Now the client can let us know where the panel is using
desktop_shell.set_panel_position, we can correctly calculate where to
put new views and how big maximized views should be.
Acked-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Panels are always assumed to be on the top edge of the output. If this
is not the case views will be placed under the panel, wherever it is,
and maximize doesn't use the correct space allocated for views.
By telling the server on which edge the panel is located, it can
correctly calculate where to put new views and how big maximized views
should be.
[Pekka Paalanen: the user of this protocol so far is Maynard.]
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
The commit 'input: Send leave and enter pair when the surface moves
under the cursor' introduced focused surface local pointer coordinates
to keep track of if a surface had been moved or transformed in a way
that the pointer posititon relative to that surface would change.
Update these coordinates also for the popup grab as otherwise every
pointer motion during a popup grab results in leave and then enter
events.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Tested-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
At the moment when surfaces are destroyed they are faded out but let's
make it configurable!
Tested-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
An error makes the client exit, which cleans up the resources anyway.
Note (Jason Ekstrand):
This is safe for two reasons. First, we should be handling object
destruction nicely anyway. Second, in each of these cases, the resources
don't have any implementation or destruction set so it has absolutely no
effect on the rest of weston whether we destroy it now or later.
This introduces a new struct, weston_layer_entry, which is now used
in place of wl_list to keep the link for the layer list in weston_view
and the head of the list in weston_layer.
weston_layer_entry also has a weston_layer*, which points to the layer
the view is in or, in the case the entry it's the head of the list, to
the layer itself.
Currently, there is a fun flicker when toggling maximization or
fullscreen on a window in mutter or more sophisicated compositors
and WMs.
What happens is that the client want so go maximized, so we
calculate the size that we want the window to resize to (640x480),
and then add on its margins to find the buffer size (+10 = 660x500),
and then send out a configure event for that size. The client
renders to that size, realizes that it's maximized, and then
says "oh hey, my margins are actually 0 now!", and so the compositor
has to send out another configure event.
In order to fix this, make the the configure request correspond to
the window geometry we'd like the window to be at. At the same time,
replace set_margin with set_window_geometry, where we specify a rect
rather than a border around the window.
Commit 9aa8ce69 forgot to set shsurf->fullscreen_output in
fullscreen_binding(), causing segfault when fullscreening using key
bindings. This patch fixes that.
https://bugs.freedesktop.org/show_bug.cgi?id=79828
Signed-off-by: Boyan Ding <stu_dby@126.com>
Fixes a crash on touch devices without a pointer, when touching
the window frame of a client.
Signed-off-by: Emilio Pozuelo Monfort <emilio.pozuelo@collabora.co.uk>
With most of the code in send_configure_for_surface, the helper
methods don't give us that much benefit, so stop using them. We
can't kill them off, as they're part of the shell interface and
used by the WM.
Currently, there's a giant bug in how xdg-shell state management
is done. If a client calls set_fullscreen and then set_maximized,
it will get two configure events:
=> set_fullscreen
<= configure(800, 600, [STATE_FULLSCREEN])
=> set_maximized
<= configure(800, 560, [STATE_FULLSCREEN, STATE_MAXIMIZED])
Since fullscreen takes precedence over maximized, the client will
render full-screen at 800x600 first, and then 800x560 next. As
a result, the surface gets the wrong size.
This is because the code that sends out configure requests is
"immediate" -- when an app calls set_maximized, we immediately
send out the configure event that would have happened if we
transitioned immediately into maximized mode.
In wl_shell, this is correct behavior. However, in xdg-shell,
this is wrong. State needs to be more carefully managed in
xdg-shell, as states aren't exclusive.
Pull all the code that sends out configure events out and send
them centrally, based on Weston's on surface state management.
This should work with both wl_shell and xdg_shell's strategies.
Currently, there's a race condition. When resizing from the left, and
a client attaches a buffer after the resize ends, you suddenly see the
buffer jump to the right, because the resize ended while multiple
attaches were in-flight. Making resize a state can fix this, as the
server can now know exactly when the resize ended, and whether a commit
was before or after that place.
We don't implement the correct tracking in this commit; that's left as
an exercise to the reader.
Additionally, clients like terminals might want to display resize popups
to display the number of cells when in a resize. They can use the hint
here to figure out whether they are resizing.
The states system, so far, has been a complicated mix of weird APIs
that solved a real race condition, but have been particularly ugly
for both compositors and clients to implement.
It's a confusing name that comes from the ICCCM. The ICCCM is best
forgotten about.
With the addition of the potential new "transient" role meaning a
parent-relative toplevel like a long-lived popup, used for e.g.
tooltips, the set_transient_for name will become even more confusing.
The check to avoid calling weston_keyboard_set_focus() for a seat that
didn't have a keyboard in restore_focus_state() was cheking the wrong
seat (the one from the previous loop). That caused a crash when
switching workspaces if there was an extra seat that didn't have a
keyboard.
https://bugs.freedesktop.org/show_bug.cgi?id=78349
Views that extend past the bottom of the output are still visible after
the workspace animation ends but before its layer is hidden. When the
layer was hidden, nothing would cause those regions to be repainted,
leading to artifacts.
https://bugs.freedesktop.org/show_bug.cgi?id=78363
With xdg_shell wl_shell starting to diverge in how they work, there's
less shared code in set_fullscreen(). The problem is that the xwayland
window manager calls into set_fulscreen() which now doesn't complete
the fullscreen transition. Add shell_interface_set_fullscreen() for
the shell interface set_fullscreen hook to use.
https://bugs.freedesktop.org/show_bug.cgi?id=78230
send_configure was originally modelled after
wl_shell_surface::send_configure, which takes these arguments. However,
the X WM and xdg_surface::configure variants don't use these arguments.
We already store the resize edges for a surface while it's being
resized, so just use the saved state in the wl_shell_surface variant.
This moves the check for shsurf->grabbed into surface_move() and
surface_resize(), which are shared with the xwayland code. This prevents
trying to resize or move an xwayland window with multiple pointers.
9c376b54ea fixed the crash when a client goes
away during a resize grab. The shsurf->resource is set to NULL in that
case and we were trying to send out events to a NULL resource. However,
xwayland shell surfaces are created by the xwayland module and don't have a
resource. We use a different function pointer for sending the configure
events that handle the events inside xwayland instead of sending protocol
events.
To fix all this, we just move the check for a NULL resource into the
functions that we use for sending configure events for wl_shell and
xdg_shell.
The geometry for visible views will keep unchanged,
weston_view_set_position() doesn't mark these views
as dirty. So there is no chance for them to reassign output, then
these views will disappear.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=72946
Signed-off-by: Xiong Zhang <xiong.y.zhang@intel.com>
When commit 07926d90 factored out the code that chooses in which layer
a surface is added to, it changed the behavior for surfaces with no
type. Instead of not adding it to any layer, the surface is added to
the current workspace.
This patch restores the old behavior.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=77527
Previously, desktop-shell would only create its internal shell_seat object
for each seat available when the desktop-shell module is loaded. This is a
problem any time seats are created dynamically. In particular, the Wayland
and RDP backends create seats on an as-needed basis and they weren't
getting picked up proprely by desktop-shell.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=77649
lower_fullscreen_surface() was removing fullscreen surfaces from
the fullscreen layer and inserting them in the normal workspace
layer. However, those fullscreen surfaces were never put back in
the fullscreen layer, causing bugs such as unrelated surfaces
being drawn between a fullscreen surface and its black view.
Change the lower_fullscreen_surface() logic so that it lowers
fullscreen surfaces to the workspace layer *and* hides the
black views. Make this reversible by re-configuring the lowered
fullscreen surface: when it is re-configured, the black view
will be shown again and the surface will be restacked in the
fullscreen layer.
https://bugs.freedesktop.org/show_bug.cgi?id=73575https://bugs.freedesktop.org/show_bug.cgi?id=74221https://bugs.freedesktop.org/show_bug.cgi?id=74222
If a client exists during a resize grab, the resource for the shell
surface being resized is destroyed. The shell surface is not destroyed
immediately, however, because of the window close animation. In that
case, the compositor would crash trying to send configure events to
the surface being resized, since it would pass a NULL pointer to
wl_resource_post_event().
The code for the resize grab was already able to handle the surface
going away, so expand it to also handle the resource going away and
fix the crash.
https://bugs.freedesktop.org/show_bug.cgi?id=77344
In order to do the window close animation, a reference for a destroyed
surface is kept. However, the reference count was also increased for
unmapped surfaces, in which case the animation wouldn't run. Since the
reference count was decremented in the animation done function, it would
never be decreased for unmapped surfaces, causing them to not be
released.
The close animation also changed how shell surfaces are released. The
destroy function for its resource was changed to not deallocate the
surface, and instead keep it around until the animation finishes and
the weston surface is destroyed. The destruction should happen in the
destroy listener for the weston surface, but it wouldn't destroy the
shell surface in the case the resource was still valid, assuming that
it would be freed in the resource destroy function.
We now carry the shell_client around with each shell_surface. This is much
more reliable than tacitly assuming that there is only one wl_shell or
xdg_shell instance bound to a particular wl_client. In particular, weston
would crash when a client bound to both wl_shell and xdg_shell even if it
only ever used one of them.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Previously, the repositioning logic would iterate the compositor's list
of layers and move the views on those layers. However, that failed in
two different ways: it didn't cover hidden workspaces and crashed when
the display was locked.
This patch changes the logic to explicit iterate over all the layers
owned by the shell. The iteration is done through a helper function,
shell_for_each_layer().
https://bugs.freedesktop.org/show_bug.cgi?id=76859https://bugs.freedesktop.org/show_bug.cgi?id=77290
When a fullscreen surface gets the maximized state, the function
reset_surface_type() is called and that causes unset_fullscreen() to be
called. That function would set the value of shsurf->fullscreen_output
to NULL. However, since the surface still has the fullscreen state, it
will be configured as a fullscreen surface again, and an attempt to
access that field would cause the compositor to crash.
Fix the crash by keeping the value of fullscreen_output around after
unset_fullscreen(). This is safe since the value is only used when a
surface has the fullscreen state and is replaced on a new request to
make the surface fullscreen.
https://bugs.freedesktop.org/show_bug.cgi?id=76867
The timer was left running after the screensaver was terminated. When
it triggered, a fade out that would in turn cause the screen to be
locked was started. Since that could happen without the compositor
emitting the idle signal, there would be no wake signal to make the
shell show the lock screen, so the system was left unresponsive
until the idle signal actually triggered.
https://bugs.freedesktop.org/show_bug.cgi?id=70923
Before commit 2f5faff7f9 when the compositor is locked it would
reset the keyboard focus on all of the seats as part of pushing the
focus_state. This was removed because it now always keeps track of the
focus_state in the workspace instead of waiting until the state is
pushed. However this had the side effect that the active surface would
retain focus when the compositor is locked. This patch just makes it
explicitly set the keyboard focus to NULL on all of the seats when
locking. This will be restored based on the workspace's state when
unlocking.
https://bugs.freedesktop.org/show_bug.cgi?id=73905
The focus_state list on a workspace only contains entries for seats
which have a keyboard focus on that workspace. For workspaces that
have no surfaces the list will be empty. That means that when a
workspace with no surfaces is switched to it would previously leave
the keyboard focus unaffected and you could still type in the surface
on the old workspace.
This patch makes it instead reset the keyboard focus to NULL for seats
without a focus_state. It does this by temporarily stealing the
compositor's list of seats while it iterates the focus_states. After
all of the focus states have been processed any seats remaining in
this temporary list have their focus reset.
https://bugs.freedesktop.org/show_bug.cgi?id=73905
Commit c85f1d45 caused the move of an unresponsive surface to be no
longer possible, since the grabbed flag would prevent the move grab
to start while the busy grab was still active.
The shell_destroy_shell_surface function only set the backing resource to
NULL, leaving an unusable surface in the popup_grab list until the surface's
fading animation finished and it could be freed. This caused a segfault if
the shell tried to forcibly break the grab during that time interval due to
the compositor losing the keyboard focus.
https://bugs.freedesktop.org/show_bug.cgi?id=77072
When a client calls the input panel (weston-keyboard e.g.)
and then goes fullscreen, the panel will not be hidden
anymore.
Signed-off-by: Manuel Bachmann <manuel.bachmann@open.eurogiciel.org>
Temporarily show minimized surfaces when switching between
surfaces with the keyboard. If the final selected one was
minimized, it will be restored.
Signed-off-by: Manuel Bachmann <manuel.bachmann@open.eurogiciel.org>
We now handle the client-side xdg_surface_set_minimized()
call, and eventually hide the target surface by moving it
to a dedicated layer.
Signed-off-by: Manuel Bachmann <manuel.bachmann@open.eurogiciel.org>
Queueing in the Presentation extension requires splitting the viewport
state into buffer state and surface state. To conveniently allow
assigning only one, the other, or both, reorganize the
weston_buffer_viewport structure.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
This provides an example of keeping a weston_surface alive after the client
destroys it. We install a destroy listener for the resource, so that we'll
be notifified when the client destroys it. Then we increase the weston_surface
refcount so that we keep the surface and initiate an animation. When
the animation finishes we can finally destroy the surface.
Rather than require that the client implement two methods for every state,
simply have one global request, change_state, and one global event,
request_change_state.
This was always a little iffy. At least it could have been a signal,
but we now have focus signal, so lets just use that. We lose
the ability to detect unresponsive clients at key event time, but we
could add that back by adding a key_signal.
Use a static assert to catch mismatch between implementation and
interface version. Fix window.c to not use XDG_SHELL_VERSION_CURRENT,
which will fail to catch version mismatches. The implementation version
must updated manually when the implementation is updated to use the new
interface.
Responsivenes is a per-client thing so we move the ping/pong functionality
to xdg_shell. Having this per-window was carries over from the EWMH
protocol, where the WM has no other way to do this. In wayland, the
compositor can directly ping the client that owns the surface.
This is used to figure out the size of "invisible" decorations, which we'll
use to better know the visible extents of the surface, which we can use for
constraining, titlebars, and more.
Remove the listener for output destroy from weston_view and instead
iterate views owned by the shell in its own output destroy listener.
This simplifies the code a bit since keeping the view listening for the
destroy on the right output was a bit complicated. This also removes the
function pointer output_destroyed from weston_view. The only user for it
was desktop shell, but now this is all handled in shell.c.
Since that signal is per output, it is necessary to track in which
output a view is in so that the signal is handled properly.
Instead, add a compositor wide output moved signal, that is handled by
the shell. The shell iterates over the layers it owns to move views
appropriately.
When xwayland creates a shell surface we don't have a resource. The
recently added shell_surface_is_wl_shell/xdg_surface() tests don't
handle that very well.
For now, we assume that a surface without a resource is created from
xwayland and is a wl_shell surface. We'll want to modify that to be a
xdg surface eventually, but for now this stops weston from crashing.
Since commit 9046d2, when destroying a surface, we remove all the
links from its children. But when the child surfaces are destroyed,
those links will be removed again, but since they were not properly
initialized, weston will crash.
Call shell_surface_set_parent instead which removes the link and
sets parent while also initializing the link, thus avoiding this
crash.
We don't have focus-follows-mouse, so it makes more sense to
maximize or fullscreen the surface that has the keyboard focus,
not the one behind the pointer.
We have to move the surface destroy listener around as we track the
currently focused surface. Introduce a helper function,
focus_state_set_focus() for this and use throughout.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=73768
Various functions that operate on a weston_surface assume the
surface has a shell_surface. That is, they unconditionally
deref the get_shell_surface() result. Hence, if for some reason
the call to get_shell_surface() returned NULL to those functions then
a segmentation fault would occur and the program would crash. So,
adding an assert(...) on the get_shell_surface() return value adds an
extra sanity check and does not change this behavior. The assert also
adds an extra benefit to the programmer by documenting that the function
expects and requires the weston_surface to have a shell_surface and
would be a program logic error, otherwise.
The assert() also silences some static analyzers about the possible
NULL deref.
Signed-off-by: U. Artie Eoff <ullysses.a.eoff@intel.com>
When the last window of the X11 compositor is closed during a fade or
while locked, we'll try to start a fade back to the lock screen. However,
if we closed the last window, there are no outputs left and the animation
will try to run with surface->output == NULL.
https://bugs.freedesktop.org/show_bug.cgi?id=73665
This is part of the fix for bug 72540. We cancel the popup grab when the
screensaver kicks in so that the screen unlock dialog can get input events.
The bigger problem is in mesa however, where we try to allocate new buffers
as cairo-gles2 does a gratuituous (but valid) eglMakeCurrent() as we
remove the tooltip or popup-menu.
Since we removed the weston_layer with the regular surfaces, EGL blocks
waiting for a frame event that never comes.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=72540
The grab stays alive as long as at least one touch point is down. If touch
point 0 is lifted while other touch points are down, the surface will jump
around when touch point 0 is put down again.
This change marks the grab as inactive once touch point 0 is lifted
and then ignores touch events until all touch points eventually are
lifted and the grab terminates.
https://bugs.freedesktop.org/show_bug.cgi?id=73750
Add a config file option to enable it, but leave it off by default. Exposay
still triggers too many lock-ups or stuck grabs and triggers under X even
when the Wayland window doesn't have keyboard focus.
We now track the child surfaces of a shell surface and the child surfaces
have a pointer back to their parent. We need to clean all this up and
NULL out the childrens parent pointers when a shell surface is destroyed.
Closes: https://bugs.freedesktop.org/show_bug.cgi?id=72931
It's possible to touch a surface to move it and let go before we get
to common_surface_move(), in which case we don't have a touch focus
when we get there. For pointers, we could click a surface, but have the
surface go away before we get to common_surface_move(), in which
case the button count is non-zero, but we don't have a surface.
In either case we crash, so let's add a check to make sure we still
have a focus surface before we try to move it.
Closes: https://bugs.freedesktop.org/show_bug.cgi?id=73448
This fixes the crash when move, rotate or resize binding is activated
while exposay effect is active.
Steps to reproduce:
- activate exposay
- try to rotate the surface with mod + right mouse button
- crash
Closes: https://bugs.freedesktop.org/show_bug.cgi?id=72885
Similar to 7c4f6cc145, if we don't have
a background image from the desktop-shell client or the touch point
for some other reason doesn't hit a surface we trigger a
segfault as we try to deref the seat->touch->focus NULL pointer.
For touch, another problem could be fudgey calibration that ends up
giving touch coordinates outside the output space.
https://bugs.freedesktop.org/show_bug.cgi?id=72839
Set up X windows that are transient for another window as transient
surfaces in shell.c. This keeps the transient windows on top of their
parent as windows are raised, lowered for fullscreened.
https://bugs.freedesktop.org/show_bug.cgi?id=69443
If we don't have a background image from the desktop-shell client or the
pointer for some other reason doesn't have a focus we trigger a
segfault as we try to deref the seat->pointer->focus NULL pointer.
https://bugs.freedesktop.org/show_bug.cgi?id=73066
This is still fairly unstable, causes lockups with fullscreen and exposay,
leaves small preview surfaces on-screen if used on the same modifier as
mod-tab. We also only need on mod-tab implementation so lets see if we
can consolidate the current and this alt-tab implementation in 1.5.
https://bugs.freedesktop.org/show_bug.cgi?id=72610
Popup windows are relative to a plain wl_surface, so that custom surfaces
can have popups. This used for the desktop-shell panel for example. Also,
popups should be immediately on top of their parent surface, as they
typically represent an extension of an UI element in the parent surface
such as a combo box or menu.
This reverts commit da704d97fa.
Conflicts:
desktop-shell/shell.c
https://bugs.freedesktop.org/show_bug.cgi?id=72547
If we don't mark the state as changed, we don't copy over next_state to
state and we fail to treat the surface as a transient. In particular,
we give it a random intial position instead of mapping it at the given
parent relative position.
https://bugs.freedesktop.org/show_bug.cgi?id=72532
Set the internal pointer for the client to NULL. This fixes a
segmentation fault at shutdown, where the shell would hang up before
and cause libwayland to call wl_client_destroy(). When the shell was
destroyed later, another call to wl_client_destroy() would cause the
crash.
https://bugs.freedesktop.org/show_bug.cgi?id=72550
Surfaces that are created by clients get their size automatically updated
by the attach/commit. Surfaces created directly by shells (such as black
surfaces) sometimes need to be manually resized. This function allows you
to do that while being somewhat less messy than messing with the internals
of weston_surface manually.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
If the saved position for a fullscreen or maximized output view is in an
output that has been unplugged, the coordinates don't make sense
anymore. In that case, invalidate them and use the initial position
algorithm when changing them back to the basic state.
Signed-off-by: Zhang, Xiong Y <xiong.y.zhang@intel.com>
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
The surface type now no longer changes and we track pending state changes in
next_state. Instead of testing type != next_type to detect changes in
state, we just look at state_changed.
An xwayland surface corresponds to a override-redirect window under X,
which is typically a menu or a popup window. They typically appear
with a keyboard and mouse grab and by nature of being override-redirect
these window can appear anywhere on screen and in the stack.
We need to resort to heuristics to decide where to place the
override-redirect in our surface stack, and for now we'll just put it on
top of everything. That's going to be correct for almost all cases of
clicking to open a menu, but we can revisit and refine if we run into
a case that needs better handling.
Just as for set_maximized() we can move the setting of the fullscreen and
state_changed flags into the common set_fullscreen() function. This
function is also used from the xwayland wm to set fullscreen windows, and
with this change that now works again.
We can set the maximized and state_changed flags in set_maximized(),
which is shared between shell_surface_set_maximized() and
xdg_surface_set_maximized().
Since internally there's no more SHELL_SURFACE_FULLSCREEN and
SHELL_SURFACE_MAXIMIZED, the surface must be set to
SHELL_SURFACE_TOPLEVEL on the respective functions.
This fixes the bug when clients start already in fullscreen mode. In
that case, they aren't set first to toplevel, and then change to
fullscreen. They are set as fullscreen directly, not receiving the
SHELL_SURFACE_TOPLEVEL type on the set_fullscreen function.
Jonny Lamb