JogAmp IRC Chat Logs

#jogamp @ irc.freenode.net - 20130130 05:05:42 (UTC)

20130130 05:05:42 -CatOut- Previous @ http://jogamp.org/log/irc/jogamp_20130129050528.html
20130130 05:05:42 -CatOut- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130130050542.html
20130130 09:18:06 * EricB1 (~Adium@anon) has joined #jogamp
20130130 09:25:10 * xranby1 (~xranby@anon) Quit (Read error: No route to host)
20130130 10:05:17 <sgothel> good morning - @Xerxes: Thx .. I got such a feeling w/ '-jar' .. but was too lazy to look it up :)
20130130 10:07:11 <sgothel> @Erik: How is the revised Bug 677 work going ? IMHO a clean commit would be nice. To do these things, Xerxes approach is best IMHO, one new branch for each feature. So you can flush it if this is not 'it' [yet].
20130130 10:11:31 * xranby (~xranby@anon) has joined #jogamp
20130130 10:26:53 <EricB1> @sgothel HiSven. I am trying to find the appropriate way to know wether or not the softKeyboard is visible. That sounds obvious, but it's not.
20130130 10:32:22 <sgothel> right - the only way NEWT 'knows' it .. is by relying on the toggle - well, it works
20130130 10:33:08 <sgothel> 'boolean Window::isKeyboardVisible()'
20130130 10:33:45 <sgothel> (sure only when using NEWT stuff)
20130130 10:34:11 <sgothel> if you find a reliable way - we can add this to NEWT Android WindowDriver
20130130 10:34:44 <sgothel> .. and add isKeyboardVisibleImpl() ..
20130130 10:35:20 <EricB1> I am digging this. There is a Configuration class that might hold the actual visibility of the keyboard. I'm testing this.
20130130 10:37:08 <sgothel> we currently trigger WindowImpl::keyboardVisibilityChanged(boolean visible)
20130130 10:37:23 <sgothel> w/ callback listener KeyboardVisibleReceiver
20130130 10:37:35 <sgothel> see WindowDriver (android) line 469
20130130 10:38:26 <sgothel> but is this only used while toggling it .. hide/visible
20130130 10:40:37 <sgothel> maybe this is enough related to those IME BACK event .. ?
20130130 10:40:47 <EricB1> If I cannot get the reliable information from the OS, I'll use keyboardVisibilityChanged.
20130130 10:40:52 <EricB1> BTW, about your last comment, returning true to notify that we handle the event, I am not sure it's the right thing to do. If MSurfaceView::dispatchKeyEventPreIme returns true, none of the regular system actions take place:
20130130 10:40:52 <EricB1> - Activity::onBackPressed() is not call anymore
20130130 10:40:52 <EricB1> - the softKeyboard does not hide
20130130 10:41:41 <EricB1> Since JOGL's key event is queued, there is no ay to know wether we are actually going to handle it or not.
20130130 10:41:44 <sgothel> so we can trigger both actions in our code - we do the same on other OS
20130130 10:42:08 <sgothel> i.e. window-closing -> we either destroy it or not ..
20130130 10:42:39 <sgothel> so we can hide the keyboard ourself .. (in this case we can return false ofc .. just a notification)
20130130 10:43:24 <sgothel> but for the BACK -> destroy, we need to capture it if WindowClosingProtocol says: DO_NOTHING_ON_CLOSE -> return true, otherwise we can return false as well - sure
20130130 10:44:23 <sgothel> in short: we only intercept and silence the signal if !softKeyboard && WindowClosingProtocol.DO_NOTHING_ON_CLOSE
20130130 10:46:24 <EricB1> Ok
20130130 10:50:37 <sgothel> don't worry about the details of sending out window / key listener stuff out, most crucial - as you stated - is whether we determine softKeyboard. I can help refining later.
20130130 10:52:30 <sgothel> @Xerxes: We will drive on Friday .. will takes us 5-6 hours ..
20130130 10:52:51 <sgothel> Train/Plane is too expensive and is not really faster
20130130 10:56:18 <sgothel> *Plain - what a shame that our train is like 5x more expensive than gasoline for car .. tsts
20130130 11:12:05 <sgothel> added node: macosx64-NV320M-10_7-jogamp-x64-jre7-sgothel-013 (same OSX machine, but using JRE7 for unit tests)
20130130 11:16:05 <xranby> sgothel: i will arrive on friday as well!
20130130 11:16:15 <xranby> i will arrive before lunch
20130130 11:16:36 <sgothel> afternoon'ish or so .. I guess, depends when we get up :)
20130130 11:16:44 <sgothel> then we can get together and do the slides :)
20130130 11:16:46 <xranby> good some time for dinner and crunch
20130130 11:16:57 <xranby> they have belgian waffles
20130130 11:17:05 <xranby> a lot of french resturants
20130130 11:17:14 <xranby> do any of you speak frensh?
20130130 11:17:27 <sgothel> nor really :) - but, we have Julien
20130130 11:18:14 <xranby> sgothel: how is the https://jogamp.org/bugzilla/show_bug.cgi?id=634 bug progressing?
20130130 11:18:31 <xranby> the debian java team is right now looking puzzed on how to package processing for debian
20130130 11:18:41 <xranby> since processing makes use of functionality after rc11
20130130 11:18:51 <xranby> introduced with 634
20130130 11:19:36 <xranby> the debian java team currently got jogamp rc11 in its experimental repository
20130130 11:19:41 <sgothel> uh .. AC is happy (see last comment) - so further enhancement is not crucial for them
20130130 11:20:49 <sgothel> the outstanding enhancement would be to add n-buffer FBO handling .. but that is just my personal thingy .. sort of, hence it's on the bottom of prio
20130130 11:21:10 <xranby> ok i posted the following response to debian java: http://lists.debian.org/debian-java/2013/01/msg00047.html
20130130 11:22:27 <sgothel> I see .. tomorrow there will be another release .. for FOSDEM
20130130 11:22:36 <xranby> good
20130130 11:22:43 <xranby> ok i better send pull requests then
20130130 11:22:49 <sgothel> yup
20130130 11:22:55 <sgothel> and I try to bring down more bugs :)
20130130 12:38:07 <sgothel> firefox plugin safety: URL about:config ; search plugins.click_to_play -> enable
20130130 13:19:59 <masterzen> sgothel: are you guys going to the fosdem beer event on Friday?
20130130 13:24:37 <xranby> [junit] ------------- Standard Output ---------------
20130130 13:24:38 <xranby> [junit] Available null device OpenAL Extensions:ALC_ENUMERATE_ALL_EXT ALC_ENUMERATION_EXT ALC_EXT_CAPTURE ALC_EXT_thread_local_context
20130130 13:24:38 <xranby> [junit] Available device OpenAL Extensions:ALC_ENUMERATE_ALL_EXT ALC_ENUMERATION_EXT ALC_EXT_CAPTURE ALC_EXT_disconnect ALC_EXT_EFX ALC_EXT_thread_local_context
20130130 13:24:38 <xranby> [junit] No extension ALC_SOFT_loopback present
20130130 13:26:08 <sgothel> we have to see - slides etc :)
20130130 13:26:16 <xranby> good news, the JOAL ALExt patch is almost complete. bad news, for some reason it fails to list one of the OpenAL-soft extensions found at the end of the OpenAl-soft extension list
20130130 13:26:23 <sgothel> @Xerxes: is that a good output .. ?
20130130 13:26:34 <sgothel> oh
20130130 13:27:06 <xranby> openal-soft/Alc/ALc.c: "ALC_EXT_thread_local_context ALC_SOFT_loopback";
20130130 13:27:52 <xranby> i was expecting ALC_SOFT_loopback to be seen after ALC_EXT_thread_local_context
20130130 13:28:02 <xranby> i will post the updated junit test that will dispplay this
20130130 13:28:11 <sgothel> ay
20130130 13:29:41 <xranby> pushed
20130130 13:30:00 <xranby> https://github.com/xranby/joal/tree/EFX
20130130 13:30:31 <sgothel> so I shall merge ?
20130130 13:30:54 <sgothel> currently fiddling w/ the OSX Java7 node ..
20130130 13:31:08 <sgothel> then I kick off builds again ..
20130130 13:31:49 <xranby> i had hoped to get this unittest to pass and find ALC_SOFT_loopback before the merge :/
20130130 13:32:37 <sgothel> as you wish - your call ..
20130130 13:33:04 <xranby> i need to test the existing joal tests to check if i have introduced any regression
20130130 13:33:06 <sgothel> you can ofc .. also do baby steps, file a bug report .. etc .. and let it fail
20130130 13:33:14 <sgothel> yup
20130130 13:37:12 <xranby> aha! problem solved
20130130 13:37:24 <xranby> [junit] Available null device OpenAL Extensions:ALC_ENUMERATE_ALL_EXT ALC_ENUMERATION_EXT ALC_EXT_CAPTURE ALC_EXT_thread_local_context ALC_SOFT_loopback
20130130 13:37:24 <xranby> [junit] Available device OpenAL Extensions:ALC_ENUMERATE_ALL_EXT ALC_ENUMERATION_EXT ALC_EXT_CAPTURE ALC_EXT_DEDICATED ALC_EXT_disconnect ALC_EXT_EFX ALC_EXT_thread_local_context ALC_SOFT_loopback
20130130 13:37:24 <xranby> [junit] begin testAlCLoopbackDeviceSOFT
20130130 13:38:42 <xranby> i am now facing the issue gouessej ran into, gluegen loaded the system install libopenal1 (that did not include all extensions) instead of the OpenAL-soft build and bundled by JogAmp
20130130 13:39:26 <xranby> sgothel: so how can we best introduce a switch to choose usage of system installed OpenAL or the one we bundle with the native joal jar?
20130130 13:40:31 <xranby> i simplu uninstalled my system installed libopenal1 and then i got the version build by JogAmp
20130130 13:40:33 <sgothel> kudos - good question, do we have something like capabilities ?
20130130 13:41:05 <sgothel> or shall we ignore the system openal lib ?
20130130 13:41:41 <sgothel> right now .. we favor the libopenal.so.1 I guess
20130130 13:42:33 <xranby> i really have no idea what the api users want :/
20130130 13:43:08 <xranby> gouessej want to be able to select it somehow
20130130 13:43:13 <sgothel> I thought 'native' openal is dead - we only use bundled openal-soft (which is not really soft -> uses other backends w/ native stuff)
20130130 13:43:38 <sgothel> so either use caps .. hmm .. or like a static initSingleton(boolean) ?
20130130 13:43:46 <sgothel> where default is -> soft
20130130 13:44:04 <xranby> yes i think default -> soft the one we bundle
20130130 13:44:32 <sgothel> for now, we may mark this as a bug .. i.e. lets get openal-soft rolling .. then fix that ..
20130130 13:44:43 <xranby> there is a bug marked
20130130 13:44:48 <sgothel> the DynamicBundle* something .. in JOAL selects the native stuff
20130130 13:45:03 <sgothel> i.e. favors libopenal.so.1 .. just comment it out
20130130 13:45:11 <xranby> https://jogamp.org/bugzilla/show_bug.cgi?id=662
20130130 13:45:17 <xranby> - Add a flag to favour OpenALSoft
20130130 13:45:22 <sgothel> yup
20130130 13:47:14 <xranby> sgothel: if i comment that line out src/java/jogamp/openal/ALDynamicLibraryBundleInfo.java how can it still find the version we bundle?
20130130 13:47:44 <sgothel> thats the magic .. it's in the JAR file
20130130 13:47:52 <sgothel> so it must be in the native jar file :)
20130130 13:48:33 <sgothel> nothing to do here .. all built-in
20130130 13:50:12 <xranby> nope.. Caused by: java.lang.RuntimeException: Couln't load native AL library
20130130 13:51:20 <sgothel> is libopenal.so in the native jar ?
20130130 13:51:35 <xranby> yes that is the bundled OpenAL-soft library
20130130 13:51:58 <sgothel> sven@risa:/usr/local/projects/JOGL/joal$ jar tf build-x86_64/joal-natives-linux-amd64.jar
20130130 13:51:58 <sgothel> META-INF/
20130130 13:51:58 <sgothel> META-INF/MANIFEST.MF
20130130 13:51:58 <sgothel> libjoal.so
20130130 13:51:58 <sgothel> libopenal.so
20130130 13:52:01 <sgothel> like that ?
20130130 13:52:28 <sgothel> i.e. that works already w. master branch
20130130 13:52:57 <xranby> yes exactly like that
20130130 13:53:46 <xranby> let me re-try and still include alLibNames.add("openal");
20130130 13:55:46 <xranby> ok that worked
20130130 13:56:54 <sgothel> yes .. ofc .. that must be in :) .. just exclude libopenal.so.1 .. and the windows derivate (dll) .. etc - sorry
20130130 13:57:34 <xranby> isPropertyDefined("joal.soft", false); then include the other jars?
20130130 13:57:41 <xranby> i mean include the other names
20130130 13:57:53 <sgothel> sure
20130130 13:58:14 <sgothel> but allow the JNLP aliasing .. guess last parameter .. so it also eats jnlp.joal.soft
20130130 13:58:37 <xranby> JNLP aliasing?=
20130130 13:58:40 <sgothel> .. Something in the DEBUG .. class I hope .. or PropertyAccess
20130130 13:59:02 <sgothel> yeah, if JNLP .. everything is considered insecure, but props starting w/ jnlp.
20130130 13:59:30 <sgothel> public static final boolean isPropertyDefined(final String property, final boolean jnlpAlias)
20130130 13:59:56 <sgothel> hope JOAL's Debug class has same, above is from GlueGen
20130130 14:00:09 <sgothel> both should derive from PropertyAccess
20130130 14:01:44 <xranby> it is isPropertyDefined(final String property, final boolean jnlpAlias)
20130130 14:03:07 <xranby> so.. hmm isPropertyDefined("joal.SystemOpenAL", true); please suggest a better name
20130130 14:03:31 <sgothel> good w/ me :)
20130130 14:03:53 <sgothel> you could do prefere system openal lib
20130130 14:04:11 <sgothel> i.e. if true: try system 1st, if fails openal-soft
20130130 14:04:17 <sgothel> otherwise: vice versa
20130130 14:04:59 <xranby> sgothel: i will favour openal-soft since it include the extensions needed by many of our users
20130130 14:05:01 <sgothel> joal.lib.openal.prefer=system ? .. well .. dunno
20130130 14:05:09 <sgothel> yes .. the default value
20130130 14:05:19 <sgothel> and julien can set prefere system ..
20130130 14:05:43 <sgothel> however .. they always try to fall back to the other not preferred case
20130130 14:13:21 <xranby> https://github.com/xranby/joal/commit/15583ca328c416bf664396025086a60afb34bd84
20130130 14:14:03 <xranby> let see maybe i have to ask for priveliges to use this
20130130 14:15:05 <xranby> oops broken commit
20130130 14:15:08 <sgothel> nope .. the PropertyAccess should handle that
20130130 14:15:17 <xranby> a missing )
20130130 14:15:27 <sgothel> however .. this is not the preferred logic .. i.e. fallback either way
20130130 14:21:42 <xranby> what is the preferred logic?
20130130 14:22:14 <sgothel> prefer openal-soft, fall-back to openal-system
20130130 14:22:28 <sgothel> option to say: prefer openal-system, fall-back to openal-soft
20130130 14:22:56 <sgothel> so the 1st found lib in that list is preferred
20130130 14:29:49 <xranby> ok https://github.com/xranby/joal/commit/89d54d1b311bd78be53bef26e89cc4e871351caf
20130130 14:36:56 <sgothel> cool .. shall I merge w/ master ?
20130130 14:37:46 <sgothel> i.e. your EFX -> master ?
20130130 14:38:15 <xranby> no :) i am now having fun with that loopback unittest
20130130 14:38:35 <sgothel> ayay
20130130 14:38:35 <xranby> if fails the bit compare comparing the recived buffer after passing it through the loop
20130130 14:39:04 <sgothel> I doubt I would hear it .. hmm old ears I have :)
20130130 14:39:41 <xranby> the nice thing about this openal-sofr extension is that it allows noise free junit testing :)
20130130 14:40:03 <sgothel> lol - don't worry, no speakers attached to nodes here :)
20130130 15:38:18 <xranby> sgothel: ok i have sent a pull request
20130130 15:38:25 <xranby> i think the work is reasonable
20130130 15:38:35 <xranby> the two manual tests we have pass
20130130 15:38:53 <xranby> and the -Djoal.SystemOpenAL flag work
20130130 15:39:02 <xranby> java -Djoal.SystemOpenAL=false -cp joal/build-i586/joal-test.jar:joal/build-i586/joal.jar:gluegen/build/gluegen-rt.jar com.jogamp.openal.test.manual.OpenALTest
20130130 15:39:13 <xranby> java -cp joal/build-i586/joal-test.jar:joal/build-i586/joal.jar:gluegen/build/gluegen-rt.jar com.jogamp.openal.Sound3DTest
20130130 15:39:14 <sgothel> kudos - great! .. I am out for a few hours .. then I get back to this .. and test
20130130 15:39:27 <sgothel> probably adding it to our nodes .. ofc!
20130130 15:40:07 <xranby> to clarify test using -Djoal.SystemOpenAL or no property defined
20130130 15:40:34 <xranby> if someone try -Djoal.SystemOpenAL=false <- wrong usage its equal to -Djoal.SystemOpenAL
20130130 15:40:52 <xranby> i will catcha bus now
20130130 15:41:40 <xranby> i have only tested on 32bit machines so yeah kudos hopefully there is no 64bit bugs
20130130 15:42:01 <xranby> but there might well be since gluegen generate different code for 64bit
20130130 16:48:23 * xranby1 (~familjen@anon) has joined #jogamp
20130130 17:02:50 * EricB1 (~Adium@anon) Quit (Quit: Leaving.)
20130130 17:44:53 <xranby1> http://forum.jogamp.org/JOAL-pullrequest-tp4027825p4028070.html
20130130 18:02:10 * hharrison (~chatzilla@anon) has joined #jogamp
20130130 18:40:37 <sgothel> perfect - @XRanby - I will merge before next build .. i.e. before I go to bed .. and try the cmake on nodes
20130130 18:45:45 * DemoscenePassiv (~Lutsche@anon) has joined #jogamp
20130130 18:54:10 <xranby1> sgothel: thumbs up, i ran a quick self-review:
20130130 18:54:14 <xranby1> JOAL will now map all extensions from the C alext.h header to be found under com.jogamp.openal.ALExt
20130130 18:54:14 <xranby1> This is an API change, before some of the extensions was found under com.jogamp.openal.AL .
20130130 18:54:14 <xranby1> Please scream if we should still support the old API. Peek at this thread for an example of code that was working using the old API on some platforms: http://forum.jogamp.org/alGenEffects-Not-Available-tp4027019p4028072.html
20130130 18:56:25 <sgothel> nope - it's a bug fix :)
20130130 18:57:25 <sgothel> I fix the OSX/Java7 node now . then head towards JOAL
20130130 18:58:45 <xranby1> OK good
20130130 18:59:34 <xranby1> how about this one
20130130 18:59:37 <xranby1> includes/openal/efx.h:ALboolean alIsEffect( ALuint eid );
20130130 18:59:58 <xranby1> gluegen generated
20130130 18:59:59 <xranby1> build-i586/gensrc/classes/com/jogamp/openal/ALExt.java: /** Entry point (through function pointer) to C language function: <br> <code> ALboolean alIsEffect(ALuint effect); </code> */
20130130 18:59:59 <xranby1> build-i586/gensrc/classes/com/jogamp/openal/ALExt.java: public byte alIsEffect(int effect);
20130130 19:00:15 <xranby1> thus the ALboolran got mapped to a byte
20130130 19:00:23 <xranby1> ALboolean
20130130 19:00:24 <sgothel> yes .. you need to make it opaque ..
20130130 19:00:54 <sgothel> like: jogl/make/config/jogl/gl3-desktop.cfg:Opaque boolean Bool
20130130 19:01:11 <sgothel> i.e. Opaque boolean ALboolean
20130130 19:01:36 <sgothel> just peek in the jogl gluegen configs .. and/or shout here .. me helping out always :)
20130130 19:01:58 <xranby1> roger..
20130130 19:02:00 <xranby1> joal/make$ grep Opaque *
20130130 19:02:00 <xranby1> joal-alc.cfg:Opaque boolean ALCboolean
20130130 19:02:00 <xranby1> joal-alext.cfg:Opaque boolean ALCboolean
20130130 19:02:00 <xranby1> joal.cfg:Opaque long ALproc
20130130 19:02:00 <xranby1> joal.cfg:Opaque boolean ALboolean
20130130 19:02:23 <xranby1> looks like it was simply missing in joal-alext.cfg
20130130 19:02:49 <sgothel> Include joal-common.cfg
20130130 19:02:49 <sgothel> ?
20130130 19:03:03 <xranby1> that might be a good refactor
20130130 19:07:28 <xranby1> sgothel: i have added the Opaque boolean ALboolean to the pull request
20130130 19:07:30 <xranby1> looks better
20130130 19:07:43 <xranby1> and work better
20130130 19:12:56 * DemoscenePassiv (~Lutsche@anon) Quit (Ping timeout: 245 seconds)
20130130 19:38:49 * xranby1 (~familjen@anon) Quit (Quit: Leaving.)
20130130 20:04:12 * EricB1 (~Adium@anon) has joined #jogamp
20130130 20:20:11 * EricB1 (~Adium@anon) Quit (Quit: Leaving.)
20130130 20:21:15 * ebrayet (~Adium@anon) has joined #jogamp
20130130 20:24:30 * ebrayet is now known as Foetus
20130130 20:24:44 * Foetus is now known as ebrayet
20130130 20:25:23 * ebrayet is now known as Foetus
20130130 20:45:45 * xranby1 (~familjen@anon) has joined #jogamp
20130130 20:46:12 <sgothel> fixing joal build (types) .. and working on nodes incl. openal-soft git branch ..
20130130 20:46:53 <xranby1> kudos
20130130 20:47:25 * xranby1 focus some more on that raspberry pi keyboard input thing
20130130 20:47:50 <sgothel> yup .. nice
20130130 23:22:43 <xranby1> sgothel: updated the NEWT-input JOGL pullrequest, added a manager to monitor all /dev/input/event* files and handle hot-swap of usb devices
20130130 23:24:03 <xranby1> will now focus on adding the most common 104 keys
20130131 00:03:34 * Foetus (~Adium@anon) Quit (Quit: Leaving.)
20130131 00:29:29 * Titanoum (~BitcoinFo@anon) Quit (Ping timeout: 255 seconds)
20130131 00:39:16 * hharrison (~chatzilla@anon) Quit (Quit: ChatZilla 0.9.89 [Firefox 19.0/20130116072953])
20130131 00:45:25 <xranby1> good night,
20130131 00:54:12 * xranby1 (~familjen@anon) Quit (Quit: Leaving.)
20130131 03:16:26 * rmk0 (~rmk0@anon) Quit (Ping timeout: 255 seconds)
20130131 05:05:55 -CatOut- Continue @ http://jogamp.org/log/irc/jogamp_20130131050555.html