Fixes an issue where an already visible scratchpad window being moved due to
'scratchpad show' leaves the entire workspace at the top of the focus stack in
the old workspace. Moving by 'focus output' back to the old workspace would
focus the entire workspace instead of just the last active container.
This makes the behavior of floating containers more consistent with i3.
The coordinates of the container are scaled when the size of the
workspace it is on changes or when the container is moved
between workspaces on different outputs.
For scratchpad containers, add a new state that preserves the dimensions
of the last output the window appeared on. This is necessary because
after a container is hidden in the scratchpad, we expect it to be in the
same relative position on the output when it reappears. We can't just
use the container's attached workspace because that workspace's
dimensions might have been changed or the workspace as a whole could
have been destroyed.
CAIRO_HINT_STYLE_FULL attempts to maximize contrast at the expense
of fidelity, this makes most fonts that haven't been hand hinted,
which makes up the majority of fonts out there, appear much worse.
In the absence of explicitly set hint style, cairo will default to
CAIRO_HINT_STYLE_SLIGHT, which attempts to improve contrast while
retaining fidelity to the original shapes, which is what we want.
When a window in the scratchpad container requests for
xdg_activation_v1, it is ignored no matter what the value of
focus_on_window_activation is.
At least allow windows in the scratchpad to set the urgent flag. When
focus_on_window_activation is set to "focus", show the parent scratchpad
where the contained requested for xdg_activation_v1.
Check whether output->damage_ring.current is empty before calling
wlr_output_attach_render(). Saves us from having to un-do that
via wlr_output_rollback().
Atm we got issue with the touch position sent to the clients. While
holding contact, leaving the initial container will continue to send
motion event to the client but with the new local position from the new
container.
This seatop goal is to send the position of the touch event, relatively
to the initial container layout position.
This function was already declared in container.h but defined in
commands/swap.c for some unknown reason. Everything in commands/ assumes
the handler context has been set appropriately by the command preludes
but this function snuck its way into seatop_* which doesn't set anything
in the handler context.
The fact that the seatop drag actions manipulate the focus without
custody of the seat means they are definitely very broken in multiseat.
With `hide_edge_borders both` (or at least `vertical`),
`window_rect.y` will equal `border_thickness` for SOME windows,
but it will be 0 for windows adjacent to top screen edge.
Therefore setting it to `border_thickness` is not sufficient.
This commit changes it to the actual y offset of content
into the container.
xdg-activation is now too strict in only allowing tokens with a seat to
activate a surface. Clients may rely on this behavior for urgency hints.
The seat argument is still useful in case the client does provide a seat
so we can activate it on the desired seat.
Fixes: 842609da64 (view: make request_activate take a seat, 2022-11-30)
fixes#7394
Test cases:
* zwlr_output_configuration_head_v1_set_adaptive_sync 0->0, no change
* 0->1, enabled
* 1->0, disabled
* 1->1, no change
Similar tests with an incapable display resulted in `"Adaptive sync
failed, ignoring"` messages as expected.
Since [1], wlr_renderer_begin() can fail. Check its return value
and bail.
This fixes an assertion error (when begin() fails and then we try
to render something) after a GPU reset.
[1]: a541c9510a
This change allows the tablet tool button to be used for floating mod
resize. In addition, it attempts to ensure that tablet tool events are
consistent such that tablet v2 events and pointer events will never be
interleaved, and such that the tool buttons count will never fall out of
sync and cause tool button emulation to break.
Some of this logic is similar to what is done for tablet tool tip, but
not quite identical, because of the complication that we have to deal
with multiple inputs that can overlap eachother.
Fixes#7036.
This might be the wrong fix, but the crash is happening because the ->data
field on an xwayland surface is NULL. A NULL data field is normal for
unmanaged surfaces, however it seems clients can do weird things: They can
create a cursor lock on a regular xwayland surface then make it unmanaged
by calling override_redirect. In this case, the xwayland server should
destroy the cursor lock, which is does, but does so in the wrong order
making it try to dereference a NULL pointer after sway has acknowledged
its new unmanaged status.
```
(gdb) bt full
0 0x000055fd91934861 in warp_to_constraint_cursor_hint (cursor=0x55fd93486c00)
at ../sway/input/cursor.c:1243
sy = 605
lx = 6.9527431433545762e-310
sx = 1272
view = 0x0
con = 0x7ffd1cdfe400
ly = -6.949595189996421e+59
constraint = 0x55fd93e7faa0
1 0x000055fd91934976 in handle_constraint_destroy (listener=0x55fd93f0fd58, data=0x55fd93e7faa0)
at ../sway/input/cursor.c:1266
sway_constraint = 0x55fd93f0fd30
constraint = 0x55fd93e7faa0
cursor = 0x55fd93486c00
2 0x00007fda8275bf6e in wl_signal_emit_mutable () at /usr/lib/libwayland-server.so.0
3 0x00007fda82e57016 in pointer_constraint_destroy (constraint=0x55fd93e7faa0)
at ../subprojects/wlroots/types/wlr_pointer_constraints_v1.c:49
4 0x00007fda82e570dc in pointer_constraint_destroy_resource (resource=0x55fd933cf8f0)
at ../subprojects/wlroots/types/wlr_pointer_constraints_v1.c:66
constraint = 0x55fd93e7faa0
5 0x00007fda8275d8ba in () at /usr/lib/libwayland-server.so.0
6 0x00007fda8275f6a9 in wl_resource_destroy () at /usr/lib/libwayland-server.so.0
7 0x00007fda82e56fb3 in resource_destroy (client=0x55fd93ea52e0, resource=0x55fd933cf8f0)
at ../subprojects/wlroots/types/wlr_pointer_constraints_v1.c:39
8 0x00007fda81d8f4f6 in () at /usr/lib/libffi.so.8
9 0x00007fda81d8bf5e in () at /usr/lib/libffi.so.8
10 0x00007fda81d8eb73 in ffi_call () at /usr/lib/libffi.so.8
11 0x00007fda8275aada in () at /usr/lib/libwayland-server.so.0
12 0x00007fda8275f01c in () at /usr/lib/libwayland-server.so.0
13 0x00007fda8275d9e2 in wl_event_loop_dispatch () at /usr/lib/libwayland-server.so.0
14 0x00007fda8275e197 in wl_display_run () at /usr/lib/libwayland-server.so.0
15 0x000055fd919264d3 in server_run (server=0x55fd919a3a80 <server>) at ../sway/server.c:320
16 0x000055fd91925457 in main (argc=1, argv=0x7ffd1cdfed98) at ../sway/main.c:411
verbose = false
debug = false
validate = false
allow_unsupported_gpu = false
config_path = 0x0
c = -1
```
sway sends wl_keyboard.enter on seat focus change and when a keyboard
active on a seat is configured. If all keyboards are removed and a
keyboard is added back without changing the focused client, no new
notify event would be sent despite having keyboard focus. This could
lead to key events without notify, which is a protocol violation.
As a quick fix, when configuring a keyboard on a seat where no keyboard
is currently active, activate the keyboard so that a focused surface
will receive a notify event.
Regressed by: e1b268af98
Closes: https://github.com/swaywm/sway/issues/7330
efd83cb8 added the rotation_angle command but it didn't insert it in
the proper place in the list, so the repeat_delay and repeat_rate
commands became unusable.
See: https://github.com/swaywm/sway/issues/4511
Adds a bool config option `primary_selection`, which explicitly
enables/disables the primary selection clipboard. Defaults to enabled.
This is implemented as a launch-only option which enables or disables the creation of the
`zwp_primary_selection_device_manager_v1` global.
Co-authored-by: Tilde Rose <t1lde@protonmail.com>
When we reload the config, we reset every input device and re-apply
configuration from the config file. This means that the keyboard keymap
is updated at least once during config reload, more if the config file
contains keyboard configuration.
When they keyboard keymap changes and is updated through wlr_seat, the
keymap ends up sent to every keyboard bound in every client, seemingly
multiple times. On an x230 of mine with a keyboard layout set in the
config file, I see 42 keymap events sent to foot on config reload.
Reduce events from keyboard configurations by skipping all but the
currently active keyboard for the seat, and by clearing the active
keyboard during input manager device reset. After this change, I only
see a single just-in-time keymap event.
Fixes: https://github.com/swaywm/sway/issues/6654
Views now maintain a reference to a launch context which, as a last
resort, is populated at map time with a context associated with its pid.
This opens the possibility of populating it before map via another
source, e.g. xdga-tokens or configuration.
This removes the need to rename the pid_workspaces when a workspace
is renamed.
It also opens the possibility of tracking other node types. Tracking
containers would allow application to be placed correctly in the
container tree even if the user has moved their focus elsewhere since
it was launched.
Any windows that have never had a title set visually behave closer to
that of an empty title, but are unformattable, as the code bails out
early on a NULL title.
Support the new dwtp (disable while trackpointing) option introduced in
libinput 1.21, allowing users to control whether the trackpoint (like
those in Thinkpads, but not only) should be disabled while using the
keyboard/touchpad.
See: https://gitlab.freedesktop.org/libinput/libinput/-/issues/731
Remove the incorrect attempt to block focus changes when an input grab
is present and replace it with the same logic used for layer_shell-based
screen lockers: restore the focus after changing it.
This fixes a use-after-free of seat->workspace if outputs are destroyed
while a screen lock is enabled.
When removing outputs, it is possible to end up in a situation where
none of the session lock client's surfaces have keyboard focus,
resulting in it not receiving keyboard events. Track the focused
surface and update it as needed on surface destroy.
Sway focuses the inactive child when focusing split containers. However,
there is currently no way to focus the parent container itself by mouse.
A user must use the keyboard to do so.
This commit maintains the current behavior, but makes it such that a
second click on the split container titlebar (i.e., after its children
are visible) focuses the split container itself.
Currently, when encountering a non-desktop display, sway offers the
output for leasing and returns without storing it in a sway specific
output type like `struct sway_output`. Additionally, running
`swaymsg -t get_outputs` doesn't show non-desktop outputs.
This commit stores the non-desktop outputs into a struct called
`sway_output_non_desktop`, and adds them to a list on `sway_root`
Without this, the `IPC_GET_TREE` ipc call would return false information
about the container's `deco_rect` and `rect` properties if
`hide_edge_borders --i3` was in effect.
This semi-colon looks like a typo. Luckily, it has no effect on the code as it's treated as an empty statement leading the switch case.
Really straightforward nitpick change, was just something I was confused by when reading over the code.
Use pango to parse font configuration early, and reject the command as
invalid if the value is invalid for pango. Since we're already parsing
the font into a `PangoFontDescription`, keep that instance around and
avoid re-parsing the font each time we render text.
Fixes: https://github.com/swaywm/sway/issues/6805
If the input device is quoted, which is common when using variables in the
config file, those quotes must be ignored here, or the input device will be
ignored.
Fixes#7029.
The "dpms" command refers to VESA Display Power Management
Signaling, a deprecated standard. It's superseded by VESA DPM.
Instead of tying out command name to a particular standard, use the
neutral term "power".
This ensures that those surprised by the deprecation of SUID operation
receive an error rather than accidentally having sway run as root.
This detection will be removed in a future release.
Try to gain SCHED_RR (round-robin) realtime scheduling privileges before
starting the server. This requires CAP_SYS_NICE on Linux systems.
We additionally register a pthread_atfork callback which resets the
scheduling class back to SCHED_OTHER (the Linux system default).
Due to CAP_SYS_NICE, setting RLIMIT_RTPRIO has no effect on the process
as documented within man 7 sched (from Linux):
Privileged (CAP_SYS_NICE) threads ignore the RLIMIT_RTPRIO limit;
as with older kernels, they can make arbitrary changes to
scheduling policy and priority. See getrlimit(2) for further
information on RLIMIT_RTPRIO
Note that this requires the sway distribution packagers to set the
CAP_SYS_NICE capability on the sway binary.
Supersedes #6992
Wlroots does not yet support the newer xdg-shell versions and now
requires the compositor to set the supported xdg-shell version during
creation. Set this to v2 for sway as well.
Fixes https://github.com/swaywm/sway/issues/7001
strncpy is useless here, is dangerous because it doesn't guarantee
that the string is NUL-terminated and causes the following warning:
../sway/criteria.c: In function ‘criteria_parse’:
../sway/criteria.c:712:25: error: ‘strncpy’ destination unchanged after copying no bytes [-Werror=stringop-truncation]
712 | strncpy(value, valuestart, head - valuestart);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Our layer shell implementation assigns every layer surface to an output
on creation. It tracks this output using the output field on the
underlying wlr_layer_surface_v1 structure. As such, much of the existing
code assumes that output is always non-NULL and omits NULL checks
accordingly.
However, there are currently two cases where we destroy a
sway_layer_surface and output is NULL. The first is when we can't find
an output to assign the surface to and destroy it immediately after
creation. The second is when we destroy a surface in response to its
output getting destroyed, as we set output to NULL in
handle_output_destroy() before we call wlr_layer_surface_v1_destroy(),
which is what calls the appropriate unmap and destroy callbacks.
The former case doesn't cause any problems, since we haven't even
allocated a sway_layer_surface at that point or registered any
callbacks. The latter case, however, currently triggers a crash (#6120)
if a popup is visible, since our popup_handle_unmap() implementation
can't handle a NULL output.
To fix this issue, keep output set until right before we free the
sway_layer_surface. All we need to do is remove some of the cleanup
logic from handle_output_destroy(), since as of commit c9060bcc12
("layer-shell: replace close() with destroy()") that same logic is
guaranteed to be happen later when wlroots calls handle_destroy() as
part of wlr_layer_surface_v1_destroy().
This lets us remove some NULL checks from other unmap/destroy callbacks,
which is nice. We also don't need to check that the wlr_output points to
a valid sway_output anymore, since we unset that pointer after disabling
the output as of commit a0bbe67076 ("Address emersions comments on
output re-enabling") Just to be safe, I've added assertions that the
wlr_output is non-NULL wherever we use it.
Fixes#6120.
The existing code gives this error when compiled with GCC 12:
../sway/server.c: In function ‘server_init’:
../sway/server.c:217:75: error: ‘%d’ directive output may be truncated writing between 1 and 11 bytes into a region of size 8 [-Werror=format-truncation=]
217 | snprintf(name_candidate, sizeof(name_candidate), "wayland-%d", i);
| ^~
../sway/server.c:217:66: note: directive argument in the range [-2147483647, 32]
217 | snprintf(name_candidate, sizeof(name_candidate), "wayland-%d", i);
| ^~~~~~~~~~~~
../sway/server.c:217:17: note: ‘snprintf’ output between 10 and 20 bytes into a destination of size 16
217 | snprintf(name_candidate, sizeof(name_candidate), "wayland-%d", i);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Because i is never negative, this is a false positive, but it is easy to
change i to unsigned to silence the error.
active_keyboard may be NULL, in which case an invalid pointer could be
passed to wlr_input_method_keyboard_grab_v2_send_modifiers. This
procedure call is unnecessary since wlroots commit 372a52ec "input
method: send modifiers in set_keyboard", so the call can simply be
removed.
Fixes#6836.
Currently, a floating window that's been fullscreened can send us
xdg_toplevel::move, and we'll enter seatop_move_floating, which lets us
drag the surface around while it's fullscreen. We don't want
this--fullscreen surfaces should always be aligned to the screen--so add
the same check that seatop_default already does when entering this mode.
Tested with Weston's weston-fullscreen demo, which sends a move request
if you click anywhere on its surface.
When REAPER submenu is closed `XCB_CLIENT_MESSAGE` with type
`NET_ACTIVE_WINDOW` is sent to set focus to parent menu.
Closes: https://github.com/swaywm/sway/issues/6324
Commit 37d7bc6998 ("transaction: Only wait for ack from visible
views") introduced a check which uses view_is_visible() to check if a view
is still visible on the screen. However view_is_visible() will early
return in case the node is in the destroying state. This is incorrect
for transactions, since a destroying view which is visible will trigger
configure events for other clients. This bug was visible when repeatedly
opening and closing two views side by side, since we ignore the
destroying node we get a frame where the still open view is shown with
the old configure values and the rest is the desktop background. The
next frame is than correct again.
Fix this by considering destroying views as visible, we correctly wait
for them and send the configure events to other views in time, fixing
the background flicker.
Fixes#6473
`popup_unconstrain` uses view coordinates to init the output box for
popups. However wlroots expects the box to be set in a toplevel surface
coordinate system, which is not always equal to view. The difference
between those is a window geometry set via xdg-shell.
GTK4 reserves some space for client-side decoration and thus has a
window with top left corner not matching to (0, 0) of a surface. The box
calculated without taking that into account was slightly shifted
compared to the actual output and allowed to position part of the popup
off screen.
SUID privilege drop is needed for the "builtin"-backend of libseat,
which copied our old "direct" backend behavior for the sake of
compatibility and ease of transition.
libseat now has a better alternative in the form of seatd-launch. It
uses the normal seatd daemon and libseat backend and takes care of SUID
for us.
Add a soft deprecation warning to highlight our future intent of
removing this code. The deprecation cycle is needed to avoid surprises
when sway no longer drops privileges.
Followup on 4e4898e90f.
If a view quickly maps and unmaps repeatedly, there will be multiple
destroyed containers with same view in a single transaction. Each of
these containers will then try to destroy this view, resulting in use
after free.
The container should only destroy the view if the view still belongs
to the container.
Simple reproducer: couple XMapWindow + XUnmapWindow in a loop followed
by XDestroyWindow.
See #6605
We currently track the focus of a seat in two ways: we use a list called
focus_stack to track the order in which nodes have been focused, with
the first node representing what's currently focused, and we use a
variable called has_focus to indicate whether anything has focus--i.e.
whether we should actually treat that first node as focused at any given
time.
In a number of places, we treat has_focus as implying that a focused
node exists. If it's true, we attempt to dereference the return value of
seat_get_focus(), our helper function for getting the first node in
focus_list, with no further checks. But this isn't quite correct with
the current implementation of seat_get_focus(): not only does it return
NULL when has_focus is false, it also returns NULL when focus_stack
contains no items.
In most cases, focus_stack never becomes empty and so this doesn't
matter at all. Since focus_stack stores a history of focused nodes, we
rarely remove nodes from it. The exception to this is when a node itself
goes away. In that case, we call seat_node_destroy() to remove it from
focus_stack and free it. But we don't unset has_focus if we've removed
the final node! This lets us get into a state where has_focus is true
but seat_get_focus() returns NULL, leading to a segfault when we try to
dereference it.
Fix the issue both by updating has_focus in seat_node_destroy() and by
adding an assertion in seat_get_focus() that ensures focus_stack and
has_focus are in sync, which will make it easier to track down similar
issues in the future.
Fixes#6395.
[1] There's some discussion in #1585 from when this was implemented
about whether has_focus is actually necessary; it's possible we could
remove it entirely, but for the moment this is the architecture we have.
cairo_image_surface_create can fail, e.g. when running out of
memory or when the size is too big. Avoid crashing in this case.
Closes: https://github.com/swaywm/sway/issues/6531
Now output_begin_destroy emits the node::destroy event similar to
workspace_begin_destroy. It currently has no listeners, since they
listen to output::disable or wlr_output::destroy instead.
We use the headless backend to create a special fallback output
used when no other output is connected. However this messes up the
"real" headless output names users have come to expect (e.g.
currently the first headless output will be named "HEADLESS-2"
instead of "HEADLESS-1").
Fix this by setting the output name with [1].
[1]: https://gitlab.freedesktop.org/wlroots/wlroots/-/merge_requests/3395
sway-bar(5) says:
> For compatibility with i3, bar mode <mode> [<bar-id>] syntax is
> supported along with the sway only bar <bar-id> mode <mode> syntax.
while the actual behavior is that `bar_cmd_mode` ignores already
selected `config->current_bar` and applies the change to all the
configured bars.
This makes it possible to hint to the renderer and backends how many
bits per channel the buffers that the compositor draws windows onto
should have. Renderers and backends may deviate from this if they
do not support the formats with higher bit depth.
