Changes

AWT GLCanvas: More strict GLDrawable realization [on AWT-EDT], skip if creation is not possible on AWT-EDT.

The Intel HD3000 OpenGL driver on Windows will deadlock @ SwapBuffers in case the drawable is created
on a thread other than the window owner thread.

We are aware of such possibilities, nevertheless the AWTEDTExecutor.singleton.invoke(..)
allowed to execute the runnable in case it cannot be invoked on AWT-EDT.
The latter is the case if the current thread is not the AWT-EDT _and_ is holding the AWT tree-lock.

With GlueGen commit 0b43b43f889ad7fc220942b0076e2001ca3cf13f, the invoke method now consumes
an argument allowing to restrict the execution to AWT-EDT only.

In such case, the drawable will be realized at a later time from the AWT-EDT.

Such a situation could be triggered if a Frame's setVisible(true)
is not issued from the AWT-EDT, as it should be!
However, to relax such use cases - we better recognize such possible dealock and avoid it.

Adapt change of commit d93c5d23e304ea20e868595748f92a5bef4f5703 to unit test (GLCanvas realization point)

Fix Bug 658 (Mesa 9.0 3.1 Intel compat quirk, 3.1 core only) ; No PROFILE_ALIASING compat -> core ; Fix setGLFunctionAvailability(..) failure path @ profile query

- Add GLRendererQuirks.GLNonCompliant, marking a GL context/profile non compliant.
  Currently: 'Mesa DRI Intel(R) Sandybridge Desktop' && 3.1 compat profile

- Fix Bug 658 (Mesa 9.0 3.1 Intel compat quirk, 3.1 core only)
  Detect case using new GLRendererQuirks.GLNonCompliant in setGLFunctionAvailability() and return 'false'.

- No PROFILE_ALIASING compat -> core
  Use true core GL profiles / context if available to ensure
  proper API behavior across platforms due to different functionality.
  E.g. don't use GL3bc if GL3 is requested.

- Fix setGLFunctionAvailability(..) failure path @ profile query
  Destroy temp context & zero result to cont. iterating through GL versions.
  This missing cleanup lead to returning the faulty GL context handle and it's mapping/usage.

Cleanup / Simplify: setGLFunctionAvailability(..) / createContextARBVersions(..) GL Version Validation

String or integer based GL version validation now happens in setGLFunctionAvailability(..)
depending on the requested profile.

Due to the 'strictMatch' argument the method fails early when unsatisfied
also allowing to simplify createContextARBVersions(..) implementation.

setGLFunctionAvailability(..): Unify GLInt and GLString version validation; Refines commit b6ae4e4dcbd740dd57de9dc3280d943e98cdaa76

- Limit getGLIntVersion() w/ non ARB ctx to 3.0 - similar to getGLVersionNumber(.., String).

- GLInt only proceeds w/ valid version

- Always test GLInt version before GLString version,
  cont. w/ GLString version if GLInt failed.

- Fail if strictMode and GLInt failed, but GLString reports GL >= 3.0

- Fail is strictMode and neither GLInt nor GLString could validate version

- Clear glError before succeeding (could be tainted by GLInt version test)

SharedResourceRunner: Always dump Exception if catched @ creation, not just in DEBUG case.

This seems to be crucial, since otherwise user is left in the dark w/o enabling debug flags.

Bug632: Test NEWT Child Window Translucency (X11/Windows) ..

- Windows: Child window is not translucent at all

- X11: Child window is translucent to parent's background,
       however - parents content is _not_ 'composed in'.

- TODO: Find whether there is a solution or not.

- Note: The child window does not change it's rel. position
        if parent moves! This is a feature, since we don't
        have impl. a layout.

NEWT WindowsWindow.c: UpdateInsets: Fix determiniation of isUndecorated - WS_SYSMENU is _not_ an indication!

Enhance PMVMatrix: Use FloatStack to implement PushMatrix and PopMatrix (Bug 657)

One FloatStack for each, MV, P and T, with initialialSize zero to save memore
and growSizes: MV = 16 arrays, P = 2 arrays, T = 2 arrays,
with array = 16 floats.

This shall save performance due to the preallocated stack when used
and growing only in the above mentioned grow intervals.

We may finetune the growSize, if required.

Update PNGJ 0.85 -> 1.12 (w/ interlace read support) ; Added PNG Interlace read tests (TestPNGTextureFromFileNEWT)

Fix GLArrayDataClient createGLSL(..) 'name' setting; Also adds equal behavior for Client/Server FFP usage.

- GLArrayDataWrapper's init() only validated and set the 'name' for VBO ARRAY_BUFFER usage and validates it,
  regardless of FFP or GLSL usage.

- GLArrayDataWrapper's init() included the no-VBO case to the above 'name' set/validate code path.

GLAutoDrawable/AnimatorBase: Add ExclusiveContextThread (ECT) feature; AnimatorBase: Add setModeBits/MODE_EXPECT_AWT_RENDERING_THREAD; FPSAnimator: Make transactions deterministic.

ExclusiveContextThread (ECT) allows user to dedicate a GLContext to a given thread.
Only the ECT will be allowed to claim the GLContext, hence releasing must be done on the ECT itself.

The core feature is accessible via GLAutoDrawable, while it can be conveniently enabled and disabled
via an AnimatorBase implementation. The latter ensures it's being released on the ECT and waits for the result.

Note that ECT cannot be guaranteed to work correctly w/ native (heavyweight) AWT components
due to resource locking and AWT-EDT access. This is disabled in all new tests per default and
noted on the API doc.

Note: 'Animator transaction' == start(), stop(), pause(), resume().

- Add ExclusiveContextThread (ECT) feature
  - GLAutoDrawable NEW:
    - Thread setExclusiveContextThread(Thread t)
    - Thread getExclusiveContextThread()

  - AnimatorBase NEW:
    - Thread setExclusiveContext(Thread t)
    - boolean setExclusiveContext(boolean enable)
    - boolean isExclusiveContextEnabled()
    - Thread getExclusiveContextThread()

- AnimatorBase: Add setModeBits/MODE_EXPECT_AWT_RENDERING_THREAD
  Allows user to pre-determine whether AWT rendering is expected before starting the animator.
  If AWT is excluded, a more simple and transaction correct impl. will be used.

- FPSAnimator: Make transactions deterministic.
  FPSAnimator previously did not ensure whether a transaction was completed.
  A deterministic transaction is required to utilize ECT.
  FPSAnimator now uses same mechanism like Animator to ensure completeness,
  i.e. Condition and 'finishLifecycleAction(..)'. Both are moved to AnimatorBase.

Tested manually on Linux/NV, Linux/AMD, Windows/NV and OSX/NV.
  - All new tests validated correctness.
  - All new tests shows an performance increase of ~3x w/ single GLWindow, where multiple GLWindows don't show a perf. increase.

Minor: GLWindow fix size validation; GLContextImpl: Remove hold ctx lock count constraints at destroy.

Add old Olympic demo from GL4Java - Make it use ES1 and suitable for FixedFunctionEmul.

build/test: newt-headless shall exclude AWT/SWT

TestNewtEventModifiersNEWTWindowAWT was incl. in newt headless

Adding GEOMETRY_SHADER support in ShaderCode, adding core GL3/GEOMETRY_SHADER unit tests. ; Simplified GLContext version number

- Adding GEOMETRY_SHADER support in ShaderCode, adding core GL3/GEOMETRY_SHADER unit tests

  Chuck Ritola reported in December 2012 that we lack support of GEOMETRY_SHADER
  and he provided a test case.

  The latter is cleaned up to use GL3 core profile features only
  tesing a pass-through and the flip-XYZ geometry shader.

  ShaderUtil is fixed.

- Simplified GLContext version number
  The OpenGL major/minor version is now hold in a VersionNumber instance
  to simplify usage. Also expose it via getGLVersionNumber() while marking
  getGLVersionMajor() and getGLVersionMinor() deprecated.

Move all Exclusive Context Thread (ECT) tests to it's own package to de-clutter core test package

Fix Bug 666: GLBase.java line 438 contained some invalid trailing characters (regression from e7064ece049705e013d80985eae698ce0ee3c4e3)

NEWT-MouseEvent getWheelRotation() API Update - Fixes Bug 659: NEWT Horizontal Scrolling Behavior (OSX, X11, Win32); Bug 639: High-Res Mouse-Wheel

- API update 'float getWheelRotation()':
  Usually a wheel rotation of > 0.0f is up, and < 0.0f is down.

  Usually a wheel rotations is considered a vertical scroll.
  If isShiftDown(), a wheel rotations is considered a horizontal scroll,
  where shift-up = left = > 0.0f, and shift-down = right = < 0.0f.

  However, on some OS this might be flipped due to the OS default behavior.
  The latter is true for OS X 10.7 (Lion) for example.

  The events will be send usually in steps of one, ie. -1.0f and 1.0f.
  Higher values may result due to fast scrolling.
  Fractional values may result due to slow scrolling with high resolution devices.

  The button number refers to the wheel number.

- Fix Bug 659: NEWT Horizontal Scrolling Behavior (OSX, X11, Win32)

  - See new API doc above

  - X11/Horiz: Keep using button1 and set SHIFT modifier

  - OSX/Horiz:
    - PAD: Use highes absolute scrolling value (Axis1/Axis2)
           and set SHIFT modifier for horizontal scrolling (Axis2)
    - XXX: Use deltaX for horizontal scrolling, detected by SHIFT modifier. (traditional)

  - Windows/Horiz:
    - Add WM_MOUSEHWHEEL support (-> set SHIFT modifier), but it's rarely impl. for trackpads!
    - Add exp. WM_HSCROLL, but it will only be delivered if windows has WS_HSCROLL, hence dead code!

  - Android:
    - Add ACTION_SCROLL (API Level 12), only used if layout is a scroll layout

    - Using GestureDetector to detect scroll even w/ pointerCount > 2, while:
         - skipping 1st scroll event (value too high)
         - skipping other events while in-scroll mode
         - waiting until all pointers were released before cont. normally
         - using View config's 1/touchSlope as scale factor

- Fix Bug 639: High-Res Mouse-Wheel
  - getWheelRotation() return value changed: int -> float
    allowing fractions, see API doc changes above.

  - Fractions are currently supported natively (API) on
    - Windows
    - OSX
    - Android

- AndroidNewtEventFactory ir refactored (requires an instance now) and
  AndroidNewtEventTranslator (event listener) is pulled our of Android WindowDriver.

LinuxMouseTracker: Fix inverted Y-axis & missing button release.

Signed-off-by: Xerxes Rånby <xerxes@zafena.se>

ShaderState: Reduce hash-map utilization at shader switch (reset uniforms/attributes); Refine API doc.

README: Add new IRC channel, mark Jabber deprecated

README: Add new IRC channel, mark Jabber deprecated

Sync LauncherUtil w/ GlueGen commit 1b9f0739ecc25105384b557afa698c42e08d4cc6 (adding main-cmdline-args)

Fix Bug 669: Recursive GLContext makeCurrent()/release()

Culprit:
  GLContext's makeCurrent() didn't clear the boolean flag 'unlockContextAndSurface'
  in case the context is already current (-> recursion).
  Above case was detected within a code block tailed by a finally block,
  which acted on mentioned flag, i.e. called lock.unlock() and hence decremented the lock count
  even though the method return w/ successful state.

  Fixed.

Added debug code:
  GLContext.release() debug code (DEBUG | TRACE_SWITCH),
  recording stack trace of last release() call, which is dumped in case no current was current.

Added 2 unit tests:
  - Simple recursive GLContext makeCurrent()/release() from within GLEventListener's display().
    Test also validates lock count and lock ownership.

  - GLAutoDrawable display() of another GLAutoDrawable
    from within GLEventListener's display(..).

NEWT/Android: Full Lifecycle for WindowDriver; Using static ViewGroup; AWTRobotUtil: More tolerant for non AWT env.; Fix adb-launch-*

- NEWT/Android WindowDriver
  - Full Lifecycle, remove refs on closeNative()
  - Respect isFullscreen()
  - Using static ViewGroup if available and surface not ready, allows running from main()

- AWTRobotUtil: More tolerant for non AWT env.
  - Check for NEWT first
  - Only use AWT iff available, which allows running on Android

- Fix adb-launch-*
  - Launch main/junit tests

Adapt to change 5060b728b26a89a14367f8fca9f4eb7968f719ca: ShaderState.uniform(..) returns false if no location available.

TestSwingAWTRobotUsageBeforeJOGLInitBug411: Fix 'chicken egg' init problem w/ AWTRobot, i.e. NativeWindow not yet initialized.

Regression since: 50f997557b91a2f014ef0c2ea848c5c326d0cfb2

Android/NewtVersionActivity: Dump avail. GLCaps / Use stderr, since log would cap 'em (too long).

Android: Allow selection of native window formats RGBA8888, RGBX8888 and RGB565; Fix HiSilicon/Vivante/Immersion.16 EGLConfig selection (zero depth buffer @ visualID)

- NEWT/Android Fix PixelFormat/NativeWindowFormat/VisualID Selection
  - Fix allows proper selection of native window formats: RGBA8888, RGBX8888 and RGB565

  - Selection is performed in 3 steps:
    1) @ Construction (non native): SurfaceHolder.setFormat( getSurfaceHolderFormat( caps ) )
    2) @ Native Surface Creation: getANativeWindowFormat( androidFormat) -> ANativeWindow_setBuffersGeometry(..)
       Note: The set native format is revalidated, i.e. read out via ANativeWindow_getFormat(..).
    3) @ EGL Creation: ANativeWindow_getFormat(..) -> fixCaps(..) - simply fixing the chosen caps.

- NEWT GLWindow.GLLifecycleHook.resetCounter:
  - Also reset GLAnimatorControl's counter, if attached.

- NEWT WindowImpl -> GLLifecycleHook.resetCounter() calls issued _after_ operation before unlock().

- JOGL/EGLGraphicsConfigurationFactory
  - Validate whether the visualID matching EGLConfig depth buffer is suitable.
    On HiSilicon/Vivante/Immersion.16: Depth buffer w/ matching visualID is zero!

- NativeWindow/Capabilities.compareTo: Fix alpha comparison

Bug 665 (part 1) - Allow dis-association of GLContext's GLDrawable ..

Changes allowing re-association (incl. null) of GLContext/GLDrawable:

- GLAutoDrawable: Refine API doc 'setContext(..)'

- GLContext: Refine API doc: 'setGLDrawable(..)' 'getGLDrawable()'

- GLContextImpl.setGLDrawable(): Handle null drawable

- GLAutoDrawableDelegate/GLAutoDrawableBase: Allow null GLContext

- GLDrawableHelper.switchContext(..)/recreateGLDrawable(): Balance GLContext.setGLDrawable(..) calls

- New GLEventListenerState, holding state vector [GLEventListener, GLContext, .. ]
  impl. relocation of all components from/to GLAutoDrawable.

- GLDrawableUtil
  - Using GLEventListenerState for swapGLContextAndAllGLEventListener(..)

+++

NEWT Window*:
  - getDisplayHandle() is 'final', no more 'shortcut' code allowed
    due to re-association incl. display handle.

  - close*:
    - close config's device (was missing)
    - null config

+++

Changes allowing reconfig of Display handle as required
to re-associate pre-existing GLContext to a 'window':

- AbstractGraphicsDevice: Add isHandleOwner() / clearHandleOwner()
  - Impl. in X11GraphicsDevice and EGLGraphicsDevice, NOP in DefaultGraphicsDevice

- DefaultGraphicsConfiguration add 'setScreen(..)'

- MutableGraphicsConfiguration
  - Make DefaultGraphicsConfiguration.setScreen(..) public

- NativeWindowFactory add 'createScreen(String type, AbstractGraphicsDevice device, int screen)'
  - Refactored from SWTAccessor

- NativeWindow x11ErrorHandler: Dump Stack Trace in DEBUG mode, always.

FBObject/GLFBODrawableImpl: Silence stderr if not in DEBUG mode relying on application handling verbosity and debug output

The following cases were dumping states on stderr:
  - FBOObject.isStatusValid() if false
    -> Only dumps if DEBUG

  - GLFBODrawableImpl.reset() if fboResetQuirk becomes true and 1st time in ClassLoader
    -> Only dumps if DEBUG

[EGL/X11]Context: Enhance DEBUG output for failed makeCurrentImpl(..)

ProxySurface: Add 'NativeSurface getUpstreamSurface()' ; ProxySurfaceImpl: Don't cache 'displayHandle'

- ProxySurface: Add 'NativeSurface getUpstreamSurface()' allowing querying direct access of a backing surface representing this instance.
  - Use case: EGLWrappedSurface
  - Default impl. returns null

- ProxySurfaceImpl: Don't cache 'displayHandle'
  - getDisplayHandle() is 'final', no more 'shortcut' code allowed
    due to re-association incl. display handle.

  - See commit b738983638703bb721ee4c9820c8ef43e2252e73

WGLContext: Enhance DEBUG output for failed makeCurrentImpl(..)

Refine GL[Auto]Drawable 'realized' state in API doc, and relax it's realized requirement in GL[Offscreen]AutoDrawableDelegate*

Compatible w/ 'before'.

TODO: Contemplate about GLDrawableFactory.createOffscreenAutoDrawable(..) whether it:
  - should better return an unrealized [auto]drawable
  - or adding another API method for such case

  Goal: Allow passing vector of [device/context/..] for use cases
        such as re-using an onscreen destructed surface and on-/offscreen hopping.

Animator: API doc cleanup, isPaused()==true if drawablesEmpty (Animator == FPSAnimator behavior), resume after add ifPaused()

- GLAnimatorControl: Refine API doc

- Animator.run():
  - if( drawablesEmpt) { pausedIssued = true; } - Same behavior as FPSAnimator

- AnimatorBase.add(..)
  - consider paused case if no drawablesEmpty, i.e. if ( isPaused() ) { resume(); }

Fix misc API docs ..

GLDrawableHelper.invoke: If drawable is !realized -> deferred execution

Fix regression of commit b738983638703bb721ee4c9820c8ef43e2252e73, possible 'pulling' display's device - check references

Bug 665 (part 3) - Allow dis-association of GLContext's GLDrawable .. - Add EGL/ES2 tests, attempt to fix wrapped EGL case

- Bug 665 (part 2) was commit 7fd5f76e1eb4bbf93fe9b1171744bd755d8f96e4

- Add EGL/ES2 tests in
  - TestGLContextDrawableSwitch01NEWT
  - TestGLContextDrawableSwitch11NEWT

- Attempt to fix wrapped EGL case (incomplete)
  - Using EGL/ES w/ non native EGL device/surface, but natively wrapped instances (most of the cases),
    a 'complicated' delegation of
       Native-Upstream -> EGL-Proxy -> EGL-Instance
    is being used heavily relying on the objects lifecycle.
    GLEventListenerState tries to roll back the realized state
    and even sets the upstream device handle,
    but this doesn't seem to be sufficient on X11.

Discussion:

It might turn out that we only can implement the survival of GLContext
and it's display device reliable w/ EGL within the GLAutoDrawable implementation,
which can hold the previous not destructed instances.

Add WindowClosingProtocol impact detail on WindowListener.windowDestroyNotify(..)

Fix Bug 678: Deliver key-char value for printable chars on all KeyEventListener (-> On Windows as well)

The following is observed, where t0 and t1 refer to subsequent different timestamps:
  NEWT delivery order:
      PRESSED (t0), RELEASED (t1) and TYPED (t1)

  WINDOWS delivery order:
      PRESSED (t0), TYPED (t0) and RELEASED (t1)
  Windows Auto-Repeat:
      PRESSED (t0), TYPED (t0)

Hence we changed the event reorder-code in NEWT to trigger NEWT-PRESSED on
Windows-TYPED for printable chars, assuring key-char values on all listener callbacks.

- KeyEvent.getKeyChar(): Removed disclaimer dedicated for Windows
- Keyevent.isActionKey(): Completed for all NEWT non-printable action keys; Added static variant
- Keyevent.isPrintableKey(): NEW: returns !isModifierKey(keyCode) && !isActionKey(keyCode) ; With static variant

- Windows WindowDriver:
  - EVENT_KEY_PRESSED handles non-printable chars only
  - EVENT_KEY_TYPE handles printable chars only
  - Native: VK_DELETE passes keyCode

- Unit tests: Wait for completion 1s -> 2s

OSX unit tests: Disable OSX 32bit tests (redundant - prepare for Java7 tests)

Add OSX Java7 build/test scripts

Fix AWTKeyAdapter: Reorder AWT events to NEWT order - also ensuring TYPED is always sent.

This foremost fixes an issue w/ OSX/Java7 and NewtCanvasAWT offscreen CALayer usage,
which utilizes AWTKeyAdapter and AWTNewtEventFactory (AWT -> NEWT) key events.

Add LinuxEventDeviceTracker used by RaspPi newt/driver/bcm/vc/iv/WindowDriver

Signed-off-by: Xerxes Rånby <xerxes@zafena.se>

LinuxEventDeviceTracker: LinuxEventDeviceManager monitor /dev/input/event*

The LinuxEventDeviceManager will start new EventDevicePoller threads
when more /dev/input/event* files become available.

Signed-off-by: Xerxes Rånby <xerxes@zafena.se>

LinuxEventDeviceTracker: gracefully manage hot-swap of USB devices.

Signed-off-by: Xerxes Rånby <xerxes@zafena.se>

LinuxEventDeviceTracker: LinuxEVKey2NewtVKey(EVKey 0-121)

Signed-off-by: Xerxes Rånby <xerxes@gudinna.com>

LinuxEventDeviceTracker: LinuxEVKey2NewtVKey(EVKey 122-248)

Signed-off-by: Xerxes Rånby <xerxes@zafena.se>

LinuxEventDeviceTracker: Map printable char using KeyEvent.isPrintableKey.

The NEWT VK namespace tries to map keyCodes
to its Unicode varianti, make use this quirk.

Signed-off-by: Xerxes Rånby <xerxes@zafena.se>

Bump Android NDK to android-ndk-r8d

Android: Adapt to cleanup of ClassLoaderUtil/LauncherUtil - GlueGen Commit b47d0d92dd222999bf38633de1cec8de6a7ad369

Android GearsES2 Launcher: Revert debug flags ..

Fix Animator resume() wait-condition (drawablesEmpty didn't wait for pausedIssued); Update TestGLWindows02NEWTAnimated (isPaused())

GLContext: *TRACE_SWITCH spacing

Java2D OGLPipeline(GLJPanel+Threading): More fine grained control about OGL threading and resource usage

- Still excluse OSX
- Respect sun.java2d.opengl property
- [Prepare] Allowing OGLPipeline thread usage w/o it's OGL resource usage

Fix bug 668 : Android keyboard input was uppercase

Fix OSX CALayer Bug 690 and Bug 691: Occasional Freeze on CVDisplayLinkStop; Layers and native GL-Context are _not_ Released ; Java Side wait for Main-Thread

- Fix Bug 690: Occasional Freeze on CVDisplayLinkStop
  - NSOpenGLLayer.disableAnimation() shall not claim the renderLock mutex,
    since the CVDisplayLink callback could be waiting for the lock.
    This waiting callback could freeze the call to CVDisplayLinkStop.

- Fix Bug 691: Layers and native GL-Context are _not_ Released
  - Following proper release cycle:

    Context unrealized:
      - JAWTWindow.detachSurfaceLayer() -> OSXUtil.RemoveCASublayer(..)
      - CGL.releaseNSOpenGLLayer(..)

    JAWTWindow.destroy()
      - MacOSXJAWTWindow.UnsetJAWTRootSurfaceLayer(..)
      - OSXUtil.DestroyCALayer(..)

  - 'Magic' CALayer release calls (w/o manual retain beforehand) at:
     - OSXUtil.RemoveCASublayer(..): [subLayer release]
     - MacOSXJAWTWindow.UnsetJAWTRootSurfaceLayer(..): [rootLayer release]
     - OSXUtil.DestroyCALayer(..): [rootLayer release]

  - 'Magic' NSOpenGLLayer's NSOpenGLContext dealloc:
     - [NSOpenGLContext clearDrawable]
     - CGLDestroyContext( [NSOpenGLContext CGLContextObj] )

- Java Side wait for Main-Thread
  - Waiting for the delegated Main-Thread on the Java side eases debugging
    and won't block the Main-Thread in native code.

  - Utilizing this for all CALayer calls

Test case: TestGLCanvasAddRemove01SwingAWT

Fix Bug 691 (part-2): Extra '[subLayer release]' is wrong, since 'CGL.releaseNSOpenGLLayer' triggers release - but very late w/ AWT usage.

OSXUtil_RemoveCASublayer0's added '[subLayer release]' in commit f6e6fab2a7ddfb5c9b614cb072c27ff697629161
is wrong, since 'CGL.releaseNSOpenGLLayer' actually does trigger it's release.
This was not seen w/ AWT tests, since it happens very later.
A NewtCanvasAWT test disclosed this error -> removed that extra release call.

The culprit for the late release w/ AWT usage was CGL.createNSOpenGLLayer's call in the current thread.
Moving it to the Main-Thread fixed the problem.

All CALayer lifecycle calls are issued on the Main-Thread now.

NSOpenGLLayer's CVDisplayLink OpenGL fitting via 'CVDisplayLinkSetCurrentCGDisplayFromOpenGLContext'
is now performed at it's context creation in 'NSOpenGLLayer::openGLContextForPixelFormat'.

The 'extra' release of the NSOpenGLLayer's NSOpenGLContext as introduced in commit f6e6fab2a7ddfb5c9b614cb072c27ff697629161
is still valid.

Fix Bug 675: NPE w/ Beans.setDesignTime(true) ; Fix GLCanvas.destroy(): Don't issue removeNotify(), delete drawable and context only!

- Fix Bug 675: NPE w/ Beans.setDesignTime(true)
  - Carefully consider Beans.isDesginTime() fixes NPE - added unit test

- Fix GLCanvas.destroy(): Don't issue removeNotify(), delete drawable and context only!
  - AWT removeNotify() shall only be issued via AWT itself, not manually
  - Add 'destroyImpl(boolean destroyJAWTWindowAndAWTDevice)' to be called by
    - GLCanvas.destroy(): destroyImpl( false );
    - GLCanvas.removeNotify(): destroyImpl( true );
  - Ensures JAWTWindow and AWTDevice are created and destroyed via the AWT callbacks
    addNotify() and removeNotify() only.

Refine unit test for Bug 691: Add onscreen/offscreen tests [if available] ; Move tests to 'acore'.

Fix Bug 691 (part-3): NSOpenGLLayer::openGLContextForPixelFormat(..) on main-thread deadlock'ed due to locked shared context

NSOpenGLLayer::openGLContextForPixelFormat(..) is performed on main-thread at 1st NSOpenGLLayer display method.
This happened irregulary, i.e. sometimes (T0) right after NSOpenGLLayer creation and attachSurfaceLayer()/AddCASublayer(..),
sometimes later (T1).

NSOpenGLLayer::openGLContextForPixelFormat(..) uses the passed shared user context.
The shared user context is locked at NSOpenGLLayer's creation (T0) and if performed at this early time
the call deadlocks due to pthread_mutex wait for the shared user context.

This fix performs NSOpenGLLayer creation and layer attachment while the shared user context
is kept unlocked and enforces NSOpenGLLayer display and hence NSOpenGLLayer::openGLContextForPixelFormat(..).

Added CGL.setNSOpenGLLayerEnabled(..) to enable/disable NSOpenGLLayer - currently not used.

- Passed AddRemove tests for GLCanvas/Swing and GLWindow/NewtCanvasAWT w/ 100 loops on Java6 and Java7 on OSX.
- Passed Instruments Leaks test w/ 10 loops on Java6 and Java7

X11/WGL: Better Exception Message for failed ARB context creation if profile > GL2 (OpenGL >= 3.0.1) was requested.

Fix TestWindowClosingProtocol01AWT: GLCanvas closing operation 'destroy' will no more call removeNotify() - see commit 3567e7e8519f82720f98b0b2ac30456cbfeddc0d

Fix Bug 644: AWT GLCanvas and GLJPanel ignored their visibility state

If !visible, GLCanvas and GLJPanel's paint* and display method shall not render
as the other GLAutoDrawable impl. do (GLWindow, SWT GLCanvas).

Uni Tests AddRemove: Remove dead code ..

TestWindowClosingProtocol01AWT: More relaxed timing reducing false positives

NEWT/OSX: Fix Memory Leak ; Fix Occasional Crash Duer to Lifecycle Ops not on Main-Thread.

- Fix Memory Leak
  - NewtWindow::dealloc -> [NewtView release]: Fixes NewtView leak
  - NewtView::dealloc -> removeTrackingRect: Removes occasional crash (double free of TrackingRect)

- Fix Occasional Crash Duer to Lifecycle Ops not on Main-Thread.
  Perform OSX WindowDriver ops on Main-Thread:
    - close0
    - changeContentView0
    - createWindow0

- Cleaned up AddRemove unit tests, added TestAddRemove03GLWindowNEWT

build-test.xml: Move 'junit.run.newt.awt' up in file; 'junit.run.newt': Exclude *AWT* and *SWT* (similar to 'junit.run.newt.headless')

Add proper copyright header w/ all remarks regarding re-distribution etc, similar to other JNI headers

OSX/Java7 Cleanup: Remove JavaNativeFoundation dependency; NativeWindow-JAWT: Remove c.fixup.jawt.version.macosx (redundant, libjawt.dylib is rpath'ed always)

Unit Tests: Use Property 'jvmJava.exe' ; Add target 'junit.run.local.java7' (if="jvmJava7.exe")

See GlueGen commit 6b86764f2e195b4046000fd5a7fcf3331ca72d21

Unit Tests: Add 'results-java7' to check-junit.sh (Completes commit fc9539d50c12f9def7a3babde7384e3d38f61721)

Bug 678 (fix), Bug 641 (API + Windows Impl.), Bug 688 (prep): Update NEWT's KeyEvent handling while distinguish keyCode (kbd layout independent) and keySym (kbd layout dependent)

API Changes:

- Virtual key codes and symbols are of type short.

- KeyEvent.keySymbol() shall return a layout dependent value (Bug 641)
  - Method returns former keyCode() value, which was layout dependent.
  - Returns 'short' value

- KeyEvent.keyCode() returns a short value, instead of int

- KeyEvent.keyCode() shall return a layout independent value (Bug 641)
  - To ease implementation, we only 'require' the scan code to be mapped to a 'US Keyboard layout',
    which allows reusing layout dependent code while preserving the goal to have a fixed physical key association

  - Implementation status:
    - Windows OK
    - X11 TODO
    - OSX: 50/50 TODO
      - Using layout independent 'action keys'
      - Using layout dependent 'printable keys'
      - returning above semantics for both, keyCode and keySym

    - Android 50/50 TODO
      - Returning the layout independent keyCode
      - Mapping probably incomplete

- KeyEvent.EVENT_KEY_TYPED and KeyListener.keyTyped(KeyEvent) (Bug 688)
  - Marked DEPRECATED
  - No more called for auto-repeat events
  - Synthesized in WindowImpl.consumeKeyEvent(..): No more injection by native- or java driver code

- NEWTEvent.eventType: int -> short
  - field, as well as all method involving eventType changed to short.

- NEWTEvent.isSystemEvent: REMOVED
  - Never used as well as never being implemented properly

Internal Changes:

- Simplified keyEvent driver code
  - Especially the Windows native driver's mapping code
    could be simplified using scanCode and MapVirtualKeyEx

- NEWT Event Factories: hashMap -> switch/case

Unit Tests:
  -
  - Added NewtCanvasAWT Offscreen Layer Tests
    important to test the AWT -> NEWT translation on OSX/CALayer:
    - TestNewtKeyCodeModifiersAWT
    - TestNewtKeyCodesAWT
    - TestNewtKeyEventAutoRepeatAWT
    - TestNewtKeyEventOrderAWT
    - TestNewtKeyPressReleaseUnmaskRepeatAWT

Unit Tests NewtEventModifiers: Attempt to make them more robust for that one particular Windows-AMD Test Node

- Adding 'waitForVisible' and 'waitForRealized' on the GLAutoDrawable before test
- Adding clearKeyboadAndMouse before test

Unit Tests NewtEventModifiers (2): Robostness

- clearKeyboadAndMouse(): Add _testMouseListener.clear(); to clear all logs!

- clearKeyboadAndMouse() after all tests
- use implicit 'clearKeyboadAndMouse()' in execOffThreadWithOnThreadEventDispatch()

Unit Tests: Add [manual] demo cases for SWT (TestGearsES2SWT, TestGearsES2NewtCanvasSWT); Add more AWT/NEWT layout scenarious.

NEWT/OSX: Fix 'locationOnScreen' usage and use proper client-area window screen position (Fixes resize -> position)

- Tested w/ NEWT GearsES2 and Java6 and Java7 on OSX

NewtCanvasAWT: Fix size determination, i.e. take newt-child's size into account if no preferred size is given.

OSX/Java7/CALayer + JAWT: Partially Fix AWT/NEWT CALayer 'out of sight' bug, where our CALayer is moved out of the visible area

- same erroneous behavior for GLCanvas and NewtCanvasAWT

- sized-frame: Set framesize and validate() it

- sized-component: Set component preferred size and call frame.pack()

- added workaround 'OffscreenLayerSurface.layoutSurfaceLayer()' to fix CALayer size, which snaps for:

  - OK initial size before setVisible: sized-frame and sized-component

  - OK resize w/ sized-frame

  - OK manual frame resize

  - Invisible: w/ sized-component after setVisible()

++

- CALayer-Sublayer (GL) has additional retain/release when added/removed
  to be on safe side.

Fix NPE of Bug 679: The ctx was released w/o fully initialized (no proper ARB ctx of expected version ?)

TestGearsES2SWT: Fix NPE

Fix GLVersionNumber (Bug 679): Don't use a subsequent version number if not valid!

Relax GLContextImpl.setGLFunctionAvailability(..)'s Version Matching if testing major < 3 versions. (Bug 679)

On OSX, we 'only' support the ctx ARB creation and hence have no !ARM fallback.

If querying a version [2.0 .. 3.0] for GL2, it is acceptable if a valid OpenGL version is lower, e.g. 1.4.

Bug 548 and Bug 679: Add 'glFlush()' requirement (OSX < 10.7.3 && NVIDIA GPU) to GLRendererQuirks and perform only if exists.

GLProfile: Adding convenient query for highest programmable core only GL profile

As suggested @ http://forum.jogamp.org/How-to-force-core-profile-tp4028307.html

However, one could always get a dedicated core profile via:
  - GLProfile.get(GLProfile.GLES2)
  - GLProfile.get(GLProfile.GL3)
  - GLProfile.get(GLProfile.GL4)

.. etc

TestNewtEventModifiers*: Robostness (uhh) Unify wait-clear result code - fix SWT event dispatching.

Fix CALayer pos/size and animation.

- Fix CALayer animation:
  - All CALayer animations are set to nil via overriding 'actionForKey'

- Fix CALayer pos/size bug:

  - Fix root and sub CALayer position to 0/0 and size on the main-thread w/o blocking.

  - If the sub CALayer implements the Objective-C NativeWindow protocol NWDedicatedSize (e.g. JOGL's MyNSOpenGLLayer),
    the dedicated size is passed to the layer, which propagates it appropriately.

  - On OSX/Java7 our root CALayer's frame position and size gets corrupted by its NSView,
    hence we have created the NWDedicatedSize protocol.

NEWT/OSX Fix: Child positioning ; NewtCanvasAWT: Change reparent time and use actual component size while setting min/pref. size at setup

- NEWT/OSX Fix: Child positioning
  - If !Offscreen and has-parent: Gather screen location by traversing through parent
    and set native position (was removed w/ commit 7d5c51b635e0795d9b170342bdebe8e7e0bbd01d since still buggy).

- NewtCanvasAWT: Change reparent time and use actual component size while setting min/pref. size at setup
  - Analog to AWT GLCanvas
  - validates and reparents at reshape(..), paint(..) and update(..)
  - reshape(..) also trigers jawtWindow.layoutSurfaceLayer()
  -

Adds a unit test for the bug 694 on gluScaleImage

Fix NEWT/AWT WindowClosing Unit Tests ; Review/Cleanup NEWT WindowClosing mechanism

Due to a NEWT WindowClosing event regression cause by NewtCanvasAWT changes
a review of our WindowClosing event mechanism was required.

Important cleanups are marked w/ '(*)' below.

I would have preferred to change the 'WindowListener.windowDestroyNotify(WindowEvent)'
method to pass a WindowCloseEvent object exposing more information like
toolkit or programmatic destruction and passing whether a 'closing' or 'nop' action
will be performed based on the WindowClosingMode.
For now I postponed this idea .. since it would change the API again,
but may reconsider it after merging the Android 'closing' patch.

- InputEvent.consumedTag -> NEWTEvent.consumedTag

- Window
  - (*) Promote setWindowDestroyNotifyAction(Runnable) to public,
        former WindowImpl.setHandleDestroyNotify(boolean).
    Using a Runnable action for WindowImpl.windowDestroyNotify(boolean)
    allows a setting defined alternative for destroy() and gets rid
    of [ab]using WindowListener.windowDestroyNotify(WindowEvent) for
    lifecycle actions. Used in:
      - GLWindow
      - GLAutoDrawableDelegate impl.

- WindowImpl
  - Respect NEWTEvent.consumedTag for WindowEvents as well

  - (*) Impl. setHandleDestroyNotify(boolean) (see above)

  - (*) destroy() simply sends out pre- and post- destruction Window events,
    where windowDestroyNotify(boolean) sends out the pre-destruction event if NOP.

  - (*) windowDestroyNotify(boolean) is public now, allowing other impl. details
    to follow proper destruction using handleDestroyNotify Runnable (-> NewtCanvasAWT).

- AWTWindowClosingProtocol:
  - addClosingListenerOneShot() -> addClosingListener()
    - calling addClosingListener() at addNotify()
    - calling removeClosingListener() at removeNotify()

  - AWTWindowClosingProtocol ctor taking NOP runnable,
    allowing to send WindowEvent.EVENT_WINDOW_DESTROY_NOTIFY at WindowClosingMode.DO_NOTHING_ON_CLOSE

  - add/remove listener on AWT-EDT

- AWTWindowAdapter
  - Add 'removeWindowClosingFrom(..)',
    allowing to remove window closing event fwd.

  - Also fwd windowClosed in window closing fwd'ing.

- NewtCanvasAWT
  - (*) Utilize AWTWindowClosingProtocol NOP runnable (see above)
    to fwd closing-NOP event to NEWT

  - (*) Unify remove/destroy code in destroyImpl(..)
    - !removeNotify -> destroy NEWT child programatic or as toolkit event
    - removeNotify || windowClosing -> destroy jawtWindow

  - (*) Remove AWTWindowAdapter/AWTParentWindowAdapter's windowClosingListener,
    since we utilize AWTWindowClosingProtocol

- DisplayImpl
  - Adding 'final void dispatchMessage(final NEWTEvent event)'
    allowing to remove the NEWTEventTask wrapping for no reason
    in enqueueEvent(..) if on EDT and waiting.

Fix Bug 677: NEWT/Android: Add support for Android's KeyEvent.KEYCODE_BACK

Original author: Eric Brayet <ericb@daysofwonder.com>
Revised by: Sven Gothel <sgothel@jausoft.com>

I took the freedom to cleanup the three original patches
from https://github.com/Pooouf/jogl.git branch 'bug_677':
  - 7449d4726633d524a3bb79efffd04cfd0ca25e58 (removed by followup patch!)
  - 68c739a4f03e46deecdbb71c125b4586aec08d63 (removes previous patch!)
  - c2813dfc325a1482d18b6fc304e4e483f5633964

Further more I was able to reduce the 'extra' code while utilizing
  - Window's isKeyboardVisible() and using keyboardVisibilityChanged(false)
    to update the hidden keyboard state.
  - Moving the key-handling code to the containing WindowDriver class
    avoiding passing a reference to the inner view.
  - Using AndroidNewtEventFactory for NEWT KeyEvent creation

+++

- Handle KeyEvent.KEYCODE_BACK w/ jogamp.newt.driver.android.WindowDriver.MSurfaceView.onKeyPreIme(..):
    if( soft keyboard is up )
      [1] Update keyboard visibility state and return NEWT KeyEvent.VK_KEYBOARD_INVISIBLE;
    else
      [2] call WindowImpl.windowDestroyNotify(true)

    [3] then cont. processing, i.e. return false;

- Turns out respecting WindowClosingMode might be
  - too complicated
  - interfere w/ Android UI behavior

- AndroidNewtEventFactory
  - createKeyEvent
    - static
    - adding boolean param 'inclSysKeys', if true, KEYCODE_BACK and KEYCODE_HOME are mapped

- Unit tests: GearsES2 + MovieCubeActivity0 shows keyboard if pressure > 0.6f
    - pressure on Android shall be between [0..1], however we have to figure out
      badly calibrated touchpads/Android device where we could experience
      pressure > 2.0f !

    - TODO: API documentation of pressure [0..1]

NEWT WindowImpl: Cleanup redundancies in destroy() and windowDestroyNotify(boolean)

- destroy()
  - screen is never null!
  - always attempt to removeScreenReference()

- windowDestroyNotify(boolean)
  - NOP if !isNativeValid() to avoid DESTROY_* events

NEWT: Harmonize MouseEvent Pressure (API Change!)

Due to high fluctuation (lack of normalized) pressure values on Android devices,
an option to query the normalized value and access to the current known maximum pressure is required.

MouseEvent:
  - getMaxPressure() returning the [self calibrated] known maximum pressure

  - getPressure(..) -> getPressure(.., boolean normalize) (API Change!)
     - return normalize ? pressure/maxPressure : pressure;

Fix AWT EDT Deadlock in AWTWindowClosingProtocol, regression of commit 808a9a27a8c1c9e0a6701a8dd81d51f8daa8129d

Adapt to GlueGen's relocation of AWTEDTExecutor - GlueGen commit 99a50b38f5650fedca0f207e03706ffa9492e50c

All AWT Related Unit Tests: Call AWT 'frame.setVisible(..)' and 'frame.pack()' on AWT-EDT - Use dedicated test TestBug572AWT for testing otherwise.