| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
|
|
|
| |
SUCCESS
|
| |
|
|
|
|
|
|
| |
an <init> exception.
Also adding current thread name.
|
| |
|
|
|
|
| |
(boolean)success instead.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
96af6c9bf2d683115996 moved the creation of the AWTGraphicsConfiguration,
GLDrawable and GLContext to the GLCanvas constructor.
This disabled recreation, removeNotify()/addNotify() after 1st addNotify(),
since the drawable/context were already set to null.
Smart recreation, ie not destroying the resources completly:
1 - drawable
2 - context
3 - AWTGraphicsConfiguration's Device (X11: Display)
is not possible, since removeNotify() is the only destroy notification
we have from AWT, which would leave the above resources open -> leak.
This is especially true for (3), since the device is created 'on the fly'.
In NEWT we use lazy creation/destroy and are able to keep the references alive.
A remodelling of AbstractGraphicsConfiguration would be required,
which indeed would be overkill.
Simple solution is to move the creation block back to addNotify()
but before 'super.addNotify()', since it needs our chosen AWTGraphicsConfiguration.
Also flagging sendReshape in addNotify().
Added test case com.jogamp.test.junit.jogl.awt.TestAWT03GLCanvasRecreate01.
|
|
|
|
| |
pause()/resume() so it becomes a non critical to multithreading, hence display() or state change needs to be synced
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
429, 428 and 405
Fixes bugs WGL pixelformat related bugs:
http://jogamp.org/bugzilla/show_bug.cgi?id=397
http://jogamp.org/bugzilla/show_bug.cgi?id=410
http://jogamp.org/bugzilla/show_bug.cgi?id=429
http://jogamp.org/bugzilla/show_bug.cgi?id=428
http://jogamp.org/bugzilla/show_bug.cgi?id=405
Tested on Window7-x86 (amd/nvidia), WinXP-x32-VirtualBox.
Solution:
Cleaned up X11/GLX code to use it as a correct boilerplate
for the new WGL selection, which now duplicates the same behavior.
X11/GLX and WGL follow the common logic:
- 1st try:
- get GLCapabilities based on users GLCapabilities
- setting recommendedIndex as preferred choice
- 2nd try:
- get all GLCapabilities available
- no preferred recommendedIndex available
If no recommendedIndex has been selected and no chooser has been passed,
we use the DefaultGLCapabilitiesChooser.
Choose the GLCapabilities if a chooser is given (or see above).
|
|
|
|
| |
class HINSTANCE is acquired at static initialization
|
|
|
|
| |
cause a SIGSEGV at JVM exit.
|
|
|
|
| |
to simplify usage.
|
|
|
|
| |
SIGSEGV on AMD
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Consider use cases with many drawables and no drawables at start,
this had to be reflected all over this patch set, implementation,
usage and test cases.
- GLAnimatorControl
- refine API doc / states
- add 'void remove(GLAutoDrawable drawable);'
- Animator*:
- using RecursiveLock 'stateSync' for all actions out of the big synchronized (animator) block:
- get status methods (thread, isPaused, ..), hence no more synchronized
- display drawables change, utilizing synced ArrayList swap
This removes the need for volatiles usage shouldPause/shouldStop within the display method.
- added blocking wait for state change for add(GLAutoDrawable)/remove(GLAutoDrawable) method
- remove flawed double checked locking in anim thread (pause/idle condition)
- thread is now a daemon thread, hence it won't hinder the JVM from shutdown
-
- Animator use change:
- Always resume after pause, except in case of final destroy -> NEWT invalidate / GLCanvas,
this considers use cases with many drawables and no drawables at start.
- GLDrawableHelper: Don't pause at implicit dispose()
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
for Canvas. ; Adding FSAA test.
Canvas/X11:
The Canvas GraphicsConfiguraton should be chosen before the native peer is being created.
Choosing AWT GraphicsConfiguration (all platforms):
Don't filter our capabilities with 'AWTGraphicsConfiguration.setupCapabilitiesRGBABits(capsChosen, gc)',
not necessary (see above) and it would remove ourrequired alpha channel.
Canvas display():
Don't render if drawable is not realized (yet).
|
| | |
|
|/
|
|
| |
made references created in double checked locks volatile.
|
| |
|
| |
|
|\ |
|
| |
| |
| |
| | |
and getChosenCapabilities() return immutable instances. Add cloneCapabilities() to create a mutable clone of an immutable set of capabilities.
|
| |
| |
| |
| | |
Unregister the shutdown hook if called manually (recommended!).
|
|/
|
|
| |
NEWT WindowsWindow GetRelativeLocation() native implementation to GDI as well.
|
|
|
|
| |
not available ..
|
|
|
|
|
|
|
|
|
|
| |
Preparation to support multiple devices on one machine,
hence adding the unitID a unique ID/index of the associated GPU, or GPU affinity.
Adding getUniqueID() to return a cached semantic unique string id for the device.
This was removed from the temp. impl in JOGL's GLContext, added unitID.
All other changes just adapt to the above.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
device.
AbstractGraphicsDevice's 'connection' and 'type' attribute is used as a unique key
to map GLProfiles and GLContext's major/profile -> major/minor/profile mapping.
Eager initialiaztion as well as lazy is supported to maintain a simple API.
This is currently tested on X11, where one app display NEWT/GL window and content
on the local and remote device.
See TestRemoteWindow01NEWT.java and TestRemoteGLWindows01NEWT.java
|
| |
|
|
|
|
| |
Add threadName to DEBUG output
|
| |
|
|
|
|
|
|
|
|
|
| |
and NewtVersion.
Adapt to GlueGen Version changes:
b735755815312b5fe2c003642de60711be1cd645 .. 556c7e70d3d57aa99b5787b1e4d8a7b1c299ed3f
Show information of all subcomponenet.
|
|
|
|
|
|
|
|
|
|
| |
support multi devices & displays.
Currently only the X11 Display connection is implemented to support multiple device connections.
Other platforms may follow.
This allows correct mapping and caching of higher level resources,
eg. ProcAddressTable, GL version mapping etc with respect to the display device.
|
| |
|
|
|
|
| |
won't exit
|
|
|
|
| |
verification
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
- Adding gluegen-gl.jar for GlueGen GL compile time parts.
- Removed dead build-gluegen.xml
- Adding com/jogamp/gluegen/runtime/opengl/* to jogl core
- Moved gluegen gl part to jogl/classes
|
|
|
|
|
|
|
|
|
|
| |
fail-fast hash cache, ..
Relates to GlueGen 6b6b9b3b81cdc85b7260664ebec547756a6be5d7, branch sgothel_wip_fixes01.
Memory object size is ptrdiff_t, hence long (64bit).
The hash value must include size as well, otherwise boundaries cannot be verified. (security)
Double check hash collisions while adding a new MemoryObject.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
46f17013c7cd59d551371edb2c1a4a57f8cbd84f (code dependencies)
This currently removes KD fullscreen for NV devices and
the whole OMX NV stream/file type detection, which renders the OMX hack useless.
However, updated EGL sync (NV proprietary) to EGL_KHR_reusable_sync and EGL_KHR_fence_sync.
|
|
|
|
| |
makeCurrentImpl/releaseImpl with X11 Error Handle
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix context scanning
---------------------
Ie on 3.0 NVidia systems, trying to create a 4.1 context leads
to a BadAlloc X11 error.
A prev patch disabled the X11 error handler, to reduce sideeffects.
In case AWT is not being used, the BadAlloc isn't catched and the JVM freezes
within the native function call.
Decorating context scanning with NativeWindow's 'setX11ErrorHandler' solves this issue.
Simplifications
-----------------
X11: always try a direct context. If this is not possible due to the display connection,
an indirect is being used anyways. Hence 'createContext(boolean direct)' -> 'createImpl()'.
X11/WGL: Simplify the context creation logic a bit.
|
|
|
|
|
|
| |
'setGLFunctionAvailability' call while scanning all available context.
.. also reorder context scanning from low -> high
|