summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* Fix Bug 643: SWT 'display.asyncExec(Runnable runnable)' runnable not ↵Sven Gothel2012-12-021-10/+9
| | | | | | | | | | | | executed on Windows Turns out that the NEWT Windows impl. didn't properly validated the client region @ WM_PAINT and hence 'exhausted' the message pipeline, i.e. never reached an IDLE state. The latter caused SWT to never reach a point where deferred asyncExec(..) Runnables got processed. Besides this SWT effect, this also caused a NEWT window on Windows to always repaint itself (?).
* NEWT WindowImpl fixes: surfaceLockCount-- if native lock fails; ↵Sven Gothel2012-12-022-28/+36
| | | | | | | | | | | | | | | | | | waitForVisible(.., fastFail:=false); waitForSize(..) @ setSize; reparent definePosition(..); - surfaceLockCount-- if native lock fails In case native lock fails, not only remove the windowLock but also decr. surfaceLockCount (proper roll back) - was a BUG! - waitForVisible(.., fastFail:=false) Don't fail fast if visibility wasn't reached. - waitForSize(..) @ setSize Wait for size change - otherwise an SWT child won't reach desired size. - reparent definePosition(..); Position might not get updated by WM events (SWT parent apparently) - Windows WindowDriver: Cleanup code a bit.
* NEWT WindowImpl: Don't issue native resize if invisible, simply use ↵Sven Gothel2012-11-301-18/+13
| | | | | | | | | | | | | | | | | defineSize(..); Don't set persitent undecorated flag if child window at creation; Use local 'screen' directly. - Don't issue native resize if invisible, simply use defineSize(..) Invisible windows may not promote size change natively, hence simply setting the size via defineSize(..) is appropriate. Latter setVisible(true) will take size into account. - Don't set persitent undecorated flag if child window at creation Even if a window is a child at creation, it maybe reparented to top-level where the default behavior is to be expected. Undecorated top-level window shall require explicit setUndecorated(true). - Use local 'screen' directly. No need to make code more complicate ..
* Fix NewtCanvasSWT's newtChild usage: Only use set newtChild if it's ready, ↵Sven Gothel2012-11-301-6/+15
| | | | i.e. SWTEDTUtil set and parented; Recognize pending resize.
* Simplify NEWT EDTUtil invoke: To start EDT Runnable maybe null - start EDT ↵Sven Gothel2012-11-305-62/+54
| | | | | | even if on EDT thread. DEBUG: Name EDTUtil impl, e.g. Default, AWT and SWT
* Bug628: Adding unit-test 'TestNewtCanvasSWTBug628ResizeDeadlock' exposing ↵Sven Gothel2012-11-292-7/+288
| | | | | | | | | | | NewtCanvasSWT asyncExec(..) bug w/ native parenting The unit test shows, that while using JOGL's SWT GLCanvas Display's asyncExec(..) works properly, but w/ NewtCanvasSWT on Windows does not. NewtCanvasSWT differs w/: - Using native parenting [Newt GLWindow to SWT Canvas] - Processing native events in own NEWT EDT, w/ own Windows dispatch hook [For the child GLWindow only]
* NewtCanvasSWT: Add DisposeListenerSven Gothel2012-11-291-3/+18
|
* SWT GLCanvas: Fix dispose bug, check for isDisposed() and add ↵Sven Gothel2012-11-291-17/+52
| | | | | | | | | | | | | | DisposeListener; Run GL tasks on current thread w/o restrictions Fix dispose bug, check for isDisposed() and add DisposeListener - Don't issue SWTAccessor.setRealized(..), since it's called implicit via super.dispose() - Check isDisposed() .. - add DisposeListener to act on parent's disposal (Shell, Composition, ..) Run GL tasks on current thread w/o restrictions + * The current thread seems to be valid for all platforms, + * since no SWT lifecycle tasks are being performed w/ this call. + * Only GL task, which are independent from the SWT threading model.
* TestSWTJOGLGLCanvas01GLn: Use display.syncExec(..) where possible, add AnimatorSven Gothel2012-11-291-5/+24
|
* SWTAccessor: Add SWT 4.3's X11-GTK version adaption (2.14, 2.24, 3.0)Sven Gothel2012-11-291-46/+170
|
* Fix regression of commit 4dd44b985fe0541be3a3bcd9045d201ed3ca2cc5: Setting ↵Sven Gothel2012-11-281-1/+1
| | | | 'sendReshape' for onscreen drawables as well.
* X11Util/Xmisc setX11ErrorHandler: Force setting X11 error handler if invoked ↵Sven Gothel2012-11-272-23/+25
| | | | | | | | | from Java ensuring to overwrite other TKs handler SWT sets it's error handler, which exits application in case of an X11 error (e.g. n/a GL3 impl.). We need to allow overwriting it, hence allowing multiple calls (NEWT, GLContext). This patch also makes native NativewindowCommon_x11ErrorHandlerEnable(..) function static.
* SWTEDTUtil/AWTEDTUtil: Fix deadlock situations ; Cleanup ↵Sven Gothel2012-11-275-118/+240
| | | | | | | | | | TestNewtCanvasSWTBug628ResizeDeadlock - Fix deadlock situation in waitUntilStopped/Idle(), skip if on AWT/SWT EDT - Use RunnableTask for sync task invocation, don't block AWT/SWT EDT. - Cleanup TestNewtCanvasSWTBug628ResizeDeadlock (works on OSX as well)
* SWTAccessor: OS_gtk_widget_unrealize optional (SWT 4.3) ; decorate ↵Sven Gothel2012-11-263-10/+40
| | | | PrivilegedAction for static initSingleton block (SWTAccessor, NewtFactory, NativeWindowFactory)
* SWTEDTUtil/AWTEDTUtil: Fix deadlock situation in waitUntilStopped(), etc - ↵Sven Gothel2012-11-253-15/+25
| | | | | | wrap task execution (or enqueing) into status-sync 'edtLock' This fixes the disparity w/ DefaultEDTUtil, i.e. aligns it's implementation/semantics.
* FPSAnimator: Add note on deamon-thread and JVM shutdown behavior.Sven Gothel2012-11-252-5/+11
|
* Fix SWTEDTUtil Bug628: Perform NEWT event dispatching on SWT-EDT, due to ↵Sven Gothel2012-11-232-12/+14
| | | | | | possible triggered locking action, i.e. display(). Do the same for AWTEDTUtil. This fix actually clarifies the annotated FIXME :)
* Bug628: Adding unit test reproducing the deadlock realiable withing first ↵Sven Gothel2012-11-233-9/+370
| | | | | | | rounds of multi-threaded display. If a NEWT Window locking operation is performed from within NEWT's EDT (native event dispatch) and a mutable SWT operation performed via NEWT's display invoke - a deadlock occurs.
* Merge pull request #53 from HaZuki/FixQuaternionSpikeSven Gothel2012-11-211-2/+2
|\ | | | | | | | | | | Fix Bug : Quaternion multiplication unexpected behavior Matches description in Bug 631 https://jogamp.org/bugzilla/show_bug.cgi?id=636
| * Fix Bug : Quaternion multiplication unexpected behaviorHaZuki2012-11-161-2/+2
| |
* | OSX CALayer Stencil/.. Fix: In case of FBO CALayer usage, use default ↵Sven Gothel2012-11-219-66/+309
|/ | | | | | | | | | caps/pixelformat w/ chosen GLProfile only Using a pixelformat w/ chosen stencil for CALayer does corrupt rendering for an unknown reason, probably due to incompatible pixelformat w/ CALayer composition. This patch simply discards any special chosen caps, while only recognizing the desired GLProfile for the FBO CALayer pixelformat.
* NewtVersionActivity: Dump information in logcatSven Gothel2012-11-131-3/+7
|
* Frustum: Clarify method names, fix point/sphere classification, add used ↵Sven Gothel2012-11-122-26/+71
| | | | | | | | | references Clarify method names: - update(..) -> updateByPMV(..), updateByPlanes(..) - isOutside(AABBox) -> isAABBoxOutside(AABBox) - .. same for point/sphere, while adding 'Location classifyType(..)'
* Frustum: Add doc. about Frustum's Plane's normal orientationSven Gothel2012-11-121-2/+16
|
* Frustum: Cleanup / update; PMVMatrix: Fix mulPMVSven Gothel2012-11-122-13/+24
| | | | | | | | | | Frustum: Cleanup / update - Remove ctor w/ PMV, use update(..) instead - avoid API explosion - Add update(Plane[]) to copy existing Frustum planes - Mention world-coordinates in update(PMV) PMVMatrix: Fix mulPMV - P*Mv in column major order is correct for Frustum
* Frustum: Clarify isOutside(AABBox) impl, add isOutside for point and sphereSven Gothel2012-11-121-6/+43
|
* Adding Ref.: Frustum-Culling, Max WagnerSven Gothel2012-11-121-10/+12
|
* Validating Frustum w/ help of Eduard White [email protected], referencing ↵Sven Gothel2012-11-122-75/+139
| | | | | | | | | | | | | | | | | the original paper Paper: Fast Extraction of Viewing Frustum Planes from the World-View-Projection Matrix http://graphics.cs.ucf.edu/cap4720/fall2008/plane_extraction.pdf Authors (in alphabetical order): Gil Gribb <[email protected]> Klaus Hartmann <[email protected]> 06/15/2001 Fix: - Column Major Order PMV (Passing to Frustum and in calculation itself) according to paper coeff. calculation of plane - Plane's signed distance function (only add d, don't multiply) - Normalization: divide by lenght, not multiply
* Frustum: Print Plane's normals properly in toString()Sven Gothel2012-11-111-1/+1
|
* Frustum: Passing Mv*P (column major order)Sven Gothel2012-11-112-34/+34
|
* Added a condition to skip updateGraphicsConfigurationARB when ↵jthedering2012-11-111-1/+2
| | | | non-hardware-accelerated capabilities are requested, because only updateGraphicsConfigurationGDI provides software rendering capabilities.
* Fix Graph SceneUIController: Remove Listener by itself in dispose()Sven Gothel2012-11-112-3/+3
|
* GLDrawableHelper disposeAllGLEventListener(): Don't use cached listener size ↵Sven Gothel2012-11-111-3/+2
| | | | / check size() > 0, since List can be modified by listener itself
* PMVMatrix: Add 'Frustum glGetFrustum()' adding same dirty/request ↵Sven Gothel2012-11-112-67/+114
| | | | | | | | | methodology as for Mvi and Mvit Allows user to derive Frustum from updated P + MV Clarify method name for clearing all update request: - disableMviMvitUpdate() -> clearAllUpdateRequests()
* Frustum: Simpler usage (ctor, update) no need for explicit compute() method. ↵Sven Gothel2012-11-111-102/+143
| | | | Don't store passing PMV reference. isInside() -> isOutside()
* AABBox: Add public direct setSize(..), skipping redundant reset() / resize(..)Sven Gothel2012-11-111-18/+31
|
* Cleanup Frustum Math Util: Independent / Compile Clean / Relocation ; ↵Sven Gothel2012-11-113-191/+194
| | | | | | | | | | | | PMVMatrix: Add getPreMultipliedPMV(..) - Independent / Compile Clean - Remove OpenMALI dependencies - Use basic float[] type and FloatUtil - Use AABBox - FIXME: May need BBox (no axis alignment ?!) - Relocation - Move to com.jogamp.opengl.math.geom (see commit 5fafc1ac360333645b807dcd8dff0c0a655ea439)
* Reorganize math code into: com.jogamp.opengl.math and ↵Sven Gothel2012-11-1134-89/+264
| | | | | | | | | | | | | | | | | | | | | | | com.jogamp.opengl.math.geom packages Note: WIP - We may relocate / reorg math package. Public relocations: com.jogamp.opengl.util -> com.jogamp.opengl.math - FixedPoint - FloatUtil com.jogamp.graph.math -> com.jogamp.opengl.math - Quaternion - VectorUtil com.jogamp.graph.geom -> com.jogamp.opengl.math.geom - AABBox VectorUtil: Introducing Vert2fImmutable and Vert3fImmutable interfaces, allowing graph Vertex instances to be used 'graph' agnostic and to document 2d/3d use-cases.
* Merge MathFloat into FloatUtilSven Gothel2012-11-116-95/+68
|
* Fix Bug 636: Quaternion multiplication unexpected behaviorTek2012-11-111-4/+4
|
* Adding basic Frustum utility class interfacing w/ PMVMatrix. Still depending ↵Eduard White2012-11-111-0/+191
| | | | on OpenMALI - hence broken.
* Bug 634 - Part 1: Fix FBObject regression of commit ↵Sven Gothel2012-11-081-4/+12
| | | | | | | | | | dbc260bd2e917ee9001461749c99da0c9cbfaf9a - init(..): Query MAX_COLOR_ATTACHMENTS: Remove erroneous samplingSink null check (it's always null) - reset(..): Create & construct samplingSink if required and missing before detaching all (due to bind) - setSamplingSink(..) returns previous samplingSink
* Fix GLAutoDrawable.dispose(): Dispose drawable even w/o context; ↵Sven Gothel2012-11-089-269/+334
| | | | | | | | | | | | | | | | | | | JAWTWindow.lockSurface(): Check AWT component's native peer - Fix GLAutoDrawable.dispose(): Dispose drawable even w/o context - It is possible to have the GLContext not being created (not made current), so drawable shall be disposed independent. - Merge Runnable 'postDisposeOnEDTAction' to dispose Runnable for clarity - GLDrawableHelper: Split disposeGL from invokeGLImpl for clarity - JAWTWindow.lockSurface(): Check AWT component's native peer - W/o a native peer (!isDisplayable()), JAWT locking cannot succeed. - On OSX OpenJDK 1.7, attempting to JAWT lock a peer-less component crashes the VM - MacOSXJAWTWindow.lockSurfaceImpl(): Remove redundant null checks
* FBObject: Rename private fields for better reading: samplesSink* -> ↵Sven Gothel2012-11-061-65/+65
| | | | samplingSink*
* Bug 634 - Part 1: FBObject, Make MSAA 'samplingSink' mutable and add ↵Sven Gothel2012-11-061-10/+33
| | | | setSamplingSink(..). Create MSAA samplingSink lazy if null.
* GLFBODrawableImpl: Following suit w/ commit ↵Sven Gothel2012-11-061-1/+1
| | | | | | | b83b068c0f426f24a58e2bd9f52de9ebd0c7876d, sync GL command stream before FBO reconfig Even though we currently have no bug experienced on this, it seems to be a good idea for highly concurrently GL driver implementations.
* Fix GLDrawableHelper.recreateGLDrawable(..): Sync GL command stream before ↵Sven Gothel2012-11-061-4/+9
| | | | | | | destruction of drawable Lack of finishing the GL command stream lead to a SIGSEGV on Windows w/ Nvidia driver where probably pending GL commands were still being processed concurrently.
* TestGLCanvasAWTActionDeadlock0[01]AWT: Disable 'restart' test on current ↵Sven Gothel2012-11-052-2/+4
| | | | | | (non AWT) thread - crashes on solaris (which is fine) Modifying AWT components shall always happen on AWT-EDT anyways.
* GLAutoDrawable: Refine API change of commit ↵Sven Gothel2012-11-058-55/+94
| | | | | | | | | | c002e04f848116922a1ed7bd96ead54961649bbd As suggested by Julien Gouesse, align 'enqueue(..)' method w/ 'invoke(..)': - public void enqueue(GLRunnable glRunnable); + public boolean invoke(boolean wait, List<GLRunnable> glRunnables);
* Analyzed OSX Deadlock w/ AWT Applet & CALayer: Andres Colubri's Test Case, ↵Sven Gothel2012-11-052-55/+127
| | | | | | | | | | | | | | | | | | | | | | Commit 77db6a5c22cb4a53cf911b4caf57127770c70968 [1] - The AWT/Applet test case on OSX utilized offscreen CALayer, which is correct. [2] - the GLEventListener.display(..) method calls Frame.setTitle(".."); which is correct, but freezes when called right after GLEventListener.init(..) BUG on OSX/CALayer: If frame.setTitle() is issued right after initialization the call hangs in at apple.awt.CWindow._setTitle(Native Method) at apple.awt.CWindow.setTitle(CWindow.java:765) [1.6.0_37, build 1.6.0_37-b06-434-11M3909] Happens w/ Oracle's JRE7u9 as well! Workaround: - Set 'justInitialized' flag in GLEventListener.init(..) - Clear 'justInitialized' flag in GLEventListener.display(..) at the end - Skip Frame.setTitle(..) in GLEventListener.display(..) if 'true == justInitialized' Somebody may send a bugreport to Oracle / OpenJDK.
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-21 10:10:23 +0000