151009 01:43:02 <sgothel> http://www.zerohedge.com/news/2015-10-08/market-watching-fed-watches-market-surge-after-fearing-market-purge .. <- BTFD ? :)
20151009 01:44:54 <sgothel> https://www.youtube.com/watch?v=0akBdQa55b4 :)
20151009 01:45:00 <sgothel> old .. I know
20151009 01:47:42 <sgothel> yeah .. Build #1467 linux/X11 all blue .. goody
20151009 01:48:16 <sgothel> Bug 1249 (regression) fixed
20151009 01:51:27 <sgothel> btw .. Bug 1249 is a good example _why_ it is much better to hide the TK/WM implementation details within the native code, instead of exposing all of the crazy details to java.
20151009 01:52:32 <sgothel> the exposure and algo from java would surely increase latency of handling due to the high amount of java/jni calls needed for such an implementation
20151009 05:05:16 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20151009050516.html