summaryrefslogtreecommitdiffstats
path: root/make/scripts/tests.sh
Commit message (Collapse)AuthorAgeFilesLines
* GLAutoDrawable: Fix GLEventListener lifecycle and expose more user control ↵Sven Gothel2012-11-041-2/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | (API Change) ; Added GLDrawableUtil A GLEventListener resides in two states, initialized and uninitialized. When added to a GLAutoDrawable, it is uninitialized. A first 'display()' will issue GLEventListener's 'init(..)' which renders it initialized. This is usually accompanied by 'reshape(..)' propagating the drawable's dimension. Destruction of the GLAutoDrawable will issue GLEventListener's 'dispose(..)' which renders it uninitialized. It turns our these means of GLEventListener controls are not sufficient in case the user requires to remove and add them during the lifecycle and rendering of their GLAutoDrawable host. GLAutoDrawable 'removeGLEventListener(..)' merely removes the GLEventListener from the list, but does not complete it's lifecycle, i.e. issues 'dispose(..)' if initialized to realease GL related resources. Hence the following essential API changes are made to complete the lifecycle: + public GLEventListener disposeGLEventListener(GLEventListener listener, boolean remove); disposing a single GLEventListener, allowing it's removal from the list being optional This is demonstrated via GLDrawableUtil.swapGLContextAndAllGLEventListener(GLAutoDrawable a, GLAutoDrawable b), see below. ++++++++ Further more the following API changes were made to expose complete control of GLEventListener to the user: - public void removeGLEventListener(GLEventListener listener); + public GLEventListener removeGLEventListener(GLEventListener listener); The return value allows simple pipelining, and also delivers information whether the passed listener was actually removed. - public GLEventListener removeGLEventListener(int index) throws IndexOutOfBoundsException; + public int getGLEventListenerCount(); + public GLEventListener getGLEventListener(int index) throws IndexOutOfBoundsException; Dropping the redundant removal by index, while adding count and get methods. + public boolean getGLEventListenerInitState(GLEventListener listener); + public void setGLEventListenerInitState(GLEventListener listener, boolean initialized); Allows retrieving and setting of listener states. All in all these API changes allows a user to experience all freedoms in dealing w/ GLEventListeners hosted by GLAutoDrawable impl. and shall be future proof. Note that we have avoided the Iterator pattern due to it's overhead of temporal objects creation. The simple indexed access allows us to implement each method as an atomic operation. +++++++++++ Further more a simple enqueue(..) method has been added, allowing to just enqueue a GLRunnable w/o provoking it's execution - as invoke(..) does. This method pleases a use case where GLRunnables are batched and shall be executed later on.. public boolean invoke(boolean wait, GLRunnable glRunnable); + public void enqueue(GLRunnable glRunnable); +++++++++++ Added GLDrawableUtil, exposes utility function to rearrange GLEventListener, modifiy GLAutoDrawable, etc. GLDrawableUtil.swapGLContextAndAllGLEventListener(GLAutoDrawable a, GLAutoDrawable b) is tested and demonstrated w/ TestGLContextDrawableSwitchNEWT. Manually tested on X11, OSX and Windows.
* NEWT Windows KeyEvent: We have to store the keyChar for typed events, since ↵Sven Gothel2012-10-311-2/+2
| | | | | | keyChar from pressed/released may be wrong (Uppercase: SHIFT-1, etc ..) Partially reverts commit: b62e1d027c289877686d6008ea8dd40e4e1541ec
* MacOSXCGLContext: Fix ShaderCode instantiation .. (duh!) ; Tested via ↵Sven Gothel2012-10-311-1/+2
| | | | reparenting TestParentingFocusTraversal01AWT
* Fix Unit Tests regarding commit 55f4405dbf9253e6ca686e44c50b2cfc1af7f3eb: ↵Sven Gothel2012-10-311-2/+2
| | | | Shader program name is valid if non-zero.
* Fix ProjectFloat (Bug 633): Adding missing offset of sliced buffer usage; ↵Sven Gothel2012-10-311-1/+2
| | | | | | reduce buffer usage (performance) in favor of float[]. Thomas De Bodt reported this error and provided the unit test.
* Add OSX CALayer OpenGL 3 (core) support: Derive pixelformat from parent ↵Sven Gothel2012-10-311-7/+3
| | | | | | | | (GL3), use GL3.2 compatible shader; Use VBO in general. Covered by: Auto unit tests: TestOffscreenLayer01GLCanvasAWT, TestOffscreenLayer02NewtCanvasAWT Manual: TestGearsES2AWT '-gl3 -layered'
* Shader: Add '#define texture2D texture' for GLSL >= 130 ; TestGearsES2AWT ↵Sven Gothel2012-10-311-2/+3
| | | | add forceGL3; TextureDraw01ES2Listener uses defaultShaderCustomization()
* Fix NEWT KeyEvent: Deliver keyChar w/ pressed and released; Deliver proper ↵Sven Gothel2012-10-311-3/+5
| | | | | | | | | | | | | | | | | | | | | | modified flags and modifier-key events; Simplify Windows key handling Preface: Modifier-keys are SHIFT, CTRL, ALT and META and they have a matching modifier-bit. - Simplify Windows key handling - Employ MapVirtualKey(..) for virtual-key to character and scancode to virtual-key mappings, allowing to drop tracking of keyCode to keyChar in java code. This also removes the platform restriction of delivering keyChar at TYPED only. - Deliver keyChar w/ pressed and released - Due to the lift restriction on the Windows platform (see above), we can deliver keyChar w/ all key events on all platforms. - Deliver proper modified flags and modifier-key events All modifier-keys deliver pressed, released and typed events with their modifier-bit set. The above is covered by unit tests, which passed on X11, Windows and OSX (manual test run).
* TestNewtKeyEventAutoRepeatAWT: Wait for 'some' KeyEvents and check count; ↵Sven Gothel2012-10-291-2/+2
| | | | TestNewtKeyPressReleaseUnmaskRepeatAWT: Enable NewtCanvasAWT test
* Complete Rename NEWT KeyEvent Unit tests, commit ↵Sven Gothel2012-10-291-6/+6
| | | | 85851c9839d620bcbbd07b6ca833f1a5901831cc
* GLRendererQuirks: Add RequiresBoundVAO (w/ impl.), GLSLBuggyDiscard (todo) ; ↵Sven Gothel2012-10-281-2/+2
| | | | | | | | | | | | | | | | GLContextImpl: Bind default VAO if having quirk RequiresBoundVAO. OSX w/ OpenGL >= 3 core context implementation requires a bound VAO for vertex attribute operations, i.e. VertexAttributePointer(..). This has been experienced on OSX 10.7.5, OpenGL 3.2 core w/ Nvidia GPU and in several forum posts. Such 'behavior' violates the GL 3.2 core specification, which does not state this requirement, hence it is a bug. (Please correct me if I am wrong!) GLContextImpl works around this quirk, by generating a default VAO and binds it at 1st makeCurrent (@creation) and deletes it at destroy. This is minimal invasive since no action is required for subsequent makeCurrent or release. We assume if a user uses and binds a VAO herself, she will mind this quirk. Note: We could enhance this workaround by quering for a currently bound VAO at makeCurrent() and bind our default if none. However, we refrain from this operation to minimize the workaround and complexity.
* Fix NEWT KeyCode: Basic KeyCode Validation on X11, Windows and OSXSven Gothel2012-10-281-1/+2
| | | | | | | - X11: Add VK_QUOTE mapping - OSX: Add single shift, ctrl alt key press; Fix mapping: Command -> Windows, Option -> ALT, add BACK_QUOTE and QUOTE.
* NEWT KeyEvent (Windows, OSX): Check whether keyCode is tracked before using ↵Sven Gothel2012-10-271-2/+2
| | | | bitfield, use more IntBitfield.put(..) return value for efficiency.
* Fix Bug 631 and enhance 601: Handle multiple keys (pressed, released, types ↵Sven Gothel2012-10-271-2/+3
| | | | | | | | | | incl. auto-repeat) - Using keyCode (bit) maps to isPressed and isAutoRepeat, allowing use of multiple keys - Enhance unit test TestKeyEventOrderNEWT w/ injecting variations of 2 diff. keys - Manual tested on X11, Windows and OSX w/ and w/o auto-repeat
* Fix Bug 601 - Auto-Repeat Behavior: Adding unit tests for typed key order w/ ↵Sven Gothel2012-10-261-2/+5
| | | | | | | | | and w/o auto repeat. Incl. fix for Windows. Auto-Repeat tests recognizes whether auto-repeat could be triggered by AWT Robot. The latter is not possible on Windows, hence manual testing was required on this platform. Impact: X11, Windows and OSX produce proper key sequence incl. auto-repeat modifier mask.
* tests.sh: TestPointsNEWTSven Gothel2012-10-231-2/+2
|
* Fix regression of commit 40d01bef2a1db44533472c37961aabbef68de644: Test for ↵Sven Gothel2012-10-231-2/+2
| | | | fourth element was invalid
* FixedFuncPipeline GL_POINTS: Fix gl_PointSize (attribute data format), Add ↵Sven Gothel2012-10-221-1/+2
| | | | | | | | | | | | | | GL_POINT_SOFT and dist/fade attenuation (Adding basic POINT unit tests) gl_PointSize (and all other uniform array elems) was not propagated due to wrong usage of GLUniformData component param. Due to efficiency, we use vec4[2] now and #defines in shader to easy readability. GL_POINT_SOFT uses gl_PointCoord to determnine inside/outside circle position while adding a seam of 10% in/out. This almost matches 'other' implementations and gives a nice smooth circle. !GL_POINT_SOFT produces a proper square (billboard). Point-Vertex shader takes dist/fade attentuation into account.
* NEWT / BCM-VC-IV: Fix transparency according to given capabilitiesSven Gothel2012-10-191-2/+2
| | | | Tested manual w/ TestGearsES2NEWT on Raspberry Pi
* FixedFuncPipeline: Cache current ShaderSelectionMode, update PMVMatrix ↵Sven Gothel2012-10-181-1/+0
| | | | according it's usage (update Mvi/Mvit only if lighting is being used)
* ImmModeSink: Add glColor3ub(), glColor4ub(); Add proper value conversion of ↵Sven Gothel2012-10-181-2/+5
| | | | imm. gl* functions; Default color padding is 1f; Make fields private.
* ImmModeSink: Fix buffer grow (+1 element @ named buffer), enable DEBUG_* via ↵Sven Gothel2012-10-161-2/+2
| | | | properties, drawIndices QUAD w/ proper range and add uint; FixedFunctionHook: drawIndices QUAD w/ proper range and add uint
* FixedFuncHook: Add ES2 alignment of certain GL functions, i.e. ↵Sven Gothel2012-10-121-2/+2
| | | | GL_QUAD_STRIP/GL_POLYGON/GL_QUADS mapping, glTexImage2D internalformat/format match.
* GLArrayDataClient.bindBuffer(gl, bind=true): checkSeal and init_vbo if ↵Sven Gothel2012-10-121-2/+2
| | | | required (similar sanity checks as enableBuffer())
* Adding unit tests for ImmModeSink and FixedFunctionPipeline (single & ↵Sven Gothel2012-10-121-3/+15
| | | | combinations, plus texture usage)
* Raise NEWT MouseButton Maximum from 6 -> 16 (API Change)Sven Gothel2012-10-081-3/+3
| | | | | | | | - Button 9 has been reported to be sent by Olamedia - Rearrange the input bit mask in InputEvent (API Change) - Raise the max. button number to 16
* Fix Bug 617 (part2): Workaround buggy GPU driver where FBO reattachment of ↵Sven Gothel2012-10-061-2/+2
| | | | | | | | | | | | | | | | | | | Colorbuffer w/ diff size leads to GL_FRAMEBUFFER_UNSUPPORTED. This occured at least on: - OS X 10.6.8 - GL_RENDERER NVIDIA GeForce 7300 GT OpenGL Engine - GL_VERSION 2.1 NVIDIA-1.6.36 Remedy is to catch the exception @ GLFBODrawableImpl.reset(..) and switch over to fallback 'reset' method: FBO reattachment -> FBO complete recreation Of course, the FBO recreation is noticable slower, but at least it seems to work on the offending system. Not tested on the offending system, but manually provoked GLException on FBOObject to trigger fallback, which is working here.
* Adding unit test for Bug 605 - WORKSFORME, i.e. no unwanted y-flip, so I can ↵Sven Gothel2012-10-061-6/+2
| | | | just assume it's the SW renderer, flipping.
* Refine a3cb6bb14f410f67fccf5ccd4cd7ecc66f448389, fix deadlock (regression)Sven Gothel2012-10-051-2/+12
| | | | | | | | The lock being claimed at validateGLDrawable() is 'offthread', i.e. may fight w/ AWT / Animator at reshape/display. Locking is moved 'down' to AWT runnable 'setRealizedOnEDTAction', which also double checks the drawable [again].
* Fix FBObject DEPTH_STENCIL renderbuffer usageSven Gothel2012-10-051-5/+5
| | | | | | If DEPTH_STENCIL was selected, recreate destroyed the common buffer when stencil was being 'reset'. To unify and simplify code, if DEPTH_STENCIL stencil == depth reference.
* Fix Bug 572 (2nd time): GLCanvas.validateGLDrawable() @ display() and ↵Sven Gothel2012-10-051-2/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | reshape() ; GLCanvas.reshape() only if drawble valid ; GLCanvas.validateGLDrawable() also test isDisplayable() ; Fix size validation ; resizeOffscreenDrawable(..) don't validate 'safe' size 1x1 - GLCanvas.validateGLDrawable() @ display() and reshape() To help users using GLCanvas w/ having a realized GLCanvas/Drawable, validateGLDrawable() is also called at reshape(). This shall ensure a valid drawable after even a non AWT-EDT issued first setVisible(). - GLCanvas.reshape() only if drawble valid Otherwise offscreen reshape attempts would happen even on unrealized drawable, which is not necessary. - GLCanvas.validateGLDrawable() also test isDisplayable() To make sure the native peer is valid, also test isDisplayable() - Fix size validation Since we have experienced odd size like 0 x -41 test each component, i.e. 0 < width && 0 < height. This is done through all JOGL/NEWT components. - resizeOffscreenDrawable(..) don't validate 'safe' size 1x1 In case method is called w/ odd size, i.e. 0 x -41, the safe size 1x1 is used. However, we cannot validate this size. Dump WARNING if odd size is detected.
* Fix SWTEDTUtil bug, where it simply doesn't start by an implicit 'invoke()' ↵Sven Gothel2012-10-051-3/+3
| | | | | | | | | | | | | - No NewtCanvasSWT resize, nor input event delivery. A new EDTUtil instance is not started automatically. Since SWTEDTUtil is attached to the DisplayImpl later in time, i.e. after it's native creation, there is no EDTUtil.invoke(..) call which started it. The not started SWTEDTUtil could not deliver any events. Fix: Start it explicitly - add API doc comment in Display.setEDTUtil(..)
* Footprint Stats ; Fix Bug 624 - Compile and generate android JAR files ↵Sven Gothel2012-10-041-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | (jogl-all-android.jar, jogl-test-android.jar) if android SDK JAR files are available. ; Split up atomic jogl-util.jar Compile and generate android JAR files (jogl-all-android.jar, jogl-test-android.jar) if android SDK JAR files are available. - See gluegen commit 55b4552aef7882c358d545d020d6f12c958ed8ed Also split up atomic jogl-util.jar - jogl-util.jar -> jogl-util.jar + jogl-util-graph.jar Footprint Stats: This demonstrates small footprint of a minimum configuration, all numbers in kilobytes (du -ksc). NOTE the min size of ~ 1MB! JOGL ALL 220 kB gluegen-rt.jar 5180 kB jogl-all.jar 4 kB libgluegen-rt.so.gz 4 kB libnativewindow_awt.so.gz 12 kB libnativewindow_x11.so.gz 100 kB libjogl_desktop.so.gz 20 kB libnewt.so.gz 5540 kB total JOGL Min X11 Min egl es1 es2 220 kB gluegen-rt.jar 296 kB atomic/jogl-core.jar 156 kB atomic/jogl-glmobile.jar 284 kB atomic/jogl-util.jar 76 kB atomic/nativewindow-core.jar 120 kB atomic/newt-core.jar 28 kB atomic/nativewindow-os-x11.jar 28 kB atomic/newt-driver-x11.jar 4 kB libgluegen-rt.so.gz 48 kB libjogl_mobile.so.gz 12 kB libnativewindow_x11.so.gz 20 kB libnewt.so.gz 1292 kB total JOGL Android - mobile egl es1 es2 224 kB gluegen-rt-android.jar 3020 kB jogl-all-android.jar 4 kB libgluegen-rt.so.gz 48 kB libjogl_mobile.so.gz 3296 kB total JOGL Min Android/Mobile Min egl es1 es2 224 kB gluegen-rt-android.jar 296 kB atomic/jogl-core.jar 156 kB atomic/jogl-glmobile.jar 8 kB atomic/jogl-os-android.jar 284 kB atomic/jogl-util.jar 76 kB atomic/nativewindow-core.jar 120 kB atomic/newt-core.jar 28 kB atomic/newt-driver-android.jar 4 kB libgluegen-rt.so.gz 48 kB libjogl_mobile.so.gz 1244 kB total
* Update PMVMatrix/GLMatrixFunc API doc and refine PMVMatrix update / ↵Sven Gothel2012-10-041-1/+2
| | | | | | | | | | | | | | | | | | | | | | get-Mvi/Mvit-Matrix operation. (Minor API change) Using bitmask for requested Mvi and Mvit matrices, same as dirty-bits to ease matching and update operation. Update of Mvi and Mvit will be performed only if it's dirty-bit and request-bit set within update(). The individual dirty bit is cleared only if it's matrix update is performed. Update is also issued at get-Mvi/Mvit-Matrix operations to ensure proper values w/o update call w/o clearing the modified-bits. update() returns true if the Mvi or Mvit matrix got updated _or_ one of the modified bits is set. update() clears the modified-bits. Adding explicit getModifiedBits() to get and clear it's state. Adding unit test. Lots of API docs ..
* Fix regression of fbe331f013608eb31ff0d8675f4e4c9881c9c48b [Remove ↵Sven Gothel2012-10-021-2/+2
| | | | | | | | | XInitThreads()] NativeWindowFactory.getDefaultToolkitLock() is no more a global singleton, but an instance which has to track/lock a single resource. Hence the decoration w/ it in GLDrawableFactory is useless and applying lock/unlock on a new instance also a bug/regression.
* Misc: tests.shSven Gothel2012-10-021-6/+5
|
* test script: remove deleted unit testSven Gothel2012-09-301-1/+0
|
* NullToolkitLock.validateLocker(): Throw exception if locking is required ↵Sven Gothel2012-09-301-1/+1
| | | | (detecting impl. bugs)
* Workaround for Bug 623: Sporadic XCB assertion failures w/ ATI proprietary ↵Sven Gothel2012-09-301-3/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | driver and w/o native X11 locking The proprietary ATI X11 driver does not handle multi-threaded [GL] clients well, i.e. triggers an XCB assertion 'from time to time'. It almost seems like that the driver either: - aliases all display connections to it's connection name, i.e. server; or - utilizes a build-in display connection w/o locking, used for some reason +++ - X11Lib: Add QueryExtension(dpy, name) allowing early driver determination w/o GL - X11Util detects 'requiresGlobalToolkitLock' and 'markAllDisplaysUnclosable' via X11 extensions. In case certain ATI extensions are available, both are set to true. - X11GLXDrawableFactory: Dropped setting 'markAllDisplaysUnclosable', using X11Util's detection (see above). - New GlobalToolkitLock to satisfy certain driver restrictions (ATI's XCB multithreading bug) - NativeWindowFactory handles new property requiresGlobalToolkitLock, in which case the new GlobalToolkitLock is being used instead of ResourceToolkitLock. - JAWTUtil ToolkitLock locks GlobalToolkitLock 1st to match new 'requiresGlobalToolkitLock' property. - Document static method requirement of X11Util, GDIUtil and OSXUtil via marker interface ToolkitProperties - ToolkitLock: New method 'validateLocked()', allowing use to validate whether the device/toolkit is properly locked and hence to detect implementation bugs. See unit test class: ValidateLockListener
* Cache XineramaIsEnabled(dpy) per display-name in X11Util minimize triggering ↵Sven Gothel2012-09-291-1/+0
| | | | | | | | | | | | | | | | ATI driver bug of sporadic XCB errors and to reduce server roundtrips. Enhances commit: e4176f4e76f519b3599ad557210def3d35266e7b X11Util.XineramaIsEnabled(dpy) hash maps the Xinerama enable result to the dpy-name. Minimize triggering ATI driver bug of sporadic XCB errors: Our multithreading code somehow triggers an _XReply in XQueryExtension beside other functions, which fails the XCB assertion: ../../src/xcb_io.c:178: dequeue_pending_request: Assertion `!xcb_xlib_unknown_req_in_deq' failed. This only appears w/ ATI driver and is probably due to a race condition in the driver (?). This change also reduces server roundtrips / querying time for the cost of 1 hash-map.
* Adding Mesa Quirk 'NoSetSwapIntervalPostRetarget': SIGSEGV on ↵Sven Gothel2012-09-281-3/+4
| | | | | | | | | | | | setSwapInterval() after changing the context's drawable w/ 'Mesa 8.0.4' dri2SetSwapInterval/DRI2 (soft & intel) Analyzing 'TestGLContextDrawableSwitchNEWT' crash at setSwapInterval -> dri2SetSwapInterval after retargeting the context (new drawable association). Turns out Mesa's dri2SetSwapInterval may have a bug. +++ GLContext TRACE_SWITCH: Add drawable handle to debug/trace output.
* NEWT X11 ScreenMode: Ignore invalid rotation eventSven Gothel2012-09-281-34/+2
|
* FBObject: Clarify reset(..), resetSamplingSink(..) ; Rename ↵Sven Gothel2012-09-271-3/+3
| | | | | | | | | | | syncFramebuffer(..) -> syncSamplingSink(..) - reset(..) adds a new argument, boolean resetSamplingSink, allowing to trigger a reset on the samplink sink as well. Use cases are documented. - made public: resetSamplingSink() - Rename syncFramebuffer(..) -> syncSamplingSink(..) to clarify semantics
* Fix Bug 616: X11: Remove XInitThreads() dependency while cleaning up device ↵Sven Gothel2012-09-271-6/+44
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | locking, resulting in a native-lock-free impl. The X11 implementation details of NativeWindow and NEWT used the X11 implicit locking facility XLockDisplay/XUnlockDisplay, enabled via XInitThreads(). The latter useage is complicated within an unsure environment where the initialization point of JOGL is unknown, but XInitThreads() requires to be called once and before any other X11 calls. The solution is simple and thorough, replace native X11 locking w/ 'application level' locking. Following this pattern actually cleans up a pretty messy part of X11 NativeWindow and NEWT, since the generalization of platform independent locking simplifies code. Simply using our RecursiveLock also speeds up locking, since it doesn't require JNI calls down to X11 anymore. It allows us to get rid of X11ToolkitLock and X11JAWTToolkitLock. Using the RecursiveLock also allows us to remove the shortcut of explicitly createing a NullToolkitLocked device for 'private' display connections. All devices use proper locking as claimed in their toolkit util 'requiresToolkitLock()' in X11Util, OSXUtil, .. Further more a bug has been fixed of X11ErrorHandler usage, i.e. we need to keep our handler alive at all times due to async X11 messaging behavior. This allows to remove the redundant code in X11/NEWT. The AbstractGraphicsDevice lifecycle has been fixed as well, i.e. called when closing NEWT's Display for all driver implementations. On the NEWT side the Display's AbstractGraphicsDevice semantics has been clarified, i.e. it's usage for EDT and lifecycle operations. Hence the X11 Display 2nd device for rendering operations has been moved to X11 Window where it belongs - and the X11 Display's default device used for EDT/lifecycle-ops as it should be. This allows running X11/NEWT properly with the default usage, where the Display instance and hence the EDT thread is shared with many Screen and Window. Rendering using NEWT Window is decoupled from it's shared Display lock via it's own native X11 display. Lock free AbstractGraphicsDevice impl. (Windows, OSX, ..) don't require any attention in this regard since they use NullToolkitLock. Tests: ====== This implementation has been tested manually with Mesa3d (soft, Intel), ATI and Nvidia on X11, Windows and OSX w/o any regressions found in any unit test. Issues on ATI: ============== Only on ATI w/o a composite renderer the unit tests expose a driver or WM bug where XCB claims a lack of locking. Setting env. var 'LIBXCB_ALLOW_SLOPPY_LOCK=true' is one workaround if users refuse to enable compositing. We may investigate this issue in more detail later on.
* Fix Bug 621, JVM Crashes During X11 Shutdown while closing all unclosed ↵x11_xinitthreadsSven Gothel2012-09-211-2/+2
| | | | | | | | | | | | | | | | | | | | display connections. This is a regression of JOGL commit f2cfb6119a3663715ed2d572643949b3bef58662 to fix Bug 610. Commit f2cfb6119a3663715ed2d572643949b3bef58662 closed all open display connections at JVM shutdown in order of their creation. This is required IF the driver is the proprietary ATI on X11. X11Util.setMarkAllDisplaysUnclosable(true) is called in this case. However, this behavior causes a SIGSEGV for other GPU driver, i.e. Nvidia :) This fix only closes the pending display connections in their respective order IF X11Util.getMarkAllDisplaysUnclosable() == true. Hence the 'new' cleanup as introduced in the offending commit is only performed on the erroneous ATI driver. Manually tested w/ TestGearsES2AWT, see it's commandline options for details.
* MacOSXCGLContext: Fix missing initialization of 'needsSetContextPBuffer' if ↵Sven Gothel2012-09-211-2/+2
| | | | not offscreen-layer.
* MacOSXCGLDrawableFactory: Cannot use shared context to retrieve quirks since ↵Sven Gothel2012-09-211-2/+2
| | | | it's transient, store it.
* Add GLRendererQuirks; Fix shared EGL/ES resources;Sven Gothel2012-09-201-7/+9
| | | | | | | | | | | | | | | | | | | | | | | Add GLRendererQuirks: - Contains centralized 'tagged' workarounds for GL renderer bugs (quirks) - Accessible via GLContext and GLDrawableFactory - Initialized in GLContext.setAvailability* - Simplify GLGraphicsConfigurationUtil.fixOffscreenGLCapabilities(..) & use Quirks - use quirks - instead of passing booleans for each config, pass factory & device Fix shared EGL/ES resources: - GLProfile needs to initialize EGLDrawableFactory's shared resources before desktop, so EGLDrawableFactory can use the fallback defaultDisplay & defaultSharedResource for host mapped sharedResources (hack). - If using defaultSharedResources for host mapped ones, do not go through initialization cycles - simply map (sharedResource + context). - EGLDrawableFactory: Use device's unique-key instead of connection only, since the latter causes a collision (EGL-connection == X11-connection).
* Fixes Bug 615 - X11: Use proper screen index/name for shared resources and ↵Sven Gothel2012-09-171-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | device denominated 'mutable' surfaces (dummy, offscreen, ..) Fix follows findings of Rob Hatcherson, but instead of parsing the display connection ourself we use the X11 macro 'DefaultScreen(display)' See <http://tronche.com/gui/x/xlib/display/opening.html#Display> "The screen number specified in the display_name argument is returned by the DefaultScreen() macro (or the XDefaultScreen() function)." Since I currently have no two-head X11 setup here, only xinerama via virtualbox, pls test / validate. Note: One Display connection may span multiple screens, i.e.: display 'lala:0.1' may span from screen 1 - 3 (non xinerama mode) Discussion: [1] How to validate whether a screen number belongs to one display connection ? We can query ScreenCount(display), however it is not clear what the range would be. [2] With 1 display connection spanning multiple screens, what is/are the proper connection string[s] ? [3] After all, it seems this ancient configuration really cannot beat a modern setup w/ XRandR having Xinerama enabled. The latter is the default anyways.
* Anotate focus tests in regards to OSX stability .. sporadically won't gain ↵Sven Gothel2012-09-171-3/+3
| | | | focus and receive keyboard input via AWTRobot