4 years agoBug 852: Validate CPU sourced data API is allowed, throw exception if not. GLContext... v2.1.0
Sven Gothel [Thu, 10 Oct 2013 11:16:39 +0000 (13:16 +0200)]
Bug 852: Validate CPU sourced data API is allowed, throw exception if not. GLContext: isCPUSourcedAvail() -> isCPUDataSourcingAvail()

4 years agoBug 852: Remove CPU sourced data API entry where not allowed (ES3 and GL core >=...
Sven Gothel [Thu, 10 Oct 2013 00:15:56 +0000 (02:15 +0200)]
Bug 852: Remove CPU sourced data API entry where not allowed (ES3 and GL core >= 3.0) ; GL2 cfg: Ignore GL2ES3 symbols (super)

Remove CPU sourced data API entries via new config 'BufferObjectOnly  <name>',
listed in 'gl-common-gpubufferonly.cfg' and included in ES3 and all GL core >= 3 interfaces.

If BufferObjectOnly is defined for a function, only the 'long offset' variant is being emitted.

Due to limitations of GlueGen's 'ExtendedInterfaceSymbolsIgnore', which only
identifies the function name and not the signature - all CPU sourced variants
are manually added to the compatibility and ES2 GL profiles via '*' files.

GLContext: Added 'isCPUSourcedAvail()' to determine whether context allows CPU sourced data,
i.e. for GL2ES1 and GLES2 ctx.

GLContext/GLProfile/GL: isGLES2() now returns false if having a ES3 context due to 'CPU sourced'


GL2 cfg: Added ignore GL2ES3 symbols of it's superclass, removing duplicated symbols.

4 years agoNEWT Reparent/Fullscreen: Add 'waitForPosition(..)' when reparenting or back from...
Sven Gothel [Wed, 9 Oct 2013 15:12:33 +0000 (17:12 +0200)]
NEWT Reparent/Fullscreen: Add 'waitForPosition(..)' when reparenting or back from fullscreen; JOGLNewtAppletBase: Reparent to pos 32/32, trying to avoid browser window focus/top stealing on X11.

4 years agoBug816 OSX CALayer: Issue w/ JSplitPane within Apple (Firefox, Safari - not appletvie...
Sven Gothel [Wed, 9 Oct 2013 05:17:37 +0000 (07:17 +0200)]
Bug816 OSX CALayer: Issue w/ JSplitPane within Apple (Firefox, Safari - not appletviewer) when move horizontal slider (vertical: ok)

Moving horizontal slider if run as applet (Firefox, Safari - not appletviewer)
doesn't move the GLCanvas even though it is resized.

4 years agoNewtCanvasAWT Focus Traversal: Remove AWT's requestFocus*() overrides in favor of...
Sven Gothel [Wed, 9 Oct 2013 03:29:08 +0000 (05:29 +0200)]
NewtCanvasAWT Focus Traversal: Remove AWT's requestFocus*() overrides in favor of FocusPropertyChangeListener requestFocusNEWTChild()

The AWT's requestFocus*() overrides were intended to receive the AWT focus (default) and
clear it afterwards to forward the focus to the NEWT component -> requestFocusNEWTChild().

This can be achieved simply by using our FocusPropertyChangeListener
and invoking requestFocusNEWTChild() when receiving the focus on the NewtCanvasAWT component.

4 years agoNEWT Reparent/Fullscreen: Fixes X11 unsuccessful return to parent window; Add reparen...
Sven Gothel [Wed, 9 Oct 2013 03:24:45 +0000 (05:24 +0200)]
NEWT Reparent/Fullscreen: Fixes X11 unsuccessful return to parent window; Add reparentWindow(..) top-level position arguments; Misc

- Fixes X11 unsuccessful return to parent window
  On X11 when returning to parent window (-> CHILD),
  we have to set the window invisible and wait for the result.
  Otherwise it sometimes happens that the WM's reparent operation fails,
  i.e. the window won't become a child of desired parent and is positioned randomly.

- Add reparentWindow(..) top-level position arguments
  .. allows bringing the child-window to top-level w/ a desired position.
  Otherwise the window would be positioned elsewhere as a top-level
  as the plain reparenting operation.

  X11 needs to set position and size _after_ making the window visible,
  otherwise WM may ignore the XConfigureWindow request.

- Reparent recreate shall always store the desired position and size
  On OSX/CALayer when recreation is being used, we need to store the pos/size
  for later creation.

- Tests: Use 'NewtAWTReparentingKeyAdapter' where possible (reparent/fullscreen)
  instead of duplicating such code.

  NewtAWTReparentingKeyAdapter: Performs reparenting and fullscreen operations
  off-thread (i.e. not on AWT/NEW EDT) while decorating the action w/
  revoking/restoring the ExclusiveContextThread (ECT).

Manually tested 'TestGearsES2NewtCanvasAWT' reparenting and fullscreen
on X11, Windows and OSX/CALayer w/ JDK 7u40 successful.

4 years agoNEWT Display: Issue EDTUtil.start() at runOnEDTIfAvail(..) even if on EDT, which...
Sven Gothel [Wed, 9 Oct 2013 00:10:19 +0000 (02:10 +0200)]
NEWT Display: Issue EDTUtil.start() at runOnEDTIfAvail(..) even if on EDT, which is to be stopped.

This case appears on e.g. OSX/CALayer (offscreen) reparenting using recreation (onscreen <-> offscreen),
i.e. display destroy/create is performed on EDT.

Misc Cleanup:

- Rename EDTUtil: restart() -> start()

- Rename Display: validateEDT() -> validateEDTStopped()

- Simplify Display.setEDTUtil(..): Remove need for redundant 'newEDTUtil' local var.

- Simplify Display.runOnEDTIfAvail(..): edtUtil is never null

4 years agoNEWT/OSX (Bug 836): Only WindowDriver.initWindow0(..) operation on main-thread must...
Sven Gothel [Tue, 8 Oct 2013 23:53:53 +0000 (01:53 +0200)]
NEWT/OSX (Bug 836): Only WindowDriver.initWindow0(..) operation on main-thread must be synchronized (wait-until-done)

- Leave WindowDriver.close0(..) w/o sync

4 years agoNEWT/OSX (Bug 836): Lifecycle operations performed on main-thread must be synchronize...
Sven Gothel [Mon, 7 Oct 2013 05:47:12 +0000 (07:47 +0200)]
NEWT/OSX (Bug 836): Lifecycle operations performed on main-thread must be synchronized (wait-until-done)

Wait-until-done (main thread):
  - WindowDriver.close0(..)
  - WindowDriver.initWindow0(..)

Otherwise a re-queued operation (i.e. CALayer attachment)
will mixup the order ..

Experienced w/ fullscreen exit.

4 years agoNEWT/Fullscreen (Bug 836): Exit re-parenting fullscreen w/ [invisible-exit-visible...
Sven Gothel [Mon, 7 Oct 2013 05:41:30 +0000 (07:41 +0200)]
NEWT/Fullscreen (Bug 836): Exit re-parenting fullscreen w/ [invisible-exit-visible] (-> like reparent); Always force focus; X11: Always use ALWAYSONTOP.

- Remove unused 'fullscreenUseSpanningMode' state

- Exit re-parenting fullscreen w/ [invisible-exit-visible] (-> like reparent)
  solves X11 issue, where the NEWT window doesn't 'return to it's parent'.
  Probably also fixes Bug 836!

- Always force focus when enter and exit FS

- X11: Always use ALWAYSONTOP
  No reason to behave different (spanning and normal-fs)

4 years agoNativeWindow/OSX: Fix RunOnThread/RunLater - Properly determine 'forkOnMain' via...
Sven Gothel [Mon, 7 Oct 2013 05:36:17 +0000 (07:36 +0200)]
NativeWindow/OSX: Fix RunOnThread/RunLater - Properly determine 'forkOnMain' via onMain && ( isOnMain || 0 < delay )

4 years agoFFMPEGMediaPlayer: Use IOUtil.decodeFromURI(..) to unescape %20
Sven Gothel [Sun, 6 Oct 2013 23:08:06 +0000 (01:08 +0200)]
FFMPEGMediaPlayer: Use IOUtil.decodeFromURI(..) to unescape %20

4 years agoBug 729, Bug 849: Honor JAWTWindow's [parent] visibility state
Sven Gothel [Sun, 6 Oct 2013 19:25:42 +0000 (21:25 +0200)]
Bug 729, Bug 849: Honor JAWTWindow's [parent] visibility state

JAWTWindow snoops HierarchyEvents for SHOWING_CHANGED to
track local component's visibility state as well as it's global visibility state.
The later is determined by it's parent's visibility change.

If 'removeNotify()' is called, component's visibility is reset to it's local visibility state.

Fixes OSX CALayer component's visibility if parent's visibility changes (Bug 729).

Fixes continuous GL rendering if parent is invisible (Bug 849).

4 years agoBug 789: glDrawArraysIndirect and glDrawElementsIndirect are using GL_DRAW_INDIRECT_B...
Sven Gothel [Sun, 6 Oct 2013 13:56:44 +0000 (15:56 +0200)]
Bug 789: glDrawArraysIndirect and glDrawElementsIndirect are using GL_DRAW_INDIRECT_BUFFER, add VBO variant and range-check

4 years agoBug 729: OSX CALayer invisible needs to have opacity 0 to not show hidden component.
Sven Gothel [Sun, 6 Oct 2013 00:28:14 +0000 (02:28 +0200)]
Bug 729: OSX CALayer invisible needs to have opacity 0 to not show hidden component.

Completes commit 3b02a219b1b9e446e87df1beb7da4266f74824fa

See unit test: TestBug816OSXCALayerPos03AWT

4 years agoBug816 OSX CALayer: getLocationOnScreenNonBlocking(..) stop traversing at top-level...
Sven Gothel [Sat, 5 Oct 2013 14:13:56 +0000 (16:13 +0200)]
Bug816 OSX CALayer: getLocationOnScreenNonBlocking(..) stop traversing at top-level heavyweight Window, fixes [J]Dialog use case.

Top-level heavyweight Window's position is it's position on screen, hence stop traversing.

Thx to Emmanuel Puybaret for pointing this out.

Adding Dialog and JDialog unit tests as inspired by Emmanuel Puybaret.

4 years agotest scripts (test Bug 729 w/ TestAWTCardLayoutAnimatorStartStopBug532)
Sven Gothel [Sat, 5 Oct 2013 11:45:40 +0000 (13:45 +0200)]
test scripts (test Bug 729 w/ TestAWTCardLayoutAnimatorStartStopBug532)

4 years agoBug 729: OSX CALayer shall honor the Component's visibility state
Sven Gothel [Sat, 5 Oct 2013 11:44:43 +0000 (13:44 +0200)]
Bug 729: OSX CALayer shall honor the Component's visibility state

A once visible CALayer (GLCanvas) must be able to become invisible w/o destruction,
e.g. as required by CardLayout's switching cards.

See unit test for Bug 532: 'TestAWTCardLayoutAnimatorStartStopBug532'

Out native 'fixCALayerLayout(..)' takes the visible state as tracked by JAWTWindow's ComponentListener
and sets our CALayer (root and sub) hidden state accordingly.

Now MacOSXJAWTWindow's layoutSurfaceLayerImpl(..) always calls down to 'fixCALayerLayout(..)'
due to update the visibility state.

4 years agoBug 816/848: Cleanup test applet layout/size ; Note Bug 848: Not reproducible after...
Sven Gothel [Fri, 4 Oct 2013 19:08:45 +0000 (21:08 +0200)]
Bug 816/848: Cleanup test applet layout/size ; Note Bug 848: Not reproducible after restart of OSX test machine!

4 years Add
Sven Gothel [Fri, 4 Oct 2013 17:01:29 +0000 (19:01 +0200)] Add

4 years agoBug 848: Add test applet ; Bug 816 Bug816AppletGLCanvas02: Remove 2nd Animator ;...
Sven Gothel [Fri, 4 Oct 2013 17:00:28 +0000 (19:00 +0200)]
Bug 848: Add test applet ; Bug 816 Bug816AppletGLCanvas02: Remove 2nd Animator ; Fix intendation (TAB -> WS)

4 years agoBug 818: Add missing html file for applet unit test of commit 51ea4e5d5d212dfef519cf8...
Sven Gothel [Fri, 4 Oct 2013 07:25:13 +0000 (09:25 +0200)]
Bug 818: Add missing html file for applet unit test of commit 51ea4e5d5d212dfef519cf824d6cbf69e41bebc7

4 years agoAWTTilePainter: Avoid NPE if Graphic2D's AffineTransform is null, use identity tranfo...
Sven Gothel [Thu, 3 Oct 2013 21:39:43 +0000 (23:39 +0200)]
AWTTilePainter: Avoid NPE if Graphic2D's AffineTransform is null, use identity tranform in such case.

4 years agoBug 818: Enable new quirk GLSLNonCompliant on OSX < 10.7 w/ NV GPU; GLJPanel: Don...
Sven Gothel [Thu, 3 Oct 2013 18:21:23 +0000 (20:21 +0200)]
Bug 818: Enable new quirk GLSLNonCompliant on OSX < 10.7 w/ NV GPU; GLJPanel: Don't use GLSL flip if quirk GLSLNonCompliant is present.

4 years agoBug 548: GLRendererQuirks.GLFlushBeforeRelease is needed on OSX < 10.7.3 w/ NV GPU...
Sven Gothel [Thu, 3 Oct 2013 17:15:25 +0000 (19:15 +0200)]
Bug 548: GLRendererQuirks.GLFlushBeforeRelease is needed on OSX < 10.7.3 w/ NV GPU [was comparing against 1.7.3 instead 10.7.3] !

4 years agoBug 816 (OSX CALayer pos): Add detailed description (pseudo code) to JAWTUtil.JAWT_OS...
Sven Gothel [Thu, 3 Oct 2013 13:35:09 +0000 (15:35 +0200)]
Bug 816 (OSX CALayer pos): Add detailed description (pseudo code) to JAWTUtil.JAWT_OSX_CALAYER_QUIRK_LAYOUT

4 years agoBug 816 (OSX CALayer pos): Fix location on 'inner CALayer' calculation
Sven Gothel [Thu, 3 Oct 2013 13:19:29 +0000 (15:19 +0200)]
Bug 816 (OSX CALayer pos): Fix location on 'inner CALayer' calculation

'inner CALayer' is the outter AWT Window client space (content).


  p0 = c.locationOnScreen();
  p0 -= c.getOutterComp.getPos();
  p0 -= c.getOutterComp.getInsets();

Where 'locationOnScreen()' is:
 p0 = 0/0;
 while( null != c ) {
   p0 += c.getPos();


JAWTWindow.getLocationOnScreenNonBlocking(..) validated against
AWT's Component.getLocationOnScreen() - OK for all use-cases.
(Validation enabled w/ DEBUG)

All unit tests manually validated on OSX 10.7 w/ jdk7u40.

4 years agoBug 816: Add 2 more unit tests (Swing JFrame w/ 2 JRootPanes, AWT Frame BorderLayout...
Sven Gothel [Thu, 3 Oct 2013 13:07:07 +0000 (15:07 +0200)]
Bug 816: Add 2 more unit tests (Swing JFrame w/ 2 JRootPanes, AWT Frame BorderLayout w/ Checkbox North, GLCanvas Center)

Thx to Emmanuel Puybaret

4 years agoBug818GLJPanelApplet: Remove unused imports
Sven Gothel [Wed, 2 Oct 2013 22:22:15 +0000 (00:22 +0200)]
Bug818GLJPanelApplet: Remove unused imports

4 years agoGLJPanel: Fix 7u40 access to property 'sun.java2d.opengl'. Property was trusted before ..
Sven Gothel [Wed, 2 Oct 2013 22:03:27 +0000 (00:03 +0200)]
GLJPanel: Fix 7u40 access to property 'sun.java2d.opengl'. Property was trusted before ..

4 years agoBug 818 (OSX 10.6.8 NVIDIA GeForce GT 330M): Add unit test (manual applet)
Sven Gothel [Wed, 2 Oct 2013 22:02:20 +0000 (00:02 +0200)]
Bug 818 (OSX 10.6.8 NVIDIA GeForce GT 330M): Add unit test (manual applet)

4 years agoAdd notice about 'GL states' regarding GLSL vertical flip (Bug 842 and Bug 826).
Sven Gothel [Wed, 2 Oct 2013 14:29:37 +0000 (16:29 +0200)]
Add notice about 'GL states' regarding GLSL vertical flip (Bug 842 and Bug 826).

4 years agoGLJPanel/AWTGLPixelBuffer: Reused BufferedImage didn't account for row-stride (regres...
Sven Gothel [Tue, 1 Oct 2013 23:12:20 +0000 (01:12 +0200)]
GLJPanel/AWTGLPixelBuffer: Reused BufferedImage didn't account for row-stride (regression of b33bdf41cf53f37203643a8551bf5d94b42a8fab)

SingleAWTGLPixelBufferProvider w/ allowing row-stride reuses the AWTGLPixelBuffer and it's BufferedImage
even w/ different width.

This leads to distortion if using the BufferedImage unhandled.

GLJPanel also set GL_PACK_ROW_LENGTH to pixelBuffer.width, which leads to an 'out-of-bounds'
exception if ReadPixels is used w/ panelwidth and panelHeight.


Introduce AWTGLPixelBuffer.getAlignedImage(width, height) which returns an aligned
BufferedImage while reusing the DataBuffer.

GLJPanel fetches a new alignedImage if required.

This allows a more efficient single buffer usage as intended, w/o the need of copying data.

4 years agotest scripts/ Remove JogAmp *LD_LIBRARY_PATH components.
Sven Gothel [Tue, 1 Oct 2013 13:36:56 +0000 (15:36 +0200)]
test scripts/ Remove JogAmp *LD_LIBRARY_PATH components.

4 years agoGLJPanel/Printing: Perform backend.reshape immediatly after printing if resized;...
Sven Gothel [Sun, 29 Sep 2013 08:25:32 +0000 (10:25 +0200)]
GLJPanel/Printing: Perform backend.reshape immediatly after printing if resized; Add unit test for resize while printing.

Perform immediatly reshape at releasePrint on AWT-EDT:
  - sendReshape = handleReshape(); // reshapeSize -> panelSize, backend reshape w/ GL reshape

4 years agoGLJPanel: Remove forced DEBUG .. at handleReshape in releasePrint (added w/ commit...
Sven Gothel [Sat, 28 Sep 2013 02:27:05 +0000 (04:27 +0200)]
GLJPanel: Remove forced DEBUG .. at handleReshape in releasePrint (added w/ commit fe284b515d984198fdbe702837809f181625a457)

4 years agoFix TileRenderer regression of commit c8abb9d2c9a8c92f0c5c42aba13e3e80c69739dc: Test...
Sven Gothel [Sat, 28 Sep 2013 02:08:59 +0000 (04:08 +0200)]
Fix TileRenderer regression of commit c8abb9d2c9a8c92f0c5c42aba13e3e80c69739dc: Test isSetup() _after_ running glEventListenerPre.

glEventListenerPre may be utilized to setup the TileRenderer.

4 years agoAWT Printing: releasePrint() issue reshape from AWT-EDT; GLJPanel: Issue 'handleResha...
Sven Gothel [Sat, 28 Sep 2013 02:03:04 +0000 (04:03 +0200)]
AWT Printing: releasePrint() issue reshape from AWT-EDT; GLJPanel: Issue 'handleReshape' if panelSize/awtSize mismatch, otherwise 'sendReshape' and exception if offscreen size doesn't match panelSize.

4 years agoTileRendererBase.GLEL.display(): Skip tile-rendering if TR is not yet setup. Sync...
Sven Gothel [Fri, 27 Sep 2013 23:52:39 +0000 (01:52 +0200)]
TileRendererBase.GLEL.display(): Skip tile-rendering if TR is not yet setup. Sync issue w/ NEWT/AWT based GLAD

NEWT based GLDrawables may trigger GLAD display() via native repaint events.

If using in conjunction w/ AWT, i.e. NewtCanvasAWT and setupPrinting(..) has been called
and it's attched to the TR .. it could happen that display tries to issue beginTile()
before the TR is being setup.

This patch mitigates this issue (while not removing it) by querying whether setup is completed.

4 years agoFix AWT Printing: Use delegated GLDrawable's size for tile-size instead of AWT-GLAD...
Sven Gothel [Fri, 27 Sep 2013 23:48:45 +0000 (01:48 +0200)]
Fix AWT Printing: Use delegated GLDrawable's size for tile-size instead of AWT-GLAD's size, since the latter uses AWT's component size.

.. the AWT component size could have been modified after setupPrinting(..).

The AWT-GLAD getWidth() and getHeight() is implemented by AWT's component
and hence may not reflect the actual GLDrawable size while printing.

4 years agoBug 816: Clarify JAWT_OSX_CALAYER_QUIRK_* semantics, add JAWT_OSX_CALAYER_QUIRK_LAYOUT
Sven Gothel [Fri, 27 Sep 2013 22:20:50 +0000 (00:20 +0200)]

4 years agoTest applets .. change font of misc test section.
Sven Gothel [Fri, 27 Sep 2013 18:25:40 +0000 (20:25 +0200)]
Test applets .. change font of misc test section.

4 years agoFix Bug 816: JAWTWindow.getLocationOnScreenNonBlocking(..) shall use JRootPane for...
Sven Gothel [Fri, 27 Sep 2013 13:31:17 +0000 (15:31 +0200)]
Fix Bug 816: JAWTWindow.getLocationOnScreenNonBlocking(..) shall use JRootPane for last position offset; AWTMisc.getInsets(..) add special JRootPane case.

Please note that we use JAWTWindow.getLocationOnScreenNonBlocking(..) to determine the
location on the screen 'only' b/c we cannot allow AWT to aquire the tree-lock!
The latter would be the case if using AWT's 'getLocationOnScreen()'.

If anybody has a more reliable implementation to achieve the same .. please provide your patch!

The following fix has been performed to fix the last issue w/ vZome.

- JAWTWindow.getLocationOnScreenNonBlocking(..) shall use JRootPane for last position offset
  - w/ vZome, the frame's position was 0/0 (invalid), instead JRootPane's position is good.
    Use JRootPane's position and stop traversing here (LW top-level).

- AWTMisc.getInsets(..) add special JRootPane case.
+     * Exception is JRootPane.
+     * Return it's parent's Window component's insets if available,
+     * otherwise return JRootPane's insets.<br>
+     * This is due to <i>experience</i> that <i>some</i> JRootPane's
+     * do not expose valid insets value.

4 years agoTest AWT Printing: Frame size contains border/insets, use as-is for scaling; Only...
Sven Gothel [Fri, 27 Sep 2013 11:49:59 +0000 (13:49 +0200)]
Test AWT Printing: Frame size contains border/insets, use as-is for scaling; Only force TEXT AA print-rendering hint.

4 years agoBug 816: Fix JAWTWindow's getLocationOnScreenNonBlocking(); Derive CALayer position...
Sven Gothel [Fri, 27 Sep 2013 11:23:39 +0000 (13:23 +0200)]
Bug 816: Fix JAWTWindow's getLocationOnScreenNonBlocking(); Derive CALayer position from AWT component's location on screen. Track fixedFrame size of root CALayer; Add Split layout to unit test, add [manual] Applet tests.

- Fix JAWTWindow's getLocationOnScreenNonBlocking()
  Skip JRootPane while traversing up to root Container.
  JRootPane would duplicate the top-level container's offset (Window insets).

- Derive CALayer position from AWT component's location on screen. Add Split layout to unit test, add [manual] Applet tests.
  AWT >= 7u40:
  - AWT position is top-left w/ insets, where CALayer position is bottom/left from root CALayer w/o insets.
  - Use getLocationOnScreenNonBlocking() to get location-on-screen w/o insets.
  - Native code: flip origin

  AWT < 7u40 still uses fixed position 0/0 for root and sub layer.

- Track fixedFrame size of root CALayer - MyCALayer:
  - Override layoutSublayers to validate root and sub-layer pos/size
  - Override setFrame to use fixedFrame, if set (similar to MyNSOpenGLLayer)

- Add Split layout to unit test, add [manual] Applet tests.
  - Thx to 'jimthev' and 'Manu' for providing Applet unit tests

4 years agoTileRenderer*: TileRendererNotify -> TileRendererListener w/ clarifying functionality...
Sven Gothel [Fri, 27 Sep 2013 05:13:24 +0000 (07:13 +0200)]
TileRenderer*: TileRendererNotify -> TileRendererListener w/ clarifying functionality (reshapeTile(..), ..); Only process GLEventListener impl. TileRendererListener; attachToAutoDrawable -> attachAutoDrawable, etc.

-TileRendererNotify -> TileRendererListener
  - Added methods:
    - void reshapeTile(TileRendererBase tr,int tileX, int tileY, int tileWidth, int tileHeight, int imageWidth, int imageHeight);
    - void startTileRendering(TileRendererBase tr);
    - void endTileRendering(TileRendererBase tr);

  allowing to clarify user code and API specification,
  i.e. TR only processes GLEventListener which impl. TileRendererListener.

  This also allows simplifying the API doc, while having a more descriptive
  reshape method focusing solely on tile rendering.

  Further more, the start/end TR methods allow certain GL related actions
  while the context is current before and after iterating through the tiles.
  This is even used for RandomTileRenderer (one tile only), to allow
  to reuse same TileRendererListener for diff TRs.

- Fix language, attach and detach usage was vice versa. We do attach an GLAutoDrawable to a TR
  - attachToAutoDrawable -> attachAutoDrawable
  - detachFromAutoDrawable -> detachAutoDrawable

- Adapted unit tests.

4 years agoDirectDataBufferInt: createBufferedImage(..) use BufferedImage type value instead...
Sven Gothel [Thu, 26 Sep 2013 23:53:18 +0000 (01:53 +0200)]
DirectDataBufferInt: createBufferedImage(..) use BufferedImage type value instead of component-count, support all int types ; Add type 'BufferedImageInt' preserving the custom image-type, add note about TYPE_CUSTOM

- createBufferedImage(..) use BufferedImage type value instead of component-count, support all int types
  - Support all integer data image-type, hence we need to pass image-type instead of component count (collision).
  - Also pass 'properties' to allow configuring all BufferedImage parameters (just in case)
  - Return BufferedImageInt to allow user to query the used image-type, see below.

- Add type 'BufferedImageInt' preserving the custom image-type, add note about TYPE_CUSTOM
  - Simply extends BufferedImage w/ custom image-type, since BufferedImage's type is TYPE_CUSTOM
    due to our custom storage type (see API doc comment).

Unit tests:
  - Testing all integer image-type's in
    - TestTiledPrintingGearsSwingAWT
    - TestTiledPrintingNIOImageSwingAWT

  - Disable all AA print-hints, all AWT text visible on X11.
    Probably has a regression on Windows / OSX .. TBD.

4 years agoFix MovieCube/Simple Demo in 'Applet' mode: Use default ctor to completly initialize...
Sven Gothel [Wed, 25 Sep 2013 03:57:43 +0000 (05:57 +0200)]
Fix MovieCube/Simple Demo in 'Applet' mode: Use default ctor to completly initialize GLMediaPlayer demo w/ def. stream etc; Fix list of command keys in applet html page.

4 years agoOSXMisc.c: Fix compilation error w/ java6's jint definition (probably clang as well)
Sven Gothel [Wed, 25 Sep 2013 00:08:42 +0000 (02:08 +0200)]
OSXMisc.c: Fix compilation error w/ java6's jint definition (probably clang as well)

Regression of commit 4b5435c68c3f12d62dadb395957362eceacfb25c

4 years agoBug 816: Fix OSX CALayer 'quirks' for AWT 1.7.0_40 - See JAWTUtil JAWT_OSX_CALAYER_QU...
Sven Gothel [Tue, 24 Sep 2013 21:13:16 +0000 (23:13 +0200)]

- Provide quirk bits for OSX CALayer depending on used JVM/AWT
  and act accordingly.

- TestBug816OSXCALayerPosAWT: Add resize by frame

4 years agoTileRendererBase: Remove redundant reshape propagation of atached GLEL, call pre...
Sven Gothel [Tue, 24 Sep 2013 18:15:51 +0000 (20:15 +0200)]
TileRendererBase: Remove redundant reshape propagation of atached GLEL, call pre/post reshape in it's display method instead.

4 years agoBug 816: Add unit test 'TestBug816OSXCALayerPosAWT'
Sven Gothel [Tue, 24 Sep 2013 09:04:24 +0000 (11:04 +0200)]
Bug 816: Add unit test 'TestBug816OSXCALayerPosAWT'

4 years agoFix Bug 840: DefaultEDTUtil.restart() shall only reuse ThreadGroup (tg) is not destro...
Sven Gothel [Sun, 22 Sep 2013 00:35:11 +0000 (02:35 +0200)]
Fix Bug 840: DefaultEDTUtil.restart() shall only reuse ThreadGroup (tg) is not destroyed, otherwise use current thread's tg.

With jdk7u40, when re-launching a NEWT applet (JOGLNewtApplet1Run),
i.e. via browser back and forth, the following exception happens:

java.lang.RuntimeException: java.lang.IllegalThreadStateException
    at com.jogamp.newt.awt.applet.JOGLNewtApplet1Run.init(
    at com.sun.deploy.uitoolkit.impl.awt.AWTAppletAdapter.init(Unknown Source)
    at sun.plugin2.applet.Plugin2Manager$ Source)
    at Source)
Caused by: java.lang.IllegalThreadStateException
    at java.lang.ThreadGroup.addUnstarted(Unknown Source)
    at java.lang.Thread.init(Unknown Source)
    at java.lang.Thread.<init>(Unknown Source)
    at jogamp.newt.DefaultEDTUtil$NEDT.<init>(
    at jogamp.newt.DefaultEDTUtil.restart(
    at jogamp.newt.DisplayImpl.runOnEDTIfAvail(
    at jogamp.newt.WindowImpl.runOnEDTIfAvail(
    at jogamp.newt.WindowImpl.setUndecorated(
    at com.jogamp.newt.opengl.GLWindow.setUndecorated(
    at com.jogamp.newt.awt.applet.JOGLNewtApplet1Run.init(
    ... 3 more

This is due to 7u40's changed ThreadGroup (tg) lifecycle, i.e. the tg is destroyed.
In such case, DefaultEDTUtil.restart() shall use the current threads tg.

4 years agoFix Bug 839: Clarify whether resetStates(..) shall clearStates() - Pass 'isInit'...
Sven Gothel [Sat, 21 Sep 2013 03:19:32 +0000 (05:19 +0200)]
Fix Bug 839: Clarify whether resetStates(..) shall clearStates() - Pass 'isInit' flag.

4 years agoAWT Printing: Fix non vertical-flipped printing, i.e. cut-off top-row's tile upper...
Sven Gothel [Fri, 20 Sep 2013 10:29:49 +0000 (12:29 +0200)]
AWT Printing: Fix non vertical-flipped printing, i.e. cut-off top-row's tile upper area, cleanup.

TestTiledPrintingGearsSwingAWT2: Provoked !flipped bug where top-row was positioned too low
due to using full size tile-height.

Cutting of the unused top-row's upper area corrects this issue.

vertical-flip mode does not expose this situation, since flipping
shifts the payload to the upper tile area.

TestTiledPrintingGearsSwingAWT2: Also tests an alternative transparent overlapping mode
without layout.

4 years agoFix Bug 826: GLJPanel: Fully restore TextureState and Viewport
Sven Gothel [Fri, 20 Sep 2013 10:18:48 +0000 (12:18 +0200)]
Fix Bug 826: GLJPanel: Fully restore TextureState and Viewport

In FBO mode save TextureState of current texture-unit,
as well as for the fbo texture-unit if the latter is a different.

In glslTextureRaster mode for verical flip (using FBO),
set the viewport to drawable size if before flipping and restore afterward - if equired.

TestGLJPanelTextureStateAWT fully tests use cases:
  - Keep texture bound w/ same or other texture-unit
  - Use 2 viewports within one drawable and keep it

4 years agoBug 839: Clarifying GLStateTracker.clearStates(..) remove 'enable' change - Part 1
Sven Gothel [Fri, 20 Sep 2013 09:37:38 +0000 (11:37 +0200)]
Bug 839: Clarifying GLStateTracker.clearStates(..) remove 'enable' change - Part 1

TODO: Only disable state tracker at GLContext.destroy()

4 years agoFix Bug 835: Use target.[sourcelevel,targetlevel,rt.jar] for all javac tasks.
Sven Gothel [Thu, 19 Sep 2013 07:20:03 +0000 (09:20 +0200)]
Fix Bug 835: Use target.[sourcelevel,targetlevel,rt.jar] for all javac tasks.

Before we bumped java language version to 1.6, i.e. target == 1.5,
we required to compile some parts w/ the host level (1.6).

This is no more required.

4 years agoTestTiledPrintingGearsSwingAWT: Enable 'bitmap' test only on Windows, issues w/ AMD...
Sven Gothel [Thu, 19 Sep 2013 06:00:26 +0000 (08:00 +0200)]
TestTiledPrintingGearsSwingAWT: Enable 'bitmap' test only on Windows, issues w/ AMD catalyst driver and pixmap surface ..

4 years agoBump to JDK/JRE 7u40.
Sven Gothel [Thu, 19 Sep 2013 02:17:51 +0000 (04:17 +0200)]
Bump to JDK/JRE 7u40.

4 years agoAWTTilePainter.setupGraphics2DAndClipBounds(): Use 'Shape getClip()'; Use double...
Sven Gothel [Thu, 19 Sep 2013 01:21:15 +0000 (03:21 +0200)]
AWTTilePainter.setupGraphics2DAndClipBounds(): Use 'Shape getClip()'; Use double precicion clip bounds 'all the way'; Explicitly scale image and clip w/ current scaled transform.

- Use 'Shape getClip()'
  Don't assume Rectangle2D, but use Shape's getBounds2D()

- Use double precicion clip bounds 'all the way'
  Remove rounding error on clip bounds w/ start value, which was _not_ using doubles.

- Explicitly scale image and clip w/ current scaled transform.
  Instead of abusing Graphics2D's clip shape to scale image size and clip-area,
  explicitly use transform both bounding boxes into transformed space,
  scale space and transform out (inversion).
  A possible NoninvertibleTransformException will be thrown while Graphics2D has not been modified.

4 years agoAWT Print Tests: Only include 600dpi tests if requested via main arg '-600dpi'
Sven Gothel [Thu, 19 Sep 2013 00:57:41 +0000 (02:57 +0200)]
AWT Print Tests: Only include 600dpi tests if requested via main arg '-600dpi'

4 years agoAWTTilePainter: Fix case with no GLOrientation, i.e. no vertical-flip. Tile location...
Sven Gothel [Wed, 18 Sep 2013 15:31:45 +0000 (17:31 +0200)]
AWTTilePainter: Fix case with no GLOrientation, i.e. no vertical-flip. Tile location and destination must follow same math as w/ vertical-flip.

.. clipping and tile-height was not considered.

4 years agoRedSquareES1: Add TileRendererBase.TileRendererNotify
Sven Gothel [Wed, 18 Sep 2013 15:13:46 +0000 (17:13 +0200)]
RedSquareES1: Add TileRendererBase.TileRendererNotify

4 years agoAWTTilePainter: Fix DEBUG message (used wrong value at println)
Sven Gothel [Wed, 18 Sep 2013 00:30:02 +0000 (02:30 +0200)]
AWTTilePainter: Fix DEBUG message (used wrong value at println)

4 years agoFix SharedResourceRunner's potential race-conditions. Use top-level synchronization...
Sven Gothel [Wed, 18 Sep 2013 00:29:24 +0000 (02:29 +0200)]
Fix SharedResourceRunner's potential race-conditions. Use top-level synchronization simplifying code and better robustness.

4 years agoAWT Printing: Move init-test of setupPrint(..) to AWT-EDT Runnable ; GLJPanel: Attemp...
Sven Gothel [Tue, 17 Sep 2013 14:09:10 +0000 (16:09 +0200)]
AWT Printing: Move init-test of setupPrint(..) to AWT-EDT Runnable ; GLJPanel: Attempt to initialize if not done yet (similar to GLCanvas)

4 years agoPrint Tests: Split 'Printable' to own class, add OffscreenPrintable using NIO Buffere...
Sven Gothel [Tue, 17 Sep 2013 01:01:11 +0000 (03:01 +0200)]
Print Tests: Split 'Printable' to own class, add OffscreenPrintable using NIO BufferedImage, adding OffscreenPrintable tests to all unit tests.

4 years agoTestTiledPrintingNIOImageSwingAWT: Add layered test ; Use 300dpi instead of 600dpi
Sven Gothel [Mon, 16 Sep 2013 01:13:16 +0000 (03:13 +0200)]
TestTiledPrintingNIOImageSwingAWT: Add layered test ; Use 300dpi instead of 600dpi

4 years agoAdd 'DirectDataBufferInt' supporting NIO based BufferedImages
Sven Gothel [Mon, 16 Sep 2013 00:59:30 +0000 (02:59 +0200)]
Add 'DirectDataBufferInt' supporting NIO based BufferedImages

- Contains convenient method to create an NIO backed BufferedImage
  'BufferedImage createBufferedImage(int width, int height, int numComponents, Point location)'

Standalone test TestTiledPrintingNIOImageSwingAWT 'prints', i.e. renders offscreen,
the Frame into
  - a traditional array-backed BufferedImage
  - an NIO-backed backed BufferedImage

4 years agoUpdating doc AWTPrinting.txt
Sven Gothel [Sun, 15 Sep 2013 21:46:35 +0000 (23:46 +0200)]
Updating doc AWTPrinting.txt

4 years agoFix AWT printing issues w/ overlapping and/or non-opaque contents ; Change AWTPrintLi...
Sven Gothel [Sun, 15 Sep 2013 21:27:16 +0000 (23:27 +0200)]
Fix AWT printing issues w/ overlapping and/or non-opaque contents ; Change AWTPrintLifecycle's lifecycle

- AWTPrintLifecycle:
  - Should decorate:
    instead of within Printable.print(..) { .. container.printAll(..); .. }

    This is due to AWT print implementation, i.e.
    AWT will issue Printable.print(..) multiple times for 'overlapping'
    or non-opaque elements!

  - Move from -> com.jogamp.nativewindow.awt
  - Make _interface_ AWT agnostic, i.e. remove Graphics2D from 'setup(..)'
  - Add 'int numSamples' to 'setup(..)' to determine the number of samples

- AWTTilePrinter:
  - Use double precision when scaling image-size and clip-rect,
    then round them to integer values.
    Otherwise AWT will use the bounding box for the clipping-rectangular.

  - Clip negative portion of clip-rect,
    this removes redundant overpaints, as well as increasing the tile count
    due to the increased clipping-size.

  - Clip the image-size in the tile-renderer according to the clip-rect.

  - DEBUG_TILES: Dump tiles to file

  - Use sub-image of final BuffereImage instead
    of adding another clipping region. This might increase performance
    if no clip-rect has been set.

  TestTiledPrintingGearsSwingAWT overlapping tests exposes
  a 'off by one' bug of the first layer's background!
  Note: The GL content seems to be correct though - maybe it's simply an AWT rounding error ..

4 years agoTileRenderer: Add and expose clipped image size; Expose size of tiles.
Sven Gothel [Sun, 15 Sep 2013 20:52:47 +0000 (22:52 +0200)]
TileRenderer: Add and expose clipped image size; Expose size of tiles.

Add clipping of the image-size and hence differentiate the image-size and
the size the tile-renderer iterates through.

The original image-size is required for the opengl reshape and rendering,
where the clipping size may restrict the range of rendering.

4 years agoAWTTilePainter: Fix null clip-rect (consider scaling); Fix non GL-oriented drawable...
Sven Gothel [Fri, 13 Sep 2013 17:47:12 +0000 (19:47 +0200)]
AWTTilePainter: Fix null clip-rect (consider scaling); Fix non GL-oriented drawable, skip vertical flip and use 1:1 y-coord.

4 years agoComplete commit 4b866d2686ab9c3fd7cf6708925b4663ad81e359: Relocate FFMPEGNatives...
Sven Gothel [Fri, 13 Sep 2013 17:44:38 +0000 (19:44 +0200)]
Complete commit 4b866d2686ab9c3fd7cf6708925b4663ad81e359: Relocate FFMPEGNatives.initIDS0() -> FFMPEGStaticNatives.initIDS0(); Cleanup up warnings and includes (clang); Forgot to commit new ffmpeg_static.h

4 years agoAdd support for clang ; Use __APPLE__ predefined macro
Sven Gothel [Fri, 13 Sep 2013 17:21:52 +0000 (19:21 +0200)]
Add support for clang ; Use __APPLE__ predefined macro

4 years agoRelocate FFMPEGNatives.initIDS0() -> FFMPEGStaticNatives.initIDS0(); Cleanup up warni...
Sven Gothel [Wed, 11 Sep 2013 18:23:32 +0000 (20:23 +0200)]
Relocate FFMPEGNatives.initIDS0() -> FFMPEGStaticNatives.initIDS0(); Cleanup up warnings and includes (clang).

4 years agoFix TileRendererBase Bug: Avoid multiple GLEL init(..) call.
Sven Gothel [Wed, 11 Sep 2013 14:40:42 +0000 (16:40 +0200)]
Fix TileRendererBase Bug: Avoid multiple GLEL init(..) call.

4 years agoGears, GearsES2, RedSquareES2: Proper debug output for init, reshape and dispose
Sven Gothel [Wed, 11 Sep 2013 14:39:54 +0000 (16:39 +0200)]
Gears, GearsES2, RedSquareES2: Proper debug output for init, reshape and dispose

4 years agoAdd 2nd 'demo' canvas to the other tiled printing tests.
Sven Gothel [Wed, 11 Sep 2013 08:26:30 +0000 (10:26 +0200)]
Add 2nd 'demo' canvas to the other tiled printing tests.

4 years agoRedSquareES2: Fix tiled aspect ratio
Sven Gothel [Wed, 11 Sep 2013 08:25:43 +0000 (10:25 +0200)]
RedSquareES2: Fix tiled aspect ratio

4 years agoTestTiledPrintingGearsSwingAWT: Add a second demo GLEventListener showing it's workin...
Sven Gothel [Tue, 10 Sep 2013 21:37:14 +0000 (23:37 +0200)]
TestTiledPrintingGearsSwingAWT: Add a second demo GLEventListener showing it's working/scaling well; Adding TileRendererBase.TileRendererNotify to RedAquareES2 having a 3rd demo case.

4 years agoGears*: Clarify frustum calculation for tiling
Sven Gothel [Tue, 10 Sep 2013 21:07:31 +0000 (23:07 +0200)]
Gears*: Clarify frustum calculation for tiling

4 years agoAWT Printing: AWTTilePainter needs to handle null clip!
Sven Gothel [Tue, 10 Sep 2013 21:06:38 +0000 (23:06 +0200)]
AWT Printing: AWTTilePainter needs to handle null clip!

4 years agoAWTPrintLifecycle: Add DEFAULT_PRINT_TILE_SIZE (512); Remove unused imports.
Sven Gothel [Sun, 8 Sep 2013 17:51:21 +0000 (19:51 +0200)]
AWTPrintLifecycle: Add DEFAULT_PRINT_TILE_SIZE (512); Remove unused imports.

4 years agoNewtCanvasAWT: Implement AWTPrintLifecycle and hence support for AWT printing.
Sven Gothel [Sun, 8 Sep 2013 17:26:20 +0000 (19:26 +0200)]
NewtCanvasAWT: Implement AWTPrintLifecycle and hence support for AWT printing.

Note: Same bug existing as w/ GLCanvas, i.e.
  'GLDrawableUtil.swapGLContextAndAllGLEventListener(gladPre, gladNew)':
  If 'gladPre' is onscreen and using MSAA (on NV/GLX),
  the ctx cannot be made current in it's new 'gladNew' location.

  Same workaround applied, i.e. use onscreen drawable while printing.

4 years agoTiledPrintingAWTBase/doc: Add description of how-to add a 'print-to-file' printer...
Sven Gothel [Sun, 8 Sep 2013 17:22:57 +0000 (19:22 +0200)]
TiledPrintingAWTBase/doc: Add description of how-to add a 'print-to-file' printer driver on OSX

4 years agobuild-*.xml: Unify 'javah' tasks where possible, i.e. per destination directory and...
Sven Gothel [Sun, 8 Sep 2013 17:17:23 +0000 (19:17 +0200)]
build-*.xml: Unify 'javah' tasks where possible, i.e. per destination directory and classpath. This hopefully saves a few milliseconds in the build :)

4 years agoAdd AWTTilePainter.dumpHintsAndScale(..), removing more duplicated code from GLCanvas...
Sven Gothel [Sun, 8 Sep 2013 17:16:09 +0000 (19:16 +0200)]
Add AWTTilePainter.dumpHintsAndScale(..), removing more duplicated code from GLCanvas/GLJPanel

4 years agoTileRenderer*: API doc fixes
Sven Gothel [Sun, 8 Sep 2013 13:43:03 +0000 (15:43 +0200)]
TileRenderer*: API doc fixes

4 years agoTileRenderer*: Fix pre-swap and post-swap in regards to endTile(..), i.e. pre-swap...
Sven Gothel [Sun, 8 Sep 2013 13:13:00 +0000 (15:13 +0200)]
TileRenderer*: Fix pre-swap and post-swap in regards to endTile(..), i.e. pre-swap only for FBO && MSAA. See TileRendererBase.reqPreSwapBuffers(..) API doc.

4 years agodoc/Implementation/AWTPrinting.txt: Adding ref to AWTPrintLifecycle
Sven Gothel [Sun, 8 Sep 2013 11:58:58 +0000 (13:58 +0200)]
doc/Implementation/AWTPrinting.txt: Adding ref to AWTPrintLifecycle

4 years agoFix some JavaDoc and API doc (reference) bugs ..
Sven Gothel [Sun, 8 Sep 2013 11:39:27 +0000 (13:39 +0200)]
Fix some JavaDoc and API doc (reference) bugs ..

4 years agoAWTPrintLifecycle: Add detailed API doc
Sven Gothel [Sun, 8 Sep 2013 10:32:56 +0000 (12:32 +0200)]
AWTPrintLifecycle: Add detailed API doc

4 years agoAggregate and cleanup '.ps' and '.pdf' artifacts (printing tests)
Sven Gothel [Sun, 8 Sep 2013 01:23:19 +0000 (03:23 +0200)]
Aggregate and cleanup '.ps' and '.pdf' artifacts (printing tests)

4 years agodoc/Implementation/AWTPrinting: Update to match latest code
Sven Gothel [Sun, 8 Sep 2013 01:17:01 +0000 (03:17 +0200)]
doc/Implementation/AWTPrinting: Update to match latest code

4 years agoTest: Don't resize frame, tweek print-matrix; AWTPrintLifecycle: Add scale and conven...
Sven Gothel [Sun, 8 Sep 2013 01:02:36 +0000 (03:02 +0200)]
Test: Don't resize frame, tweek print-matrix; AWTPrintLifecycle: Add scale and convenient AWT container traversal context; GLCanvas/GLJPanel properly handle existing MSAA and req. AA;

- Test: Don't resize frame, tweek print-matrix
  - Use scaleComp72 to scale the frame to fit on page,
    i.e. global print matrix

  - Use scaleGLMatXY = 72.0 / glDPI
    to locally scale on the GL drawable as being passed to AWTPrintLifecycle.setup(..)

  - Hence frame stays untouched/stable, no need for 'offscreen' print test,
    which is removed.

- AWTPrintLifecycle: Add scale and convenient AWT container traversal context
  Use a simple decoration for all AWTPrintLifecycle impl. components within a container:

    final AWTPrintLifecycle.Context ctx = AWTPrintLifecycle.Context.setupPrint(frame, g2d, scaleGLMatXY, scaleGLMatXY);
    try {
    } finally {

- GLCanvas/GLJPanel properly handle existing MSAA and req. AA;
  - GLCanvas: Workaround bug where onscreen MSAA cannot switch to offscreen FBO,
              i.e. stay 'onscreen'
  - GLJPanel: Use new offscreen FBO if MSAA is requested and not yet used.

- GLJPanel.Offscreen.postGL(): always swapBufer(), was missing for !GLSL swapping

Results GLCanvas / GLJPanel:
  - Good scaling
  - Stable behavior / visibility
  - High DPI mode works

4 years agoTiledPrintingAWTBase: Fix scaling - Fit frame to page, add MSAA RenderingHints test...
Sven Gothel [Sat, 7 Sep 2013 17:59:15 +0000 (19:59 +0200)]
TiledPrintingAWTBase: Fix scaling - Fit frame to page, add MSAA RenderingHints test; setupPrint(Graphics2D): Query RenderingHints to use MSAA rendering

- AWTPrintLifecycle.setupPrint(Graphics2D): Query RenderingHints to use MSAA rendering
  - Impl. in GLCanvas
  - TODO GLJPanel (would need a new offscreen buffer)

- TiledPrintingAWTBase:
  - Fix scaling - Fit frame to page
  - add MSAA RenderingHints test

- GLCanvas: Remove dumpStack() DEBUG output

4 years agoTileRenderer*: Fix FBO MSAA use-case, i.e. call swapBuffers() before endTile(); Enhan...
Sven Gothel [Sat, 7 Sep 2013 17:47:28 +0000 (19:47 +0200)]
TileRenderer*: Fix FBO MSAA use-case, i.e. call swapBuffers() before endTile(); Enhance unit tests for MSAA, also add TileRendererBase.TileRendererNotify to GearsES2

GL[Auto]Drawable.swapBuffers() must be called before endTile().

This is especially important if using multisampling offscreen FBO drawables,
where swapBuffers() triggers the <i>downsampling</i> to the readable sampling sink.

Otherwise, we will be 'one tile behind' !

4 years agoFix TileRenderer*: Allow general usage w/ any GL profile, only image-buffer requires...
Sven Gothel [Sat, 7 Sep 2013 14:57:47 +0000 (16:57 +0200)]
Fix TileRenderer*: Allow general usage w/ any GL profile, only image-buffer requires >= GL2ES3; Always set pack-alignment, Set glReadBuffer(..) >= GL2ES3

- Allow general usage w/ any GL profile, only image-buffer requires >= GL2ES3
  Due to GL2ES3.GL_PACK_ROW_LENGTH and image-width != tile-width

- Always set pack-alignment
  Forgot for tile-buffer

- Set glReadBuffer(..) >= GL2ES3
  Required if using FBO offscreen, i.e. MSAA mode. git info: FAQ, tutorial and man pages.