The timestamp comes from the GSource, meaning it's a more accurate
representation of when the frame started to be dispatched compared to
getting the current time in any callback.
Currently unused.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1285
In certain scenarios, the frame clock needs to handle present feedback
long before the assumed presentation time happens. To avoid scheduling
the next frame to soon, avoid scheduling one if we were presented half a
frame interval within the last expected presentation time.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1285
This adds a current unused, apart from tests, frame clock. It just
reschedules given a refresh rate, based on presentation time feedback.
The aiming for it is to be used with a single frame listener (stage
views) that will notify when a frame is presented. It does not aim to
handle multiple frame listeners, instead, it's assumed that different
frame listeners will use their own frame clocks.
Also add a test that verifies that the basic functionality works.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1285
When a transition is created for the allocation change, it will delay
the new allocation box getting set depending on transition details.
This, however, means that e.g. the 'needs_allocation' flag never gets
cleared if a transition is created, causing other parts of the code to
get confused thinking it didn't pass through a layout step before paint.
Fix this by calling clutter_actor_allocate_internal() with the current
allocation box if a transition was created, so that we'll properly clear
'needs_allocation' flag.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1345
41130b08eb added a fix for culling subsurfaces with geometry scale.
Unfortunately it only did so for the opaque regions, not for clip and
unobscured regions, as the effect was hidden by bug that was only
fixed by 3187fe8ebc.
Apply the same fix to clip and unobscured regions and use the chance
to move most of the slightly hackish geometry scale related code
into a single place.
We need to scale slightly differently in the two cases, indicated by
the new `ScalePerspectiveType` enum, as the scale is dependent on the
perspective - once from outside, once from inside of the scaled actor.
Closes https://gitlab.gnome.org/GNOME/mutter/-/issues/1312
Since we now always return a resource scale, we can remove the boolean
return value from clutter_actor_get_resource_scale() and
_clutter_actor_get_real_resource_scale(), and instead simply return the
scale.
While at it, also remove the underscore from the
_clutter_actor_get_real_resource_scale() private API.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1276
Add private API to ClutterBackend to set a fallback resource scale
available to Clutter. This API will be used for "guessing" the
resource-scale of ClutterActors in case the actor is not attached to a
stage or not properly positioned yet.
We set this value from inside mutters MetaRenderer while creating new
stage-views for each logical monitor. This makes it possible to set the
fallback scale to the scale of the primary monitor, which is the monitor
where most ClutterActors are going to be positioned.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1276
The portal API requires a screencast session only for absolution motion
with remote desktop, other methods including relative motion do not
require a screencast session.
There is no reason to be more strict than the API actually is, check for
a screencast session only when required, like for absolute motion events
and touch events.
Tested with https://gitlab.gnome.org/snippets/1122
Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1307
There are a couple of places in gnome-shell where we aren't interested
in which workspace is active, but whether a given workspace is active.
Of course it's easy to use the former to determine the latter, but we
can offer a convenience property on the workspace itself almost for
free, so let's do that.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1336
These tests were written (and copy-pasted) before ClutterActor
had an actual background-color property. As a preparation to
the removal of ClutterRectangle, replace all these rectangles
with plain actors and background colors.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1332
The property is deprecated and the current implementation simply
redirects it to ClutterActor::background-color, so remove it.
Also update the tests to set the background color directly.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1332
It is deprecated in favor of the 'z-position' property, and
the implementation itself redirects to the z-position, so
just drop it and replace all get|set_depth calls to their
z-position counterparts.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1332
We were setting the pipeline colour to all white (1.0, 1.0, 1.0, 1.0)
and so the default layer combine function multiplied each pixel
(R, G, B, A) by all ones. Obviously multiplying by one four times per
pixel is a waste of effort so we remove the colour setting *and* set
the layer combine function to a trivial shader that will ignore whatever
the current pipeline colour is set to. So now we do **zero** multiplies
per pixel.
On an i7-7700 at UHD 3840x2160 this results in 5% faster render times
and 10% lower power usage (says intel_gpu_top). The benefit is probably
much higher for virtual machines though, as they're no longer being
asked to do CPU-based math on every pixel of a window.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1331
The previous commit removed checks for intermediate focus states which
would make tests randomly fail, because of their time dependence. What
can be tested however is that if there is no other window available that
would accept the focus, that the focus remains at 'none', after the
focused window has been closed. This newly introduced test checks the
focus directly after closing the window (and syncing) and after the time
it would have taken for the queue to finish. The first check has a
similar timing issue as the removed focus checks in the other tests, but
the test will never accidentally fail, because regardless of whether the
queue has finished or not, the focus is always expected to be 'none'.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1329
While c3d13203 ensured that the test-client has actually closed the
window before testing for the focus change, it also made another timing
related issue with the tests more likely to happen. Serveral tests
assert that the focus is set to 'none' after the focussed window has
been closed when the window below does not accept focus. This however
can never be reliably tested, because closing the window triggers
timeout based iteration of a queue of default focus candidate windows.
This starts after the window has been closed and might finish before the
clients have finished synchronizing. This issue is more likely to
trigger the shorter the queue is and the more test clients there are
that could delay the synchronization.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1329
This avoids some issues which could happen on some setups[0] due to
meta-native-renderer.c:dummy_power_save_page_flip →
meta_onscreen_native_swap_drm_fb implicitly turning of the primary
plane (by destroying the KMS framebuffer assigned to it):
* drmModeObjectSetProperty could return an "Invalid argument" error
between setting a non-empty cursor with drmModeSetCursor(2) and
enabling the primary plane again:
Failed to DPMS: Failed to set connector 69 property 2: Invalid argument
(This was harmless other than the error message, as we always re-set
a mode on the CRTC after setting the DPMS property to on, which
enables the primary plane and implicitly sets the DRM property to on)
* drmModeSetCursor(2) could return an "Invalid argument" error between
setting the DPMS property to on and enabling the primary plane again:
Failed to set hardware cursor (drmModeSetCursor failed: Invalid argument), using OpenGL from now on
[0] E.g. with the amdgpu DC display code.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1240
In commit 4c1fde9d MetaCullable related code was moved out of
MetaShapedTexture into MetaSurfaceActor. While generally desirable,
this removed drawing optimizations in MetaShapedTexture for partial
redraws. The common case for fully obscured actors was still supposed
to work, but it was now discovered that it actually did not.
This commit revert parts of 4c1fde9d: it reintroduces clipping
to MetaShapedTexture but leaves all culling and actor related logic
in MetaSurfaceActor.
Thanks to Daniel van Vugt for uncovering the issue.
Fixes https://gitlab.gnome.org/GNOME/mutter/-/issues/850
Fixes https://gitlab.gnome.org/GNOME/mutter/-/issues/1295https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1326
When trying to find a default focus window, the code iterates through a
queue of candidates with a timeout between each candidate. If the window
the current timeout is waiting for gets destroyed, this process just
stops instead of trying the next window in the queue.
This issue was made more likely to be triggered with the previous change
to the closed-transient-no-input-parents-queued-default-focus-destroyed
test due to the introduction of a wait, which can introduce a
delay between the two destroy commands.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1325
Some tests were not waiting for the test client to actually issue
destroy commands before checking their effect on the window focus.
Similarly when mutter is supposed to change the focus based on a delay
by sending a WM_TAKE_FOCUS to the client, this also could fail without
synchronization with the client before checking the result.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1325
We delegate the answer through CoglDriverVtable::is_hardware_accelerated
since this is properly a property of the renderer, and not something the
cogl core should know about. The answer given for the nop driver is
admittedly arbitrary, yes it's infinitely fast but no there's not any
"hardware" making it so.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1194
Event delivery with grabbing popups stay essentially the same within
the grabbing client, we still must honor the implicit grab as long as
there is one.
This is however not the case, the popup_grab_focus() function ignores
the button state. The popup_grab_button() function will already indirectly
re-sync the focus surface after the last button is released, so checking
for button state in popup_grab_focus() is sufficient to make the implicit
grab honored with popup grabs involved.
Fixes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1275https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1270
When a touch sequence was rejected, we'd update the event timestamps of
incoming touch events to help with implementing grabs. This was done by
sending a ClientMessage with a counter, and comparing the counter to
decide whether we're seing a replayed event or not.
This had the unforseen consequence that we would potentially end up
destroying all actors including the stage, since, when mutter receives a
ClientMessage event, it would assume that it's a WM_PROTOCOLS event, and
handle it as such. The problem with this approach is that it would
ignore fact that there might be other ClientMessage types sent to it,
for example the touch synchronization one. What could happen is that the
touch count value would match up with the value of the WM_DELETE_WINDOW
atom, clutter would treat this as WM_PROTOCOLS:WM_DELETE_WINDOW, which
it'd translate to clutter_actor_destroy(stage).
Destroying the stage in such a way is not expected, and caused wierd
crashes in different places depending on what was going on.
This commit make sure we only treat WM_PROTOCOLS client messages as
WM_PROTOCOLS client messages effectively avoiding the issue.
This fixes crashes such as:
#0 meta_window_get_buffer_rect (window=0x0, rect=rect@entry=0x7ffd7fc62e40) at core/window.c:4396
#1 0x00007f1e2634837f in get_top_visible_window_actor (compositor=0x297d700, compositor=0x297d700) at compositor/compositor.c:1059
#2 meta_compositor_sync_stack (compositor=0x297d700, stack=<optimized out>, stack@entry=0x26e3140) at compositor/compositor.c:1176
#3 0x00007f1e263757ac in meta_stack_tracker_sync_stack (tracker=0x297dbc0) at core/stack-tracker.c:871
#4 0x00007f1e26375899 in stack_tracker_sync_stack_later (data=<optimized out>) at core/stack-tracker.c:881
#5 0x00007f1e26376914 in run_repaint_laters (laters_list=0x7f1e2663b7d8 <laters+24>) at core/util.c:809
#6 run_all_repaint_laters (data=<optimized out>) at core/util.c:826
#7 0x00007f1e26b18325 in _clutter_run_repaint_functions (flags=flags@entry=CLUTTER_REPAINT_FLAGS_PRE_PAINT) at clutter-main.c:3448
#8 0x00007f1e26b18fc5 in master_clock_update_stages (master_clock=0x32d6a80, stages=0x4e5a740) at clutter-master-clock-default.c:437
#9 clutter_clock_dispatch (source=<optimized out>, callback=<optimized out>, user_data=<optimized out>) at clutter-master-clock-default.c:567
#10 0x00007f1e27e48049 in g_main_dispatch (context=0x225b8d0) at gmain.c:3175
#11 g_main_context_dispatch (context=context@entry=0x225b8d0) at gmain.c:3828
#12 0x00007f1e27e483a8 in g_main_context_iterate (context=0x225b8d0, block=block@entry=1, dispatch=dispatch@entry=1, self=<optimized out>) at gmain.c:3901
#13 0x00007f1e27e4867a in g_main_loop_run (loop=0x24e29f0) at gmain.c:4097
#14 0x00007f1e2636a3dc in meta_run () at core/main.c:666
#15 0x000000000040219c in main (argc=1, argv=0x7ffd7fc63238) at ../src/main.c:534
and
#0 0x00007f93943c1f25 in raise () at /usr/lib/libc.so.6
#1 0x00007f93943ab897 in abort () at /usr/lib/libc.so.6
#2 0x00007f9393e1e062 in g_assertion_message (domain=<optimized out>, file=<optimized out>, line=<optimized out>, func=0x7f93933e6860 <__func__.116322> "meta_x11_get_stage_window",
#3 0x00007f9393e4ab1d in g_assertion_message_expr ()
#4 0x00007f939338ecd7 in meta_x11_get_stage_window (stage=<optimized out>) at ../mutter/src/backends/x11/meta-stage-x11.c:923
#5 0x00007f939339e599 in meta_backend_x11_cm_translate_device_event (x11=<optimized out>, device_event=0x55bc8bcfd6b0) at ../mutter/src/backends/x11/cm/meta-backend-x11-cm.c:381
#6 0x00007f939339f2e2 in meta_backend_x11_translate_device_event (device_event=0x55bc8bcfd6b0, x11=0x55bc89dd5220) at ../mutter/src/backends/x11/meta-backend-x11.c:179
#7 0x00007f939339f2e2 in translate_device_event (device_event=0x55bc8bcfd6b0, x11=0x55bc89dd5220) at ../mutter/src/backends/x11/meta-backend-x11.c:208
#8 0x00007f939339f2e2 in maybe_spoof_event_as_stage_event (input_event=0x55bc8bcfd6b0, x11=0x55bc89dd5220) at ../mutter/src/backends/x11/meta-backend-x11.c:284
#9 0x00007f939339f2e2 in handle_input_event (event=0x7fff62d60490, x11=0x55bc89dd5220) at ../mutter/src/backends/x11/meta-backend-x11.c:309
#10 0x00007f939339f2e2 in handle_host_xevent (event=0x7fff62d60490, backend=0x55bc89dd5220) at ../mutter/src/backends/x11/meta-backend-x11.c:413
#11 0x00007f939339f2e2 in x_event_source_dispatch (source=<optimized out>, callback=<optimized out>, user_data=<optimized out>) at ../mutter/src/backends/x11/meta-backend-x11.c:467
#12 0x00007f9393e6c39e in g_main_dispatch (context=0x55bc89dd03e0) at ../glib/glib/gmain.c:3179
#13 0x00007f9393e6c39e in g_main_context_dispatch (context=context@entry=0x55bc89dd03e0) at ../glib/glib/gmain.c:3844
#14 0x00007f9393e6e1b1 in g_main_context_iterate (context=0x55bc89dd03e0, block=block@entry=1, dispatch=dispatch@entry=1, self=<optimized out>) at ../glib/glib/gmain.c:3917
#15 0x00007f9393e6f0c3 in g_main_loop_run (loop=0x55bc8a042640) at ../glib/glib/gmain.c:4111
#16 0x00007f9393369a0c in meta_run () at ../mutter/src/core/main.c:676
#17 0x000055bc880f2426 in main (argc=<optimized out>, argv=<optimized out>) at ../gnome-shell/src/main.c:552
Related: https://gitlab.gnome.org/GNOME/mutter/-/issues/338
Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/951https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1317
Make the clutter_input_device_get_actor() API public and remove
clutter_input_device_get_pointer_actor() in favour of the new function.
This allows also getting the "pointer" actor for a given touch sequence,
not only for real pointer input devices like mice.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1275
Add a method to ClutterSeat that allows peeking the list of input
devices and allow looping through devices a bit faster. The API left is
private so we can make use of peeking the GList internally, but don't
have to expose any details to the outside, which means we'd have to
eventually stick with a GList forever to avoid breaking API.
Since we now have the peek_devices() API internally, we can implement
ClutterSeats public list_devices() API using g_list_copy() on the list
returned by peek_devices().
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1275
As explained in the last commits, we'll let gnome-shell take care of
this since freezing and thawing needs to be decoupled from the effect
starting and ending.
So stop freezing the MetaWindowActor when starting the effect and
thawing the actor when ending the effect.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1250
As explained in the last commit, gnome-shell needs to be able to thaw
window actor updates during its size-change effect is active.
So make meta_window_actor_freeze() and meta_window_actor_thaw() public
API, which will allow the shell to freeze and thaw actor updates itself.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1250
The size-change animation in gnome-shell needs to sync the window actors
geometry during the animation, it currently does this by notifying the
compositor that the animation was finished before it actually is.
This causes a few bugs in Mutter though, since it will now emit the
"effects-completed" signal on the window actor even though they aren't
completed.
To fix that, we need to decouple freezing and thawing of actor updates
from window effects and allow gnome-shell to thaw actor updates before
it notifies Mutter that the effect is completed.
The first step for this is allowing to sync the actor geometry while an
effect is active, this should be redundant since effects which actually
need to inhibit those updates will freeze the actor anyway. Also a
geometry change happening while another effect is active will kill the
old effect anyway because MetaPluginManager kills all the active window
effects before starting a new one; so the new size-change effect for any
geometry change is going to kill the current effect.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1250
Trying to figure out what this comment was actually about, it turns out
that MSC means Media Stream Counter, and as mentioned in an article[0]
is related to DRI3 and the X11 Present extension. Anyway, the comment
has been there raising questions for some years now, I think we can
remove it.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
The ID and name are just moved into the instance private, while the rest
is moved to a `MetaCrtcModeInfo` struct which is used during
construction and retrieved via a getter. Opens up the possibility to
add actual sub types.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
Just as with MetaOutput, instead of the home baked "inheritance" system,
using a gpointer and a GDestroyNotify function to keep the what
effectively is sub type details, make MetaCrtc an abstract derivable
type, and make the implementations inherit it.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
Instead of the home baked "inheritance" system, using a gpointer and a
GDestroyNotify function to keep the what effectively is sub type
details, make MetaOutput an abstract derivable type, and make the
implementations inherit it.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
It's used for two things: avoid reading configs, and actual hotplug
update mode. The former requires the suggested position to be (-1, -1)
to trick the monitor configuration generator to skip using the suggested
position even if hotplug update mode is set to TRUE. The latter should
use the actual hotplug mode coordinates.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
Now set as a property during construction. Only actually set by the
Xrandr backend, as it's the only one currently not supporting all
transforms, which is the default.
While at it, move the 'ALL_TRANFORMS' macro to meta-monitor-tranforms.h.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
The output info is established during construction and will stay the
same for the lifetime of the MetaOutput object. Moving it out of the
main struct enables us to eventually clean up the MetaOutput type
inheritence to use proper GObject types.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
MetaCrtcInfo and MetaOutputInfo did not represent information about
MetaCrtc and MetaOutput, but the result of the monitor configuration
assignment algorithm, thus rename it to MetaCrtcAssignment and
MetaOutputAssignment.
The purpose for this is to be able to introduce a struct that actually
carries information about the CRTCs and outputs, as retrieved from the
backend implementations.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
That is is_presentation, is_primary, is_underscanning and backlight.
The first three are set during CRTC assignment as they are only valid
when active. The other is set separately, as it is untied to
monitor configuration.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
It was used during configuration to ensure that we always dealt with
every output and CRTC. Do this without polluting the MetaOutput and
MetaCrtc structs with intermediate variables not used by the
corresponding types themself.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1287
The current code assumes that the actor will always have the same
size and position of the background texture, but part of the implicit
contract of being a ClutterContent is being able to render itself
at any given actor, at any given size.
For example, if the current code is given an actor with 0x0+100+100
as geometry, and no clipped region, it'll render not the whole
background, but the 0x0+100+100 rectangle of the background. In
practice, the actor geometry acts like a "clip mask" over the
background texture, due to the assumption that the actor will
always have the same size of the monitor.
Make the calculation of the texture slices relative to the actor
box.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1302
MetaBackgroundActor is still necessary for culling purposes,
but now the actual rendering of the background is delegated
to MetaBackgroundContent, as well as the sizing information.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1302
MetaBackgroundContent is a ClutterContent implementation
that can render a background to any attached actor. Right
now, it preserves all the properties and the rendering
model of MetaBackgroundActor.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1302
X11 window stacking operations are by nature prone to race conditions.
For example, we might queue a "raise above" operation, but before it
actually takes place, the sibling the window was to be rased above, is
withdrawn.
In these cases we'd log warnings even though they are expected to
happen. Downgrade these warnings to debug messages, only printed when
MUTTER_VERBOSE is set.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1300
Test that the stage-views list of ClutterActor is correct when moving an
actor, reparenting it, or hiding an actor up the hierarchy. Also test
that the "stage-views-changed" signal works as expected.
Don't test actor transforms for now because those aren't supported yet.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1196
When the stage views the stage is shown on are changed, ClutterStage
currently provides a clutter_stage_update_resource_scales() method
that allows invalidating the resource scales of all actors. With the new
stage-views API that's going to be added to ClutterActor, we also need a
method to invalidate the stage-views lists of actors in case the stage
views are rebuilt and fortunately we can re-use the infrastructure for
invalidating resource scales for that.
So since resource scales depend on the stage views an actor is on,
rename clutter_stage_update_resource_scales() and related methods to
clutter_stage_clear_stage_views(), which also covers resource scales.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1196
When an app disappears after some data from it has been copied to the
clipboard, the owner of the clipboard selection becomes a new memory
selection source. The initial reference this new selection source is
never unref'ed, which leads to this being leaked on the next clipboard
selection owner change.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1293
Using XDG_CONFIG_HOME allows users to place their keyboard configuration into
their home directory and have them loaded automatically.
libxkbcommon now defaults to XDG_CONFIG_HOME/xkb/ first, see
https://github.com/xkbcommon/libxkbcommon/pull/117
However - libxkbcommon uses secure_getenv() to obtain XDG_CONFIG_HOME and thus
fails to load this for the mutter context which has cap_sys_nice.
We need to manually add that search path as lookup path.
As we can only append paths to libxkbcommon's context, we need to start with
an empty search path set, add our custom path, then append the default search
paths.
The net effect is nil where a user doesn't have XDG_CONFIG_HOME/xkb/.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/936
We would get the MetaDisplay from the backend singleton before creating
the MetaCompositor, then in MetaCompositor, get the backend singleton
again to get the stage. To get rid of the extra singleton fetching, just
pass the backend the MetaCompositor constructors, and fetch the stage
directly from the backend everytime it's needed.
This also makes it available earlier than before, as we didn't set our
instance private stage pointer until the manage() call.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1289
Also fix a test that dependends on a specific element order in a list
that wasn't defined to have any particular order.
The frames per second is decreased from 30 to 10, to make the test less
flaky when running in CI.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1289
The shadow was disabled for the X11 client as it was far to unreliable
when comparing sizes.
It seems that the Wayland backend has been somewhat unreliable as well,
where some race condition causing incorrect sizes thus a flaky test.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1288
A "show" command calls gtk_window_show() and gdk_display_sync(), then
returns. This means that the X11 window objects are guaranteed to have
been created in the X11 server.
After that, the test runner will look up the window's associated
MetaWindow and wait for it to be shown.
What this doesn't account for is if mutter didn't get enough CPU time to
see the new window. When this happens, the 'default-size' stacking test
sometimes failed after hiding and showing the X11 window.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1288
When we created the DMA buffer backed CoglFramebuffer, we handed it over
to CoglDmaBufHandle which took its own reference. What we failed to do
was to release our own reference to it, effectively leaking it.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1283
The stream will clean up the buffers, so let it do that before we
destroy them under its feet. Note that it'll only do this after the
following PipeWire commit:
commit fbaa4ddedd84afdffca16f090dcc4b0db8ccfc29
Author: Wim Taymans <wtaymans@redhat.com>
Date: Mon Jun 1 15:36:09 2020 +0200
stream: allow NULL param and 0 buffers in disconnect
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1283
wait_reconfigure ensures that the whole configure back and forth
completes before continuing. Doing this after every state change ensures
that we always end up with the expected state, thus fixes flakyness of
the restore-position stacking test.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1279
This cannot be made to work reliably. Some factoids:
- Internal devices may be connected via USB.
- The ACPI spec provides the _PLD (Physical location of device) hook to
determine how is an USB device connected, with an anecdotal success
rate. Internal devices may be seen as external and vice-versa, there is
also an "unknown" value that is widely used.
- There may be non-USB keyboards, the old "AT Translated Set 2 Keyboard"
interface does not change on hotplugging.
- Libinput has an internal series of quirks to classify keyboards as
internal of external, also with an "unknown" value.
These heuristics are kinda hopeless to get right by our own hand. Drop
this external keyboard detection in the hope that there will be something
more deterministic to rely on in the future (e.g. the libinput quirks
made available to us directly or indirectly).
Fixes: https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/2378
Related: https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/2353https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1277
Move Wayland support (i.e. the MetaWaylandCompositor object) made to be
part of the backend. This is due to the fact that it is needed by the
backend initialization, e.g. the Wayland EGLDisplay server support.
The backend is changed to be more involved in Wayland and clutter
initialization, so that the parts needed for clutter initialization
happens before clutter itself initialization happens, and the rest
happens after. This simplifies the setup a bit, as clutter and Wayland
init now happens as part of the backend initialization.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1218
On X11 we don't update the texture in certain circumstances, such as if
the surface is a fullscreen unredirect, or doesn't have a Pixmap. On
Wayland we only want to avoid updating the texture if there is no
texture, but as this is handled implicitly by MetashapedTexture, we
don't need to try to emulate the X11-y conditions in the generic layer
and instead just have the implementations handle update processing
themself.
This doesn't have any functional changes, but removes a vfunc from
MetaSurfaceActorClass.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1218
We failed to use the buffer age when monitors were rotated, as when they
are, we first composite to an offscreen framebuffer, then later again to
the onscreen. The buffer age checking happened on the offscreen, and an
offscreen being single buffered, they can't possible support buffer
ages.
Instead, move the buffer age check to check the actual onscreen
framebuffer. The offscreen to onscreen painting is still always full
frame, but that will be fixed in a later commit.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1237
Will be used for logging to identify what view a log entry concerns. For
the native and nested backend this is the name of the output the CRTC is
assigned to drive; for X11 it's just "X11 screen", and for the legacy
"X11 screen" emulation mode of the nested backend it's called "legacy
nested".
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1237
There's no reason to notify the surface that its geometry changed when
the visibility of the actor changes. This is only needed to update the
outputs of the surface, so do that directly instead.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1235
We started listening to notify::mapped with commit
5eb5f72434 in order to emit
wl_surface.leave events consistently when a surface gets hidden. This
caused a problem with the ClutterClones used in the overview, since
those temporarily map and unmap the windows for painting, spamming
wl_surface.leave and enter events to all surfaces.
We can easily fix that by also treating mapped clones as mapped, which
means the surface should also be on a wl_output when the overview is
shown.
Fixes https://gitlab.gnome.org/GNOME/mutter/-/issues/1141https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1235
All existing users of clutter_actor_has_mapped_clones() actually want to
know whether the actor is being cloned by a visible clone, it doesn't
matter to them if that clone is attached to an actor somewhere else in
the tree or to the actor itself.
So make clutter_actor_has_mapped_clones() a bit more convenient to use
and also check the clones of the parent-actors in that function.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1235
We started listening to "notify::position" on surface actors with commit
08e4cb54. This commit was done to fix a regression from commit cf1edff9,
which forgot to handle some cases like the actual WindowActor and not
the SurfaceActor (which is a child of the WindowActor) moving (that was
fixed by listening to MetaWindows "position-changed" signal). Also that
commit introduced meta_wayland_surface_update_outputs_recursively(),
which updates the outputs of all (sub-)surfaces in case any position
changed and made sure subsurfaces also get their outputs updated in case
the parent actor moved.
Connecting to the "notify::position" signal, which the above commit also
did is now superflous though because position changes will queue a
relayout and the actors allocation will change during the next
allocation cycle, notifying the "allocation" property which we also
listen to.
So save some resources and stop listening to that signal.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1235
We don't have enough Xlib code in mutter ...
Joking aside, it can be useful to make the cursor invisible
without hiding it, for example for replacing the actual cursor
with an actor in gnome-shell; the real cursor should still
update the focus surface in that case, and not sneak into
screenshots or -casts.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1244
We're iterating inside the PipeWire loop when detecting PipeWire errors,
and shouldn't destroy the PipeWire objects mid-iteration. Avoid this by
first disabling the stream src (effectively stopping the recording),
then notifying about it being closed in an idle callback. The
notification eventually makes the rest of the screen cast code clean up
the objects, including the src and the associated PipeWire objects, but
will do so outside the PipeWire loop iteration.
Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1251https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1251
In the native backend, the MetaRenderer manages the view by creating one
per CRTC, but until now the MetaStageX11 managed the view for the X11
backend. This caused some issues as it meant meta_renderer_get_views()
not returning anything, and that the view of the X11 screen not being a
MetaRendererView, while in the other backends, all views are.
Fix this by moving the view management responsibility to
MetaRendererX11Cm, and have MetaStageX11 only operate on it via
meta_renderer_x11_cm_*() API. The MetaRendererX11Cm takes care of making
sure the view is always added to the list in the renderer, and turning
X11 screen sizes into "layouts" etc.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1251
"Legacy" is a misleading name, it's just how the native backend and the
X11 backend behaves differently. Instead rename it to 'add_view()' and
add the sanity check to the caller.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1251
A DMA buffer might not be able to scanout, and in that case the import
with GBM_BO_USE_SCANOUT will fail. Handle that by failing to scanout,
effectively falling back to compositing.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1261
Since PIDs are inherently insecure because they are reused after a
certain amount of processes was started, it's possible the client PID
was spoofed by the client.
So make sure users of the meta_window_get_pid() API are aware of those
issues and add a note to the documentation that the PID can not be
totally trusted.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1180
Since the PID of a window can't change as long as the window exists, we
can safely cache it after we got a valid PID once, so do that by adding
a new `window->client_pid` private property.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1180
The shell uses the PID of windows to map them to apps or to find out
which window/app triggered a dialog. It currently fails to do that in
some situations on Wayland, because meta_window_get_pid() only returns a
valid PID for x11 clients.
So use the client PID instead of the X11-exclusive _NET_WM_PID property
to find out the PID of the process that started the window. We can do
that by simply renaming the already existing
meta_window_get_client_pid() API to meta_window_get_pid() and moving
the old API providing the _NET_WM_PID to meta_window_get_netwm_pid().
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1180
They all checked that the remote session service talked with the
correct peer, and some of them did check that there is an associated
screencast session.
Add a new check for the session being started (as it's state is
decoupled with screencast session availability) and move all checks
to a function that is called from all input-oriented DBus methods.
Fixes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1254https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1258
It was removed in 3.34 as part of 6ed5d2e2. And we thought that was the
only thread that might exist and use X11. But the top gnome-shell crasher
in 3.36 seems to suggest otherwise.
We don't know what or where the offending thread is, but since:
1. We used XInitThreads for years already prior to 3.34; and
2. Extensions or any change to mutter/gnome-shell could conceivably use
threads to make X calls, directly or indirectly,
it's probably a good idea to reintroduce XInitThreads. The failing assertion
in libx11 is also accompanied by a strong hint:
```
fprintf(stderr, "[xcb] Most likely this is a multi-threaded client " \
"and XInitThreads has not been called\n");
```
https://bugs.launchpad.net/bugs/1877075
Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1252https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1256
When the wallpaper image is larger than the monitor resolution we already
use mipmapping to scale it down smoothly in hardware. We use
`GL_TEXTURE_MIN_FILTER` = `GL_LINEAR_MIPMAP_LINEAR` for the highest quality
scaling that GL can do. However that option is designed for 3D use cases
where the mipmap level is changing over time or space.
Since our wallpaper is not changing distance from us we can improve the
rendering quality even more than `GL_LINEAR_MIPMAP_LINEAR`. To do this we
now set `GL_TEXTURE_MAX_LEVEL` (if available) to limit the mipmap level or
blurriness level to the lowest resolution (highest level) that is still
equal to or higher than the monitor itself. This way we get the benefits
of mipmapping (downscaling in hardware) *and* retain the maximum possible
sharpness for the monitor resolution -- something that
`GL_LINEAR_MIPMAP_LINEAR` alone doesn't do.
Example:
Monitor is 1920x1080
Wallpaper photo is 4000x3000
Mipmaps stored on the GPU are 4000x3000, 2000x1500, 1000x750, ...
Before: You would see an average of the 2000x1500 and 1000x750 images.
After: You will now only see the 2000x1500 image, linearly sampled.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1003
It's very useful to have common functions for easily creating a monitor
test setup for all kinds of tests, so move create_monitor_test_setup()
and check_monitor_configuration() and all the structs those are using to
monitor-test-utils.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1243
We're going to move some structs from monitor-unit-tests.c to
monitor-test-utils.h and some names are currently clashing with the
struct names here, so rename those to be specific to the
MonitorStoreUnitTests.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1243
check_monitor_test_clients_state() is a function that's only meant to be
used in the monitor-unit-tests, and since we're going to move the
functions for creating MonitorTestSetups into a common file, this
function is going to be in the way of that. So move the checking of the
test client state outside of check_monitor_test_clients_state().
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1243
We're going to move the functions for building MonitorTestSetups to the
common monitor-test-utils.c file.
To make building test setups a bit more straightforward in case no
TestCaseExpect is wanted, change create_monitor_test_setup() to take a
MonitorTestCaseSetup instead of a MonitorTestCase as an argument.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1243
Commit e06daa58c3 changed the tested values to use corresponding valid
enum values instead of negative ones. Unfortunately that made one value
become a duplicate of an existing one and also in part defeated the original
intention of checking the implementation of
`meta_output_crtc_to_logical_transform`.
Use `meta_monitor_transform_invert` to fix both shortcomings.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1242
One of the important classes in Mutter's handling of client textures is
the `MetaShapedTexture`. This commit adds a few gtk-doc comments which
explain its purpose, with special attention to the viewport methods.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1210
Since we're now connecting to one more signal of MetaWaylandOutput, keep
signal connections in one place and move connecting the
"output-destroyed" signal to surface_entered_output() and disconnecting
it to surface_left_output().
This also allows us to use the "outputs_to_destroy_notify_id" as a
simple set and rename it to "outputs".
While at it, also use g_hash_table_destroy() instead of
g_hash_table_unref() since destroy is more clear than unref and does the
same thing in this case.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1230
When hotplugging a new monitor, we recreate all the MetaWaylandOutputs
and need to emit leave events to the surfaces for the old wl_outputs and
enter events for the newly created ones.
There's a race condition though: We might update the monitors a surface
is on (and thus emit enter/leave events for the wl_outputs) before the
Wayland client is registered with the new wl_output (ie. the
bind_output() callback of MetaWaylandOutput was called), which means we
don't send an enter event to the client in surface_entered_output().
Since MetaWaylandSurface now has the MetaWaylandOutput in its outputs
hashtable, it thinks the client has been notified and won't send any
more enter events.
To fix that, make MetaWaylandOutput emit a new signal "output-bound"
when a client bound to the output and make all surfaces which are on
that output listen to the signal. In the signal handler compare the
newly added client to the client the surface belongs to, and if it's the
same one, send an enter event to that client.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1230
The "output-destroyed" signal is used for notifying MetaWaylandSurfaces
that an output they are shown just got invalid (for example because a
monitor hotplug happened).
While we delay the destroying of outputs by 10 seconds since commit
1923db97 because of a race-condition, it doesn't make sense to wait 10
seconds until we let surfaces know that an output was destroyed.
So move the emission of the "output-destroyed" signal to
make_output_inert(), which is called before we start the 10 seconds
delay.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1230
When tiling, we want to set the tile monitor. To not have to do this
from the call site, make meta_window_tile() fall back to the current
monitor if nothing set it prior to the call.
This will make it more convenient for test cases to test tiling
behavior.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1171
The test tests that (for both X11 and Wayland) that:
* The client unmaximizes after mapping maximized to a predictable size
* That the client unmaximizes to the same size after toggling maximize
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1171
This makes sure that a client has properly responded to a configure
event it itself triggered. In practice, this is just two 'wait'
commands, with a 'dispatch' in between, which is needed because a single
one does not reliably include the two way round trip happening when e.g.
responding to a unmaximize configure event triggered by a unmaximize
request.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1171
The 'assert_size' command checks that the size of the window, both
client side and compositor side, corresponds to an expected size set by
the test case.
The size comparison can only be done when the window is using 'csd', in
order for both the client and server to have the same amount of
understanding of the title bar. For ssd, the client cannot know how
large the title bar, thus cannot verify the full window size.
Sizes can be specified to mean the size of the monitor divided by a
number. This is that one can make sure a window is maximized or
fullscreened correctly.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1171
Gtk is quite buggy and "fluid" in how it handles the shadow margins for
windows under X11. The "size" of the window fluctuate between including and
excluding a shadow margin in a way that causes issues, as there are no
atomic update of any state going on.
In order to avoid running into those particular issues now, lets get rid
of shadows so the margins are always zero, when the client is using the
X11 backend.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1171
To get some kind of consistency between what 'resize' means for the
compositor and the client, make the size correspond to the "frame rect"
of the window, i.e. the window geometry in the Wayland case, and the
window size including the titlebar in the X11 case.
This is so that the window size later can be reliably compared both in
the compositor and in the client using the same expected dimensions.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1171
When toying with the test client to try to reproduce issues (e.g.
writing commands on stdin to create and manipulate windows), when you
write a command incorrectly you'll get a warning printed to standard
out. The problem, however, is that it doesn't include a line break in
the end, meaning when you type the correct command, it won't be on a new
line.
Fix this minor annoyance by adding line breaks to all warning messages.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1171
The test client could already understand the resize command, but they
could not be added to metatests as the command was not properly plumbed
via the test runner. Establish the plumbing for the resize command so
that resize tests can be added.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1171
This removes ClutterAnimation and related tests. ClutterAnimation has
been deprecated for a long time, and replacements exist and are used by
e.g. GNOME Shell since a while back.
This also disables a few relatively unrelated interactive tests, as they
rely on ClutterAnimation to implement some animations they use to
illustrate what they actually test.
As interactive tests currently are more or less untestable due to any
interaction with them crashing, as well as they in practice means
rewriting the tests using non-deprecated animation APIs, they are not
ported right now. To actually port the interactive tests, it needs to be
possible to fist interact with them.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1192
When memfd_create isn't used, the file isn't sealed. Therefore, we
should skip test_readonly_seals on the fallback case. This fixes
compilation error on FreeBSD 12, which does not support memfd_create.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1229
We were iterating through evcodes, but using API that expects Clutter button
numbers. Instead of transforming those to Clutter numbers to have those translated
back, use the inner seat API that already takes evcodes.
Fixes stuck buttons keys after a virtual device is destroyed while those are
pressed.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1239
Move to center uses all monitors for calculating work area.
This can lead to an unexpected behaviour on some monitor
configurations resulting in current window being split between
monitors. We should move window to the center of the active display.
Closes https://gitlab.gnome.org/GNOME/mutter/-/issues/1073
Inhibiting remote access means any current remote access session is
terminated, and no new ones can be created, until remote access is
uninhibited. The inhibitation is ref counted, meaning there can be more
than one inhibitor.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1212
When resizing an X11 window with client side decorations, the shadow is
clipped by the frame bounds so that we don't need to paint the shadow
under the opaque areas covered by the window and its frame.
When the X11 client uses the EMWH synchronization mechanism (like all
gtk-3 based clients), the actual window may not be updated so that the
actual window and it frame may be behind the expected window frame
bounds, which gives the impression of de-synchronized shadows.
To avoid the issue, keep a copy of the frame bounds as a cache and only
update it when the client is not frozen so that the clipping occurs on
the actual content.
Closes: https://gitlab.gnome.org/GNOME/mutter/issues/1178https://gitlab.gnome.org/GNOME/mutter/merge_requests/1214
It takes coordinates in stage coordinate space, and will result in
a screen cast stream consisting of that area, but scaled up by the scale
factor of the view that overlaps with the area and has the highest scale
factor.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1207
Will be used by the stage to not paint the overlays. We skip all
overlays since overlays are only ever used for pointer cursors when the
hardware cursors cannot or should not be used.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1207
These phase callbacks are not intended to be inovked when something
secondary is painting the stage, such as a screen cast stream, or
similar. Thus, only invoke the callbacks when there is a view associated
with the paint context, which will not be the case for offscreen
painting.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1207
If there is a paint context available (i.e. for the phases that are
during the actual stage paint), pass it along the callbacks, so that
the callback implementations can change their operation depending on the
paint context state.
This also means we can get the current view from the paint context,
instead of the temporarily used field in the instance struct.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1207
A paint flag affects a paint operation in ways defined by the flags.
Currently no flags are defined, so no semantical changes are defined
yet. Eventually a flag aiming to avoid painting of cursors is going to
be added, so that screen cast streams can decide whether to include a
cursor or not.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1207
If drmModePageFlip() or custom_page_flip_func fails, process_page_flip() was
forgetting to undo the ref taken for that call. This would leak page_flip_data.
The reference counting works like this:
- when created, ref count is 1
- when calling drmModePageFlip, ref count is increased to 2
- new: if flip failed, ref count is decreased back to 1
- if calling schedule_retry_page_flip(), it takes a ref internally
- if calling mode_set_fallback(), it takes a ref internally
- all return FALSE paths have an explicit unref
- return TRUE path has an explicit unref
This issue was found by code inspection and while debugging an unrelated issue
with debug prints sprinkled around. I am not aware of any end-user visible
issues being fixed by this, as the leak is small and probably very rare.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1209
When testing a laptop with intel and DisplayLink devices, attempting to set the
DL output as the only active output resulted in GNOME/Wayland freezing. The
main event loop was running fine, but nothing on screen would get updated once
the DL output become the only one. This patch fixes that issue.
DisplayLink USB 3 devices use an out-of-tree kernel DRM driver called EVDI.
EVDI can sometimes fail drmModePageFlip(). For me, the flip fails reliably when
hotplugging the DL dock and when changing display configuration to DL only.
Mutter has a workaround for failing flips, it just calls drmModeSetCrtc() and
that succeeds.
What does not work reliably in the fallback path is Mutter keeping track of the
pageflip. Since drmModePageFlip() failed, there will not be a pageflip event
coming and instead Mutter queues a callback in its stead. When you have more
than one output, some other output repainting will attempt to swap buffers and
calls wait_for_pending_flips() which has the side-effect of dispatching any
queued flip callbacks. With multiple outputs, you don't get stuck (unless they
all fail the exact same way at the same time?). When you have only one output,
it cannot proceed to repaint and buffer swap because the pageflip is not marked
complete yet. Nothing dispatches the flip callback, leading to the freeze.
The flip callback is intended to be an idle callback, implemented with a
GSource. It is supposed to be called as soon as execution returns to the main
event loop. The setup of the GSource is incomplete, so it will never dispatch.
Fix the GSource setup by setting its ready-time to be always in the past. That
gets it dispatched on the next cycle of the main event loop. This is now the
default behavior for all sources created by meta_kms_add_source_in_impl().
Sources that need a delay continue to do that by overriding the ready-time
explicitly.
An alternative solution could have been to implement GSource prepare and check
callbacks returning TRUE. However, since meta_kms_add_source_in_impl() is used
by flip retry code as well, and that code needs a delay through the ready-time,
I was afraid I might break the flip retry code. Hence I decided to use
ready-time instead.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1209
According to the XSetSelectionOwner libX11 documentation:
[...] If the owner window it has specified in the request is later
destroyed, the owner of the selection automatically reverts to None,
but the last-change time is not affected.
This is indeed visible through the selection_timestamp field in
XFixesSelectionNotify events.
Use this to check whether the selection time is recent-ish (thus
likely coming from an explicit XSetSelectionOwner request) and honor
the client intent by setting a "NULL" owner. If the selection time
is too old, it's definitely an indication of the owner client being
closed, the scenario where we do want the clipboard manager to take
over.
This fixes two usecases:
- X11 LibreOffice / WPS clear the selection each time before copying
its own content. Mutter's clipboard manager would see each of those
as a hint to take over, competing with the client over selection
ownership. This would simply no longer happen
- Password managers may want to clear the selection, which would be
frustrated by our clipboard manager.
There's a slight window of opportunity for the heuristics to fail
though, if a X11 client sets the selection and closes within 50ms, we
would miss the clipboard manager taking over.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1206
The X11 selection source was being preserved after unsetting its
ownership. This is no leak as it would be eventually replaced by
another source, or destroyed on finalize. But it's pointless to
keep it.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1206
Test the two modes of MetaAnonymousFile, MAPMODE_SHARED and
MAPMODE_PRIVATE and make sure they don't leak data to other FDs when
writing to an FD provided by `meta_anonymous_file_get_fd` even though
the data of both FDs is residing in the same chunk of memory.
We do all the reading tests using mmap instead of read() since using
read() on shared FDs is going to move the read cursor of the fd. That
means using read() once on the shared FD returned by
meta_anonymous_file_get_fd() in MAPMODE_PRIVATE breaks every subsequent
read() call.
Also test the fallback code of MetaAnonymousFile in case `memfd_create`
isn't used for the same issues.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1012
Add MetaAnonymousFile, an abstraction around anonymous read-only files.
Files can be created by calling meta_anonymous_file_new(), passing the
data of the file. Subsequent calls to meta_anonymous_file_open_fd()
return a fd that's ready to be sent over the socket.
When mapmode is META_ANONYMOUS_FILE_MAPMODE_PRIVATE the fd is only
guaranteed to be mmap-able readonly with MAP_PRIVATE but does not
require duplicating the file for each resource when memfd_create is
available. META_ANONYMOUS_FILE_MAPMODE_SHARED may be used when the
client must be able to map the file with MAP_SHARED but it also means
that the file has to be duplicated even when memfd_create is available.
Pretty much all of this code was written for weston by Sebastian Wick,
see https://gitlab.freedesktop.org/wayland/weston/merge_requests/240.
Co-authored-by: Sebastian Wick <sebastian@sebastianwick.net>
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1012
Instead of having everything clumped at MetaWaylandDataManager,
split the primary selection to its own struct. This manager is
handled separately from wl_data_device_manager and other selection
managers, so they would be able to interoperate between them, even.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1193
This is still an openly defined struct, as we will need accessed
by "subclasses". Same principle applies than with the
MetaWaylandDataSource refactor, this is not meant to introduce
functional changes, so just go with it.
On the bright side, the interactions are now clearer, so it could
be made saner in the future.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1193
The split wasn't 100% clean, and some extra private API had to be
added for it (but well, looking at the API, it's already evident
there's a cleanup/streamlining task due). This is meant to be a
refactor with no functional changes, so just go with it.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1193
We already have a signal callback that translates selection ownership changes to
data_device/primary .selection events. Given both will be run when a data source
is being replaced, and this event emission being deleted is kinda short sighted
in that in only knows about Wayland, rely entirely on MetaSelection::owner-changed
emission.
Fixes spurious .selection(null) events being sent when a compositor-local source
takes over the selection without the focus changing (eg. screenshot to clipboard).
Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1160https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1204
We are meant to send a .cancelled event after the drop is performed
in certain situations, but only for version>3 clients. Since this is
all version 3 business, only set the drop_performed flag for v3
clients. This drops the need to perform version checks at the time
of cancelling (which is present for other usecases in v1).
Fixes emission of wl_data_source.cancelled for v1 clients.
Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1177https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1203
For the cases where we read a fixed size from the selection (eg. imposing
limits for the clipboard manager), g_input_stream_read_bytes_async() might
not read up to this given size if the other side is spoonfeeding it content.
Cater for multiple read/write cycles here, until (maximum) transfer size is
reached.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1198
Flushing the X11 selection output stream may happen synchronously or
implicitly, in which case there is not a task to complete. Check there
is actually a task before returning errors. We additionally set the
pipe_error flag, so future operations will fail with an error, albeit
with a more generic message.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1198
If a write_async() comes up while we are flushing on the background,
the task will be queued, but not deemed a reason on itself to keep
flushing (and finish the task) after a property delete event.
To fix this, do not ever queue up write_async tasks (this leaves
priv->pending_task only used for flush(), so the "flush to end"
behavior in the background is consistent). We only start a
background flush if there's reasons to do it, but the tasks are
immediately finished.
All data will still be ensured to be transfered on flush/close,
this makes the caller in this situation still able to reach to it.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1198
It does not make sense to check for the stream not being closed,
this might happen multiple times during the lifetime of the stream
for a single transfer. We want to notify the INCR transfer just
once.
Check for the explicit conditions that we want, that the remaining
data is bigger than we can transfer at once, and that we are not
yet within the INCR transfer.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1198
The stream automatically flushes after data size exceeds the
size we deem for INCR chunks, but we still try to copy it all.
Actually limit the data we copy, and leave the rest for future
INCR chunks.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1198
INCR transfers are mandated to finish with a final 0-size XChangeProperty
roundtrip after the final data chunk. Actually honor this and ensure we
iterate just once more for this.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1198
This seemed to work under the assumption that a flush() call can
only result in one INCR roundtrip. This is evidently not true, so
we should hold things off until all pending data is actually flushed.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1198
If say we want 32bit data, but have 2 bytes stored, we would simply
ignore flush requests. Allow (and don't clear) the needs_flush flag
if we have less than the element size accumulated.
Instead handle this in can_flush(), so it's triggered whenever we
have enough data to fill 1 element, or if the stream is closing
(seems a broken situation, but triggered by the caller).
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1198
XMaxRequestSize/XMaxExtendedRequestSize are documented to return
the maximum size in 4-byte units, whereas we are comparing this
to byte lenghts. We can afford 4x the data here.
Since I don't know the payload size of the XChangeProperty request,
be generous and allot 400 bytes for it, we have some to spare.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1198
When closing the lid of a laptop, we reconfigure all the monitors in order
to update the CRTCs and (if enabled) the global UI scaling factor.
To do this, we try first to reuse the current configuration for the usable
monitors, but if we have only monitor enabled and this one is on the laptop
lid we just end up creating a new configuration where the primary monitor is
the laptop one (as per find_primary_monitor() in MetaMonitorConfigManager),
but ignoring the user parameters.
In case the user selected a different resolution / scaling compared to the
default one, while the laptop lid is closed we might change the monitors
layout, causing applications to rescale or reposition.
To avoid this, when creating the monitors configuration from the current
current state, in case we have only one monitor available and that one is
the laptop panel, let's just reuse this configuration.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1200
Try to bypass compositing if there is a fullscreen toplevel window with
a buffer compatible with the primary plane of the monitor it is
fullscreen on. Only non-mirrored is currently supported; as well as
fullscreened on a single monitor. It should be possible to extend with
more cases, but this starts small.
It does this by introducing a new MetaCompositor sub type
MetaCompositorNative specific to the native backend, which derives from
MetaCompositorServer, containing functionality only relevant for when
running on top of the native backend.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
We need to coordinate with MetaCompositor during pre-paint so that we
have control over whether MetaLater callbacks happen first, or the
MetaCompositor pre-paint logic.
In order to do so, make MetaLater listen to a new signal "pre-paint" on
MetaCompositor, that is called MetaCompositors own pre-paint handling.
This fixes an issue where the top window actor was calculated after the
MetaCompositor pre-paint handling, meaning the top actor being painted
was out-of-date.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
Since the order of destruction during MetaDisplay tear down is a bit
unordered, there are pieces that try to destruct its compositing
dependent pieces (i.e. queued MetaLater callbacks) after MetaCompositor
has been cleaned up, meaning we need to put some slightly awkward NULL
checks to avoid crashing.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
MetaCompositor is the place in mutter that manages the higher level
state of compositing, such as handling what happens before and after
paint. In order for other units that depend on having a compositor
instance active, but should be initialized before the X11 implementation
of MetaCompositor registers as a X11 compositing manager, split the
initialization of compositing into two steps:
1) Instantiate the object - only construct the instance, making it
possible for users to start listening to signals etc
2) Manage - this e.g. establishes the compositor as the X11 compositing
manager and similar things.
This will enable us to put compositing dependent scattered global
variables into a MetaCompositor owned object.
For now, compositor management is internally done by calling a new
`meta_compositor_do_manage()`, as right now we can't change the API of
`meta_compositor_manage()` as it is public. For the next version, manual
management of compositing will removed from the public API, and only
managed internally.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
While at it, fix some style inconsistencies, for now use a single
singleton struct instead of multiple static variables, and
other non-functional cleanups. Semantically, there is no changes
introduced.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
This will check whether the current backing buffer is compatible with
the primary plane of the passed CoglOnscreen. Since this will extend the
time before a buffer is released, the MetaWaylandBufferRef is swapped
and orphaned if a new buffer is committed before the previous one was
released. It'll eventually be released, usually by the next page flip
callback.
Currently implemented for EGLImage and DMA-BUF buffer types.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
Advertising support for modifiers means we will most likely not not be
able to scan out client buffers directly, meaning it just as likely that
we won't be able to scan out even fullscreen windows without atomic KMS.
When we have atomic support, we should advertise support for modifiers
if atomic is used to drive the CRTCs, as we by then can check whether we
can scan out directly, place in an overlay plane, etc.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
While this is fairly incomplete, as to check things fully we need to use
TEST_ONLY in atomic to try out a complete assignment on the device, but
this works well enough for legacy non-modifier cases.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
Instead of always swapping buffers and flipping the back buffer, make it
possible to scan out a provided buffer directly without swapping any EGL
buffers.
A buffer is passed as an object implementing the empty CoglScanout
interface. It is only possible to do this in the native backend; and the
interface is implemented by MetaDrmBufferGbm. When directly scanned out,
instead of calling gbm_surface_lock_front_buffer() to get the gbm_bo and
fbid, get it directly from the MetaDrmBufferGbm, and use that to create
the page flip KMS update.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
Better to have the relevant object figure out whether it is a good
position to be unredirectable other than the actor, which should be
responsible for being composited.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
This removes the MetaWindowX11::priv pointer. It is replaced with a
meta_window_x11_get_private() helper function, and another method to get
the client rect without going through MetaWindowX11Private.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
Surface buffers are created with meta_drm_buffer_new_acquire(), taking a
gbm_surface acquiring the gbm itself, and meta_drm_buffer_new_take()
that takes over ownership of a passed gbm_bo.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
Currently a buffer use count always reaches zero before it is replaced.
This is due to the fact that at the point a new buffer is attached, the
last potential user releases it (the stage) since the currently
displayed frame has a composited copy of the buffer.
This may however change, if a buffer is scanned out directly, meaning it
should not be released until the page flip callback is invoked.
Prepare for this by making the buffer reference a heap allocated struct,
enabling us to keep a pointer to it longer than the buffer is attached.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/798
The CRTC level transform (not necessarily the hw transform) must be
taken into account when calculating the position of the CRTC in the
stage coordinate space, when placing the hw cursor, otherwise we'll
place the cursor as if the monitor was not rotated.
This wasn't a problem in the past, as with rotation, we always used the
OpenGL cursor, so the issue newer showed.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1199
The port to per CRTC views was incomplete; we still used the logical
monitor layout as the stage view layout, while still using one view per
CRTC.
This worked fine for most cases, e.g. regular monitors, tiled or
non-tiled, transformed or non-transformed. Where it broke, however, was
when a monitor consists of multiple CRTCs. We already have the layout a
CRTC corresponds to on the stage kept with the CRTC metadata, so use
this directly.
Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1170https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1199`
The CRTC level transform (i.e. not necessarily the one set on the
hardware) is what is relevant for calculating the layout the CRTC will
have on the stage, so only use the one that can be handled by the
hardware for the CRTC assignment.
This makes the CRTC layout valid for tiled monitors.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1199
Previously the tile coordinate was used to offset a CRTC scanout
coordinate within a larger framebuffer. Since 3.36 we're always
scanning out from (0, 0) as we always have one framebuffer per CRTC; we
instead use the tile coordinate to calculate the coordinate the tile has
in the stage view. Adapt calculation to fulfil this promise instead of
the old one.
This also corrects the tiled custom monitor test case.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1199
test_client_new might return early if conditions are not met, leaving some
allocated data around without freeing it.
Since we're not using the client before, there's no need to initialize it early
and just initialize it when it's going to be returned.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1195
Dereference the loop variable rather than the original list head. This
fixes a regression introduced in 4413b86a3 ("backends: Replace
ClutterDeviceManager usage in favor of ClutterSeat", 2019-10-04) which
broke button scrolling with trackballs.
Closes:https://gitlab.gnome.org/GNOME/mutter/-/issues/1120
Most visible with xwayland-on-demand, at the time of setting things up
for X11 selections, we don't forward the current state. This makes the
first started X11 app oblivious to eg. the current clipboard.
Syncing selections up at the time of initializing the X11 selection
stuff ensures that doesn't happen.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1186
On VT switch, the devices are removed, which means for Wayland disabling
the keyboard.
When the keyboard is disabled, the associated `xkb_state` is freed and
recreated whenever the keyboard is re-enabled when switching back to the
compositor VT.
That means the `xkb_state` for Wayland is lost whereas the same for
clutter is kept, which causes to a discrepancy with locked modifiers on
VT switch.
To avoid that issue, preserve the XKB info only to dispose it when the
keyboard is eventually finalized.
Closes: https://gitlab.gnome.org/GNOME/mutter/issues/344https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1185
The motion events of tablets for example need to be mapped on the
selected screen area if the input device is configured to use only a
part of the active logical monitor.
To achieve this behavior each motion event is transformed using the
transformation matrix set for the input device.
Closes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1118
At some point we crossed the streams... In a short timespan we had
1f00aba92c merged, pushing WacomDevice to a common parent object,
and dcaa45fc0c implementing device grouping for X11.
The latter did not rely on the former, and just happened to
merge/compile without issues, but would promptly trigger a crash
whenever the API would be used.
Drop all traces of the WacomDevice internal to MetaInputDeviceX11.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1183
Currently we check whether a window is alive everytime it's focused.
This means that an application that doesn't respond to the check-alive
event during startup always showing the "application froze" dialog,
without the user ever trying to interact with it.
An example where this tends to to happen is with games, and for this
particular scenario, it's purely an annoyance, as I never tried to
interact with the game window in the first place, so I don't care that
it's not responding - it's loading.
To avoid these unnecessary particular "app-is-frozen" popups, remove the
alive check from the focus function, and instead move it back to the
"meta_window_activate_full()" call. To also trigger it slightly more
often, also add it to the path that triggers the window focus when a
user actively clicks on the window.
This means that we currently check whether a window is alive on:
* Any time the window is activated. This means e.g. alt-tab or
selecting the window in the overview.
* The user clicks on the window.
Note that the second only works for an already focused window on
Wayland, as on X11, we don't refocus it. This particular case isn't
changed with this commit, as we didn't call meta_window_focus() to begin
with here.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1182
This fixes an issue where a non-maximized screen casted window would be
stretched to fill the whole screen cast stream, instead of just the crop
that corresponds to the current window size.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1174
Picking now only happens on allocated actors, but the
callback in the actor-pick test is not waiting for the
stage to run an allocation cycle. Ideally, we'd wait
for this cycle, but for now, forcing an allocation works
as well.
Allocate the overlay actor in the actor-pick test.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1169
Normally we bail out in `sync_actor_geometry()`. The comment there
states:
```
Normally we want freezing a window to also freeze its position; this allows
windows to atomically move and resize together, either under app control,
or because the user is resizing from the left/top. But on initial placement
we need to assign a position, since immediately after the window
is shown, the map effect will go into effect and prevent further geometry
updates.
```
The signal for the initial sync originates in `MetaWindow` though and predates
`xdg_toplevel_set_maximized`, which again calls `meta_window_force_placement`,
triggering the signal too early. As a result, Wayland clients that start up
maximized have a wrong map animation, starting in the top-left corner.
In order to fix this without changing big parts of the geometry logic and risking
regressions, force the initial sync again before mapping.
Solution suggested by Jonas Ådahl.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1164
IBusInputContext/ClutterInputFocus/GtkIMContext all go for offset+len
for their ::delete-surrounding signals, with offset being a signed int
(neg. to delete towards left of selection, pos. to delete towards right
of selection) and len being an unsigned int from the offset (and
presumably, skipping the current selection).
The text-input protocols however pass in this event two unsigned integers,
one being the length of text to delete towards the left of the selection,
and another the length of text to delete towards the right of the selection.
To translate properly these semantics, positive offsets shouldn't account
for before_length, and negative offset+len shouldn't account for after_length.
The offset/length approach may of course represent deletions that are
detached from the current cursor/selection, we simply delete the whole range
from the cursor/selection positions then.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/517
The input method can assign a negative value to
clutter_input_method_delete_surrounding() to move the cursor to the left.
But Wayland protocol accepts positive values in delete_surrounding() and
GTK converts the values to the negative ones in
text_input_delete_surrounding_text_apply().
https://gitlab.gnome.org/GNOME/mutter/issues/539
GObject recommends to break references to other objects on dispose
instead of finalize, also we want to release the pressed virtual buttons
as early as possible if we know the object is getting destroyed.
So release the pressed buttons and unref our virtual
MetaInputDeviceNative when the dispose vfunc is called, which also
allows us to release the buttons immediately from javascript instead of
waiting for the garbage collector by calling run_dispose() on the
object.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1157
This allows us to screencast any window continuously, even
without it being visible. Because it's still being painted,
clients continue to receive frame callbacks, and people
are happy again.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1129
Just like what's done for monitor screencasting. Unfortunately, there's
no mechanism to share fences with PipeWire clients yet, which forces
us to guarantee that a frame is completed after blitting.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1129
MetaScreenCastWindowStreamSrc connects to the "damaged" signal of
MetaWindowActor. This signal is not exactly tied to the paint cycle
of the stage, and a damage may take quite a while to arrive when
a client doesn't want to draw anything. For that reason, the window
screencast can start empty, waiting for a damage to arrive.
Ensure at least one frame is recorded when enabling the window stream.
Fixes: https://gitlab.gnome.org/GNOME/mutter/issues/1097https://gitlab.gnome.org/GNOME/mutter/merge_requests/1129
cogl_framebuffer_push_rectangle_clip() acts on the current modelview
matrix. That means the result of clipping then translating will be
different of the result of translating then clipping.
What we want for window screencasting is the former, not the latter.
Move the translation code (and associated) to after clipping.
Fixes: https://gitlab.gnome.org/GNOME/mutter/issues/1097https://gitlab.gnome.org/GNOME/mutter/merge_requests/1129
When calculating the transform we should apply to the cursor sprite
before uploading it to the cursor plane, we must also take into
account non upright mounted LCD panels.
Otherwise the cursor ends up 90 degrees rotated on devices where the
LCD panel is mounted 90 degrees rotated in its enclosure.
This commit fixes this by calling meta_monitor_logical_to_crtc_transform
in get_common_crtc_sprite_transform_for_logical_monitors to adjust the
transform for each Monitor in the LogicalMonitor.
Fixes: https://gitlab.gnome.org/GNOME/mutter/-/issues/1123https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1153
Support for them appears to be way less common than e.g. png, which is
currently the preferred format from Firefox, Chromium, Libreoffice and others.
Adopt to that fact.
As a side effect, this works around a bug observed when copying images in
Firefox on Wayland.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1141
When resizing a window interactively, we'll set a grab operation and a
grab window, among other things. If we're resizing (including setting
initial size, i.e. mapping) another window, that didn't change position,
don't use the gravity of the grab operation when resizing our own
window.
This fixes an issue with jumpy popup position when moving a previously
mapped gtk popover.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/999
The transactional KMS API has been modelled after atomic KMS. Atomic KMS
currently doesn't support forwarding cursor hotspot metadata, thus it
was left out of the transactional KMS API having the user set the simply
create a plane assigment with the cursor sprite assigned to a cursor
plane using regular coordinates.
This, however, proved to be inadequate for virtual machines using
"seamless mouse mode" where they rely on the cursor position to
correspond to the actual cursor position of the virtual machine, not the
cursor plane. In effect, this caused cursor positions to look "shifted".
Fix this by adding back the hotspot metadata, right now as a optional
field to the plane assignment. In the legacy KMS implementation, this is
translated into drmModeSetCursor2() just as before, while still falling
back to drmModeSetCursor() with the plane coordinates, if either there
was no hotspot set, or if drmModeSetCursor2() failed.
Eventually, the atomic KMS API will learn about hotspots, but when
adding our own atomic KMS backend to the transacitonal KMS API, we must
until then still fall back to legacy KMS for virtual machines.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1136
When calculating the resource scale of a clone source, we might end up
in situations where we fail to do so, even though we're in a paint. A
real world example when this may happen if this happens:
* A client creates a toplevel window
* A client creates a modal dialog for said toplevel window
* Said client commits a buffer to the modal before the toplevel
If GNOME Shell is in overview mode, the window group is hidden, and the
toplevel window actor is hidden. When the clone tries to paint, it fails
to calculate the resource scale, as the parent of the parent (window
group) is not currently mapped. It would have succeeded if only the
clone source was unmapped, as it deals with the unmapped actor painting
by setting intermediate state while painting, but this does not work
when the *parent* of the source is unmapped as well.
Fix this by inheriting the unmapped clone paint even when calculating
the resource scale.
This also adds a test case that mimics the sequence of events otherwise
triggered by a client. We can't add a Wayland client to test this, where
we actually crash is in the offscreen redirect effect used by the window
dimming feature in GNOME Shell.
Fixes: https://gitlab.gnome.org/GNOME/mutter/-/issues/808https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1147
For HiDPI pointer cursors backed by Wayland surfaces, the hotspot must
be placed using integers on the logical pixel grid. In practice what
this means is that if the client loads a cursor sprite with the buffer
scale 2, and it's hotspot is not dividable by 2, it will be rounded
down to an integer that can. E.g. a wl_surface with buffer scale 2 and a
cursor image with hotspot coordinate (7, 7) will have the coordinate
(3.5, 3.5) in surface coordinate space, and will in practice be rounded
down to (3, 3) as the hotspot position in wl_pointer only takes
integers.
To not potentially shift by 1 pixel on HiDPI monitors when switching
between wl_surface backend cursor sprites and built-in ones, make the
built in one emulate the restrictions put up by the Wayland protocol.
This also initializes the theme scale of the xcursor sprite instances to
1, as they may not have been set prior to being used, it'll only happen
in response to "prepare-at" signals being emitted prior to rendering.
Fixes: https://gitlab.gnome.org/GNOME/mutter/issues/1092https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1107
We checked that the content size was appropriately painted in the stage,
but didn't take into account that the size of the sampled texture
region, meaning that when stage views were scaled, we'd think that we
would draw a texture scaled, as e.g. a 200x200 sized texture with buffer
scale 2 would have the size 100x100. When stage views were not scaled,
we'd apply a geometry scale meaning it'd end up as 200x200 anyway, thus
pass the check, but when stage views are scaled, it'd still be painted
as a 100x100 shaped texture on the stage, thus failing the
are-we-unscaled test.
Fix this by comparing the transformed paint size with the sampled size,
instead of the paint size again, when checking whether we are being
painted scaled or not. For example, when stage views are scaled, our
200x200 buffer with buffer scale 2, thus content size 100x100 will
transform to a 200x200 paint command, thus passing the test. For
non-scaled stage views, our 200x200 buffer with buffer scale 2 thus
content size 100x100 will also transform into a 200x200 paint command,
and will also pass the check, as the texture sample region is still
200x200.
Fixes: https://gitlab.gnome.org/GNOME/mutter/issues/804https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1124
A user may have configured an output to be panning, e.g. using xrandr
--output <output> --mode <mode> --panning <size>. Respect this by making
the logical monitor use the panning size, instead of the mode. This
makes e.g. makes the background cover the whole panning size, and panels
etc will cover the whole top of the panned area, instead of just the top
left part covering the monitor if having panned to (0, 0).
No support is added to configuring panning, i.e. a panned monitor
configuration cannot be stored in monitors.xml.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1085
This is so that cogl-trace.h can start using things from cogl-macros.h,
and so that it doesn't leak cogl-config.h into the world, while exposing
it to e.g. gnome-shell so that it can make use of it as well. There is
no practical reason why we shouldn't just include cogl-trace.h via
cogl.h as we do with everything else.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1059
The upper layers (OSDs basically) want to know the monitor that a
tablet is currently assigned to, not the monitor just as configured
through settings.
This broke proper OSD positioning for display-attached tablets since
commit 87858a4e01, as the MetaInputMapper kicks in precisely when
there is no configured monitor for the given device.
Consulting both about the assigned output will make OSDs pop up
again in the right place.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/971
We used to inhibit all pad actions while the OSD is shown, but one we
would actually want to handle are mode switches while the OSD is open.
So it has an opportunity to catch up to the mode switch.
This lets MetaInputSettings reflect the mode switch (eg. when querying
action labels), so the OSD has an opportunity to update the current
actions.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/975
Commit cda9579034 fixed a corner case when setting the initial workspace
state of transient windows, but it still missed a case:
should_be_on_all_workspaces() returns whether the window should be on all
workspaces according to its properties/placement, but it doesn't take
transient relations into account.
That means in case of nested transients, we can still fail the assert:
1. on-all-workspaces toplevel
2. should_be_on_all_workspaces() is TRUE for the first transient's parent,
as the window from (1) has on_all_workspaces_requested == TRUE
3. should_be_on_all_workspaces() is FALSE for the second transient's
parent, as the window from (2) is only on-all-workspace because
of its parent
We can fix this by either using the state from the root ancestor
instead of the direct transient parent, or by using the parent's
on_all_workspaces_state.
The latter is simpler, so go with that.
https://gitlab.gnome.org/GNOME/mutter/issues/1083
This class sits between ClutterInputDevice and the backend implementations,
it will be the despositary of features we need across both backends, but
don't need to offer through Clutter's API.
As a first thing to have there, add a getter for a WacomDevice. This is
something scattered across and somewhat inconsistent (eg. different places
of the code create wacom devices for different device types). Just make it
here for all devices, so users can pick.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1109
Most people just see a harmless warning when applying this setting to
all touchpads (which this patch fixes). But tap[-and-drag] is supposed
to remain enabled for display-less Wacom tablets, despite configuration
changes.
Fix this by using the mapping function, so the setting is forced on for
wacom devices. This happens on a per-device basis, so the warning is
gone too.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1145
This fixes a case that was overlooked in
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1036 - when we
have a geometry scale > 1 and Wayland subsurfaces that have an offset
to their parent surface (which is often the case when the toplevel surface
includes decoration/shadows etc.), we have to add extra offset to their
opaque regions so they match their 'visible' location.
This is necessary as `meta_cullable_cull_out_children` moves the coordinate
system during culling, but does not know about geometry scale.
Also, remove the redundant check for `window_actor` - we only hit this code
path if a `window_actor` culls out its children.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1108
Some tablets like the Cintiq 24HDT have several mode switch buttons
per group. Those are meant to jump straight to a given mode, however
we just handle cycling across modes (as most other tablets have a
single mode switch button per group).
So spice up the mode switch handling so we handle multiple mode
switch buttons, assigning each of them a mode. If the device only
has one mode switch button, we do the old-fashioned cycling.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/970
This error was just logged but not raised. Do as the code comment said
and raise a pipe error at that moment, and for subsequent operations
on the output stream (although none besides close() should be expected
after propagating the error properly).
Related: https://gitlab.gnome.org/GNOME/mutter/issues/1065
When a page flip fails with a certain error code, we've treated this as
a hint that page flipping is broken and we should try to use mode
setting instead.
On some drivers, it seems that this error is also reported when there
was no mode set, which means we'll have no cached mode set to use in the
fallback. The lack of prior mode set tends to happen when we hit a race
when the DRM objects change before we have the time to process a hotplug
event.
Handle the lack a missing mode set in the flip fallback path, with the
assumption that we'll get a hotplug event that'll fix things up for us
eventually.
Closes: https://gitlab.gnome.org/GNOME/mutter/issues/917
Both IBus and ClutterInputFocus work in character offsets for the cursor
position in the preedit string. However the zwp_text_input protocol does
define the preedit string cursor offset to be in bytes.
Fixes client bugs in representing the caret within the preedit string,
as we were clearly giving the wrong offset.
Fixes: https://gitlab.gnome.org/GNOME/gtk/issues/2517https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1132
We send configure events for state changes e.g. for `appears-focused`,
etc. What we don't want to do is to do this for popup windows, as in
Wayland don't care about this state.
When the focus mode was configured to "sloppy focus" we'd get
`appears-focused` state changes for the popup window only by moving the
mouse cursor around, and while a popup may care about focus, it does not
care about related appearance, as there is no such state in xdg_popup.
What these state changes instead resulted in was absolute window
configuration events, intended for toplevel (xdg_toplevel) windows. In
the end this caused the popup to be positioned aginst at (0, 0) of the
parent window, as the assumptions when the configuration of the popup
was acknowledged is that it had received a relative position window
configuration.
Fix this by simply ignoring any state changes of the window if it is a
popup, meaning we won't send any configuration events intended for
toplevels for state changes. Currently we don't have any way to know
this other than checking whether it has a placement rule. Cleaning up
MetaWindow creation is left to be dealt with another day.
Fixes: https://gitlab.gnome.org/GNOME/mutter/issues/1103https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1122
If the CRTCs the cursor is visible on do not share a common scale
and transform, we can't use the cursor hardware plane as we only have one.
We therefore fall back to software / gl cursor.
The check for that currently happens after we tried to upload the cursor image
to the hardware plane though.
This is made worse by the fact that in the scaling step, where we scale the
cursor image to the desired size, until now we expected a valid common scale -
otherwise scaling the image by an uninitialized float.
Make sure we bail out early during the scale/upload step if we don't have common
scales and transforms - to avoid that bug and save some unnecessary work.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1125
Make sure it is only the special modifier (hardcoded to 1 currently)
which is being pressed (not counting locked modifiers) before notifying
that the special modifier is pressed, as we are interested in it being
pressed alone and not in combination with other modifier keys.
This helps in two ways:
- Pressing alt, then ctrl, then releasing both won't trigger the locate
pointer action.
- Pressing alt, then ctrl, then down/up to switch workspace won't interpret
the last up/down keypress as an additional key on top of the special ctrl
modifier, thus won't be forwarded down to the focused client in the last
second.
Closes: https://gitlab.gnome.org/GNOME/mutter/issues/812https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1014
If you first press a key that triggers the "special modifier key" paths
(ctrl, super), and then press another key that doesn't match (yet?) any
keybindings (eg. ctrl+alt, super+x), the second key press goes twice
through process_event(), once in the processing of this so far special
combination and another while we let the event through.
In order to keep things consistent, handle it differently depending on
whether we are a wayland compositor or not. For X11, consider the event
handled after the call to process_event() in process_special_modifier_key().
For Wayland, as XIAllowEvents is not the mechanism that allows clients see
the key event, we can just fall through the regular paths, without this
special handling.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1014
There is a race where an output can be used as a fullscreen target, but
it has already been removed due to a hotplug. Handle this gracefully by
ignoring said output in such situations.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1120
To keep consistent and avoid confusion, rename the function:
`meta_window_x11_buffer_rect_to_frame_rect()`
to:
`meta_window_x11_surface_rect_to_frame_rect()`
As this function doesn't deal with the `window->buffer_rect` at all.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1091
The code in `build_and_scan_frame_mask` predates the introduction of the
`MetaShapedTexture` API to get the texture width hand height.
Use the new `meta_shaped_texture_get_width/height` API instead of using
the CoGL paint texture.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1091
For X11 clients running on Wayland, the actual texture is set by
Xwayland.
The shape, input and opaque regions, however are driven by X11
properties meaning that those may come at a different time than the
actual update of the content.
This results in black areas being visible at times on resize with
Xwayland clients.
To make sure we update all the regions at the same time the buffer is
updated, update the shape, input and opaque regions when the texture is
committed from when the Xwayland surface state is synchronized.
That fixes the remaining black areas being sometimes visible when
resizing client-side decorations windows on Xwayland.
Closes: https://gitlab.gnome.org/GNOME/mutter/issues/1007https://gitlab.gnome.org/GNOME/mutter/merge_requests/1091
For X11 clients running on Xwayland, the opaque, input and shape regions
are processed from different properties and may occur at a different
time, before the actual buffer is eventually committed by Xwayland.
Add a new API `update_regions` to window actor to trigger the update of
those regions when needed.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1091
Commit 7dbb4bc3 cached the client area when the client was frozen.
This is not sufficient though, because the buffer size might still be
lagging waiting for the buffer from Xwayland to be committed.
So instead of caching the client size from the expected size, deduce the
client area rectangle from the surface size, like we did for the frame
bounds in commit 1ce933e2.
This partly reverts commit 7dbb4bc3 - "window-actor/x11: Cache the
client area"
https://gitlab.gnome.org/GNOME/mutter/issues/1007https://gitlab.gnome.org/GNOME/mutter/merge_requests/1091
Listen for GPU hotplug events to initialize their cursor support.
This fixes one reason for why DisplayLink devices may not be using a hardware
cursor. Particularly, when a DisplayLink device is hotplugged for the first
time such that EVDI creates a new DRM device node after gnome-shell has already
started, we used to forget to initialize the cursor support.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1097
Extract the code to initialize a single GPU cursor support into its own
function. The new function will be used by GPU hotplug in the future.
This is a pure refactoring without any behavioral changes.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1097
For every stream src, we created and attached a GSource. Upon stream
src destruction, we g_source_destroy():ed the GSource. What
g_source_destroy() does, hawever, is not really "destroy" it but only
detaches it from the main context removing the reference the context had
added for it via g_source_attach(). This caused the GSource to leak,
although in a detached state, as the reference taken on creation was
still held.
Fix this by also removing our own reference to it when finalizing.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1106
PipeWire will be unable to dequeue a buffer if all are already busy.
This can happen for valid reasons, e.g. the stream consumer not being
fast enough, so don't complain in the journal if it happens.
https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1115
While we will always have cursor planes, as we'll currently create fake
ones when real ones are missing (See #1058), eventually we will run into
situations where we can't create fake ones, for example for atomic KMS
drivers that don't advertise any cursor planes.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1079
If we don't force the placement, we enter the constrain machinery with
the position (0, 0), meaning we always get the "current work area" setup
to correspond to whatever logical monitor was at that position.
Avoid this by doing the same as "meta_window_force_placement()" and set
"window->calc_placement" to TRUE while move-resizing, causing the
move-resize to first calculate the initial position.
Fixes: https://gitlab.gnome.org/GNOME/mutter/issues/1098https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/1110
This commit completes the implementation of `xdg_wm_base` version 3,
which introduces support for synchronized implicit and explicit popup
repositioning.
Explicit repositioning works by the client providing a new
`xdg_positioner` object via a new request `xdg_popup.reposition`. If the
repositioning is done in combination with the parent itself being
reconfigured, the to be committed state of the parent is provided by the
client via the `xdg_positioner` object, using
`xdg_positioner.set__parent_configure`.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/705
This sets the `is_reactive` flag on the window placement rules, causing
the popups to be reconfigured as they are affected by environmental
changes, such as the parent moving in a way making the popup partially
offscreen.
As with synchronization, the implementation is dormant, as the
version of the advertised global isn't bumped yet, as the new protocol
version is not yet fully implemented.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/705
This commits adds support on the MetaWindow and constraints engine side
for asynchronously repositioning a window with a placement rule, either
due to environmental changes (e.g. parent moved) or explicitly done so
via `meta_window_update_placement_rule()`.
This is so far unused, as placement rules where this functionality is
triggered are not yet constructed by the xdg-shell implementation, and
no users of `meta_window_update_placement_rule()` exists yet.
To summarize, it works by making it possible to produce placement rules
with the parent rectangle a window should be placed against, while
creating a pending configuration that is not applied until acknowledged
by the client using the xdg-shell configure/ack_configure mechanisms.
An "temporary" constrain result is added to deal with situations
where the client window *must* move immediately even though it has not yet
acknowledged a new configuration that was sent. This happens for example
when the parent window is moved, causing the popup window to change its
relative position e.g. because it ended up partially off-screen. In this
situation, the temporary position corresponds to the result of the
movement of the parent, while the pending (asynchronously configured)
position is the relative one given the new constraining result.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/705
MetaGravity is an enum, where the values match the X11 macros used for
gravity, with the exception that `ForgetGravity` was renamed
`META_GRAVITY_NONE` to have less of a obscure name.
The motivation for this is to rely less on libX11 data types and macros
in generic code.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/705
A placement rule placed window positions itself relative to its parent,
thus converting between relative coordinates to absolute coordinates,
then back to relative coordinates implies unwanted restrictions for
example when the absolute coordinate should not be calculated againts
the current parent window position.
Deal with this by keeping track of the relative position all the way
from the constraining engine to the move-resize window implementation.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/705
To organize things a bit better, put the fields related to the placement
rule state in its own anonymous struct inside MetaWindow. While at it,
rename the somewhat oddly named variable that in practice means the
current relative window position.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/705
After popup placement rules have gone through the constraints engine has
ended up resulting in an actual move, pass the window configuration down
the path using relative coordinates, as that is what the next layer
(xdg-shell implementation) actually cares about.
In the future, this will also be helpful when the configured position is
not against the current state of the parent.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/705
A placement rule is always about placing a window relative to its
parent. In order to eventually place it against predicted future parent
positions, make the placement rule processing output relative
coordinates, having the caller deal with turning them into absolute.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/705
meta_window_wayland_finish_move_resize() inhibited window moves to be
finished if there was a resize grab active at the time, in order to
handle window resizing. Change this to only affect the grabbed window
itself, so that e.g. a popup can be positioned according to a pending
configuration while there is an active resize grab.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/705
This is made a signal, so the upper layers (read: gnome-shell) may
decide what services to spawn. The signal argument contains a task
that will resume MetaX11Display startup after it is returned upon.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/945
This is used by GDK and the X11 bits, but may also be used for
other initialization services we might need to run along with
Xwayland initialization.
However, as the -initfd argument in Xwayland is a fairly new
feature, add some meson build-time checks so that the feature
is handled transparently while allowing to explicitly set/unset
it.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/945
This shape region culling was wrongly implemented in f5a28aa9, as it
does not take frame offsets into account, and is also redundant, as
we already set the opaque region of the underlying surface accordingly.
The other parts were implemented in ac7aa114, the reason given in
the commit message:
```
Wayland clients do this through the opaque region in the surface
actor. However X11 clients were considered fully transparent for
culling purposes, which may result in mutter painting other bits
of the background or other windows that will be painted over in
reality.
```
is wrong though - culling on X11 actors works just fine and did only
not work in Wayland sessions because of a bug that got fixed in
19814497.
In conclusion the whole part appears to be redundand and some testing
done suggests the same. Drop it.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1082
If a opaque region is explicitly set we should not consider the surface
opaque, as that implies e.g. a shape region is set.
If no opque region is set but the texture does not have an alpha channel,
we can savely cull it out.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1082
We want sysprof's exact datadir for compatability with
platforms where software is installed into their own
individual immutable prefix's. Such that, mutter's prefix will
never equate to sysprof's. This depends on a MR in sysprof [0]
which adds datadir to its pkgconfig files, as these files will always
have the proper path we want.
This adds version a constraint on sysprof_dep, as datadir was added to
the .pc in this version.
[0]: https://gitlab.gnome.org/GNOME/sysprof/merge_requests/19https://gitlab.gnome.org/GNOME/mutter/merge_requests/957
Given that on Wayland we are pretty much guaranteed to finish MetaX11Display
setup after the MetaCompositor is enabled, we may drop the
meta_compositor_manage() x11 initialization bits, and move them into the
MetaX11Compositor subclass where it's actually needed.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/944
We artificially made Xwayland initialization synchronous, as we used
to rely on MetaX11Display and other bits during meta_display_open().
With support for Xwayland on demand and --no-x11, this is certainly
not the case.
So drop the main loop surrounding Xwayland initialization, and turn
it into an async operation called from meta_display_init_x11(). This
function is turned then into the high-level entry point that will
get you from no X server to having a MetaX11Display.
The role of meta_init() in Xwayland initialization is thus reduced
to setting up the sockets. Notably no processes are spawned from here,
deferring that till there is a MetaDisplay to poke.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/944
This ATM completes the task right away, but we will want to do
further things here that are asynchronous in nature, so prepare
for this operation being async.
Since the X11 backend doesn't really need this, make it go on
the fast lane and open the MetaX11Display right away, the case
of mandatory Xwayland on a wayland session is now handled
separately.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/944
With Xwayland initialization going async, these errors will seep
into the parts controlled by g_test*(), resulting in the harmless
errors about DBus names not acquired turned fatal.
Set an error log handler, and specifically ignore those.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/944
It might not be available right on initialization time if X11 is started
asynchronously. As this is a requirement for our tests, ensure it is there
before proceeding with the test.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/944
This used to be set on meta_compositor_manage(), but only if there is a
MetaX11Display. Given meta_display_init_x11() is Wayland only, and we can
always assume compositing to be enabled, just have it invariably set after
the X server is up.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/944
Even though cogl_framebuffer_flush() was supposed to be enough,
it ends up creating streams with odd visual glitches that look
very much like unfinished frames.
Switch back to cogl_framebuffer_finish(), which is admittedly
an overkill, but it's what works for now. There is anedoctal
evidence showing it doesn't incur in worse performance.
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1086
Much like monitor streaming, implement window streaming by
making the window actor draw itself with a paint context
that used the passed framebuffer.
Now that all MetaScreenCastStreamSrc subclasses implement
blit_to_framebuffer, remove the conditional check from
meta_screen_cast_stream_src_blit_to_framebuffer().
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1086
