First, inactivate shared gl contexts known by each sink pad.
Then, destroy the gl context known by the glmixer.
Finally, re-activate shared gl contexts.
This is to satisfy the fact that no shared gl context must be current
when an opengl context is destroyed.
Moreover the application may hang or crash without those steps.
Before, the window size was given at its creation. Now, it's done at
the drawing step because it's only relevant when there is a glimagesink
element in the pipeline.
glmixer can be seen as a glfilter except it handles N requested
sink pads.
Each sink pad and the src pad are video/x-raw-gl.
glmixer is responsible for managing different framerates from inputs.
It uses OpenGL context sharing. It means that each input is in its
own OpenGL context shared together and shared with the OpenGL context
of the ouput gl chain.
Also add a glmosaic which is an example of implementation of glmixer.
For now glmosaic is a cube but it will be fixed in the next commits.
For now the glmixer has some weird behaviours in some configurations
but it will be improved in the next commits.
The autotools builds is temporarly broken since those changes
have been made on win32.
Before, a gstgldisplay was instancied by the gl src in terms of gl chain.
And then the next element got it through the first gstglbuffer.
Now, this is done though queries.
All glelements get their ref on a gstgldisplay in READY state.
This rewrite is mainly a first step to be able to share OpenGL context hold
by the gstgldisplay using more complex glelements.
For example, with a glvideomixer. The associated gstgldisplay of each gl chain
of the sink pads will share their OpenGL context.
A texture is not destroyed when when we are done with it.
This texture is just added to the texture pool in order to be
re-used. In this case no OpenGL code is executed so we do not need to
request gl thread.
Add a pkg-config check for opengl and if not found assume opengl-es. If user has
none of both one still get build error later on (there is no pkg-config for
opengl-es).
Add more files to EXTRA dist and build the opengles variant if selected.
Simmilar changes could be done for the winCE backend.
Default implementation of NSOpenglView::update is not safe because it
just calls update on the opengl context whereas we are not in the gl thread.
Also fix the white flickering when resizing, because now we need to call
the draw callback manually when resizing.
- All gstglwindow members are now modified only in the gl thread
to avoid thread concurrency
- OpenGL context is now properly clean
- fix a couple of things in implementation of xoverlay interface
It works with both gst-launch and a cocoa app (non-embedded and embedded)
But there is still some problems:
- sometimes crash when closing
- flickering when resizing
- embedded mode not perfect
I will first make the CMake build work with cocoa backend
in order to generate a XCode project.
Then it should be easier to fix those issues.
Fixes bug #584877
Before this commit calling "gst_x_overlay_set_xwindow_id" more
than one time, had no effect.
It mainly affects the glimagesink implementation.
But on win32 (and CE), some stuff has to be done to
release the old parent.
And add a switchxoverlay example where the user
can click on left/right part of the main window to
switch the xoverlay.
The external opengl context must be specify when creating
our OpenGL context (glx) or just after (wgl).
When calling glXCreateContext or wglShareLists, the
external opengl context must not be current.
Then our gl context can be current in the gl thread while
the external gl context is current in an other thread.
See tests/examples/clutter/cluttershare.c
In OpenGL 2.x for Embedded System, a lot of basic scene/draw functions
have been removed. It means that everything is made using vertex and
fragment shaders.
I have also added a gstglwindow backend for winCE that uses EGL
(Native Platform Graphics Intercace) (which is a full part of
OpenGL ES specification). It remove the use of wgl/glx functions.
Althought the XEvent's xclient.data.l array is an array of
longs they will be constrained to 32 bit by the X11 protocol.
On 64 bit architectures use two elements of the array to store
one pointer.
This fixes segfaults that happen at least for every example
on startup.
This reverts commit 96e4ab18c2cf9876f6c031b9aba6282d0bd45a93.
You should have asked first. And you would have been told "no",
because it causes people on development branches to do a huge
amount of extra work.
Althought the XEvent's xclient.data.l array is an array of
longs they will be constrained to 32 bit by the X11 protocol.
On 64 bit architectures use two elements of the array to store
one pointer.
This fixes segfaults that happen at least for every example
on startup.
