| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
cleanup in X11
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
visibility:
- set pointer visible or invisible
confined:
- confine pointer to window, or not
warp:
- set mouse position within the window
Implemented for X11, tested manually with TestGearsES2NEWT (see code for action keys).
TODO: Windows, MaxOSX and Android (limited)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Note: surface's format _is_ the nativeVisualID of EGL.
More ANativeWindow (ANW) coop:
- add acquire/release for ANW
- validate ANW format
- Add fixCaps: 'format -> GLCaps' validation
Test:
- Constructor: Start w/ surface format RGB_565 (default)
- createNative: Filter EGLConfig w/ surface format
- Move surfaceHandle fetching from surfaceCreated -> surfaceChanged,
where the format is available.
- surfaceChanged: use fixedCaps validation
|
|
|
|
|
|
|
|
|
| |
parent,
a white window rectangle remains.
.. we also need to understand the absolute screen position better,
ie. when required and when not (at window creation currently).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
'defer' 1st argument
- Adapt to GlueGen's Lock ChangeSet: e4baba27507ce78e64a150ec6f69fb96f5721a34
- All java callbacks for native have 'defer' 1st argument.
This allows enqueuing resulting events to the EDT if required,
ie. the native thread may not be 'compatible' (MacOSX).
- MacOSX-Native: enqueue key/mouse events and defer:=true for all java callbacks
Since we are comming from a 3rd-party thread (AWT/NSApp-MainThread)
we shall not abuse it.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Feature related:
- Added always-on-top
- Added translucency
- Child Window Position
- AWT parent: manual traverse up the tree and calc position on screen
(Problem: the parent view rect is not at the proper position,
but covers the whole frame)
EDTUtil related:
- Works now w/ AWT ot headless (again)
- OSX native JNI callbacks gathering JNIEnv properly
and attaches/detaches thread.
- AWT case: using AWT-Event which properly dispatches our cocoa events
- MainThread (headless) case: Fork off thread w/ main class
and kick off NSApp run().
This leads to same behavior as w/ AWT case.
- Using DefaultEDTUtil
- Cleanup MainThread (implements EDTUtil)
- Currently not used as EDTUtil (osx), just as launcher
- Removed EDTUtil impl code, reuse DefaultEDTUtil
- Cleanup AWTEDTUtil (implements EDTUtil)
- Currently not used as EDTUtil (osx)
|
|
|
|
|
|
| |
reparent/fullscreen)
Window position is not deterministic enough and slows down processing while sync on it
|
|
|
|
| |
setFullscreenEWMH
|
| |
|
| |
|
| |
|
|
|
|
| |
window-decoration/insets size
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- FullScreen
- lock parent window if child
- X11: more sophisticated EWMH FS usage
- X11: set window 'Above' before FS and at focus
- allow window WM default position at window creation
- default position { -1, -1 } as hint to native WM
to gather a suitable default position
- wait until user-pos or WM-pos reached
- reconfigureWindow*
- allow -1 values for pos/size to mark no-change
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
changed ; Enhanced tests; Catch NV/XRANDR/GL bug
X11 fixes
- X11Screen properly uses it's display's connection
decorated in lock/unlock (for ScreenMode etc)
Ensure Screen's size is set if screenMode changed
- ScreenImpl's ScreenModeListener updates it's screen size
so 'external' changes will be detected.
Enhanced tests
- Verify more data rel. ScreenMode
Catch NV/XRANDR/GL bug
- Read TestScreenMode01NEWT/TestScreenMode01bNEWT comments
|
|
|
|
|
|
|
|
|
|
|
|
| |
- don't assume size/pos change - hence don't set window's values, but wait for satisfaction
- don't send resize events on our own, just rely on the event mechanism
- fullscreen: don't wrap action around invisibility from Java, Win7 flashes otherwise.
Clients who benefit from it (X11) impl. it natively.
- fullscreen exit: validate pos/size in case of a child window, like reparenting.
Otherwise the container might gets confused (eg. AWT).
|
|
|
|
| |
no reason to supress this information, it's sent only once (like X11)
|
|
|
|
| |
redundant DBG_PRINT on _NET_FRAME_EXTENTS
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
WindowImpl::reconfigureWindowImpl
Implement proper Inset usage (window decoration size)
- Insets are either polled (updateInsets()) or event driven (insetsChanged())
- Insets are used for size/pos calculations from Java side
- Natural size/pos in NEWT is client-area, ie w/o Insets
- Adding setTopLevelPosition()/setTopLevelSize() for top-level values,
ie including insets
WindowImpl::reconfigureWindowImpl
- Use flags to pass down the requested action to the native implementation
- Impl. all native actions: visible, decoration, reparent, resize, fullscreen
- Always use size/pos in client-area space, impl. shall use Insets to tranform them
- Remove double-setting of (reparent/fullscreen), which where introduced due to buggy impl. code
- Fix return from fullscreen position: Was overwritten with FS position (0/0)
- Fix decoration change: Remove visible toggle - not required, and actually disturbing
X11Windows/WindowsWindow: Added/Fixed Insets impl.
Tests (manual):
- TestSharedContextVBOES2NEWT utilizies proper window layout using Insets
- TestParenting03bAWT uses window layout for reparenting
|
|
|
|
| |
write access
|
|
|
|
|
|
|
| |
-> 0x0
Windows: setPosition was invoking setSize (new size propagation) even w/ nop size,
let WM event wmSize invoke setSize.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
NativeCode:
"XChangeProperty( dpy, w, _NET_WM_STATE, XA_ATOM, 32, PropModeReplace, [ FULLSCREEN, ABOVE ] .."
shall only be issued if entering FS, otherwise [ FULLSCREEN, ABOVE ] will be set on the client
side at leaving FS, which is not intended.
This caused a size reset to fullscreen size/mode.
WindowImpl:
Used WindowImpl.this.width and WindowImpl.this.height, where w and h should be used,
the FS and pre-FS values.
|
|
|
|
|
|
|
|
|
|
| |
Use EDT X11 Display connection (DPY) for whole display/screen/window lifecycle,
but the user utilization (OpenGL, ..).
Only using the same DPY for creation and event dispatching allows catching
WM_DELETE_ATOM 'ClientMessage's.
Sync X11Window.c w/ commit 4dbb8731219212e27c9afb769a1c62b32bd230a6
- remove 'test' code .. use orig lines
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
DisplayRelease0; Using 'EDT' suffix for display arguments
CloseDisplay in same order as creation (ATI)
- This enhanced the erroneous bug 515 (b54497155815852744adb657816cb4057948dae2) situation
with closing the display connections. However, some SIGSEGV still slipped through.
Adding DisplayRelease0
- Intended for cleaning up resources. Currently a NOP.
Using 'EDT' suffix for display arguments
- To mark the semantics of the display connection, which may be for window or EDT now.
|
|
|
|
|
|
|
|
|
|
|
|
| |
9ed513e9a9616f6028084df4c650c8caf31ea49d
In case of exessive destroy/create (the NEWT reparenting test cases),
some dpyEDT events are slipping through the event dispatcher.
This fix uses issues more XSync on both Display connection in case of 'requestFocus'
and 'closeWindow'.
'requestFocus' also uses the dpyEDT to issue the XSetInputFocus(..), since it's EDT related.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
libxcb 1.7 bug 20708
See https://jogamp.org/bugzilla/show_bug.cgi?id=502
Since the libX11/xcb code doesn't seem to be fixed anytime soon
a better usable workaround is required than using a system property
to enable 'over locking'.
It turns out that the race condition is related to the parallel
X11 Display connection usage of GLX/OpenGL and event dispatching.
This workaround utilizes 2 X11 Display handles, one for windowing/OpenGL
and one for event dispatching.
This approach allows us to cont. multithreading use w/o locking the display
and works on both implementations, the old bug-free libX11 and the 'new' buggy one.
Downside is the little resource overhead of the 2nd X11 Display connection .. well.
- Removes the property: 'nativewindow.x11.mt-bug'
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
separation ; android cleanup
remaining all-in-one jnlp's / jars:
jogl-all-awt.jnlp -> jogl.all.jar
jogl-all-noawt.jnlp -> jogl.all-noawt.jar
jogl-all-mobile.jnlp -> jogl.all-mobile.jar
native for all above: jogl-all-natives-linux-amd64.jar
jogl.all-android.apk jogl.all-android.jar
more may follow for each supported platfrom
++++
- newt: proper 'driver' separation
- all drivers reside now in jogamp.newt.driver.*
- remove intptr.cfg / use gluegen's
|
|
|
|
|
|
| |
dependency, not OpenGL
The native impl packages utilize EGL, not opengl, to query the proper pixelformat.
|
|
|
|
|
|
|
|
|
| |
trace), cleanup
- X11 Error Handler: if throwing JVM stack trace or fatal JVM error
query proper JNIEnv for running thread and attach thread to JVM if necessary.
- NEWT/X11: Proper XEvent polling documentation, cleanup window creation event mask
|
| |
|
|
|
|
|
|
|
| |
.. (remove local copy)
- Use them for gluegen code generation
- Use them for native compilation (cc)
|
|
|
|
| |
after Java callback
|
| |
|
|
|
|
| |
modifier properly.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add WindowListener.windowDestroyed()
To expose a proper window lifecycle, ie destroy-notify and destroyed,
this notification is added.
This will be used at least in unit tests, where we verify destruction.
Remove WindowImpl.windowDestroyed():
This native hook (planned to be called by native destroy notification)
is unreliable or not supported for all platforms.
NEWT relies on the pre destroy native hooks and handles the final
destroy notification itself.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This solves the issue when an applet is started/stop and started again,
or another applet runs in the same JVM.
Also soves the issue for multiple JVMs.
RegisteredClassFactory can be instanced to manage one shared window class,
currently in use for GDI's dummy window and NEWT.
A class base name and a window proc handle must be passed in the factory cstr.
Before registering, the class is tested if already exists,
eg another applet in the same JVM.
If registration fails, the class name will iterate until successful or MAX_INT reached,
eg if multiple JVMs are running.
Added NativeWindow Common Native Code.
|
| |
|
| |
|
| |
|
|
|
|
| |
NativeWindow X11.
|
|
|
|
| |
NEWT WindowsWindow GetRelativeLocation() native implementation to GDI as well.
|
|
|
|
|
|
|
|
| |
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.
|
|
|
|
|
|
|
|
|
|
| |
Lifecycle.reparentActionPre()/reparentActionPost() -> pauseRenderingAction()/resumeRenderingAction()
for a more generic use, ie reparenting and screen mode change.
ScreenMode change: No more visibility/fullscreen changes, no more locking,
just pause/resume animation.
X11 ScreenMode set: move from thread/wait to simple polling over time (timeout)
|
|
|
|
|
|
|
|
|
|
| |
DEBUG strings w/ thread name
nativewindow.TraceLock -> nativewindow.debug.ToolkitLock.TraceLock
Sync Xmisc (DummyWindow) with NEWT's creation
test scripts: awt and non-awt usage
|
| |
|
| |
|