Now that we have the display's context to hook into, we can use it to
construct other GL contexts and don't need a GdkSurface vfunc anymore.
This has the added benefit that backends can have different GdkGLContext
classes on the display and get new GLContexts generated from them, so
we get multiple GL backend support per GDK backend for free.
I originally wanted to make this a vfunc on GdkGLContextClass, but
it turns out all the abckends would just call g_object_new() anyway.
Instead of
Display::make_gl_context_current()
we now have
GLContext::clear_current()
GLContext::make_current()
This fits better with the backends (we can actually implement
clearCurrent on macOS now) and makes it easier to implement different GL
backends for backends (like EGL/GLX on X11).
We also pass a surfaceless boolean to make_current() so the calling code
can decide if a surface needs to be bound or not, because the backends
were all doing whatever, which was very counterproductive.
GLES doesn't support the GL_BGRA + GL_UNSIGNED_INT_24_8 hack that
we use on desktop OpenGL to upload textures directly in the cairo
pixel format. This adds the required conversions to all the places
that currently need it.
We also add a data_format to the internal gdk_gl_context_upload_texture()
function to make it clearer what the format are. Currently it is always
the cairo image surface format, but eventually we want to support other
formats so that we can avoid some of the unnecessary conversions we do.
Also, the current gdk_gl_context_upload_texture() code always converts
to a cairo format and uploads that like we did before. Later commits
will allow this to use other upload formats that gl supports to avoid
conversions.
gdk_gl_context_has_framebuffer_blit() and gdk_gl_context_has_frame_terminator()
were only used by by GDK/Win32, and they do not provide performance advantages
in GTK master, so clean up the code a bit by dropping them.
We need to use GL_BGRA instead of GL_RGBA when doing glReadPixels() on
EGL on Windows (ANGLE) so that the red and blue bits won't be displayed
inverted.
Also fix the logic where we determine whether to bit blit or redraw
everything.
That way, we can store the right region there: The actual painted area
instead of the exposed area (which is way too small).
Also, the GL context is the only user of this data, so storing it there
seems way smarter.
This is a way to query the damaged area of the backbuffer.
The GL renderer uses this to compute the extents of that damage region
(computed via buffer age) and use them to minimize the area to redraw.
This changes the semantics of GL rendering to "When calling
gdk_window_begin_frame() with a GL context, the area by
gdk_gl_context_get_damage() needs to be redrawn and every other pixel of
the backbuffer is guaranteed to be correct.
After gdk_window_end_frame() on a GL-drawn window, the whole backbuffer
must be correct.
We can always glXBufferSwap() now because of this.
... instead of a gl context.
This requires some refactoring in the way we mark the shared context as
drawing: We now call begin_frame/end_frame() on it and ignore the call
on the main context.
Unfortunately we need to do this check in all vfuncs, which sucks. But I
haven't found a better way.
This way, we can query the GL context's state via
gdk_gl_context_is_drawing().
Use this function to make GL contexts as attached and grant them access
to the front/backbuffer for rendering.
All of this is still unused because GL drawing is still disabled.
No visible changes as GL rendering is disabled at the moment.
What was done:
1. Move window->invalidate_for_new_frame to glcontext->begin_frame
This moves the code to where it is used (the GLContext) and prepares it
for being called where it is used when actually beginning to draw the
frame.
2. Get rid of buffer-age usage
We want to let the application render directly to the backbuffer.
Because of that, we cannot make any assumptions about the contents the
application renders outside the clip area.
In particular GskGLRenderer renders random stuff there but not actual
contents.
3. Pass the actual GL context
Previously, we passed the shared context to end_frame, now we pass the
actual GL context that the application uses for rendering. This is so
that the vfuncs could prepare the actual contexts for rendering (they
don't currently).
4. Simplify the code
The previous code set up the final drawing method in begin_frame.
Instead, we now just ensure the clip area is something we can render
and decide on the actual method in end_frame.
This is both more robust (we can change the clip area in between if we
want to) and less code.
This allows us to decide when the R and B color channels should be
flipped with a much better granularity.
For instance, when using GLX_EXT_texture_from_pixmap to create a GL
texture from a surface we don't need to swap the R and B channels, as
the internal representation of the texture data will already have the
appropriate colors.
We also don't need to flip color channels when blitting from a texture.
On some platforms we can ask the GL context machinery to create a GLES
context, instead of a GL one.
In order to ask for a GLES context at GdkGLContext realization time, we
use a bit field like we do for forward compatible, or debug contexts.
The 'use-es' bit also changes the way we select a default version,
because OpenGL and OpenGLES versions differ.
https://bugzilla.gnome.org/show_bug.cgi?id=743746
We want to have the ability to fall back to legacy GL contexts when
creating them. In order to do so, we need to store the legacy bit on the
GdkGLContext, as well as being able to query it.
Setting the legacy bit from outside GDK is not possible; we cannot
create GL contexts in 3.2 core profile *and* compatibility modes at the
same time, and if we allowed users to select the legacy mode themselves,
it would break the creation of the GdkWindow's paint GL context.
What we do allow is falling back to legacy GL context if the platform
does not support 3.2 core profiles — for instance, on older GPUs or
inside virtualized environments.
We are also going to use the legacy bit internally, to choose which GL
API we can use when drawing GL content.
https://bugzilla.gnome.org/show_bug.cgi?id=756142
Users of the GdkGLContext API should be allowed to set properties on the
shim GdkGLContext instance prior to realization, so that the
backend-specific implementation can use the value of those properties
when creating the windowing system specific resources.
The main three options are:
• a major/minor version tuple, to request a specific GL version
• a debug bit, to request a "debug context", which provides additional
validation and run time checking
• a forward compatibility bit, to request a context that does not
have deprecated functionality
See also:
- https://www.opengl.org/registry/specs/ARB/glx_create_context.txthttps://bugzilla.gnome.org/show_bug.cgi?id=741946
One of the major requests by OpenGL users has been the ability to
specify settings when creating a GL context, like the version to use
or whether the debug support should be enabled.
We have a couple of requirements in terms of API:
• avoid, if at all possible, the "C arrays of integers with
attribute, value pairs", which are hard to write and hard
to bind in non-C languages.
• allow failing in a recoverable way.
• do not make the GL context creation API a mess of arguments.
Looking at prior art, it seems that a common pattern is to split the
construction phase in two:
• a first phase that creates a GL context wrapper object and
does preliminary checks on the environment.
• a second phase that creates the backend-specific GL object.
We adopted a similar pattern:
• gdk_window_create_gl_context() creates a GdkGLContext
• gdk_gl_context_realize() creates the underlying resources
Calling gdk_gl_context_make_current() also realizes the context, so
simple GL users do not need to care. Advanced users will want to
call gdk_window_create_gl_context(), set up the optional requirements,
and then call gdk_gl_context_realize(). If either of these two steps
fails, it's possible to recover by changing the requirements, or simply
creating a new GdkGLContext instance.
https://bugzilla.gnome.org/show_bug.cgi?id=741946
As the alignments, strides and image formats may be different across
platforms, make the texture upload a vfunc to allow backends to override
the GL commands for uploading textures for the software implementation for
gdk_gl_texture_from_surface(), if necessary.
Suggested by Alex to avoid copying non-trivial portions of code which would
then add maintainenace burden.
https://bugzilla.gnome.org/show_bug.cgi?id=740795
If buffer age is undefined and the updated area is not the whole
window then we use bit-blits instead of swap-buffers to end the
frame.
This allows us to not repaint the entire window unnecessarily if
buffer_age is not supported, like e.g. with DRI2.
To properly support multithreaded use we use a global GPrivate
to track the current context. Since we also don't need to track
the current context on the display we move gdk_display_destroy_gl_context
to GdkGLContext::discard.
This adds the new type GdkGLContext that wraps an OpenGL context for a
particular native window. It also adds support for the gdk paint
machinery to use OpenGL to draw everything. As soon as anyone creates
a GL context for a native window we create a "paint context" for that
GdkWindow and switch to using GL for painting it.
This commit contains only an implementation for X11 (using GLX).
The way painting works is that all client gl contexts draw into
offscreen buffers rather than directly to the back buffer, and the
way something gets onto the window is by using gdk_cairo_draw_from_gl()
to draw part of that buffer onto the draw cairo context.
As a fallback (if we're doing redirected drawing or some effect like a
cairo_push_group()) we read back the gl buffer into memory and composite
using cairo. This means that GL rendering works in all cases, including
rendering to a PDF. However, this is not particularly fast.
In the *typical* case, where we're drawing directly to the window in
the regular paint loop we hit the fast path. The fast path uses opengl
to draw the buffer to the window back buffer, either by blitting or
texturing. Then we track the region that was drawn, and when the draw
ends we paint the normal cairo surface to the window (using
texture-from-pixmap in the X11 case, or texture from cairo image
otherwise) in the regions where there is no gl painted.
There are some complexities wrt layering of gl and cairo areas though:
* We track via gdk_window_mark_paint_from_clip() whenever gtk is
painting over a region we previously rendered with opengl
(flushed_region). This area (needs_blend_region) is blended
rather than copied at the end of the frame.
* If we're drawing a gl texture with alpha we first copy the current
cairo_surface inside the target region to the back buffer before
we blend over it.
These two operations allow us full stacking of transparent gl and cairo
regions.
