| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
| |
|
|
| |
internal APIs, critical array is not required, hence redundant.
|
| |
|
|
|
|
|
|
|
| |
- attributes +NSOpenGLPFANoRecovery +NSOpenGLPFAAccelerated
- use SurfaceUpdateListener() to notify layer
- swapBufferImpl() adapt to 0038e2d41825c22bdd18a7b86a8229a3fab674a3
- pbuffer: don't enforce POT tex-size to surface size, but pbuffer only
- FIXME: Check POT pbuffer/tex-size for X11/Win32/.. !
-
|
| |
|
|
|
|
|
|
|
| |
- Use proper OSX types for NS/CGL prototypes (gluegen) and impl.
- Impl layeredSurface (native):
- OSXUtil: NSView backing creation
- OSXUtil: AttachJAWTSurfaceLayer
- CGL: NSOpenGLLayer type impl. and hook
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
jogamp.<module> (2/2) - edit files
- com.jogamp.opengl.impl -> jogamp.opengl
- com.jogamp.opengl.util.glsl.fixedfunc.impl -> jogamp.opengl.util.glsl.fixedfunc
- com.jogamp.nativewindow.impl -> jogamp.nativewindow
- com.jogamp.newt.impl -> jogamp.newt
This sorts implementation details from the top level, ie skipping the public 'com',
allowing a better seperation of public classes and implementation details
and also reduces strings.
This approach of public/private seperation is also used in the OpenJDK.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
glMap*Buffer*; GLX/WGL/CgGL: All runtime dynamic; Misc ..
TODO: Compile and test on MacOSX ..
Fix:
=====
Multithreading/Locking:
See jogl/doc/Implementation/MultiThreading.txt
- Locking layer is not platform agnostic, ie GLContextImpl, GLDrawableImpl, ..
and NEWT: Window/Display
- No more use of JAWT global lock necessary, removed.
- No need for X11 Display lock, on the contrary,
this made the NV driver hang.
- Use common window/surface lock
- All NativeWindow surfaceLock's are recursive now
glMapBuffer: If size is 0, don't do cont with the native call.
glMapBufferRange: Fix capacity.
glNamedBufferDataEXT: Track the size.
glMapNamedBufferEXT: Manual impl. - use the tracked size
glXGetVisualFromFBConfig, glXChooseFBConfig, glXChooseVisual: Instead of
ignoring and implement a renamed version (*Copied), we just use ManualImplementation
for the proper copy-result code.
DesktopGLDynamicLookupHelper: Initialize _hasGLBinding* attributes
in the determing loadGLJNILibrary() method, which is called by super().
Otherwise static init will overwrite them after the super() call.
X11GLXDrawableFactory: Don't release anything at shutdown (removed sharedContext.destroy()),
since this caused a freeze/SEGV sometimes.
Fixed NEWT's reparentWindow() functionality incl NewtCanvasAWT usage.
- Native: if not visible, don't focus, etc
- NewtCanvasAWT: Use the container size to start with
- Run the command on the EDT
Using GlueGen's new DynamicLibraryBundle utility:
- X11, Windows and MacOSX OpenGL adapted to DynamicLibraryBundleInfo.
- X11GLXDynamicLookupHelper -> X11GLXDynamicLibraryBundleInfo
- Remove all path from lib names.
- GL order: libGL.so.1, libGL.so, GL
- shallLinkGlobal: true -> to server some 'old' DRI systems
-> http://dri.sourceforge.net/doc/DRIuserguide.html
- shallLookupGlobal: false
- Try both : glXGetProcAddressARB and glXGetProcAddress
- Using bootstrap: GLX.glXGetProcAddress(long glxGetProcAddressHandle, String glFuncName)
Found the issue with LIBGL_DRIVERS_PATH, ie if not set
no valid GL instance can be found (ie ATI fglrx/DRI).
This may happen if using a differen user than the desktop user
for whom the env var is set within some /etc/X11/Xsession.d/ script.
Enhancements:
=============
GLBufferSizeTracker: Use IntIntHashMap and add DirectState size tracking.
GLBufferStateTracker: Use IntIntHashMap.
GLStateTracker: Use IntIntHashMap.
GLDynamicLookupHelper: More generic (global loading/lookup and GetProcAddress function name list),
remove redundant code.
FIXME:
MacOSXCGLDynamicLookupHelper:
- Not tested
- Not using NSImage lookup anymore as recommended by OSX API Doc,
so dlsym is used always (to be tested)
WindowsWGLDynamicLookupHelper:
- Not tested
GLX/WGL/CgGL is all runtime-dynamic as now, ie loaded and looked-up at runtime,
no compile time dependencies to GL anymore, nor a need to specify CgGL.
Split up WGL in GDI and WGL, to allow proper dynamic runtime linkage of OpenGL32
while using static binding to GDI32
NEWT events generated by native code are enqueued and not send directly.
This should ease locking mechanisms .. if any are necessary.
NEWT: More platform specific code moved to *Impl method,
simplifying the generic code of the superclass and impl protocol.
Cleanup:
=========
Replace all InternalBufferUtil's with com.jogamp.common.nio.Buffers
Removed all InternalBufferUtil's from repository
Removed GLContextImpl notion of 'optimized' surface locking,
where the surface gets unlocked during makeCurrent/release.
This just makes no sense and would impact multithreading in a horrible way.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Fixes for OpenGL 3.2
- More strict define/function checking
- Better documentation in source code: extension relation
- GL extension marker exclusion is done by GlueGen now
- Adapted gluegen cfg files
Added OpenGL 3.2 extensions for GL3
Revalidation of all OpenGL 2.0/3.0/3.1 and 3.2 extensions
GL2GL3 based on OpenGL 2.0/3.0 headers
EGL shares common egl-common.cfg file
Removed experimental java system property 'jogl.GLContext.3_2',
try a 3.2 context in case of GL3.
X11GLXContext:
- Always try to make a created context current,
to verify it's usability. This is different than under WGL.
- Temporary removed usage: GLX.GLX_CONTEXT_PROFILE_MASK_ARB
NV driver bug.
|
|
|
git-svn-id: file:///usr/local/projects/SUN/JOGL/git-svn/svn-server-sync/jogl/trunk@1959 232f8b59-042b-4e1e-8c03-345bb8c30851
|
