#jogamp @ irc.freenode.net - 20131210 05:05:57 (UTC)


20131210 05:05:57 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20131209050556.html
20131210 05:05:57 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20131210050557.html
20131210 05:11:54 * hija (~hija@anon) Quit (Quit: hija)
20131210 05:41:50 * kermyt (~kermyt@anon) Quit (Ping timeout: 246 seconds)
20131210 06:12:16 * xranby (~xranby@anon) Quit (Ping timeout: 246 seconds)
20131210 06:16:19 * kermyt (~kermyt@anon) has joined #jogamp
20131210 06:24:45 * kermyt (~kermyt@anon) Quit (Ping timeout: 272 seconds)
20131210 06:26:16 * xranby (~xranby@anon) has joined #jogamp
20131210 06:40:11 * kermyt (~kermyt@anon) has joined #jogamp
20131210 06:57:20 * kermyt (~kermyt@anon) Quit (Ping timeout: 272 seconds)
20131210 08:05:57 * monsieur_max (~maxime@anon) has joined #jogamp
20131210 08:16:10 * eclesia (~husky@anon) has joined #jogamp
20131210 08:16:17 <eclesia> good morning
20131210 08:30:00 <xranby> hija: /System/Library/Frameworks/gluegen-rt.Framework/gluegen-rt <-- this indicate you may still have an old Apple provided copy of java3d gluegen-rt inside your system folder, remove it
20131210 08:30:15 <xranby> based on the unmaintained code
20131210 08:30:35 <xranby> repository hosted by apple/oracle/sun
20131210 08:31:41 <xranby> file a bug to apple if their uninstaller forgot to remove these files
20131210 08:32:42 <xranby> good morning
20131210 08:46:41 * kermyt (~kermyt@anon) has joined #jogamp
20131210 09:25:14 <monsieur_max> hi people :)
20131210 09:27:04 <monsieur_max> i found a bug occuring under windows, a # sign into the path would prevent jogamp to work
20131210 09:28:38 <eclesia> same for ':' . windows limitations
20131210 09:28:45 <monsieur_max> if i recall correctly, this was working with one of the 2.1.2 rc ( or maybe before )
20131210 09:29:05 <monsieur_max> but with the 2.1.2, it stopped
20131210 09:29:20 <monsieur_max> eclesia: the problem is that it was working before ...
20131210 09:46:24 <xranby> monsieur_max: is this bug: https://jogamp.org/bugzilla/show_bug.cgi?id=902 ?
20131210 09:47:56 <xranby> are you using IOUtil.toURISimple(new File(String filepath)) to generate the URI?
20131210 09:48:52 <monsieur_max> xranby: well no, jogamp would simply crash on startup, nothing to do with my code
20131210 09:49:02 <monsieur_max> ( at least it seems so )
20131210 09:49:34 <xranby> try run the movie cube example using the -file argument
20131210 09:49:53 <xranby> and check if it can open a file from a path contining #
20131210 09:51:01 <xranby> monsieur_max: is jogamp located inside a path containing # or is the file locaten inside a path containing # ?
20131210 09:51:49 <monsieur_max> jogamp
20131210 09:52:21 <xranby> oh then it may be a bug in gluegen
20131210 09:53:04 <monsieur_max> xranby: http://pastebin.com/cY2Ce17X
20131210 09:53:40 <monsieur_max> to me, it seems so
20131210 09:55:04 <xranby> monsieur_max: please try the current 2.1.3rc
20131210 09:55:20 <xranby> it contains a lot of gluegen fixes for the uri's
20131210 09:55:30 <monsieur_max> ok
20131210 09:55:36 <monsieur_max> i'll give it a try
20131210 09:55:42 <monsieur_max> ( in a few hours )
20131210 09:55:47 <xranby> http://jogamp.org/git/?p=gluegen.git;a=summary
20131210 09:56:04 <xranby> http://jogamp.org/deployment/archive/master/gluegen_753-joal_504-jogl_1170-jocl_887-signed/archive/
20131210 09:59:13 <xranby> it may be that we need to do reverse %23 -> # similar to how we reverse #20 -> space
20131210 09:59:19 <xranby> %20 -> space
20131210 10:00:02 <xranby> all in all it is an uri to filesystem namespace conversion problem
20131210 10:00:47 <xranby> some chars cant be found in RFC 2396 and some chars are not allowed in filesystems
20131210 10:07:48 <xranby> monsieur_max: the reason why jogamp switched from using URL's to URI is to reduce network lookups working with URL had some serve sideeffects
20131210 10:07:51 <xranby> like: http://michaelscharf.blogspot.se/2006/11/javaneturlequals-and-hashcode-make.html
20131210 10:08:02 <xranby> java.net.URL.equals and hashCode make (blocking) Internet connections....
20131210 10:08:50 <xranby> by using URI's we gain faster bootup time and no slow internet lookups
20131210 10:09:11 <xranby> but we have to work around all the URI namespace issues
20131210 10:09:17 <xranby> thus in the end it will be good (tm)
20131210 10:10:08 <monsieur_max> haha :) ok thanks for the insight
20131210 10:10:57 * hija (~hija@anon) has joined #jogamp
20131210 10:41:27 <xranby> hija: hi!
20131210 10:41:36 <hija> hey xranby
20131210 10:41:41 <xranby> (09:30:03) xranby: hija: /System/Library/Frameworks/gluegen-rt.Framework/gluegen-rt <-- this indicate you may still have an old Apple provided copy of java3d gluegen-rt inside your system folder, remove it
20131210 10:41:42 <xranby> (09:30:18) xranby: based on the unmaintained code
20131210 10:41:42 <xranby> (09:30:39) xranby: repository hosted by apple/oracle/sun
20131210 10:41:42 <xranby> (09:31:45) xranby: file a bug to apple if their uninstaller forgot to remove these files
20131210 10:42:05 <hija> oh, didnt see that
20131210 10:42:21 <xranby> hija: apple used to include the outdated sun java3d with their jre installations
20131210 10:43:01 <hija> well its fixed now anyways
20131210 10:43:11 <hija> I don't know what caused it exactly
20131210 10:43:41 <sgothel> good morning ..
20131210 10:43:44 <hija> I think eclipse was getting confused because I had 2 jogls and 2 glugens in my workspace
20131210 10:43:49 <hija> morning sven
20131210 10:43:54 <sgothel> yes - the '#' is out of bounds for URIs !
20131210 10:44:19 <sgothel> b/c '#' is the anchor .. and really confuses URI handling (java classes)
20131210 10:45:47 <sgothel> anymore SWT patches in the pipe for this week ?
20131210 10:47:14 * xranby (~xranby@anon) Quit (Ping timeout: 272 seconds)
20131210 10:50:40 * xranby (~xranby@anon) has joined #jogamp
20131210 12:36:25 <xranby> monsieur_max: https://jogamp.org/bugzilla/show_bug.cgi?id=908
20131210 12:37:20 <monsieur_max> oh ok ...
20131210 12:37:57 <xranby> thus is looks like a jre issue..
20131210 12:38:11 <xranby> it may be an idea to fix this in openjdk
20131210 12:38:48 * rmk0_ is now known as rmk0
20131210 12:38:59 <xranby> in ordet to allow the File constructor to work
20131210 12:39:01 <xranby> java.lang.IllegalArgumentException: URI has a fragment component at java.io.File.<init>(Unknown Source)
20131210 12:39:18 * rmk0 is now known as Guest41901
20131210 12:39:36 <monsieur_max> let's say that i'm not totally happy with this ending
20131210 12:40:03 <xranby> monsieur_max: the problem is not inside jogamps codebase
20131210 12:40:08 <monsieur_max> yeah, i know :)
20131210 12:40:59 <monsieur_max> i don't blame you guys, first because this is not your fault, and because you're doing an awesome job
20131210 12:41:11 <monsieur_max> but the fact is that it'll raise issues
20131210 12:41:39 <monsieur_max> you can't prevent users to act strangely and put damn # signs in their folder names
20131210 12:41:41 <xranby> # is reserved by URI & URL logic to do something.html#fragment
20131210 12:41:44 <xranby> hmm
20131210 12:41:48 <monsieur_max> yep
20131210 12:42:08 <monsieur_max> at first i was quite suprirsed to see that you can actually enter this signe in a folder name
20131210 12:42:12 <monsieur_max> *sign
20131210 12:42:19 * Guest41901 (~rmk0@anon) Quit (Quit: leaving)
20131210 12:42:45 * rmk0_ (~rmk0@anon) has joined #jogamp
20131210 12:45:06 <xranby> http://tools.ietf.org/html/rfc3986#page-16
20131210 12:54:19 <xranby> monsieur_max: try use the lastest 2.1.3rc in any case
20131210 12:54:31 <xranby> monsieur_max: actually # work on my linux system
20131210 12:55:14 <monsieur_max> i'll give it a try tonight, i'm at work right now
20131210 12:55:38 <monsieur_max> no problem for me to use the latest version of jogamp, the game is not released yet :)
20131210 12:55:45 <monsieur_max> latest is goodest
20131210 12:55:51 <monsieur_max> ;)
20131210 12:59:50 * rmk0_ (~rmk0@anon) Quit (Quit: leaving)
20131210 13:01:12 * rmk0 (~rmk0@anon) has joined #jogamp
20131210 14:28:38 * xranby (~xranby@anon) Quit (Ping timeout: 240 seconds)
20131210 14:30:01 * xranby (~xranby@anon) has joined #jogamp
20131210 14:45:58 * xranby (~xranby@anon) Quit (Ping timeout: 240 seconds)
20131210 14:48:54 * xranby (~xranby@anon) has joined #jogamp
20131210 15:06:37 * bbbruce (~bx@anon) Quit (Ping timeout: 272 seconds)
20131210 15:18:19 * xranby (~xranby@anon) Quit (Ping timeout: 240 seconds)
20131210 15:33:22 * xranby (~xranby@anon) has joined #jogamp
20131210 15:42:59 * xranby (~xranby@anon) Quit (Read error: Operation timed out)
20131210 15:43:36 * xranby (~xranby@anon) has joined #jogamp
20131210 15:51:10 <[Mike]> sgothel: Did you see my message? :)
20131210 16:10:00 * bbbruce (~bx@anon) has joined #jogamp
20131210 16:56:53 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20131210 17:04:03 * eclesia (~husky@anon) has left #jogamp
20131210 18:02:39 * monsieur_max (~maxime@anon) has joined #jogamp
20131210 18:22:29 * kermyt (~kermyt@anon) Quit (Ping timeout: 248 seconds)
20131210 18:23:41 * kermyt (~kermyt@anon) has joined #jogamp
20131210 18:56:09 * hija (~hija@anon) Quit (Quit: hija)
20131210 19:47:42 * xranby (~xranby@anon) Quit (Read error: Operation timed out)
20131210 19:58:38 * kermyt (~kermyt@anon) Quit (Ping timeout: 246 seconds)
20131210 20:01:23 * kermyt (~kermyt@anon) has joined #jogamp
20131210 20:03:58 * xranby (~xranby@anon) has joined #jogamp
20131210 20:55:50 * xranby (~xranby@anon) Quit (Ping timeout: 245 seconds)
20131210 21:02:25 * xranby (~xranby@anon) has joined #jogamp
20131210 21:22:05 * xranby (~xranby@anon) Quit (Ping timeout: 245 seconds)
20131210 21:25:29 * xranby (~xranby@anon) has joined #jogamp
20131210 21:57:42 * hija (~hija@anon) has joined #jogamp
20131210 22:08:58 <[Mike]> on a side note I have a patch for https://jogamp.org/bugzilla/show_bug.cgi?id=362
20131210 22:10:52 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20131210 22:15:31 <[Mike]> just have to cleanup my test-case. Besides compression, the size calculation is incorrect for a mipmap sequence of 2x1 -> 1x1, For RGB 2x1 mipmap has the size of 6bytes and the 1x1 of 3bytes. Since the next mipmap size calculation uses 4 as its divisor: 6/4 != 3 (Mipmap level 6 didn't match expected data size (expected 1, got 3)).
20131210 22:27:01 * void256 (~void@anon) has joined #jogamp
20131210 23:05:00 * kermyt (~kermyt@anon) Quit (Ping timeout: 245 seconds)
20131210 23:07:39 * kermyt (~kermyt@anon) has joined #jogamp
20131210 23:58:27 <sgothel> @Mike .. Hi, which note / message for me ?
20131210 23:59:21 <sgothel> I appreciate your fix, w/ unit test - yes, please. thank you, great.
20131211 00:00:03 <[Mike]> Hi sven, its about commit 538a4184 "Fix NEWT WindowImpl reparent-recreate w/ GLEventListenerState: Bug introduced w/ commit e2506d7"
20131211 00:00:24 <sgothel> and .. where is that message hidden ?
20131211 00:00:43 <sgothel> (me not doing github review .. need a personal email at least)
20131211 00:00:55 <[Mike]> oh I just messaged here on irc
20131211 00:01:19 <sgothel> you know our log ? then pls give me the URL :)
20131211 00:01:40 <sgothel> hint: the time in the log is an URL to itself
20131211 00:02:22 <[Mike]> pretty cool http://jogamp.org/log/irc/jogamp_20131209050556.html#l134
20131211 00:05:01 <[Mike]> I didn't fully explain myself as I wanted to discuss it a little. Mainly I am unsure the correct functionality.
20131211 00:05:26 <sgothel> so this is about wasVisible state in WindowImpl ?
20131211 00:06:29 <[Mike]> yes and NewtCanvasAWT reparenting
20131211 00:07:24 <[Mike]> the wasVisible will never be true as the newtCanvas sets its child visibility to false.
20131211 00:07:30 <[Mike]> before the reparent
20131211 00:07:46 <sgothel> right
20131211 00:09:13 <[Mike]> before the commit WindowImpl.java:1095 read as destroy (true), now it reads as destroy (wasVisible) which will always be false.
20131211 00:10:25 <sgothel> which is OK, nothing to preserve ..
20131211 00:10:42 <[Mike]> yes. I need it to preserve though :)
20131211 00:10:51 <sgothel> duh .. now I get what you mean ..
20131211 00:10:56 <sgothel> nice, pls file a bug
20131211 00:11:18 <[Mike]> Will do. Thanks!
20131211 00:11:21 <sgothel> guess we need the preserve determination in GLWindow itself
20131211 00:11:24 <sgothel> good call
20131211 00:11:36 <sgothel> i.e. not relying on 'wasVisible' ..
20131211 00:11:54 <sgothel> isNativeValid() is ok though ..
20131211 00:12:02 <[Mike]> yes
20131211 00:12:18 <sgothel> have to think why I chose this one, and not simply test whether the GLContext is natively valid ..
20131211 00:12:29 <sgothel> probably due to late hours work (then .. or now :)
20131211 00:13:50 <sgothel> which destroy(..) call in reparent is it (in todays source) ? line ?
20131211 00:14:31 <sgothel> 1339 ?
20131211 00:15:27 <sgothel> do you have a unit test ?
20131211 00:15:39 <sgothel> or can you make one based on the many we have ?
20131211 00:16:11 <sgothel> (pls state the 'trigger' of reparent destruction - OSX off-/onscreen via NEWTCanvasAWT ?)
20131211 00:16:35 <[Mike]> yes today's source: 1301
20131211 00:17:42 <sgothel> ^^ and the rest of my Qs .. pls
20131211 00:19:25 <[Mike]> No unit test, I will create one. The trigger seems to be on multiple monitors, swapping out one GLWindow for another in NEWTCanvasAWT.
20131211 00:19:53 <sgothel> so it's 1301 + 1339 for full reparenting w/ NEWT assuming forceDestroyCreate==true (i.e. OSX on-/offscreen)
20131211 00:20:06 <[Mike]> Also using windows
20131211 00:20:10 <[Mike]> win7
20131211 00:20:17 <[Mike]> No osx :)
20131211 00:20:42 <sgothel> forceDestroyCreate: how come ?
20131211 00:20:55 <[Mike]> I think forceDestroyCreate is false
20131211 00:21:01 <sgothel> NewtFactory.isScreenCompatible(newParentWindow, screen) ? really ?
20131211 00:21:27 <sgothel> i.e. on Windows it's always the same .. hmm
20131211 00:21:51 <[Mike]> yes forceDestroyCreate is false.
20131211 00:22:31 * hija (~hija@anon) Quit (Quit: hija)
20131211 00:23:06 <[Mike]> I use a laptop to develop, setting an externally connected monitor as the primary device causes the incompatibility.
20131211 00:23:24 <sgothel> hmm :)
20131211 00:23:29 <sgothel> ok .. so the index then
20131211 00:23:44 <sgothel> good we can test w/ the 'force' flag
20131211 00:24:02 <sgothel> pls elaborate on all this ..
20131211 00:24:24 <sgothel> then I will do the fix today/tomorrow
20131211 00:24:37 <sgothel> today .. it's already tomorrow :)
20131211 00:24:40 <[Mike]> Sure I can also attach the newt/jogl logs.
20131211 00:24:57 <sgothel> great thx
20131211 00:25:24 <[Mike]> Still today here, tomorrow shortly.
20131211 00:25:31 <[Mike]> :)
20131211 00:26:47 <sgothel> didn't think thats possible in Europe .. hmm west spain/france .. memories :)
20131211 00:37:49 <sgothel> hmm, we check native creation in GLAutoDrawableBase: if( null != context && context.isCreated() ) { ..
20131211 00:37:55 <sgothel> have to think about that ..
20131211 00:39:31 <sgothel> right, it's more the case of "will become visible again (wasVisible)."
20131211 00:39:39 <sgothel> i.e. not whether it _is_ valid ..
20131211 00:40:24 <sgothel> hence need to inject some logic, maybe a hint - or not set invisible in NewtCanvasAWT .. hmm
20131211 00:51:07 * void256 (~void@anon) Quit (Remote host closed the connection)
20131211 00:52:33 * hija (~hija@anon) has joined #jogamp
20131211 00:57:12 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=915 < @Mark What shall we do ? Closing this one .. or working hard for a workaround ?
20131211 00:57:20 <sgothel> is this already filed w/ Mesa ?
20131211 00:57:38 <sgothel> Or can we say, the desktop GL3 ctx is 'good enough' ?
20131211 00:57:51 <sgothel> Renamed to: BugĀ 915 - GLES2 and GLES3 contexts rejected on Mesa/AMD where ES3 is available since 2.1.1
20131211 00:58:40 <sgothel> I do favor the latter (closing, desktop GL3 good enough - will hopefully be fixed in Mesa/AMD)
20131211 00:59:04 * hharrison (~chatzilla@anon) has joined #jogamp
20131211 00:59:15 <sgothel> Good morning!
20131211 01:00:29 <hharrison> good evening!
20131211 01:00:49 <hharrison> Did you get my pull request? (www fixup)
20131211 01:01:14 <sgothel> ah .. jogl .. forgot, will do before going to bed, thx
20131211 01:02:13 <hharrison> Not a big deal, just nice I finally have a real website to point at
20131211 01:02:34 <hharrison> and a much more interesting screenshot
20131211 01:13:54 <hharrison> anyways, of for now
20131211 01:13:59 * hharrison (~chatzilla@anon) Quit (Quit: ChatZilla 0.9.90.1 [Firefox 26.0/20131202182626])
20131211 01:48:54 * hija (~hija@anon) Quit (Quit: hija)
20131211 01:52:37 <sgothel> @Mike: I am almost done w/ the patch .. bugreport ?
20131211 02:22:05 * xranby (~xranby@anon) Quit (Ping timeout: 245 seconds)
20131211 02:27:52 * xranby (~xranby@anon) has joined #jogamp
20131211 02:28:17 <sgothel> @Xerxes: Can you file a bug report for your issue you posted here .. will test later
20131211 03:11:13 <[Mike]> sgothel: I had to step away (dinner). Creating the bug report now.
20131211 04:36:12 <sgothel> you may pull, compile and test ..
20131211 04:36:28 <sgothel> currently doing a local test here .. will build later
20131211 04:36:39 <sgothel> (i.e. jenkins build/test)
20131211 04:37:35 <[Mike]> will do
20131211 04:46:07 <sgothel> no regressions here (local test on linux)
20131211 05:05:57 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20131211050557.html