After the introduction of the possibility to resize tiled windows,
it is a sensible decision to make windows aware of their tiling
match. A tiling match is another window that is tiled in such a
way that is the complement of the current window.
The newly introduced behavior attepts to make tiling as smooth as
possible, with the following rules:
* Windows now compute their tile match when tiling and, if there's
a match, they automatically complement the sibling's width.
* Resizing a window with a sibling automatically resizes the sibling
too to be the complement of the window's width.
* It is not possible to resize below windows' minimum widths.
https://bugzilla.gnome.org/show_bug.cgi?id=645153
Now that tiled windows are resizable, the user may grow a tiled
windows until it covers the entire work area. As this makes the
window state mostly indistinguishable from maximization, avoid
subtle differences by properly maximizing the window in that case.
https://bugzilla.gnome.org/show_bug.cgi?id=645153
Currently tiled windows are not resizable and their size is fixed
to half the screen width. Adjust the code to work with fractions
other than half, and allow users to adjust the split by dragging
the window edge that is not constrained by a monitor edge.
Follow-up patches will improve on that by resizing neighboring
tiled windows by a shared edge, and making the functionality
available to client-side decorated windows implementing the
new edge constraints protocol.
Now that the preview tile mode has been split from the window's
tile_mode property, it is much more natural to pass the requested
tile_mode to the tile() function instead of setting it externally
and calling the function to apply the state.
The existing semantics of the tile_mode property are terribly confusing,
as it depends on some other property whether it represents the requested
or current mode. Clear this up by just using separate variables for the
two. As it is unlikely that we will ever support more than one tile
preview, we can track the requested mode globally instead of adding
another per-window variable.
https://bugzilla.gnome.org/show_bug.cgi?id=645153
update_num_workspaces() is a no-op when the number of workspaces
did not actually change. That is fine, except that we still want
to initialize the _NET_NUMBER_OF_DESKTOPS hint on startup to not
break components like nautilus-desktop that rely on it.
https://bugzilla.gnome.org/show_bug.cgi?id=760651
Automatic maximization is done when a window is almost the size of the
work area of a monitor. This makes no sense to try when there is no
monitor available, so skip trying to do this when headless.
https://bugzilla.gnome.org/show_bug.cgi?id=787637
When we are headless, treat this as if the window is always not monitor
sized. This might cause windows to temporarly become redirected while
being headless, but this is harmless as when a new monitor is
connected, we'll recalculate weather it should be redirected or not.
https://bugzilla.gnome.org/show_bug.cgi?id=787637
When dynamic workspaces are used, we pick up an existing NUMBER_OF_DESKTOP
hint in meta_screen_init_workspaces() to properly restore workspaces
on restart. Unconditionally setting that hint to 1 *before* reading it
breaks that, and we end up shifting all windows to the first workspace.
This reverts commit 8532b10290.
https://bugzilla.gnome.org/show_bug.cgi?id=760651
Also adds a soft assert to meta_window_is_on_primary_monitor() to make
it easier to spot when callers might want to handle headless
in a certain way.
https://bugzilla.gnome.org/show_bug.cgi?id=730551
For now we abuse of meta_window_get_flatpak_id not to break the APIs,
so that it's working seamlessly in gnone shell too.
Rename flatpak_id to sandboxed_app_id internally to get prepared to the future
API.
https://bugzilla.gnome.org/show_bug.cgi?id=788217
If a non-latin based keyboard layout is active, for example Cyrillic,
keybindings won't work unless we resolve the bound keysyms on a
secondary latin based layout. So, to make keybindings work on non-latin
based layouts, detect if a keymap doesn't have all of the basic latin
letters (a-z) and resolve from an additional US layout as well.
https://bugzilla.gnome.org/show_bug.cgi?id=787016
Add the infrastructure to resolve keybinding symbols from multiple
layouts. It is still unused, but will be, when the primary layout does
not have the required latin keysyms in it.
https://bugzilla.gnome.org/show_bug.cgi?id=787016
Make sure to call set_number_of_spaces_hint in meta_screen_new.
_NET_NUMBER_OF_DESKTOPS is required by nautilus-desktop to correctly
get the desktop workarea.
https://bugzilla.gnome.org/show_bug.cgi?id=760651
This avoids updating state (such as position, size etc) when going
headless. Eventually, when non-headless, things will be updated again,
and not until then will we be able to update to a valid state.
https://bugzilla.gnome.org/show_bug.cgi?id=730551
If there are no active logical monitors, don't try to dereference a
NULL one to get a preferred output winsys id. Instead just set an
invalid one.
https://bugzilla.gnome.org/show_bug.cgi?id=730551
Add MetaFraction, which consists of two integers, the numerator an the
denominator. The utility function to convert a double to a MetaFraction
comes from gstreamer.
https://bugzilla.gnome.org/show_bug.cgi?id=784199
When resolving what keycodes a key binding resolves to, only look up
key codes from the current layout group. Without this, unwanted
overlaps may occur. For example when a keymap has both a dvorak and a
qwerty layout on different layout groups, one keybinding may be bound
on multiple keys, arbitrarily "shadowing" another.
https://bugzilla.gnome.org/show_bug.cgi?id=786408
Add a mechanism to MetaWaylandSurface that inhibits compositor's own
shortcuts when the surface has input focus, so that clients can receive
all key events regardless of the compositor own shortcuts.
This will help with implementing "fake" active grabs in Wayland and
XWayland clients.
https://bugzilla.gnome.org/show_bug.cgi?id=783342
Moved from g-s-d's media keys plugin, where it was called "video-out",
since it requires changing the current monitor configuration and we
want to remove the old DBus API.
This implementation is intentionally simple and not really meant for
more than debugging and validating the various configurations. A
better user experience will be introduced in gnome-shell with a custom
keybinding handler.
The default value includes <Super>P in addition to the standard keysym
for historical reasons.
https://bugzilla.gnome.org/show_bug.cgi?id=781906
When updating the main monitor, make sure to update the toplevel main
monitor before trying to use that as the main monitor for non-toplevel
windows (such as popups). Without this, when the main monitor is
updated as a side effect to monitors being changed (for example due to
a hot plug event, or coming back from being suspended) the
main monitor pointer may, after 'monitors-changed' has completed, point to
freed memory resulting in undefined behaviour.
https://bugzilla.gnome.org/show_bug.cgi?id=784867
This is used to request key focus on the close dialog whenever
a window that is frozen would receive key focus. Also, ensure
that the dialog gets focus when first shown if the window was
meant to receive input.
https://bugzilla.gnome.org/show_bug.cgi?id=762083
Otherwise the ClutterEventFilter will consider these handled, and not
forward these to Clutter. This gets necessary for key handling if we
mean to implement the close dialog with Clutter UI.
https://bugzilla.gnome.org/show_bug.cgi?id=762083
Just like we do for buttons, with a few twists. These have 2 directions
mappable to different keycombos, and are affected by the current mode
in their group.
https://bugzilla.gnome.org/show_bug.cgi?id=782033
Moved from g-s-d's media keys plugin, where it was called
"video-rotate", since it requires changing the current monitor
configuration and we want to remove the old DBus API.
https://bugzilla.gnome.org/show_bug.cgi?id=781906
Since commit 6b5cf2e, we keep override redirect windows on a layer
above regular windows in the clutter actor scene graph. In the X
server, and thus for input purposes, these windows might end up being
stacked below regular windows though, e.g. because a new regular
window is mapped after an OR window.
Fix this disconnect by re-stacking OR windows on top when syncing the
window stack with the compositor.
https://bugzilla.gnome.org/show_bug.cgi?id=780485
To be able to render the pointer cursor sprite at sub-(logical)-pixel
positions, track the pointer position using floats instead of ints.
This also requires users of the cursor sprite rect to deal with
floating points, when e.g. finding the logical monitor etc.
https://bugzilla.gnome.org/show_bug.cgi?id=765011
When moving a window between two non-adjecent logical monitors, don't
try to tile a window when the window position is outside of any logical
monitor.
https://bugzilla.gnome.org/show_bug.cgi?id=783630
When the number of (static) workspaces decreases, we relocate windows
from extra workspaces before removing them. As removing a non-empty
workspace is not allowed, we assert that it doesn't contain any windows
before removing it.
However that assert is
- pointless, because meta_workspace_remove() already asserts that
the workspace is empty
- wrong, because even empty workspaces contain windows that are set
to show on all workspaces
Simply drop the assert to avoid a crash when trying to remove a workspace
while on-all-workspaces windows are present.
https://bugzilla.gnome.org/show_bug.cgi?id=784223
While it doesn't make sense to set a window as transient to
itself, our existing check whether making a window transient
doesn't cover it, so it's still possible to create an infinite
loop.
https://bugzilla.gnome.org/show_bug.cgi?id=783502
For size change animations, plugins rely on the size change effect being
followed by size changed signal (or effects being kill altogether).
However unless the move_resize operation included the STATE_CHANGED flag,
the size changed event emitted when the compositor syncs the window
geometry only happens when the operation resulted in an actual change.
To avoid animations getting stuck in that case, make sure to include the
flag when tiling a window.
https://bugzilla.gnome.org/show_bug.cgi?id=783293
A single keysym can resolve to multiple keycodes. Instead of only using
the first one and ignoring the others, we store all codes in
MetaResolvedKeyCombo and then handle all of them in keybinding
resolution. If we already have bound a keycode for a keybinding with a
specific keysym then this can get overwritten by a new keybinding with a
different keysym that resolves to the same keycode. Now that we resolve
and bind all keycodes for a keysym this might happen more often; in that
case warn but still overwrite, but only for the first keycode for each
keysym. If a secondary (i.e. all non-first keycodes) is already indexed
we just ignore that; this should resemble the old behavior where we
only took the first keycode for any keysym as close as possible.
https://bugzilla.gnome.org/show_bug.cgi?id=781223
Call meta_compositor_size_change_window while tiling in order
to emit the size-change signal. Since the untiling action is
considered a unmaximize size change, treat tiling as a maximize
size change for consistency.
https://bugzilla.gnome.org/show_bug.cgi?id=782968
When terminating mutter running as a display server, don't try to resize
maximized windows when unmanaging, as at this point, they will have no
MetaWaylandSurface. Originally this was done instead of setting the
net_wm_state to not mess with future window managers, but when we're a
Wayland compositor, this does not matter.
https://bugzilla.gnome.org/show_bug.cgi?id=782156
This is an interface that can be used to implement the "application
is not responding" dialog. One instance is created per window, which
is initially hidden, and can be shown/hidden on demand.
https://bugzilla.gnome.org/show_bug.cgi?id=711619
The UI scaling depends on whether the framebuffers are scaled. Enable
the caller to determine the what scale its UI should be drawn in, in
relation to the stage coordinate space by calling this function. A new
singal "ui-scaling-factor-changed" is added in order to liston for for
changes.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
This commit adds support for rendering onto enlarged per logical
monitor framebuffers, using the scaled clutter stage views, for HiDPI
enabled logical monitors.
This works by scaling the mode of the monitors in a logical monitors by
the scale, no longer relying on scaling the window actors and window
geometry for making windows have the correct size on HiDPI monitors.
It is disabled by default, as in automatically created configurations
will still use the old mode. This is partly because Xwayland clients
will not yet work good enough to make it feasible.
To enable, add the 'scale-monitor-framebuffer' keyword to the
org.gnome.mutter.experimental-features gsettings array.
It is still possible to specify the mode via the new D-Bus API, which
has been adapted.
The adaptations to the D-Bus API means the caller need to be aware of
how to position logical monitors on the stage grid. This depends on the
'layout-mode' property that is used (see the DisplayConfig D-Bus
documentation).
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Logical monitors in a configuration must be adjecent to each other,
meaning there will be at least one pixel long side touching some other
logical monitor.
The exception to this is when there is only one logical monitor, which
cannot be adjecent to any other.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
When a state changed, e.g. a window went from unfullscreen to
fullscreen, always sync the window geometry, otherwise a compositor
application (e.g. gnome-shell) might end up with an unfinished window
state transition effect.
Without always syncing, the compositor plugin will see a 'size-change'
event, as a result of the state change, but if the size didn't change,
it would never see the 'size-changed' event. If an effect, for example
gnome-shell's fullscreen effect, is triggered on 'size-change' it might
rely on the actual size change to not get stuck. This commit allows it
to have this dependency.
This fixes a bug where a fullscreen effect gets "stuck" when a window
goes fullscreen without changing the window geometry.
https://bugzilla.gnome.org/show_bug.cgi?id=780292
Wayland windows can be zero sized until clients attach a buffer, but
our rectangle code doesn't deal with this case well, in particular,
meta_screen_get_monitor_for_rect() might end up choosing the wrong
monitor for a zero sized rectangle since
meta_rectangle_contains_rect() considers a zero sized rectangle at the
right or bottom edges of another rectangle (the monitor's) to be
contained there.
Since out size limits constraint will enforce a minimum size of 1x1,
we might as well enforce that when setting up the constraint info so
that the correct monitor gets chosen and the single monitor constraint
doesn't move these windows to the wrong one.
https://bugzilla.gnome.org/show_bug.cgi?id=772525
Wayland windows are initially zero sized until clients commit the
first buffer. Despite being invisible, clients are allowed to request
such windows to be fullscreened on a specific output before they
attach the first buffer which means we need to be able to move them.
meta_window_move_to_monitor() doesn't handle this case because these
windows' initial monitor is a placeholder since their initial
coordinates are 0,0+0+0, which results in us using a rectangle as
old_area for meta_window_move_between_rects() that might be to the "right"
of the window causing the move to go further out of the visible
screen's coordinates. This is later "corrected" by the constraints
system but the window might end up in the wrong monitor.
To fix this, we can make meta_window_move_between_rects() accept a
NULL old_area, meaning that we move the window to the new_area without
trying to keep a relative position.
https://bugzilla.gnome.org/show_bug.cgi?id=772525
The reason for the display to be closed may be meta_screen_new()
returning NULL, in which case we don't have a screen to free.
Avoid a segfault on exit by adding a proper check.
https://bugzilla.gnome.org/show_bug.cgi?id=778831
Split up the X11 backend into two parts, one for running as a
Compositing Manager, and one for running as a nested Wayland
compositor.
This commit also cleans up the compositor configuration calculation,
attempting to make it more approachable.
https://bugzilla.gnome.org/show_bug.cgi?id=777800
This signal provides the necessary information to let gnome-shell trigger
updates of pad leds/oleds whenever a pad group switches mode, and the
actions associated to buttons do too.
https://bugzilla.gnome.org/show_bug.cgi?id=776543
And add specific private methods to notify about tablet mapping and mode
switches. The signal allows the mutter side to trigger OSDs in a generic
way.
https://bugzilla.gnome.org/show_bug.cgi?id=771098
As all the relevant backends are expected to provide
ClutterPadButtonEvents, it makes no sense to split the information,
plus all other event fields are now available and might be needed
in the future.
https://bugzilla.gnome.org/show_bug.cgi?id=771098
Function "handle_raise_or_lower (src/core/keybindings.c)" is called
when running 'raise-or-lower' on a window. This function iterates
through all the windows in the stack to determine if our window is
already on top or obscured. The problem is that the window stack
includes windows in another workspaces and also windows that are
minimized.
https://bugzilla.gnome.org/show_bug.cgi?id=705200
Add private API for overriding the compositor configuration, i.e. the
compositor type (X11 WM or Wayland compositor) and backend type. This
will make it possible to add a special test backend used by src/tests/.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Put the monitor xinerama index in a separate struct that is attached to
the logical monitor using g_object_set/get_qdata(). Eventually this
should be moved to some "X11 window manager" object, but lets keep it
in MetaScreen until we have such a thing.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Don't store logical monitor specific state in an array where the index
from the monitor manager is used as index locally. Instead just use
table associating a logical monitor with a monitor specific state
holder, and store the state in there. This way we don't have the
workspace implementation relying on implementation details of other
units.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Instead of storing the logical monitors in an array and having users
either look up them in the array given an index or iterate using
indices, put it in a GList, and use GList iterators when iterating and
alternative API where array indices were previously used.
This allows for more liberty regarding the type of the logical monitor.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Rewrite check_fullscreen_func to not use indexes (and
offset-index-as-pointer) tricks. This also removes the usage of an API
constructing temporary logical monitor arrays carrying indices.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Move the last piece of monitor grid getter API to the monitor manager
away from MetaScreen. The public facing API are still there, but are
thin wrappers around the MetaMonitorManager API.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
The method used for getting the current logical monitor (the monitor
where the pointer cursor is currently at) depends on the backend type,
so move that logic to the corresponding backends.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Turning a rectangle into a logical monitor also has nothing to do with
the screen (MetaScreen) so move it to MetaMonitorManager which has that
information.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Let the backend initialize the cursor tracker, and change all call
sites to get the cursor tracker from the backend instead of from the
screen. It wasn't associated with the screen anyway, so the API was
missleading.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Instead of keeping around array indexes, keep track of them by storing
a pointer instead. This also changes from using an array (imitating the
X11 behaviour) to more explicit storing.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
To complement the current API which takes an index referencing a
logical monitor in the logical monitor array, add API that takes a
direct reference to the logical monitor itself. The intention is to
replace the usage of the index based API with one that doesn't rely on
internal implementation details.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
It was just pointer to the actual list; having to synchronize a list of
logical monitors with the actual monitors managed by the backend is
unnecessary.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
The fullscreen monitors state is set given a set of xinerama monitor
identification numbers. When the monitor configuration changes (e.g. by
a hotplug event) these are no longer valid, and may point to
uninitialized or unallocated data. Avoid accessing
uninitialized/unallocated memory by clearing the fullscreen monitor
state when the monitor configuration changes.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
In preparation for further refactorizations, rename the MetaMonitorInfo
struct to MetaLogicalMonitor. Eventually, part of MetaLogicalMonitor
will be split into a MetaMonitor type.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
We currently only focus unfocused windows on button press if no
modifiers (or just ignored modifiers) are in effect. This behavior
seems surprising and counter-intuitive so let's do it for any modifier
combination instead.
https://bugzilla.gnome.org/show_bug.cgi?id=746642
There's no reason to keep this ~15 year old piece of code around as
well as the preference handling that would only make sense if this
hunk was actually enabled.
https://bugzilla.gnome.org/show_bug.cgi?id=746642
A window's unconstrained_rect is essentially just the target rectangle
we hand to meta_window_move_resize_internal() except it's not updated
until the window actually moves or resizes.
As such, for wayland client resizes, since they're async, using
window->unconstrained_rect right after calling move_resize_internal()
to update the grab anchor position on unmaximize doesn't work as it
does for X clients.
To fix this, we can just use the target rectangle for the grab
anchor. Note that comment here was already wrong since it says we
should be taking constraints into account and yet the code used the
unconstrained rect anyway.
https://bugzilla.gnome.org/show_bug.cgi?id=770345
In order for the compositor plugin to be able to animate window size
changes properly we need to let it know of the starting and final
window sizes.
For X clients this can be done synchronously and thus with a single
call into the compositor plugin since it's us (the window manager)
who's in charge of the final window size.
Wayland clients though, have the final say over their window size
since it's determined from the client allocated buffer.
This patch moves the meta_compositor_size_change_window() calls before
move_resize_internal() which lets the compositor plugin know the old
window size and freezes the MetaWindowActor.
Then we get rid of the META_MOVE_RESIZE_DONT_SYNC_COMPOSITOR flag
since it's not needed anymore as the window actor is frozen and that
means we can use meta_compositor_sync_window_geometry() as the point
where we inform the compositor plugin of the final window size.
https://bugzilla.gnome.org/show_bug.cgi?id=770345
And remove the wayland-specific handling. This works for both Wayland and
X11 (provided the compositor receives pad events through a passive grab
there).
https://bugzilla.gnome.org/show_bug.cgi?id=773779
We kind of rely on the ::show-pad-osd handler to destroy the
previous actor. Just prevent the emission of multiple signals
till the actor has been destroyed.
https://bugzilla.gnome.org/show_bug.cgi?id=771067
Commit fcc7501eb8 had the side-effect of
stacking fullscreen windows below docks which went unnoticed since we
don't use docks in GNOME anymore.
Instead of re-introducing the fullscreen layer, which we don't need
otherwise, we can fix this issue by ensuring we stack docks below all
other windows when the monitor they're on is marked fullscreen. This
has the added benefit that the visibility rule for 3rd party docks
becomes the same as gnome-shell's chrome.
https://bugzilla.gnome.org/show_bug.cgi?id=772937
mutter would remove focus from a toplevel when showing one of its
transient window which is not on top and not focused.
When using xdg_popup without grab as allowed in xdg_shell v6, the popup
wouldn't be focused, and if an intermediate event occurs before the
popup is shown, it's not placed on top either, which could randomly
trigger a loss of focus in the corresponding toplevel window.
Remove that special case, it doesn't make much sense to globally unset
focus when mapping a new window.
https://bugzilla.gnome.org/show_bug.cgi?id=773210
The frame rect will at this point not be set for Wayland popups, since
the popup is placed and constrained before the actual buffer will be
attached. To still be able to calculate a proper monitor to be used for
constraining, use the ConstraintInfo::current dimensions instead, since
they will have the expected size. This should not cause any issues with
present paths since when a window is otherwise placed, it usually
doesn't change monitor calculation result.
This fixes opening a popup menu that would be positioned on the left
edge of a not-left-most monitor, for example a 'File' menu on a window
maximized on a second monitor.
https://bugzilla.gnome.org/show_bug.cgi?id=773141
The order doesn't only affect the visual layout, but also which action
cancels the dialog (and therefore responds to Escape). It is completely
surprising that this triggers a destructive action like force-quit, so
swap the actions to wait when the dialog is cancelled.
https://bugzilla.gnome.org/show_bug.cgi?id=737109
GNOME Shell's window matching currently fails frequently with Flatpak
applications, as one of the primary hints used to link windows with
.desktop files - the WM_CLASS - no longer matches when flatpak renames
the exported .desktop file. Luckily, Flatpak provides us with a fail-safe
way to map from the PID to the corresponding application ID, so expose an
appropriate method that allows GNOME Shell to reliably match windows to
the corresponding Flatpak app.
https://bugzilla.gnome.org/show_bug.cgi?id=772614
In order to kill a window, on both X11 and wayland we first try to
kill(3) the corresponding process, so we can add the newly added
get_client_pid() method to share that code.
https://bugzilla.gnome.org/show_bug.cgi?id=772613
It is often useful to identify the client process that created
a particular window, however the existing meta_window_get_pid()
method relies on _NET_WM_PID, which is only available on X11 and
depends on applications to set it correctly (which may not even
be possible when the app runs in its own PID namespace as Flatpak
apps do). So add a get_client_pid() method that uses windowing
system facilities to resolve the PID associated with a particular
window.
https://bugzilla.gnome.org/show_bug.cgi?id=772613
When a modal transient is unmanaging, most likely the parent of the
modal transient should be focused.
In Wayland, a MetaWindow is created when a shell surface role (like
xdg_toplevel) is created, but a window cannot be shown until a buffer
is attached. If a client would create two modal transients and make
them both have the same parent, but only one get a buffer attached
(i.e. shown), when unmanaging the modal transient that was showing,
when finding a new focus candidate, the stacking code will ignore the
not-to-be-shown buffer-less modal transient when finding a good
candidate for focusing. In the case described here, this means it will
find the parent of the unmanaging modal transient.
This newly chosen candidate will then be passed to meta_window_focus();
meta_window_focus() will then try to find any modal transient to focus
instead, will find the one without any buffer, then fail to focus it
because it cannot be mapped, thus making meta_window_focus() not focus
anything. Since meta_window_focus() didn't change any focus state, the
assert in meta_window_unmanage() checking that the unmanaging window
isn't focused anymore will be hit, causing mutter to abort.
For now, fix this by checking whether the modal transient can actually
be focused in meta_window_focus(). For X11 client windows, a window
will be defined to be focusable always, but for Wayland client windows,
a window will be determined focusable only if it has a buffer attached.
In the future, we should probably do a more thorough refactorization of
focus handling to get rid of any X11 - Wayland differences.
https://bugzilla.gnome.org/show_bug.cgi?id=757568
Even without a compositor grab, key events may still be expected to
be processed by the compositor and not applications, for instance
when using ctrl-alt-tab to keynav in the top bar. On X11, focus is
moved to the stage window in that case, so that events are processed
before they are dispatched by the window manager. On wayland, we need
to handle this case ourselves, so make sure to not pass key events to
wayland in that case, and move the key focus back to the stage when
appropriate.
https://bugzilla.gnome.org/show_bug.cgi?id=758167
For some reason, when a modal dialog was made an attaching
transient-for, if the window wasn't "constructing", it would be
unmanaged and rely on some side effect to be recreated. This side
effect is not triggered for Wayland clients, thus if one happen to set
a surface as "modal" via gtk_surface.set_modal before
xdg_toplevel.set_parent, it'd be unmanaged and never show up.
Instead, simply just set the tranciency anyway for Wayland clients.
This makes GTK+ clients that set_modal() before set_transient_for()
work.
https://bugzilla.gnome.org/show_bug.cgi?id=770324
Windows from Xwayland still needs to use the Wayland path, but is
represented an MetaWindowX11, thus the abstraction introduced in
"window: Make meta_window_has_pointer() per protocol implemented"
is wrong. Lets turn back time, and reconsider how this can be
abstracted more correctly in the future.
This reverts commit 9fb891d216.
Add support for assigning a window a custom window placement rule used
for calculating the initial window position as well as defining how a
window is constrained.
The custom rule is a declarative rule which defines a set of parameters
which the placing algorithm and constrain algorithm uses for
calculating the position of a window. It is meant to be used to
implement positioning of menus and other popup windows created via
Wayland.
A custom placement rule replaces any other placement or constraint
rule.
https://bugzilla.gnome.org/show_bug.cgi?id=769936
There may be external/compositor-specific reasons to trigger the
pad OSD. Expose this call so the pad OSD can be triggered looking
up the right settings, monitor, etc...
This API will be used from the gnome-shell pad OSD implementation, in order
to show the actions that currently apply to every button/ring/strip in the
tablet.
When launching a GNOME session from a text-mode VT, the logind session
type is unlikely to be set to either "wayland" or "x11". We search for a
supported session type first with logind and then with
$XDG_SESSION_TYPE. As a fallback, we also test $DISPLAY in case of a
"tty" logind session to support starting through xinit. Ideally, such
setups should set XDG_SESSION_TYPE=x11.
If no supported session type is found, we throw an error.
https://bugzilla.gnome.org/show_bug.cgi?id=759388
They are already effectively interchangeable so this should reduce
pointless casts.
Just like in GDK though, we need to keep the old definition for
instrospection to be able to include the struct's fields.
Add support for drawing a stage using multiple framebuffers each making
up one part of the stage. This works by the stage backend
(ClutterStageWindow) providing a list of views which will be for
splitting up the stage in different regions.
A view layout, for now, is a set of rectangles. The stage window (i.e.
stage "backend" will use this information when drawing a frame, using
one framebuffer for each view. The scene graph is adapted to explictly
take a view when painting the stage. It will use this view, its
assigned framebuffer and layout to offset and clip the drawing
accordingly.
This effectively removes any notion of "stage framebuffer", since each
stage now may consist of multiple framebuffers. Therefore, API
involving this has been deprecated and made no-ops; namely
clutter_stage_ensure_context(). Callers are now assumed to either
always use a framebuffer reference explicitly, or push/pop the
framebuffer of a given view where the code has not yet changed to use
the explicit-buffer-using cogl API.
Currently only the nested X11 backend supports this mode fully, and the
per view framebuffers are all offscreen. Upon frame completion, it'll
blit each view's framebuffer onto the onscreen framebuffer before
swapping.
Other backends (X11 CM and native/KMS) are adapted to manage a
full-stage view. The X11 CM backend will continue to use this method,
while the native/KMS backend will be adopted to use multiple view
drawing.
https://bugzilla.gnome.org/show_bug.cgi?id=768976
Make it possible to force mutter to start as a X11 compositing/window
manager. This is needed when intending to start mutter as an X11 window
manager while running inside a Wayland session, for example when
intending to debug it in Xephyr.
https://bugzilla.gnome.org/show_bug.cgi?id=768976
Introduce two new clutter backends: MetaClutterBackendX11 and
MetaClutterBackendNative. They are so far only wrap ClutterBackendX11
and ClutterBackendEglNative respectively, but the aim is to move things
from the original clutter backends when needed.
https://bugzilla.gnome.org/show_bug.cgi?id=768976
This layer isn't really being used and in fact, it causes
meta_stack_get_default_focus_window() to return a fullscreen window
even if the naturally topmost window in the stack isn't a fullscreen
one.
Note that commit a3bf9b01aa changed how
we choose the default focus window from the MRU to the topmost in the
stack.
https://bugzilla.gnome.org/show_bug.cgi?id=768221
When restarting (X compositor only, obviously), we want to keep
the same window focused. There is code that tries to do this by
calling XGetInputFocus() but the previously focused window will
almost certainly not still be focused by the time we get to the
point where we call XGetInputFocus(), and in fact, probably was
no longer correct after the previous window manager exited, so
the net result is that we tend to focus no window on restart.
A better approach is to leave the _NET_ACTIVE_WINDOW property
set on the root window during exit, and if we find it set when
starting, use that to initialize focus.
https://bugzilla.gnome.org/show_bug.cgi?id=766243
Emit a signal so that interested parties can recreate their FBOs and
queue a full scene graph redraw to ensure we don't end up showing
graphical artifacts.
This relies on the GL driver supporting the
NV_robustness_video_memory_purge extension and cogl creating a
suitable GL context. For now we only make use of it with the X backend
since the only driver with which this is useful is NVIDIA.
https://bugzilla.gnome.org/show_bug.cgi?id=739178
printf string precision counts bytes so we may end up creating invalid
UTF-8 strings here. Instead, use glib's unicode aware methods to clip
the title.
https://bugzilla.gnome.org/show_bug.cgi?id=765535
Stacking hidden X windows below the guard window is a necessity to
ensure input events aren't delivered to them. Wayland windows don't
need this because the decision to send them input events is done by us
looking at the clutter scene graph.
But, since we don't stack hidden wayland windows along with their X
siblings we lose their relative stack positions while hidden. As
there's no ill side effect to re-stacking hidden wayland windows below
the X guard window we can fix this by just doing it regardless of
window type.
https://bugzilla.gnome.org/show_bug.cgi?id=764844
If we try to send notify event (either from surface_state_changed()
or from meta_window_wayland_move_resize_internal()),
we will crash, because we don't have a sufrace anymore.
There's no reason why to resize the window that is being
unmanaged anyway.
https://bugzilla.gnome.org/show_bug.cgi?id=751847
meta_parse_accelerator() considers 0 length accelerator strings as
valid, meaning that the keybinding should be disabled. Unfortunately,
it doesn't initialize the MetaKeyCombo so if the caller doesn't
initialize it either, we end up using random values and possibly
grabbing random keys.
https://bugzilla.gnome.org/show_bug.cgi?id=766270
Before this commit, on Wayland, the buffer rect would have the size of
the attached Wayland buffer, no matter the scale. The scale would then
be applied ad-hoc by callers when a sane rectangle was needed. This
commit changes buffer_rect to rather represent the surface rect (i.e.
what is drawn on the stage, including client side shadow). The users of
buffer_rect will no longer need to scale the buffer_rect themself to
get a usable rectangle.
https://bugzilla.gnome.org/show_bug.cgi?id=763431
Since g_array_append_val isn't smart enough to do a proper upcast, we
have to do it manually, lest we get junk.
This fixes various RAISE_ABOVE: window not in stack: 0x8100c8003
warnings that appear on 32-bit systems.
Each wl_surface.commit with a newly attached buffer should result in
one wl_buffer.release for the attached buffer. For example attaching
the same buffer to two different surfaces must always result in two
wl_buffer.release events being emitted by the server. The client is
responsible for counting the wl_buffer.release events and be sure to
have received as many release events as it has attached and committed
the buffer, before reusing it.
https://bugzilla.gnome.org/show_bug.cgi?id=762828
On the X11 backend we don't track the pointer position in
priv->current_x/y which remain set to zero. That means we never set
the clutter stage cursor if point 0,0 isn't covered by any monitor
since we return early.
Commit 4bebc5e5fa introduced this to
avoid crashing on the prepare-at handlers when the cursor position
doesn't fall inside any monitor area but we can handle that higher up
in the stack. In that case, the sprite's scale doesn't matter since
the cursor won't be shown anyway so we can skip setting it.
https://bugzilla.gnome.org/show_bug.cgi?id=763159
CSD X11 clients and Wayland clients don't have a window frame drawn by
the compositor to flash. So instead of flashing the whole screen when
configured to just flash the window, flash just the window region.
https://bugzilla.gnome.org/show_bug.cgi?id=763284
To support invoking the system bell on Wayland we shouldn't have paths
that fallback to X11. Let the X11 caller deal with the absence of
libcanberra, and change API to not take any X events.
https://bugzilla.gnome.org/show_bug.cgi?id=763284
The libsn API provides its timestamps in the "Time" X11 type, which is
usually is a typedef for "unsigned long". The type of the "timestamp"
parameter of StartupNotificationSequence is a signed 64 bit integer.
When building on an architecture where a "unsigned long" is not 64 bit,
we'd then pass a 32 bit unsigned integer via a va_list where a signed 64
bit integer is expected causing va_arg to read past the passed 32 bit
unsigned integer.
Fix this by ensuring that we always pass the expected type via the
va_list. Also change the internal timestamp type from time_t (which
size is undefined) to gint64, to avoid any potential overflow issues.
https://bugzilla.gnome.org/show_bug.cgi?id=762763
If a MetaLater callback queued another MetaLater with a scheduling
later than the one currently being invoked, make it so that the newly
scheduled callback will actually be invoked.
The fact that it doesn't already do this is a regression from
cd7a968093.
https://bugzilla.gnome.org/show_bug.cgi?id=755605
As of "core: start as wayland display server when
XDG_SESSION_TYPE=wayland" it is no longer possible to run a nested
mutter Wayland session on top of another Wayland session. This patch
adds a command line argument to make it possible to force mutter to
start as a nested compositor.
https://bugzilla.gnome.org/show_bug.cgi?id=758658
This is kind of in a middle ground at the moment. Even though it
handles sequences not coming from libsn, they're added nowhere at
the moment, we'll rely on the app launch context being in the x11
side at the moment.
Also, even though we do create internal sequence objects, we keep
exposing SnStartupSequences to make gnome-shell happy, we could
consider making this object "public" (and the sequence objects with
it), things stay private at the moment.
https://bugzilla.gnome.org/show_bug.cgi?id=762268
If a broken or naughty application tries set up its windows to create
a loop in the transient relationship, mutter will hang, looping forever
in meta_window_foreach_ancestor()
To avoid looping infinitely at various point in the code, check for a
possible loop when setting the transient relationship and deny the
request to set a window transient for another if that would create a
loop.
Bugzilla: https://bugzilla.gnome.org/show_bug.cgi?id=759299
In order to reuse some vector math for pointer confinement, move out
those parts to its own file, introducing the types old types
"MetaVector2" and "MetaLine2" outside of meta-barrier-native.c, as well
as introducing MetaBorder which is a line, with a blocking direction.
https://bugzilla.gnome.org/show_bug.cgi?id=744104
The new tiling code, instead of based around "tiling states", is instead
based around constrained edges. This allows us to have windows that have
three constrained edges, but keep one free-floating, e.g. a window tiled
to the left has the left, top, and bottom edges constrained, but the
right edge can be left resizable.
This system also is easily extended to support corner tiling. We also,
using the new "size state" system, also keep normal, tiled, and
maximized sizes independently, allowing the maximize button to bounce
between maximized and tiled states without reverting to normal in
between. Dragging from the top will always restore the normal state,
though.
https://bugzilla.gnome.org/show_bug.cgi?id=751857
In case a window is hidden when we're ordered to make it transient to
a different parent we must re-evaluate its visibility status or we'll
get into an inconsistent state where the parent is visible and the
child isn't.
https://bugzilla.gnome.org/show_bug.cgi?id=759297
This seems like a more generally useful and intuitive behavior. Note
that, in X sessions, this is what already happened in practice since
meta_display_begin_grab_op() calls meta_window_grab_all_keys() which,
on X11, does meta_window_focus().
https://bugzilla.gnome.org/show_bug.cgi?id=756789
Don't update the stack until after setting the window->transient_for
field. Updating before will cause the stack transient-for constraint to
be missing until the next time constraints are applied.
https://bugzilla.gnome.org/show_bug.cgi?id=755606
Wine removes the minimize func from its Motif hints on full-screen
windows, because, as the Win32 API literally says, the minimize button
is indeed not visible on full-screen windows.
Given that this code was added to prevent minimizing a panel by
accident, I don't necessarily think that it's relevant anymore.
https://bugzilla.gnome.org/show_bug.cgi?id=758186
Unsetting it in meta_display_handle_event() will make the pointer
emulation checks fail on TOUCH_END event handlers across clutter
actors, the sequence should still be considered as pointer emulating
at that time.
As we don't have a way to hook this post clutter event handling,
instead unset/reset it lazily on the next pointer emulating TOUCH_BEGIN
event, the checks would already fail on other sequences, even if the
pointer emulating touch ended earlier. The only extra thing we need
to take care about is sequence collision, at which point it's safe to
just unset the stored sequence if its new incarnation isn't flagged/
deemed as pointer emulating.
https://bugzilla.gnome.org/show_bug.cgi?id=756754
When managing window, we queue showing the window.
Under wayland, if we commit surface quickly enough,
the showing is unqueued and commit procedure takes care
of mapping and placing the window. In the oposite case,
queue is processed before client sets all we need and
then we have wrong size of window, which leads to broken placement.
Therefore force placement in queue only if the window should already
be mapped. If it is not mapped, we don't care where it is anyway.
https://bugzilla.gnome.org/show_bug.cgi?id=751887
We have been ignoring those buttons since 3.16 after they had been
broken in the default theme for a couple of versions. As nobody
appears to miss them, it's time to remove them for good.
Displaying all Wayland windows with the XID of 0x0 makes it hard
to figure out what is going on ... use the recently-added
window->stamp to show Wayland windows as W1/W2/W3...
This commits refactors cursor handling code and plugs in logic so that
cursor sprites changes appearance as it moves across the screen.
Renderers are adapted to handle the necessary functionality.
The logic for changing the cursor sprite appearance is done outside of
MetaCursorSprite, and actually where depends on what type of cursor it
is. In mutter we now have two types of cursors that may have their
appearance changed:
- Themed cursors (aka root cursors)
- wl_surface cursors
Themed cursors are created by MetaScreen and when created, when
applicable(*), it will extend the cursor via connecting to a signal
which is emitted everytime the cursor is moved. The signal handler will
calculate the expected scale given the monitor it is on and reload the
theme in a correct size when needed.
wl_surface cursors are created when a wl_surface is assigned the
"cursor" role, i.e. when a client calls wl_pointer.set_cursor. A
cursor role object is created which is connected to the cursor object
by the position signal, and will set a correct texture scale given what
monitor the cursor is on and what scale the wl_surface's active buffer
is in. It will also push new buffers to the same to the cursor object
when new ones are committed to the surface.
This commit also makes texture loading lazy, since the renderer doesn't
calculate a rectangle when the cursor position changes.
The native backend is refactored to be triple-buffered; see the comment
in meta-cursor-renderer-native.c for further explanations.
* when we are running as a Wayland compositor
https://bugzilla.gnome.org/show_bug.cgi?id=744932
Before, it used to be in the screen, but now,
meta_cursor_reference_from_theme can never fail. Move it to where we
load the images from the cursor name.
This was introduced in commit c6793d477a
to prevent window self-maximisation. It turns out that that bug seems
to have been fixed meanwhile in a different way since the reproducer
in https://bugzilla.gnome.org/show_bug.cgi?id=461927#c37 now works
fine with this special handling removed.
In fact, failing to set window->fullscreen immediately when loading
the initial set of X properties causes us to create a UI frame for a
window that sets _NET_WM_STATE_FULLSCREEN.
This, in turn, might cause the fullscreen constrain code to fail if
the window also sets min_width/min_height size hints to be the monitor
size since the UI frame size added to those makes the rectangle too
big to fit the monitor. If the window doesn't set these hints, we
fullscreen it but the window will get sized such that the UI frame is
taken into account while it really shouldn't (see the reproducer
above).
https://bugzilla.gnome.org/show_bug.cgi?id=753020
Since commit 14b0a83f64 we store the
main window monitor instead of computing it every time. This means
that we must now ensure that it's updated before trying to use it
which we do from meta_screen_resize_func() or else we'll crash on an
assertion later on when removing a monitor:
assertion failed: (which_monitor < workspace->screen->n_monitor_infos)
https://bugzilla.gnome.org/show_bug.cgi?id=752674
They otherwise fall through paths that enable bypass_clutter, this
is necessary so they can be picked by captured-event handlers
along the actor hierarchy.
https://bugzilla.gnome.org/show_bug.cgi?id=752248
Since we scale surface actors given what main output their toplevel
window is on, also scale the window geometry coordinates and sizes
(window->rect size and window->custom_frame_extents.top/left) in order
to make the window geometry represent what is being rendered on the
stage.
https://bugzilla.gnome.org/show_bug.cgi?id=744934
The main monitor of a window is maintained as 'window->monitor' and is
updated when the window is resized or moved. Lets avoid calculating it
every time it`s needed.
https://bugzilla.gnome.org/show_bug.cgi?id=744934
A much less hacky version of maximize / unmaximize is reimplemented
in terms of this, but it could also eventually be used for fullscreen /
unfullscreen, and tile / untile.
The only time we ever execute this code is when we're minimizing or
hiding a window, in which case we should respect stacking order.
This fixes weird "bugs" where windows from the same app magically pop up
over other windows.
This is an extremely niche feature, and conflicts with the rest of our
interface being consistent about not allowing resizing while tiled or
maximized.
A window may be hidden even if not minimized itself, for instance
when an ancestor is minimized. As meta_window_focus() will refuse
to actually focus the window in that case, don't pick it in the first
place.
https://bugzilla.gnome.org/show_bug.cgi?id=751715
Going from fullscreen to unfullscreen involves a frame border size, so
in order to properly interpret the saved rect size, we need to make sure
that the frame borders are fully up to date.
The "calc showing" operation is queued in a few places alongside MetaWindow
creation, we should be ignoring these until there is a buffer to show.
https://bugzilla.gnome.org/show_bug.cgi?id=750552
window->is_alive isn't initialized explicitly so it defaults to FALSE
meaning that if the first ping fails we'd short circuit and not show
the delete dialog as we should.
We could initialize the variable to TRUE but in fact we don't even
need the variable at all since our dialog management is enough to
manage all the state we need, i.e. we're only interested in knowing
whether we're already displaying a delete dialog.
This does change our behavior here since previously we wouldn't
display the dialog again if the next ping failed after the dialog is
dismissed but this was arguably a bug too since in that case there
wouldn't be a way to kill the window after waiting for a while and the
window kept being unresponsive.
https://bugzilla.gnome.org/show_bug.cgi?id=749711
When running as an X11 compositor we do this for every event we see on
the X event stream. As a wayland compositor we don't go through that
code path but since we see all events we can easily do this on motion
events.
In fact, we don't even need this caching when we're a wayland
compositor since we can always find where the pointer is without a
round trip but we're sharing the current monitor logic with the X
path so let's keep it as is for now.
https://bugzilla.gnome.org/show_bug.cgi?id=748478
The existing private get_monitor_neighbor() function returns a
MetaMonitorInfo, which is private as well. Add a public wrapper
that returns a monitor index instead, as we do for other public
monitor-related methods.
https://bugzilla.gnome.org/show_bug.cgi?id=633994
Since the frame is the window that's redirected, there's no reason for
it to match the root window. There *is*, however, a big incentive to
match the window's visual, since not doing so might trigger automatic
redirection.
On a specific platform, we construct a depth-32 root window, and stick a
depth-24 child window inside it. The frame ends up being created
depth-32, not depth-24, so we get automatic redirection.
clutter currently never emits activated or deactivated signals on
the stage object when using the EGL backend. Since the stage never
gets activated, accessibility tools, like orca, don't work.
This commit makes mutter take on the responsibility, by tracking
when the stage gains/loses focus, and then synthesizing stage
CLUTTER_STAGE_STATE_ACTIVATED state events.
A limitation of this approach is that clutter's own notion of
the stage activeness won't reflect mutter's notion of the
stage activeness. This isn't a problem, in practice, and can
be addressed in the medium-term after making changes to
clutter.
https://bugzilla.gnome.org/show_bug.cgi?id=746670
With all input events being handled through clutter, this only confuses
things, and most nominally, coerces touch events through places we didn't
intend to, like the window frame.
This makes again all touch events only handled in the passive grab on X11,
while the rest stays pointer (emulated) only.
https://bugzilla.gnome.org/show_bug.cgi?id=745335
In order to switch to the correct surface actor scale given the monitor
the surface is on, without relying on the client committing a new state
given some other side effect, sync the surface actor state when the main
monitor associated with the corresponding window changed.
https://bugzilla.gnome.org/show_bug.cgi?id=744933
This just exposes the type and the singleton getter necessary to make
it available to introspection. We'll expose more functionality as it
becomes needed.
https://bugzilla.gnome.org/show_bug.cgi?id=743745
We're locked to frame sync anyway, so it doesn't make sense to try to
redraw early. In casual testing, this seems to actually make things
faster, as well.
When looking for space to place a new window, other non-minimized
windows on the same workspace should be taken into account. However
the current check does not work correctly when the placed window is
located on all workspaces, so handle that case explicitly.
https://bugzilla.gnome.org/show_bug.cgi?id=743217
When running as a dispay server pointer barriers are a server side
feature and requires no client interaction of any sort. This patch
implements pointer barriers that can be used when running as a display
server on the native backend. Running as a display server using the X11
backend is currently not supported.
https://bugzilla.gnome.org/show_bug.cgi?id=706655
This patch removes the X11 specific code from MetaBarrier and creates an
abstraction layer MetaBarrierImpl. The existing X11 implementation is
moved to a new GObject MetaBarrierImplX11 implementing the abstract
interface MetaBarrierImpl which is instantiated by MetaBarrier when
supported.
While at it, move it to backends/ and properly name the files.
https://bugzilla.gnome.org/show_bug.cgi?id=706655
MetaKeyCombo is about the *unresolved* keybinding, which can either be a
"keysym" (<Ctrl>F) or a "keycode" (<Ctrl>0x21). When we resolved the
keysym to a keycode, we stuffed it back in the same MetaKeyCombo, which
confused about what the "keycode" field was for. Thus, we often stomped
on the user's explicit choice if they chose a keycode binding value.
To solve this, create a separate structure, the "devirtualized key combo"
or MetaKeyDevirtCombo, which contains a resolved keycode from the
keysym, and a devirtualized modifier mask. The MetaKeyCombo is now
always a "source" value, and the MetaKeyDevirtCombo is now always what
the user chose.
This also lets us significantly clean up the overlay and ISO key binding
paths.
The reason MetaKeyCombo has a keycode value at all is *not* to store the
devirtualized keycode from the keysym, but instead to allow people that
type in "0x55" into the preference. Everything except the overlay-key
respected this. Make the overlay-key binding respect this.
Break down the beautiful core/ui abstraction barrier by inserting
a pointer to MetaWindow into a MetaUIFrame. I'm a scoundrel, I know.
We'll use this very soon to destroy meta_core_get.
Rest in peace you magnificent format, love-child of arcane X11 drawing
API and markup craze, you will not be missed.
We do remember however the bravery of a many men and women, who fearlessly
descended into the guts of your intrinsics and turned ugliness into beauty;
their work will still be spoken of when you will long have been forgotten.
https://bugzilla.gnome.org/show_bug.cgi?id=741917
This reverts commit ec8ed1dbb0.
1) It turns out to add a momentary flicker from the transition
between the login screen and user session
2) It actually isn't needed anymore since bug 733026
https://bugzilla.gnome.org/show_bug.cgi?id=740377
The set/unset branches of meta_display_update_pointer_emulating_sequence()
have been split and put directly where it makes sense. The pointer emulated
sequence will be updated before processing the CLUTTER_TOUCH_BEGIN, and
after processing the CLUTTER_TOUCH_END, this way the checks on this hold
true during all the sequence lifetime.
https://bugzilla.gnome.org/show_bug.cgi?id=738411
Returning FALSE here gets the button release event propagated to the
client on wayland, which is unexpected after xdg_surface.move/resize()
have been called.
https://bugzilla.gnome.org/show_bug.cgi?id=738888
Commit 2f9c601 accidentally changed the logic here, changing the grab
behavior when not using raise-on-click. Fix this.
Spotted-by: Adam Goode <adam@spicenitz.org>
Fullscreen windows look weird when they are overlapped by system chrome,
which currently happens when another window is stacked above. We used to
auto-minimize fullscreen windows in that case, which proved to be both
unreliable and unpopular. So instead, keep the system chrome hidden even
when the fullscreen window is not stacked at the top.
https://bugzilla.gnome.org/show_bug.cgi?id=693991
The constructor would collect windows that are sticky before initializing its state
which would lead to a crash in the case of windows with struts which trigger a work
area recalculation where mutter would assume, due to uninitialized state, that an
existing work area has to be freed.
https://bugzilla.gnome.org/show_bug.cgi?id=738384
There's a small window before a window that is being unmanaged is
unregistered with the display. The MetaScreen::window-left-monitor
and MetaWorkspace::window-removed emissions fall right into that
window, so code that runs in that time may well be out of our
control; we can make sure that the method it can use to get an
updated list of windows no longer contains the destroyed window
though, which is a much better option than expecting everyone to
filter the list themselves.
In Wayland, popup window types are not override-redirect, and thus can
steal window focus away from their parent window when clicked on.
This means that we need to make sure their appearance is properly
propagated to the parent windows so the parent windows don't lose their
focus while they're propagated.
Since we now directly expose the reverses bindings directly, we
don't have to have this special-case in do_choose_window.
More importantly, if the backwards binding is pressed and has the Shift
key included, this will actually revert it
This doesn't matter for Alt-Tab in gnome-shell, which already replaces
it with a better Alt-Tab replacement, but it does matter for Alt-Esc,
which switches between windows directly.
Windows are relocated before their workspace is removed, however this
is only necessary for windows that are *only* on that workspace; for
windows on all workspaces, that step is annoying as it will unset the
sticky state requested by the user.
https://bugzilla.gnome.org/show_bug.cgi?id=737625
The workspace MRU lists are updated when windows are managed/unmanaged
or change workspaces. However those updates obviously only apply to
existing workspaces - new workspaces will always start out with an empty
MRU list, despite sticky windows already being "on" that workspace.
As we now assert that the list contains all windows located on the
workspace, we need to initialize it correctly to avoid a crash.
https://bugzilla.gnome.org/show_bug.cgi?id=737581
We only grant requests to set the demands-attention hint if the window
is at least partially obscured; so for non-minimized windows on the
active workspace, we check if any other window on the same workspace
that is higher in the stack overlaps.
However in the case of a sticky window, window->workspace is NULL, so
we end up considering any non-sticky window on a different workspace.
At this point we have already established that the window is showing
on the active workspace, so use that to filter for windows that may
overlap.
Since the introduction of set_workspace_state(), window->workspace
will always be NULL when on_all_workspaces is set - passing that
to a workspace function that does not validate its input will then
result in a crash.
Use the get_workspace() function instead, which will always return
a valid workspace.
Since commit 2eec11b445, windows without a __NET_WM_DESKTOP property
that should be on all workspaces are not added to the active workspace;
this is correct, however not adding them to any workspace is not ...
set_workspace_state () returns early when the desired sticky state
and workspace match the current property values, assuming that the
corresponding MRU lists are already correct in that case.
However that might not be the case when we are setting the initial
state, so don't take the shortcut in that case.
https://bugzilla.gnome.org/show_bug.cgi?id=737178
A window may either be sticky because it has been requested as such,
or because it is placed on a non-primary monitor (and the corresponding
preference is set). While we do take the latter into account, we
currently override the sticky state later during initialization;
be a bit more careful there to get the initial state right.
https://bugzilla.gnome.org/show_bug.cgi?id=737178
Commit 1af0033368 made a subtle change
regarding how XKeysymToKeycode behaves. It does a depth first search
while XKeysymToKeycode is documented to do a breadth first search:
"this function looks in each column of the core keyboard mapping in
turn and returns the lowest numbered key that matches in the lowest
numbered group" - from the XKB library documentation
Looping over all keycodes for each layout and level index makes us go
back to the previous behavior.
https://bugzilla.gnome.org/show_bug.cgi?id=737134
With the change to how hidden windows are stacked, the position
of the guard window with respect to the hidden windows got flipped
and the guard window was at the bottom of everything; fix it to
be on top of the hidden windows.
https://bugzilla.gnome.org/show_bug.cgi?id=737233
Commit 2f229c3928 removed the code to compute the above-tab
keycode and replaced it with a simple constant from linux/input.h.
We obviously cannot depend on linux headers on non-linux systems,
so provide a fallback definition in that case (which is expected
to work assuming the system is using the Xorg xf86-input-keyboard
driver).
https://bugzilla.gnome.org/show_bug.cgi?id=737135
The titlebar rect is in window coordinates, while screen regions are
obviously not. Fix by translating into screen coordinates before
testing for overlaps.
https://bugzilla.gnome.org/show_bug.cgi?id=736915
Going through GSD's settings was done in context of patches that
did not land; it is simpler and more consistent with GTK+ to use
the corresponding XSetting instead.
Restacking the frame for a window while unmanaging the window is
harmless, but for undecorated (in particular, client-side-decorated)
windows, this causes problems because the window is typically
destroyed by the client immediately after withredrawing the window.
Skip windows flagged as being unmanaged when assembling the new
stack and when comparing the old order to the new stack.
Add a stacking test for this.
We shouldn't scale the cursor size in mutter we g-s-d exports
the correct size on hidpi so use gtk-cursor-theme-size.
This way we also catch changes on resolution updates.
https://bugzilla.gnome.org/show_bug.cgi?id=729337
This reverts commit 4fe66ce0a9.
This is wrong ... we should not scale the cursor size but read
the cursor xsettings that gets exported by gsd. Also this won't update on
resolution changes.
https://bugzilla.gnome.org/show_bug.cgi?id=729337
When restacking the last window alone, we would trigger this off-by-one
error. This would throw us off the end of the array, causing lower_below
warnings for nonsensical values.
Since the last window already is lowered below everything else, we
shouldn't need to lower it.
We have a quite accurate view of the X stack, so there's no good reason to ask
the X server to do restacking that has no effect. (Restackings that have no
effect on either X windows or Wayland windows were generally optimized out in
the synchronization code, but in other cases like moving an X window that is
only beneath Wayland windows to the top of the stack we would make such
requests.)
Removing such requests:
- Is a small efficiency win in itself
- Allows us to immediately go ahead and apply Wayland changes to the verified stack
- Prevents queued Wayland changes piling up waiting for an X event that will never
be received, since the X server will not send confirmation of no-op restacks.
Since such operations may still have an effect on the relative stacking of X
and Wayland windows, we need to continue applying them to the local stack.
https://bugzilla.gnome.org/show_bug.cgi?id=736559
Now that all actual stack shuffle is handled inside stack-tracker.c, we can make
meta_stack_tracker_record_[raise_above/lower_below] internal to that file and
remove the unused meta_stack_tracker_record_lower().
https://bugzilla.gnome.org/show_bug.cgi?id=736559
stack.c:sync_stack_to_xserver had both code for assembling the desired stack, and
code for enforcing the desired stack on the actual stack of X and Wayland windows;
the latter part is properly the domain of stack-tracker.c; moving the code to
apply the stack there both simplifies it and keeps stack.c more manageable.
https://bugzilla.gnome.org/show_bug.cgi?id=736559
There was still code in stack.c to handle skipping override-redirect windows,
but since quite a while ago, meta_stack_add() is not called for OR windows
since they are outside our stacking control. Add an assertion and remove
unnecessary code.
https://bugzilla.gnome.org/show_bug.cgi?id=736559
stack.c kept it's own record of the last stacking it requested, so that
restacking could be done with minimal moves, but we already have a better
view of the stacking order with the stack tracker, so use that instead.
This allows eliminating the special case for the first restack.
https://bugzilla.gnome.org/show_bug.cgi?id=736559
Since MetaStackTracker is the code that knows about the current X stacking order
and the relationship between X windows and Wayland windows, it's cleaner to
encapsulate stack manipulation in MetaStackTracker rather than have the calling
code make the X calls and only call into MetaStackTracker to inform it about
the changes.
https://bugzilla.gnome.org/show_bug.cgi?id=736559
The step where we requery the stacking order from the server than combine
it in an arbitrary fashion with Wayland windows can be eliminated by observing
that we are the final authority for Wayland window stacking - so if we
apply each X event that we receive from the X server to our stack in a
way that leaves the X windows ordered in the same way as on the server,
and apply events that we have stored locally in a way that doesn't affect
the ordering of X windows, than we have a fully correct ordering of windows.
Ordering this in the order of first applying the X event and then applying the
local portion also means that as long as we had an up-to-date view of the X
stack the composite operation will be identical to what was requested.
https://bugzilla.gnome.org/show_bug.cgi?id=736559
Putting X windows and pointers to MetaWindows into a union had a number of
problems:
- It caused awkward initialization and conditionalization
- There was no way to refer to Wayland windows (represented by
MetaWindow *) in the past, which is necessary for the MetaStackTracker
algorithms
- We never even cleaned up old MetaStackWindow so there could be
records in MetaStackWindow pointing to freed MetaWindow.
Replace MetaStackWindow with a 64-bit "stack ID" which is:
- The XID for X Windows
- a "window stamp" for Wayland windows - window stamps are assigned
for all MetaWindow and are unique across the life of the process.
https://bugzilla.gnome.org/show_bug.cgi?id=736559
Add a private hook for the test framework to get XSyncAlarmEvent events -
this will be used to implement XSyncCounter based synchronization
so that the test framework can deterministically wait until Mutter
has seen actions performed by an X11 client.
https://bugzilla.gnome.org/show_bug.cgi?id=736505
Setting the scaling factor immediately after calling clutter_init()
avoids creating the stage at one size, then later resizing it to
a different size.
https://bugzilla.gnome.org/show_bug.cgi?id=736279
We can enter weird states where get_default_window is called during
window unmanagement, before the window has been fully removed from
the stack. Make sure these windows are *never* returned from
get_default_window, as focusing them can cause an assertion fail,
or worse.
Clutter events include the layout index codified into modifier_state,
unlike XI2 device events, which means that we need to mask it out so
that we can match successfully.
A lot of applications assume that the window is fully positioned when it
gets the MapNotify, especially simple applications. Make sure that the
window is only mapped through the calc_showing logic.
O-R windows appear in workspace->windows, which aren't relocatable,
so we can't simply check if the workspace is empty after relocating
all normal windows, since those windows remain.
Make sure that the only windows we have are those that are
on_all_workspaces.
This is bad behavior, and can also cause us to get in an infinite loop
if an OR window is mapped and unmapped in quick succession. This
sequence causes a MapNotify followed by an UnmapNotify, and when
processing the events, we'll call XMapWindow, XUnmapWindow, which will
put another set of MapNotify, UnmapNotify events in our queue, which we
then process by calling XMapWindow, XUnmapWindow, and so it goes
forever, or at least some scheduler uncorks us by making us call
XMapWindow when the window is already mapped, or XUnmapWindow when the
window is already unmapped.
We can stop this madness by simply making sure never to call neither
XMapWindow or XUnmapWindow on OR windows, which is the correct thing to
do anyway.
We commonly used the generic, undetailed signal 'changed' to track
changes to preferences. Since we crash on unknown preference types,
this can be dangerous if somebody adds a new setting that has a
type we're unfamiliar with, and something else changes it.
Instead of crashing, just fizzle out doing nothing.
In gnome-shell, we have a feature where the user can unmaximize windows
by dragging them from the panel above the window. With accurate
anchoring, this looks really weird as the cursor is now "detached" from
the window. Detect this case and put the cursor in the middle of the
window titlebar instead.
This seems to be more accurate with what we currently see in
GNOME. Without having the app expose this information to us, it might be
a better idea to use the default frame style for this information instead.
The cursor was anchored wrongly when trying to unmaximize windows from
dragging them from the top of the screen because of a few think-o's and
some code that forgot to be updated when we moved to the frame rect
coordinate system.
The cursor is still offset for windows that start dragging from the top
panel. This is technically correct, but looks wrong. We'll fix this in
the next commit.
The initialization sequence before was quite icky, initializing Clutter
in a few different places depending on what was going on.
Put that all back into main.c
This removes our Xwayland dependency in the native path. The direct
grabs are still there for the X11 backend and are a bit disgusting,
but that's OK. We can refactor it out later.
This introduces some pretty lousy hackery because it depends on
https://github.com/xkbcommon/libxkbcommon/pull/10 , and I really
don't want to wait on that to squash this dep.
Now that the internal mutter bindings and gnome-shell stopped using
META_KEY_BINDING_REVERSES, and after moving the 'adding shift reverses
the keybinding action' logic to gnome-control-center, we can remove
META_KEY_BINDING_REVERSES from mutter.
Plugin API is broken as this constant is removed from the exported
headers. ABI is broken as using this flag is now a noop.
https://bugzilla.gnome.org/show_bug.cgi?id=732385
Currently the bindings for {switch,cycle}.* actions are created with the
META_KEY_BINDING_REVERSES flag so that <shift>+binding triggers the
reverse action. However, gnome-control-center does not know about this
kind of implicit bindings, and, for example, cannot warn when the user
tries to setup a conflicting <shift>+xxx binding.
These backward <shift> bindings are being explicitly set in
gsettings-desktop-schemas, so the META_KEY_BINDING_REVERSES annotation
can be removed for them from mutter.
https://bugzilla.gnome.org/show_bug.cgi?id=732385
MetaKeyBinding can be marked as being reversed
(META_KEY_BINDING_IS_REVERSED), but MetaKeyHandlerFunc callbacks
cannot check whether this flag was set or not on the MetaKeyBinding
which triggered the callback.
https://bugzilla.gnome.org/show_bug.cgi?id=732295
The existing workspace management code is quite hairy, with plenty of
logic inline in all of window.c, workspace.c, and screen.c, making it
hard to understand or make changes to, since you might forget to change
several of the other places the code was around.
Rewrite the internal workspace management logic so that it's
centralized and all in window.c. Document the invariants we need to
maintain, and ensure that these invariants are properly kept, with
asserts in various places.
Extensive testing on gnome-shell did not bring up any issues, and this
is a considerable cleanup.
This will be used to replace some of the hooks that are used to call
into window.c, so that the workspace index property is properly kept up
to date.
We can't name the property "index" since it causes conflicts with the
meta_workspace_index method. This should really be called
meta_workspace_get_index, but oh well.
I accidentally broke this in commit a119ea9. The code was considerably
more complicated than it needs to be, so let's replace it with a
g_list_find and nothing more.
Scanning over the hash table of XIDs is a terrible idea. Not only were
we excluding Wayland windows, but we were also looking at alarms and
barriers, too. We were lucky that that only contained GObjects where
our checks would work.
