| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
| |
need for explicit call
- OffscreenLayerSurface.layoutSurfaceLayer() removed, no more required
- JAWTWindow adds a ComponentListener, which issues FixCALayerLayout() at resized, moved and shown.
- MyNSOpenGLLayer no more requires fix*Size() methods
- MyNSOpenGLLayer::setDedicatedSize() need no explicit CATransaction, performed by caller.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
RunOnMainThread(waitUntilDone:=true,..) can deadlock the main-thread if called from AWT-EDT,
since the main-thread may call back to AWT-EDT while injecting a new main-thread task.
This patch revises all RunOnMainThread CALayer usage, resulting in only one required left:
- OSXUtil.AddCASublayer() w/ waitUntilDone:=false
Hence the CALayer code has no more potential to deadlock main-thread/AWT-EDT.
OSXUtil.AddCASublayer() must be performed on main-thread, otherwise the
CALayer attachment will fail - no visible rendering result.
+++
Note: A good trigger to test this deadlock is to magnify/zoom
the OSX desktop (click background + ctrl-mouse_wheel)
before running some unit tests.
TestGLCanvasAWTActionDeadlock01AWT and TestAddRemove02GLWindowNewtCanvasAWT
also have the potential to trigger the mentioned deadlock.
|
|
|
|
| |
DPYs; JAWTWindow: Add JAWT info in toString()
|
|
|
|
| |
808a9a27a8c1c9e0a6701a8dd81d51f8daa8129d
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
mechanism
Due to a NEWT WindowClosing event regression cause by NewtCanvasAWT changes
a review of our WindowClosing event mechanism was required.
Important cleanups are marked w/ '(*)' below.
I would have preferred to change the 'WindowListener.windowDestroyNotify(WindowEvent)'
method to pass a WindowCloseEvent object exposing more information like
toolkit or programmatic destruction and passing whether a 'closing' or 'nop' action
will be performed based on the WindowClosingMode.
For now I postponed this idea .. since it would change the API again,
but may reconsider it after merging the Android 'closing' patch.
- InputEvent.consumedTag -> NEWTEvent.consumedTag
- Window
- (*) Promote setWindowDestroyNotifyAction(Runnable) to public,
former WindowImpl.setHandleDestroyNotify(boolean).
Using a Runnable action for WindowImpl.windowDestroyNotify(boolean)
allows a setting defined alternative for destroy() and gets rid
of [ab]using WindowListener.windowDestroyNotify(WindowEvent) for
lifecycle actions. Used in:
- GLWindow
- GLAutoDrawableDelegate impl.
- WindowImpl
- Respect NEWTEvent.consumedTag for WindowEvents as well
- (*) Impl. setHandleDestroyNotify(boolean) (see above)
- (*) destroy() simply sends out pre- and post- destruction Window events,
where windowDestroyNotify(boolean) sends out the pre-destruction event if NOP.
- (*) windowDestroyNotify(boolean) is public now, allowing other impl. details
to follow proper destruction using handleDestroyNotify Runnable (-> NewtCanvasAWT).
- AWTWindowClosingProtocol:
- addClosingListenerOneShot() -> addClosingListener()
- calling addClosingListener() at addNotify()
- calling removeClosingListener() at removeNotify()
- AWTWindowClosingProtocol ctor taking NOP runnable,
allowing to send WindowEvent.EVENT_WINDOW_DESTROY_NOTIFY at WindowClosingMode.DO_NOTHING_ON_CLOSE
- add/remove listener on AWT-EDT
- AWTWindowAdapter
- Add 'removeWindowClosingFrom(..)',
allowing to remove window closing event fwd.
- Also fwd windowClosed in window closing fwd'ing.
- NewtCanvasAWT
- (*) Utilize AWTWindowClosingProtocol NOP runnable (see above)
to fwd closing-NOP event to NEWT
- (*) Unify remove/destroy code in destroyImpl(..)
- !removeNotify -> destroy NEWT child programatic or as toolkit event
- removeNotify || windowClosing -> destroy jawtWindow
- (*) Remove AWTWindowAdapter/AWTParentWindowAdapter's windowClosingListener,
since we utilize AWTWindowClosingProtocol
- DisplayImpl
- Adding 'final void dispatchMessage(final NEWTEvent event)'
allowing to remove the NEWTEventTask wrapping for no reason
in enqueueEvent(..) if on EDT and waiting.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Fix CALayer animation:
- All CALayer animations are set to nil via overriding 'actionForKey'
- Fix CALayer pos/size bug:
- Fix root and sub CALayer position to 0/0 and size on the main-thread w/o blocking.
- If the sub CALayer implements the Objective-C NativeWindow protocol NWDedicatedSize (e.g. JOGL's MyNSOpenGLLayer),
the dedicated size is passed to the layer, which propagates it appropriately.
- On OSX/Java7 our root CALayer's frame position and size gets corrupted by its NSView,
hence we have created the NWDedicatedSize protocol.
|
|
|
|
| |
SWT event dispatching.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
where our CALayer is moved out of the visible area
- same erroneous behavior for GLCanvas and NewtCanvasAWT
- sized-frame: Set framesize and validate() it
- sized-component: Set component preferred size and call frame.pack()
- added workaround 'OffscreenLayerSurface.layoutSurfaceLayer()' to fix CALayer size, which snaps for:
- OK initial size before setVisible: sized-frame and sized-component
- OK resize w/ sized-frame
- OK manual frame resize
- Invisible: w/ sized-component after setVisible()
++
- CALayer-Sublayer (GL) has additional retain/release when added/removed
to be on safe side.
|
|
|
|
| |
NativeWindow-JAWT: Remove c.fixup.jawt.version.macosx (redundant, libjawt.dylib is rpath'ed always)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
main-thread deadlock'ed due to locked shared context
NSOpenGLLayer::openGLContextForPixelFormat(..) is performed on main-thread at 1st NSOpenGLLayer display method.
This happened irregulary, i.e. sometimes (T0) right after NSOpenGLLayer creation and attachSurfaceLayer()/AddCASublayer(..),
sometimes later (T1).
NSOpenGLLayer::openGLContextForPixelFormat(..) uses the passed shared user context.
The shared user context is locked at NSOpenGLLayer's creation (T0) and if performed at this early time
the call deadlocks due to pthread_mutex wait for the shared user context.
This fix performs NSOpenGLLayer creation and layer attachment while the shared user context
is kept unlocked and enforces NSOpenGLLayer display and hence NSOpenGLLayer::openGLContextForPixelFormat(..).
Added CGL.setNSOpenGLLayerEnabled(..) to enable/disable NSOpenGLLayer - currently not used.
- Passed AddRemove tests for GLCanvas/Swing and GLWindow/NewtCanvasAWT w/ 100 loops on Java6 and Java7 on OSX.
- Passed Instruments Leaks test w/ 10 loops on Java6 and Java7
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
'CGL.releaseNSOpenGLLayer' triggers release - but very late w/ AWT usage.
OSXUtil_RemoveCASublayer0's added '[subLayer release]' in commit f6e6fab2a7ddfb5c9b614cb072c27ff697629161
is wrong, since 'CGL.releaseNSOpenGLLayer' actually does trigger it's release.
This was not seen w/ AWT tests, since it happens very later.
A NewtCanvasAWT test disclosed this error -> removed that extra release call.
The culprit for the late release w/ AWT usage was CGL.createNSOpenGLLayer's call in the current thread.
Moving it to the Main-Thread fixed the problem.
All CALayer lifecycle calls are issued on the Main-Thread now.
NSOpenGLLayer's CVDisplayLink OpenGL fitting via 'CVDisplayLinkSetCurrentCGDisplayFromOpenGLContext'
is now performed at it's context creation in 'NSOpenGLLayer::openGLContextForPixelFormat'.
The 'extra' release of the NSOpenGLLayer's NSOpenGLContext as introduced in commit f6e6fab2a7ddfb5c9b614cb072c27ff697629161
is still valid.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Layers and native GL-Context are _not_ Released ; Java Side wait for Main-Thread
- Fix Bug 690: Occasional Freeze on CVDisplayLinkStop
- NSOpenGLLayer.disableAnimation() shall not claim the renderLock mutex,
since the CVDisplayLink callback could be waiting for the lock.
This waiting callback could freeze the call to CVDisplayLinkStop.
- Fix Bug 691: Layers and native GL-Context are _not_ Released
- Following proper release cycle:
Context unrealized:
- JAWTWindow.detachSurfaceLayer() -> OSXUtil.RemoveCASublayer(..)
- CGL.releaseNSOpenGLLayer(..)
JAWTWindow.destroy()
- MacOSXJAWTWindow.UnsetJAWTRootSurfaceLayer(..)
- OSXUtil.DestroyCALayer(..)
- 'Magic' CALayer release calls (w/o manual retain beforehand) at:
- OSXUtil.RemoveCASublayer(..): [subLayer release]
- MacOSXJAWTWindow.UnsetJAWTRootSurfaceLayer(..): [rootLayer release]
- OSXUtil.DestroyCALayer(..): [rootLayer release]
- 'Magic' NSOpenGLLayer's NSOpenGLContext dealloc:
- [NSOpenGLContext clearDrawable]
- CGLDestroyContext( [NSOpenGLContext CGLContextObj] )
- Java Side wait for Main-Thread
- Waiting for the delegated Main-Thread on the Java side eases debugging
and won't block the Main-Thread in native code.
- Utilizing this for all CALayer calls
Test case: TestGLCanvasAddRemove01SwingAWT
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
EGL/ES2 tests, attempt to fix wrapped EGL case
- Bug 665 (part 2) was commit 7fd5f76e1eb4bbf93fe9b1171744bd755d8f96e4
- Add EGL/ES2 tests in
- TestGLContextDrawableSwitch01NEWT
- TestGLContextDrawableSwitch11NEWT
- Attempt to fix wrapped EGL case (incomplete)
- Using EGL/ES w/ non native EGL device/surface, but natively wrapped instances (most of the cases),
a 'complicated' delegation of
Native-Upstream -> EGL-Proxy -> EGL-Instance
is being used heavily relying on the objects lifecycle.
GLEventListenerState tries to roll back the realized state
and even sets the upstream device handle,
but this doesn't seem to be sufficient on X11.
Discussion:
It might turn out that we only can implement the survival of GLContext
and it's display device reliable w/ EGL within the GLAutoDrawable implementation,
which can hold the previous not destructed instances.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Don't cache 'displayHandle'
- ProxySurface: Add 'NativeSurface getUpstreamSurface()' allowing querying direct access of a backing surface representing this instance.
- Use case: EGLWrappedSurface
- Default impl. returns null
- ProxySurfaceImpl: Don't cache 'displayHandle'
- getDisplayHandle() is 'final', no more 'shortcut' code allowed
due to re-association incl. display handle.
- See commit b738983638703bb721ee4c9820c8ef43e2252e73
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Changes allowing re-association (incl. null) of GLContext/GLDrawable:
- GLAutoDrawable: Refine API doc 'setContext(..)'
- GLContext: Refine API doc: 'setGLDrawable(..)' 'getGLDrawable()'
- GLContextImpl.setGLDrawable(): Handle null drawable
- GLAutoDrawableDelegate/GLAutoDrawableBase: Allow null GLContext
- GLDrawableHelper.switchContext(..)/recreateGLDrawable(): Balance GLContext.setGLDrawable(..) calls
- New GLEventListenerState, holding state vector [GLEventListener, GLContext, .. ]
impl. relocation of all components from/to GLAutoDrawable.
- GLDrawableUtil
- Using GLEventListenerState for swapGLContextAndAllGLEventListener(..)
+++
NEWT Window*:
- getDisplayHandle() is 'final', no more 'shortcut' code allowed
due to re-association incl. display handle.
- close*:
- close config's device (was missing)
- null config
+++
Changes allowing reconfig of Display handle as required
to re-associate pre-existing GLContext to a 'window':
- AbstractGraphicsDevice: Add isHandleOwner() / clearHandleOwner()
- Impl. in X11GraphicsDevice and EGLGraphicsDevice, NOP in DefaultGraphicsDevice
- DefaultGraphicsConfiguration add 'setScreen(..)'
- MutableGraphicsConfiguration
- Make DefaultGraphicsConfiguration.setScreen(..) public
- NativeWindowFactory add 'createScreen(String type, AbstractGraphicsDevice device, int screen)'
- Refactored from SWTAccessor
- NativeWindow x11ErrorHandler: Dump Stack Trace in DEBUG mode, always.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
RGB565; Fix HiSilicon/Vivante/Immersion.16 EGLConfig selection (zero depth buffer @ visualID)
- NEWT/Android Fix PixelFormat/NativeWindowFormat/VisualID Selection
- Fix allows proper selection of native window formats: RGBA8888, RGBX8888 and RGB565
- Selection is performed in 3 steps:
1) @ Construction (non native): SurfaceHolder.setFormat( getSurfaceHolderFormat( caps ) )
2) @ Native Surface Creation: getANativeWindowFormat( androidFormat) -> ANativeWindow_setBuffersGeometry(..)
Note: The set native format is revalidated, i.e. read out via ANativeWindow_getFormat(..).
3) @ EGL Creation: ANativeWindow_getFormat(..) -> fixCaps(..) - simply fixing the chosen caps.
- NEWT GLWindow.GLLifecycleHook.resetCounter:
- Also reset GLAnimatorControl's counter, if attached.
- NEWT WindowImpl -> GLLifecycleHook.resetCounter() calls issued _after_ operation before unlock().
- JOGL/EGLGraphicsConfigurationFactory
- Validate whether the visualID matching EGLConfig depth buffer is suitable.
On HiSilicon/Vivante/Immersion.16: Depth buffer w/ matching visualID is zero!
- NativeWindow/Capabilities.compareTo: Fix alpha comparison
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Horizontal Scrolling Behavior (OSX, X11, Win32); Bug 639: High-Res Mouse-Wheel
- API update 'float getWheelRotation()':
Usually a wheel rotation of > 0.0f is up, and < 0.0f is down.
Usually a wheel rotations is considered a vertical scroll.
If isShiftDown(), a wheel rotations is considered a horizontal scroll,
where shift-up = left = > 0.0f, and shift-down = right = < 0.0f.
However, on some OS this might be flipped due to the OS default behavior.
The latter is true for OS X 10.7 (Lion) for example.
The events will be send usually in steps of one, ie. -1.0f and 1.0f.
Higher values may result due to fast scrolling.
Fractional values may result due to slow scrolling with high resolution devices.
The button number refers to the wheel number.
- Fix Bug 659: NEWT Horizontal Scrolling Behavior (OSX, X11, Win32)
- See new API doc above
- X11/Horiz: Keep using button1 and set SHIFT modifier
- OSX/Horiz:
- PAD: Use highes absolute scrolling value (Axis1/Axis2)
and set SHIFT modifier for horizontal scrolling (Axis2)
- XXX: Use deltaX for horizontal scrolling, detected by SHIFT modifier. (traditional)
- Windows/Horiz:
- Add WM_MOUSEHWHEEL support (-> set SHIFT modifier), but it's rarely impl. for trackpads!
- Add exp. WM_HSCROLL, but it will only be delivered if windows has WS_HSCROLL, hence dead code!
- Android:
- Add ACTION_SCROLL (API Level 12), only used if layout is a scroll layout
- Using GestureDetector to detect scroll even w/ pointerCount > 2, while:
- skipping 1st scroll event (value too high)
- skipping other events while in-scroll mode
- waiting until all pointers were released before cont. normally
- using View config's 1/touchSlope as scale factor
- Fix Bug 639: High-Res Mouse-Wheel
- getWheelRotation() return value changed: int -> float
allowing fractions, see API doc changes above.
- Fractions are currently supported natively (API) on
- Windows
- OSX
- Android
- AndroidNewtEventFactory ir refactored (requires an instance now) and
AndroidNewtEventTranslator (event listener) is pulled our of Android WindowDriver.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Windows: Child window is not translucent at all
- X11: Child window is translucent to parent's background,
however - parents content is _not_ 'composed in'.
- TODO: Find whether there is a solution or not.
- Note: The child window does not change it's rel. position
if parent moves! This is a feature, since we don't
have impl. a layout.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Test*NewtEventModifiers for SWT (TestNewtEventModifiersNewtCanvasSWT)
- Misc OSX/SWT: OSXUtil.RunOnMainThread(..) refinement
- 'waitUntilDone' is implemented on Java site via lock/wait on RunnableTask to not freeze OSX main thread.
- Fix Test*NewtEventModifiers for SWT (TestNewtEventModifiersNewtCanvasSWT)
- Deal with SWT's requirement to run the SWT event dispatch on the TK thread,
which must be the main thread on OSX.
We spawn off the actual test-action into another thread,
while dispatching the events until the test-action is completed.
- AWTRobot: Add 'void requestFocus(Robot robot, Object obj, int x, int y)'
- Use waitForIdle() only if programmed in Robot (Deadlock w/ OSX SWT)
- Required for SWT usage (see above)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
GLCapabilities on X11 (feature complete)
To allow custom GLCapabilities, we had to use native parenting on X11 w/ a new child window.
The desired visualID chosen by the users GLCapabilities is passed to the new child window.
The redraw drops must be caused by the original GDK or the new child GDK window.
Now we use a plain X11 child window similar to NEWT's X11 window and NewtCanvasSWT,
which doesn't expose this bug.
(Note: SWTAccessor/GLCanvas still contains the uncommented GDK code path for further inspection, if desired)
Also added SWTNewtEventFactory to test event handling on the SWT GLCanvas w/ GearsES2.
TestSWTJOGLGLCanvas01GLn tests custom GLCapabilities now.
SWTEDTUtil has been moved to private: com.jogamp.newt.swt -> jogamp.newt.swt.
|
| |
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
| |
PrivilegedAction for static initSingleton block (SWTAccessor, NewtFactory, NativeWindowFactory)
|
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
|
|
| |
Propagate drawable change to MacOSXCGLContext where either context/NSView or context/NSOpenGLLayer
association needs to get updated.
Fixes drawable/context switch.
|
|
|
|
|
|
|
|
| |
unboxed
No reason for ever using new Boolean when the constants are available.
Signed-off-by: Harvey Harrison <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
| |
92398025abdabb2fdef0d78edd41e730991a6f94 GlobalToolkitLock for create/destroy
Turns out on it has no effect and ATI prop. driver still has XCB failures at this point.
|
|
|
|
|
|
|
|
|
| |
XInitThreads())
X11Util/Native: Fix X11Util_initialize0() arguments were wrong and code still invoked XInitThreads() .. woops;
Added missing included "jogamp_nativewindow_x11_X11Util.h" incl. it's generation via javah,
which was the culprit of not detecting it at compile time.
This is a fix for commit fbe331f013608eb31ff0d8675f4e4c9881c9c48b
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Utilizing a GlobalToolkitLock in general to lock the display connection results in deadlock
situations where locked surfaces signal other [offscreen] surfaces to render.
We have to see whether we find a better solution, for now sporadic XCB assertion still happen.
But it is preferrable to point to the root cause, then to jumping through hoops to complicate locking
or even to deadlock.
Locking:
- X11GLXGraphicsConfigurationFactory add missing device locking in:
- getAvailableCapabilities
- chooseGraphicsConfigurationStatic
- Newt/X11Window: Discard display events after window close.
Relax ATI XCB/threading bug workaround:
- ToolkitProperties: requiresGlobalToolkitLock() -> hasThreadingIssues()
- NativeWindowFactory: Don't use GlobalToolkitLock in case of 'threadingIssues' the impact is too severe (see above)
- NativeWindowFactory: Add getGlobalToolkitLockIfRequired(): To be used for small code blocks.
If having 'threadingIssues' a GlobalToolkitLock is returned, otherwise NullToolkitLock.
- X11GLXContext: [create/destroy]ContextARBImpl: Use 'NativeWindowFactory.getGlobalToolkitLockIfRequired()' for extra locking
Misc Cleanup:
- *DrawableFactory createMutableSurface: Also create new device if type is not suitable
- *DrawableFactory createDummySurfaceImpl: Pass chosenCaps and use it (preserves orig. requested user caps)
|
|
|
|
| |
(Fix regression of commit 92398025abdabb2fdef0d78edd41e730991a6f94)
|
|
|
|
| |
(detecting impl. bugs)
|
| |
|
|\ |
|
| | |
|
| | |
|
| |
| |
| |
| | |
Signed-off-by: Xerxes Rånby <[email protected]>
|
| |
| |
| |
| | |
'nativewindow.debug.X11Util.ATI_HAS_NO_MULTITHREADING_BUG' to disable GlobalToolkitLock workaround for testing.
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
| |
dependency (signal is async); X11Util: XSync w/ discard reusable Display when queue-in.
|
|
|
|
| |
server roundtrips.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
NativeWindowFactory's TYPE_* strings, not NEWT's; NEWTFactory: Use default NEWT package name if rel. 'path'.
Use relative sub-package names in NativeWindowFactory's TYPE_* strings, not NEWT's
Otherwise NEWT depends solely on NativeWindowFactory strings
- Default subpackages denominate a relative path, i.e. start with a dot: '.egl', '.windows', '.x11'
- Custom name may be absolute, eg: 'my.company.special.drivers.chip4'
NEWTFactory: Use default NEWT package name if relative 'path'.
- If NativeWindowFactory type starts w/ dot (rel. path), simply prepend the default NEWT package prefix
otherwise use complete package name as-is.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
NSOpenGLContext::setView(NULL || incomplete-view) ; Add missing [ctx release] in MyNSOpenGLLayer ; Misc
NSOpenGLContext::setView(NULL || incomplete-view) was called on 2 occasions:
[1] - MacOSXCGLContext native createContext
[2] - NSOpenGLLayer internals
For [1], we simply don't call setView(..) ourselfs in case view is NULL or incomplete (invisible)
For [2], we need to wrap the class 'MyNSOpenGLContext:NSOpenGLContext' and capture setView(NULL)
++
Add missing [ctx release] in MyNSOpenGLLayer, otherwise resource won't get dealloc'ed
+++
Misc:
- MacOSXCGLContext. contextRealized(true): set pbuffer -> ctx, otherwise 1st makeCurrent call will not catch it
- MacOSXOnscreenCGLContext: don't add ContextUpdater to invisible ProxySurface's (dummy window)
|