Changes

Fix commit a393e45613d87101dbb13763df263c2f9291d2d0: jogl's cg native lib is jogl_cg (jar packaging)

NEWT WindowImpl.runOnEDTIfAvail(..): Run task from current thread if owning the windowLock. Avoiding deadlock when cmds issued from within locked code path.

This allows e.g. GLEventListener::display(..) { .. glWindow.setSize(100, 100); .. }

NativeSurface Cleanup (API Change) - Adapt to GlueGen Lock cleanup commit: 834b9e530e652b7ff7c5e222720bce3ad2b11c5f

- adapt to GlueGen Lock cleanup
- remove isSurfaceLocked(), use 'null != getSurfaceLockOwner()' instead

Misc:

- remove unused priv./impl. methods
- add @Override

Misc cleanup: add @Override

Misc cleanup: Add @Override, remove trailing whitespace

Misc cleanup: Add @Override, remove trailing whitespace

Capabilities cleanup: make most get*() final, impl. Comparable to CapabilitiesImmutable, Misc ..

- Comparable<CapabilitiesImmutable> to CapabilitiesImmutable:
  "interface CapabilitiesImmutable extends VisualIDHolder, WriteCloneable" ->
  "interface CapabilitiesImmutable extends VisualIDHolder, WriteCloneable, Comparable<CapabilitiesImmutable>"

- CapabilitiesImmutable and GLCapabilitiesImmutable get-methods final (most)

- GLCapabilitiesImmutable.compareTo: add doubleBuffered (between stereo and samples)

Misc:
  - Add @Override
  - remove trailing whitespace,

GLDrawable* cleanup: Add @Override, remove trailing whitespace, ..

GLContext* TRACE_SWITCH: Add GLContext instance hash value to allow tracking.

SWT GLCanvas: Fix destroy(), Remove local concurrency hack & out-of-process locking ; Revise threading code; Proper name for SWT unit tests.

SWT GLCanvas:
  - Fix destroy()
    drawable.setRealized(false); is being called within dispose() method

  - Remove local concurrency hack
    [MT-0] The hack: 'final GLContext ctx = getContext(); a = null != ctx ? ctx.something() : 0;'
    is thread safe locally, however, w/o covering the hole use case of the caller
    it makes no sense to add thread safe code here - would be only an illusion.
    In case any of the methods are called outside of a locked state
    extra care should be added. Maybe we shall expose locking facilities to the user.
    However, since the user shall stick to the GLEventListener model while utilizing
    GLAutoDrawable implementations, she is safe due to the implicit locked state.

  - Removing out-of-process locking
    [MT-1] Claiming the recursive lock in the dispose/display/.. methods
    and _then_ issuing a complex off-thread GL task could lead to deadlock.
    The GL task could involve calling GLEventListener methods, which itself
    could try to manipulate the GLCanvas -> deadlock.
    Similar to [MT-0] we may need to either find a proper locking mechanism
    or simply ignore it and reduce functionality.
    TBH .. the number of scenarious of the mentioned deadlock are very limited
    and exotic.

  - Revise threading code
    [MT-2] Besides the other MT-* remarks, the logic whether to spawn off
    the GL task and determination which thread to use is too complex and redundant.
    (See isRenderThread(), runInGLThread() and runInDesignatedGLThread())

  - Proper name for SWT unit tests.
    Reflect the semantics.

SWT/AWT GLCanvas multithreading annotations (see commit 1a91ec5c8b6fd9d9db7bc115569c369fe7b38e9b) ; AWT GLCanvas remove dead code.

NEWT GLWindow multithreading fix and annotations (see commit bc7503c77892a9e14b10e8b8e9ce48b148c6fa4c).

NEWT GLWindow multithreading fix:
  - Add required locking of display(), otherwise a drawable/context destruction of another threads
    could lead to [still] use asynced data.

AWT/SWT GLCanvas: Remove volatile of context instance, use drawable instance's volatile 'feature'

Bug #589 - Enhance snapshot/textIO test: Add offscreen mode; Add AWT onscreen/offscreen-layer resize test.

Minor cleanup

NativeSurface.unlockSurface(): Change fail-fast policy to fail safe tolerant

- This policy allows more simple destruction handling w/o validating on the top level.

- Hence 'unlockSurface()' shall not throw any exception.

- 'lockSurface()' keeps unchanges, clarified w/ explicit 'throws' declaration,
  ie will fail-fast.

GLContextImpl/GLDrawableImpl: More fail-safe, cleanup, mark some methods final

GLContextImpl: Cleanup
  - release(): simplify the conditions
  - destroy(): allow locked twice before (destroy case)

GLDrawableImpl:
  - fail safe: swapBuffers
  - final methods: getFactoryImpl, toHexString, getGLProfile, getRequestedGLCapabilities,
                   getFactory, isRealized, lockSurface, unlockSurface

JAWTUtil/MacOSXWindowSystemInterface-pbuffer: Cleanup / More DEBUG info

JAWTUtil.getJAWT(..):
  - better local var names
  - one time init
  - add DEBUG trace
JAWTUtil.<init>:
  - add offscreenLayer info

MacOSXWindowSystemInterface-pbuffer.m:
  - remove warning: access of protected vars from outside class
  - better local var names
  - more DEBUG info (pbuffer texture)

Fix Bug #589 (JAWT Offscreen-Layer resize) and Offscreen-Layer setSwapInterval() deadlock; Reuse JAWT instance; Cleanup

- Fixes
  - OSXUtil.CreateCALayer*(..): Pass layer target size (if known).
    This fixes Bug #589

  - MacOSXWindowSystemInterface-pbuffer.m:
     - ALL: displayLink NULL check

     - setSwapInterval(..): lock only for variable setting,
       could deadlock when start/stop CVDisplayLink

  - JAWTWindow.destroy(): use 'surfaceLock' instead of 'synchronized'

- Cleanup / Performance
  - JAWTWindow.lockSurface(): Reuse JAWT instance

  - MacOSXJAWTWindow: AttachJAWTSurfaceLayer0(..) -> SetJAWTRootSurfaceLayer0(..)
                      Reflects semantic better.

- DEBUG
  - JAWTWindow.updateBounds(..) notify of bounds change

Fix Bug 597: XDisplayString(0) -> SIGSEGV; Handle n/a DISPLAY

It works now as follows:

Note: Remote connection tested using ssh to localhost or an actual remote
machine.

Test w/ proper X11 authentication, i.e. DISPLAY and xauth:

> cd projects/JOGL/jogl-bugs/597-headless/
> echo $DISPLAY
localhost:10.0

> xclock
'works'

T1.1 OK> java plop
jogamp.opengl.x11.glx.X11GLXDrawableFactory@f581593

T1.2 OK> java -Djava.awt.headless=true plop
jogamp.opengl.x11.glx.X11GLXDrawableFactory@6f34a7d8

+++

Test _without_ proper X11 authentication, i.e. DISPLAY and xauth:

> xclock
No protocol specified
No protocol specified
Error: Can't open display: :0.0

T2.1 OK> java plop
No protocol specified
No protocol specified
Exception in thread "main" java.lang.InternalError: Can't connect to X11 window
server using ':0.0' as the value of the DISPLAY variable.
   (Stack trace from:
        GLProfile.initProfilesForDefaultDevices(..) ->
        NativeWindowFactory.initSingleton(..) ->
        JAWTUtil.<clinit>->
        sun.awt.X11.XToolkit.<clinit> ->
        sun.awt.X11GraphicsEnvironment.initDisplay(Native Method)

T2.2 OK> java -Djava.awt.headless=true plop
No protocol specified
No protocol specified
...
jogamp.opengl.x11.glx.X11GLXDrawableFactory@394a8cd1

Test case cleanup

- Gears w/ 640x480 size and proper aspect ratio frustum
- query type of drawable.getNativeSurface() instead of auto-drawable itself
- RedSquareES2 allow setting no-rotation and aspect-ratio for frustum
- TestInitConcurrentNEWT: Don't use vsync

GLAutoDrawable: Refine API doc; Use new abstract impl. GLAutoDrawableBase (GLWindow, ..); Add new GLAutoDrawableDelegate.

- Refine API doc
  - 'void setContext(GLContext)' -> 'GLContext setContext(GLContext)'
  - Add note to createContext(GLContext) override

-Use new abstract impl. GLAutoDrawableBase, used by:
  - GLWindow
  - GLAutoDrawableDelegate
  - GLPbufferImpl

- Add new GLAutoDrawableDelegate incl. unit test

GLAutoDrawable* refinement of abstraction / generalization - API Change!

- GLAutoDrawable (compat change - recompile):
  - 'void invoke(boolean wait, GLRunnable glRunnable)' -> 'boolean invoke(boolean wait, GLRunnable glRunnable)'
    Allows notifying caller whether the task has been executed or at least enqueued.

- GLAutoDrawable add 'GLEventListener removeGLEventListener(int index)'
  - This allow one to remove a specific GLEventListener and reusing it (return value).

- GLDrawableImpl remove 'destroy()' to favor 'setRealized(false)'
  - Using more common code of GLAutoDrawableBase, i.e. GLPbufferImpl can use defaultDestroyOp().
  - Removes redundancy of methods

- GLAutoDrawableBase/Delegate
  - better 'default' names to emphasize it's purpose, adding API doc

  - includes more generic functionality
    - defaultWindowDestroyNotify()
    - defaultDestroyOp()

  - TestGLAutoDrawableDelegateNEWT demonstrates a simple example w/ all window events handled.

- Fix TestParenting01cSwingAWT's threading use (gl disturbance thread)

NativeWindow/Newt X11ErrorHandler enhancement / unification - don't throw exceptions. Handles also XAWT BadMatch X_SetInputFocus.

X11ErrorHandler code now dumps proper information about the opcode and error message and the running Java thread.

Having propery "nativewindow.debug.X11Util.XErrorStackDump" or "nativewindow.debug=all' set,
a stack trace is dumped.

Since the X11ErrorHandler may catch an XAWT error: BadMatch X_SetInputFocus,
we cannot throw an exception and better keep running.

NEWT/KD Adapt to 'new' KD Window Creation API, use EGLConfig

GLContext adds FBO availability to profile mapping; Fix GL[Auto]Drawable/GLContext re-association (switch) incl. unit test.

- GLContext adds FBO availability to profile mapping

- GLContext added 'GLDrawable setGLDrawable(GLDrawable readWrite, boolean setWriteOnly)'
  allowing to set the write GLDrawable. This method enables switching context/drawable.

Fix GL[Auto]Drawable/GLContext re-association (switch) incl. unit test
  Commit eed8508ae1132e5f45f788e9cb3f3d5a1050ac70 impl. of GLAutoDrawable's setContext(..)
  enabled proper setting of the GLAutoDrawable context incl. updating the context's drawables.
  Test covers:
  - remove/set (GLContext, GLEventListener) of GL[Auto]Drawable
  - switch (GLContext, GLEventListener) of 2 GLAutoDrawables

FIX JOGL's index.html - Reminder to myself: Always check web pages befor pushing them :)

Enhance Bootsrapping of JOGL around 37% - 40% (1st start in new JVM) - GLProfile and GLContext*

GLProfile: Enhance bootsrapping performance of loading GL*Impl class
  - Offthread classloading of all GL*Impl via reflection at startup
    reduces startup time here around 12% (800ms down to 700ms).

GLContext*: Enhance bootsrapping performance of querying available GL profiles
  - Add PROFILE_ALIASING mode, defaults to true - can be disabled w/ property 'jogl.debug.GLContext.NoProfileAliasing'
  - PROFILE_ALIASING:
      If true (default), bootstrapping the available GL profiles
      will use the highest compatible GL context for each profile,
      hence skipping querying lower profiles if a compatible higher one is found.
        Linux x86_64 - Nvidia: 28%,  700ms down to 500ms
        Linux x86_64 - AMD   : 40%, 1500ms down to 900ms

- GL*Impl:
  - make fields final: glProfile, _context, buffer*Tracker and glStateTracker
  - allow null _context/glProfile in initialization (bootstrapping)

- JoglVersion.getDefaultOpenGLInfo(..)
  - add arg: 'boolean withCapabilitiesInfo', allowing to suppres the list of caps

test scripts windows: use new jogl-all.jar and jogl-test.jar

C code: remove warnings

Fix SWT GLCanvas threading. Note: On OSX _only_ it's main thread is valid!

Fix commit dfee8c58d4915f78f57545c26a492668b2b68a87

EGLGraphicsDevice adds desctruction callback and nativeDisplayID; EGLDisplayUtil adds creation of EGLGraphicsDevice.

Due to EGL's location in JOGL, supporting destruction of an EGLGraphicsDevice
is solved (hack) temporary by passing an eglTerminate callback to it's ctor.

Using EGLGraphicsDevice's close() method to also issue eglTerminate() simplifies the code.

In future we shall move EGL to nativewindow!

GLCapabilities*: Add 'isFBO()' and 'setFBO(boolean)', allowing upcoming impl. to select offscreen FBO (GLDrawable, ..)

Rename jogamp.newt.driver.kd.* classes to allow OpenKODE testing by passing
-cp jar/atomic/newt-driver-kd.jar -Dnativewindow.ws.name=jogamp.newt.driver.kd

Rename jogamp.newt.driver.kd.* classes to allow OpenKODE testing by passing
-cp jar/atomic/newt-driver-kd.jar -Dnativewindow.ws.name=jogamp.newt.driver.kd

Bug 599 - FBObject / Offscreen Support - Part 1

- New FBObject implementation handling FBO and it's attachments  *** API CHANGE: Util -> Core ***
  while it's size and sample-count can be reconfigured on the fly.
    - com.jogamp.opengl.util.FBObject -> com.jogamp.opengl.FBObject
    - agnostic to texture unit
    - separate attachments using OO hierarchy reflecting FBO
    - handling MSAA and blitting
    - no FBO destruction for reconfig (attach/detach)

- New GLFBODrawableImpl impl. an FBObject based GLDrawable
   - Instantiated by a dummy native surface (onscreen and invisible)
     hooked up to a dummy GLDrawable, which is the delegation for context creation.

   - Utilizies ProxySurface.UpstreamSurfaceHook for dummy surface
     avoiding specialization for native platforms.

   - TODO: Allow to utilize common surface interface as a
     dummy-surface to supporting API seperation of
     windowing/GL. The latter allows impl. of createGLDrawable(NativeSurface)
     with FBO.

- New OffscreenAutoDrawable (extends GLAutoDrawableDelegate)
  for all offscreen drawables. Shall replace GLPbuffer.

- New GLCapabilities*.isFBO() / setFBO(boolean) to request FBO offscreen,
  similar to isPBuffer(). Rule: if both are requested, FBO shall be favored.

- GLContext adds raw FBO availability query (min. FBO avail),
  FBObject contains fine grained queries (TODO: Move parts to GLContext for efficiency).

- Add framebuffer tracking, allowing fast querying:
  - GLBase/GLContext:
      public int getBoundFramebuffer(int target);
      public int getDefaultDrawFramebuffer();
      public int getDefaultReadFramebuffer();

  - GLContextImpl
      public final void setBoundFramebuffer(int target, int framebufferName)
      .. called by GL impl bind framebuffer

- GL: getDefaultDrawFramebuffer(), getDefaultReadFramebuffer()
  Adding default framebuffer queries being issued by
  GL.glBindFramebuffer(target, 0) w/ a default framebuffer, o.e. zero.
  This allows a transparent use of a custom FBO even in case the applications
  attempts to reset FBO to zero.
    Value flow: GL <- GLContext <- GLDrawable,

- GLCapabilities handle fbo/pbuffer seperate, don't disable the other

- GLContext/GL track read/write framebuffer to be queried by FBObject
  to determine whether to bind/unbind  a framebuffer

- Test cases for multiple FBO w/ and w/o MSAA

Other Features:

- New interface ProxySurface.UpstreamSurfaceHook,
  allowing to hook an upstream surface of unknown type
  providing lifecycle and information (size, ..) callbacks.
  Used for all new dummy NativeSurface impl and SWT GLCanvas.

  - GLContext -> GLDrawable propagation context/drawable lifecycle
    via ProxySurface.UpstreamSurfaceHook allowing dynamic resources
    to react (create, init, ..)
    - contextRealized()
    - contextMadeCurrent()

- SurfaceChangeable -> MutableSurface
  currently only contains setting the surface handle.

  TODO: May need to move ProxySurface.UpstreamSurfaceHook -> MutableSurface.UpstreamSurfaceHook,
        allowing other impl. classes (NEWT OffscreenWindow) to utilize the new
        upstream hookup mechanism - will allow FBO/Dummy window to work.

- SWT GLCanvas using ProxySurface.UpstreamSurfaceHook for proper size
  propagation.

- New GLAutoDrawable::getUpstreamWidget(), allowing GLEventListener
  to fetch the owning Java side UI element (NEWT, SWT, AWT, ..).

- GLDrawableFactory: Removed createOffscreenSurface() - unused and not GL related

- EGLDrawableFactory handles device/profile avail. mapping
  while actually creating context/drawable.
  This allows us to learn whether the ES context is software/hardware as well as FBO avail.

- EGLDrawable: Removed secret buckets of EGL configs :)
  Employ native surface (X11, WGL, ..) to EGL 'mapping' in
  EGLDrawableFactory utilizing new EGLUpstreamSurfaceHook (implements ProxySurface.UpstreamSurfaceHook).

Other Bugs:

- Add CTX_OPTION_DEBUG to ctx/extension cache key since only a debug ctx
  may expose the ARB debug capability.
  This bug caused lack of ARB/AMD debug functionality.

- Fix GLProfile deadlock (debug mode, w/ EGL/ES, no X11),
  dump availability information _after_ lock.

- ImmModeSink draw(): Use GL's glDrawElements(..), don't cast for GL2ES1.
  Fixes use for GL2ES2.

- Fix KeyEvent.getKeyChar() comment (-> only stable for keyTyped(..))

Misc:

- Refined alot of API doc

- New GLExtensions holds commonly used GL extension strings,
  allows better referencing and usage lookup.

- Move GL (interface) decl. to GLBase

- GLBuffers: Cleanup API doc (format, types)

- TextureIO: Add PAM and PPM static suffix identifier

- GLCapabilities getNumSamples() returns 0 if sampleBuffers is disabled, this seems to be more natural.

- finalized a lot

Fix NEWT exception handling in event dispatching: Catch and fwd to caller

Exceptions caused by NEWTEvent processing (on it's EDT)
were not propagated to the caller (diff thread).
Hence the EDT were brought down and the caller may have waited forever (deadlock).

Catch exception if caller waits and throw exception at waiting point.

GLCapabilities Native Aquisition: Set alpha bits at last - due to it's auto setting by setSampleBuffers(true) and setBackgroundOpaque(false)

This bug lead to X11 GLCapabilities rgba: 8/8/8/1 - which ofc is invalid. Sideeffect was a bad selected GLXFB configuration
and the GLContext couldn't be made current.

Patch sets alpha bits reflecting reality carefully after opaque/samples. Added API doc note.

Minor edits

Fix TestFocus01SwingAWTRobot failure duer to RedSquareES2's setFullscreen() toggle mouse adapter :)

Moved the fullscreen toggle mouse adapter to main test class.

Fix OSX regression of commit 20bf031db719f7baa4c6e74734fc999061e08fe2 - handling w/ non NSView handles (pbuffer)

tests.sh minor edit

Fix OSX OffscreenLayerSurface (OLS) regressions (pbuffer based)

- Use pbuffer (still), don't set FBO (invisible)

- OLS (only impl is JAWTWindow now) stores the attached layer handle
  created and attached by the GLContext implementation,
  so 'others' may detach it -> NewtCanvasAWT

- NewtCanvasAWT.removeNotify() needs to ask the OLS to detach the layer
  since it's parent will be gone.

- MacOSXCGLContext destroy allows a removed OLS (see above)

Fix TestSWTAccessor02GLn regression .. forgot to: canvas.addControlListener(swtUpstreamHook);

Fix Bug 606 - New AWT threading implementation breaks .. ; Fix GLAutoDrawable multi-threading w/ proper pattern (hope so)

Considering code changes and remarks:
  3ed491213f8f7f05d7b9866b50d764370d8ff5f6
  1a91ec5c8b6fd9d9db7bc115569c369fe7b38e9b
  3334a924309a9361a448d69bc707d4cce416b430
  4f27bcecf7484dc041551f52a5c49e2884cb3867

It seems necessary to have

- recursive locking employed for all semantic actions which changes drawable & context (and the Window resource)

- to avoid deadlock, we have to ensure the locked code segment will not spawn
  off to another thread, or a thread holds the lock, spawns of an action requiring the lock. .. sure

- other read-only methods (flags, ..) shall at least utilize a safe local copy of a volatile field
  if further use to produce the result is necessary.

- flags like sendReshape require to be volatile to guarantee it's being processed

Patch impacts: AWT/SWT GLCanvas, GLAutoDrawableBase [and it's specializations]
and hopefully closes any loopholes of missing a cache hit, etc.

If you review this and find optimizations, i.e. removing a lock due to semantics etc,
don't hold back and discuss it, please.

Align TestSWTAccessor02GLn w/ TestSWTJOGLGLCanvas01GLn, dropping the canvas cstr on SWT thread,
which seems not to be required.

Adding TODO note: Optimize/Fix NIO caching of glMapBuffer/glUnmapBuffer

Fix GraphicsConfigurationFactory: Map factory to device-type _and_ capabilities-type; Add a pre-set nativeVisualID to chooseGraphicsConfiguration(..)

Map factory to device-type _and_ capabilities-type:
  - Allows using different GraphicsConfigurationFactory implementations for different capabilities-types.
    Previous impl. failed to use an OpenGL agnostic CapabilitiesImmutable for 'chooseGraphicsConfiguration(..)'
    since only the GL aware factory was mapped. The latter failed since it expected a GLCapabilitiesImmutable.

  - The passed capabilities-type as well as device-type given at getFactory(..)
    is traversed top-to-down to find a most suitable factory:
        For-All devT := getTopDownDeviceTypes(deviceType)
            For-All capsT := getTopDownCapabilitiesTypes(capabilitiesType)
               f = factory.get(devT, capsT);
               if(f) { return f; }
            end
        end

Add a pre-set nativeVisualID to chooseGraphicsConfiguration(..)
  - In situations where a native visualID is already chosen [by external means for example],
    but we still need to query a matching GraphicsConfiguration - we require to pass
    a non VisualIDHolder.VID_UNDEFINED nativeVisualID.

    We had a hack implemented before within some implementations and their static calls,
    however an agnostic mechanism is required to implement new NativeSurface/Window's
    platform agnostic.

  - X11GLXGraphicsConfigurationFactory: respect a pre-set xvisualID

  - X11GLXDrawableFactory.createProxySurfaceImpl(..) queries the given windowHandle's
    visualID and 'chooses' the configuration accordingly. If the visualID is undefined
    an exception is thrown, since window is invalid.
    These mechanics are implicit for Windows and OSX.

Fix X11GLXGraphicsConfiguration.updateGraphicsConfiguration():
  - Skip any action if a valid X11GLCapabilities is already chosen, i.e. w/ visualID.
    Otherwise choose a suitable configuration incl. visualID.
    The latter is quite impossible and invalid, since visualID must be defined at window creation time
    and the update method is issued with a valid window.

X11 - Misc:
  - Added 'int jogamp.nativewindow.x11.X11Lib.GetVisualIDFromWindow(..)'

  - All returned visualID's are of type 'int'

SWT Update: SWT GLCanvas creates lazy when resource is ready; Create new NewtCanvasSWT allowing to parent NEWT windows natively.

SWT GLCanvas creates lazy when resource is ready
  - Ensures drawable and context are created when size > zero and native visualID is valid.
    The latter is platform dependent.
  - Note that you cannot utilize custom GLCapabilities w/ this one,
    since the configurations is already realized - use NewtCanvasSWT.

Create new NewtCanvasSWT allowing to parent NEWT windows natively:
  - Similar to NewtCanvasAWT
  - Allows attaching / detaching NEWT windows

NewtCanvasAWT: Public setNEWTChild(..) fixed

Added test cases for the above - tested on Linux, OSX and Windows w/ SWT

Note: As usual for OSX, add -XstartOnFirstThread

Details:

- NEWT Display has new method: 'EDTUtil setEDTUtil(EDTUtil)'
  allowing to set a custom event dispatch utility.
  We use this to set our SWTEDTUtil for using NEWT w/ SWT
  complying w/ SWT threading constraints.

Add property: 'jogl.disable.opengles' to disable querying and using OpenGL ES

This might be required by a few older buggy ES implementations.

Also assists to discable ANGLE is not properly detected (?)
on Windows 32bit - the latter causes SEGV within FF and Chrome.

TODO: Fix ANGLE detection and usage within broser

NOTE: ANGLE works fine standalone ..

Fix X11 Display Connection leak w/ new GLAutoDrawableBase code when used w/ offscreen drawables, reported by Mark Raynsford

New common GLAutoDrawableBase missed to close the AbstractGraphicsDevice
in case it has been created and dedicated for the passed GLDrawable.
This detailed knowledge is only known to the creator, hence it is passed
in the constructor and is being passed through all specializations.

Further more the new X11/GLX impl. of GLDrawableFactory's
'createMutableSurfaceImpl' always creates it's own private X11 display connection
to avoid locking / threading issues. Since the old implementation reused the
shared display connection which is prone to threading issues, this bug was not visible before.

Also fixed the unit test TestNEWTCloseX11DisplayBug565,
now correctly validating that no display connection is left over
after a new cycle of create/destroy of onscreen and offscreen drawables.

Use armv6 optimization for armhf to stay compatible with the new Raspberry Pi Raspbian distribution.

Adapt to GlueGen ARM Build Changes 422d7a5eb53fca6642ebf4e8910d8b0311bb2597

Note: Patch set not clean since build*xml files contain 'NEWT Broadcom' driver changes of upcoming commit
      due to pervious uncommitted work. Will be fixed w/ next commit.

NEWT: Adding support for BCM VC IV (Broadcom VideoCodec 4) and Linux console mouse tracker

Rasperry PI uses the 'BCM VC IV' GPU via console as it's default configuration.
This driver enables direct support for JOGL/NEWT.

Due to the lack of detection (TODO) users have to specify the Java property:
  -Dnativewindow.ws.name=jogamp.newt.driver.bcm.vc.iv

- Autodetection should be included in 'NativeWindowFactory._getNativeWindowingType()'
  while adding a new TYPE: 'TYPE_BCM_VC_IV'.

- Autodetection may need to detect whether an X11 Display runs and the installed
  EGL library uses it (instead of the default console one)

This work is authored in coop w/ Xerxes Rånby <xerxes@zafena.se>!

Include NEWT driver BCM_VC_IV to 'all' JAR (completes 0ca481381b51b4082ac2b3bbae80cfaf5b60c3b8)

EGLDisplayUtil: Workaround (latest) PVR 540 EGL regression where 3nd EGLDisplay's eglInitialize(..) fails ; Fix EGLDrawableFactory.getEGLSurface()

- EGLDisplayUtil: Workaround (latest) PVR 540 EGL regression where 3nd EGLDisplay's eglInitialize(..) fails
  In this case and if eglGetDisplay(..) fails w/ a non EGL_DEFAULT_DEVICE, fall back to EGL_DEFAULT_DEVICE - always.
  This workaround actually simplifies handling both cases.

- Fix EGLDrawableFactory.getEGLSurface()
  Tests whether a given NativeSurface w/ EGLGraphicsDevice and EGLGraphicsConfiguration
  has a valid EGL Surface. Only if true, reuse the whole NativeSurface,
  otherwise construct the missing pieces (device, config and use a WrappedSurface for EGL).

EGLGraphicsConfigurationFactory: Fix platform w/o GLX but X11 window handle.

Without GLX but X11 windows, there is no pre-configured X11GLXGraphicsConfigurationFactory set,
hence fetch the fallback X11GraphicsConfigurationFactory via:
  GraphicsConfigurationFactory.getFactory(com.jogamp.nativewindow.x11.X11GraphicsDevice.class, CapabilitiesImmutable.class)

TODO: Cleanup cases for other platform - simplify!

Fix FFMPEGMediaPlayer: Use GL_ALPHA (texture format intern/data) instead of GL_RGBA/GL_ALPHA ; Load dedicated native libav/libffmpeg

- ES2 spec does not allow GL_RGBA/GL_ALPHA.

- Load dedicated native libs (libav/libffmpeg) since distributions don't create
  symlink e.g. libavutil.so -> libavutil.so.53

Robostness: GLDrawableImpl's contextMadeCurrent()/contextRealized() ; GLFBODrawableImpl.contextMadeCurrent(false), OffscreenAutoDrawable.setSize(..)

GLDrawableImpl's contextMadeCurrent()/contextRealized():
  - Catch exception which may appear during callback and cont. w/ GLContextImpl's release()/destroy()
    while throwing catched exception at end.

GLFBODrawableImpl.contextMadeCurrent(false):
  - Detect null Colorbuffer ASAP and throw exception

OffscreenAutoDrawable.setSize(..):
  - Catch exceptions at 1) GLFBODrawableImpl.setSize(..) and
    2) GLContext.release() .. throw it in proper order.

Robustness: FBObject or reset(..)

- Add property 'jogl.fbo.force.min' to simulate minimum FBO caps

- Check new size in reset(..)

- Ignore the status in attachColorbuffer(..), attachRenderbufferImpl(..) and attachTexture2D(..)
  when recreating the resources in detachAllImpl(..)/reset(..).
  The status is validated after recreation of all resources in detachAllImpl(..)

Fix EGLDrawableFactory.getEGLSurface(..): Catch WrappedSurface case w/ EGLDevice and EGLGraphicsConfiguration - Regression of 43a473b2005d7f59a7f4f5b8bc7ca9ae88b4e894

Do not create a WrappedSurface around a given WrappedSurface
if the latter is intended for EGL, i.e. uses EGLDevice and EGLGraphicsConfiguration
even though the surface handle is not valid [yet].

For this case we assume the WrappedSurface has a lifecycle hook
as it is used for the dummy drawable.

Otherwise we would have a recursive WrappedSurface, ie with
EGLDrawableFactory.createOnscreenDrawableImpl( dummySurface ),
where dummySurface is a WrappedSurface w/ EGL pbuffer drawable lifecycle hook.

Commit 43a473b2005d7f59a7f4f5b8bc7ca9ae88b4e894 didn't take the above case into account.

Robostness FBObject / GLReadBufferUtil: Ignore pre-existing GL errors - remove GL error checking in FBObject bind/unbind.

- User GL code caused errors shall not fail impl.

- FBObject bind/unbind GL error checking is almost useless due to it's simple code,
  would only catch user GL code errors, which should be ignored here.

- MultisampleDemoES2: Only enable GL_MULTISAMPLE if available, i.e. validate passed multisample flag