Dependencies are added automatically, and we no longer get warnings
like:
clutter/clutter/meson.build:628: DEPRECATION: Library mutter-clutter-4
was passed to the "libraries" keyword argument of a previous call to
generate() method instead of first positional argument. Adding
mutter-clutter-4 to "Requires" field, but this is a deprecated behaviour
that will change in a future version of Meson. Please report the issue
if this warning cannot be avoided in your case.
The texture tower can return no texture e.g. if the calculated level is
negative. This was handled before, but regressed with
e1370ee209. This fixes a potential crash
observed occasionally when starting Firefox nightly using the Wayland
backend in overview mode.
If a KMS device has the DRM_CAP_DUMB_PREFER_SHADOW and a software based
GL driver is used, always use a shadow fb. This will speed up read backs
in the llvmpipe OpenGL implementation, making blend operations faster.
Fixes: https://gitlab.gnome.org/GNOME/mutter/issues/106
Now that everything is settled, from the initialization
process to the subclasses to moving code to the compositor,
MetaWindowActor can be a proper abstract class that cannot
be instantiated.
Thus, make MetaWindowActor an abstract class.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/368
This vfunc was added as a was to work around the convoluted
initialization process. Now that we figured it out and moved
the MetaWindowActor-specific initialization to constructed(),
we can override that.
Remove post_init() and use GObject.constructed() entirely.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/368
MetaWindowActor breaks layering isolation by accessing
and injecting itself into compositor->windows. This is
a bad practice, and effecticely makes returning the
new actor useless, since we doesn't even use the return
value.
Move window actor creation to under MetaCompositor and
stop violating (too badly) the resposabilities of each
component. This moves meta_window_actor_new() into
meta_compositor_add_window().
Also, move the remaining initialization code to the
GObject.constructed vfunc.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/368
Document the roles of MetaSurfaceActor and MetaWindowActor,
and when their subclasses are used.
(And this is actually the first real documentation under
src/compositor/README!)
https://gitlab.gnome.org/GNOME/mutter/merge_requests/368
MetaWindowActor handles sending _NET_WM_FRAME_* X atoms to
clients - even pure Wayland clients.
Now that we have Wayland- and X11-specific implementations of
MetaWindowActor, we can delegate this to MetaWindowActorX11,
and allow pure Wayland apps to not even connect to
MetaSurfaceActor:repaint-scheduled.
Do that by moving all the X11-specific code to the X11-specific
MetaWindowActorX11 class. Add vfuncs to MetaWindowActorClass
that are necessary for the move, namely:
* pre_paint() and post_paint()
* post_init()
* frame_complete()
* set_surface_actor()
* queue_frame_drawn()
https://gitlab.gnome.org/GNOME/mutter/merge_requests/368
Those are stub specialized classes for MetaWindowActor. This will
help ensuring that we do not execute X11-specific code paths on
pure Wayland clients.
The relationship between the window actor and the surface is the
following:
* Wayland: MetaWindowActorWayland + MetaSurfaceActorWayland
* X11: MetaWindowActorX11 + MetaSurfaceActorX11
* Xwayland: MetaWindowActorX11 + MetaSurfaceActorWayland
It is not possible to have MetaWindowActorWayland backed by a
MetaSurfaceActorX11 surface.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/368
We will introduce specialized MetaWindowActors for X11
and Wayland in the future, so it needs to be derivable.
Make it a derivable class, and introduce a private field.
The MetaWindowActorClass definition is in the private
header in order to prevent external consumers of Mutter
to create MetaWindowActor implementations of their own.
That is, MetaWindowActor is only internally derivable.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/368
EGLStream textures are imported as GL_TEXTURE_EXTERNAL_OES and reading
pixels directly from them is not supported. To make it possible to get
pixels, create an offscreen framebuffer and paint the actor to it, then
read pixels from the framebuffer instead of the texture directly.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/362
When a texture is transformed in any way (e.g. Wayland buffer
transforms), we cannot just fetch the pixels from the texture directly
and be done with it, as that will result in getting the untransformed
pixels.
To properly get the pixels in their right form, first draw to an
offscreen framebuffer, using the same method as when painting on the
stage, then read from the framebuffer into a cairo image surface.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/362
Fixes: https://gitlab.gnome.org/GNOME/mutter/issues/408
While for normal textures, GL_TEXTURE_2D should be used, when it's an external
texture, binding it using GL_TEXTURE_2D results in an error.
Reading the specification for GL_TEXTURE_EXTERNAL_OES it is unclear whether
getting pixel data from a texture is possible, and tests show it doesn't result
in any data, but in case it would eventually start working, at least bind the
correct target for now.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/362
Don't just set the internal format to the dummy format "any", as that causes
code intended to be unreachable code to be reached. It's not possible to
actually know the internal format of an external texture, however, so it might
not actually correspond to the real format.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/362
DRM_EVENT_CONTEXT_VERSION is the latest context version supported by
whatever version of libdrm is present. Mutter was blindly asserting it
supported whatever version that may be, even if it actually didn't.
With libdrm 2.4.78, setting a higher context version than 2 will attempt
to call the page_flip_handler2 vfunc if it was non-NULL, which being a
random chunk of stack memory, it might well have been.
Set the version as 2, which should be bumped only with the appropriate
version checks.
https://bugzilla.gnome.org/show_bug.cgi?id=781034
This makes the build less verbose, as all .gir generation except for
clutters didn't pass --quiet to g-ir-scanner, making it output long
linking commands. Do this by adding a common introspection_args
variable.
While at it, put -U_GNU_SOURCE in there too, as it was always passed
everywhere as without it the scanner would log warnings.
This is the last remaining feature necessary to achieve
parity with the Autotools build.
A few changes were made to the install locations of the
tests, in order to better acomodate them in Meson:
* Tests are now installed under a versioned folder (e.g.
/usr/share/installed-tests/mutter-4)
* The mutter-cogl.test file is now generated from an .in
file, instead of a series of $(echo)s from within Makefile.
Notice that those tests need very controlled environments
to run correctly. Mutter installed tests, for example, will
failed when running under a regular session due to D-Bus
failing to acquire the ScreenCast and/or RemoteScreen names.
When running installed tests, the working directory for Cogl
tests is /usr/libexec/installed-tests/mutter-cogl-4/conform,
which isn't writable by normal users.
To avoid the adding stray hidden files to the current directory,
adapt the runner script to fallback to $(mktemp) - which is
available on all platform we care about - and avoid adding
hidden files everywhere.
Presumably glReadPixels itself can be more performant with pixel format
conversions than doing a fix-up conversion on the CPU afterwards. Hence,
pick required_format based on the destination rather than the source, so
that it has a better chance to avoid the fix-up conversion.
With CoglOnscreen objects, CoglFramebuffer::internal_format (the source
format) is also wrong. It is left to a default value and never set to
reflect the reality. In other words, read-pixels had an arbitrary
intermediate pixel format that was used in glReadPixels and then fix-up
conversion made it work for the destination.
The render buffers (GBM surface) are allocated as DRM_FORMAT_XRGB8888.
If the destination buffer is allocated as the same format, the Cogl
read-pixels first converts with glReadPixels XRGB -> ABGR because of the
above default format, and then the fix-up conversion does ABGR -> XRGB.
This case was observed with DisplayLink outputs, where the native
renderer must use the CPU copy path to fill the "secondary GPU"
framebuffers.
This patch stops using internal_format and uses the desired destination
format instead.
_cogl_framebuffer_gl_read_pixels_into_bitmap() will still use
internal_format to determine alpha premultiplication state and multiply
or un-multiply as needed. Luckily all the formats involved in the
DisplayLink use case are always _PRE and so is the default
internal_format too, so things work in practise.
Furthermore, the GL texture_swizzle extension can never apply to
glReadPixels. Not even with FBOs, as found in this discussion:
https://gitlab.gnome.org/GNOME/mutter/issues/72
Therefore the target_format argument is hardcoded to something that can
never match anything, which will prevent the swizzle from being assumed.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/313
This function gets hit even today on relatively modern Intel systems (I
have a Haswell Desktop with Mesa 18.2.4) if the pixel format is right.
Presumably it makes things slower for no longer a reason.
According to cb146dc515, this
functionality was refactored into a workaround path in 2012. The commit
message mentions the problem existing before Mesa 8.0.2. The number
refers to https://bugs.freedesktop.org/show_bug.cgi?id=46631 .
The use case where I hit this is when improving support for DisplayLink
video outputs. These are used through a "secondary GPU", and since
DisplayLink does not have a GPU, Mutter uses the CPU copy path with Cogl
read-pixels[1]. If the DisplayLink framebuffer was allocated as
DRM_FORMAT_XRGB8888 (the only format it currently handles correctly),
mesa_46631_slow_read_pixels_workaround would get hit. The render buffer is
the same format as the framebuffer, yet doing the copy XRGB -> XRGB ends
up being slower than XRGB -> XBGR which makes no sense.
This patch is not sufficient to fix the XRGB -> XRGB copy performance,
but it is required.
This patch reverts CoglGpuInfoDriverBug into what it was before
cb146dc515.
[1] This is not actually true until
https://gitlab.gnome.org/GNOME/mutter/merge_requests/278 is
merged.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/313
The actor-shader-effect test actors are 50px wide, but we check the 51st
pixel. This went along undetected until "clutter: Avoid rounding
compensation when invalidating 2D actors" because the paint volumes were
made slightly bigger and the shaders paint all over them (I guess nobody
noticed those actors being actually ~52px wide).
Update the test to check the middle of the opposite edge, so we keep neatly
rounded numbers.
The test does a clutter_actor_set_scale_full() call that only updates
the scale center (i.e. no changes to scale-x/y), but expects to receive
notifications of actor scale changes.
Since "Revert "Revert "ClutterActor: Optimize away idempotent
scale/position updates"" these are optimized away, so just drop the
assumption.
The depth buffer is marked as invalid when 1) the framebuffer is just created,
and 2) whenever GL_DEPTH_TEST is enabled on it. This will ensure the
framebuffers attached depth buffer (if any) is properly cleared before it's
actually used, while saving needless clears while depth testing is disabled
(the default).
https://bugzilla.gnome.org/show_bug.cgi?id=782344
