We oftne use -1 to indicate unset values. In case of output (x, y), we
would consider the fields set if they are not both -1. This means that
(0, -1) and (-1, 0) are valid coordinates, but (-1, -1) is not.
We support negative output positioning, so we cannot use -1 to mean
unset. Zero is also not an option as that would disallow reverting a set
position back to (0, 0).
INT_MAX is an unreasonable output position, so use it to indicate unset
values, and only use the value when both are set.
The function `load_include_configs` already changes the directory to the
one containing the parent config. Therefore, `load_include_config`
trying to assemble the "full" path leads to repetition of path segments,
making the `realpath` call fail with ENOENT.
Just calling `realpath` on the path itself from the directory with the
parent configuration is sufficient, so there is no point in passing
`parent_dir` to `load_include_config`.
problem: an invalid usage of the command resize set will cause sway to crash because it doesn't check for an invalid height.
solution: validate height along with width.
This ensures these functions are always called (even when a library
such as wlroots or libc perform the fork) and removes the need to
manually call them.
In [1] we re-introduced a debug flag to enable wl_drm. Time has
passed and Xwayland + VA-API + amdvlk now all support
linux-dmabuf-v1.
[1]: https://github.com/swaywm/sway/pull/7916
According to the spec, the closed event should be sent when the surface
is no longer shown, because the output may have been destroyed or the
user may have asked for it to be removed. In such cases, the clients
should destroy the resource.
This fixes mako not being able to show notifications if the assigned
output was destroyed while a notificataion was still visible
Fixes: 188811f808 ("scene_graph: Port layer_shell")
Destroying the wlr_renderer in a callback to its own renderer_lost event
is unsafe due to wl_signal_emit*() still accessing it after it was
destroyed.
Delegate recreation of renderer to an idle callback and ensure that only
one such idle callback is scheduled at a time by storing the returned
event source.
Change begin_destroy to remove event listeners before the final destroy,
since otherwise event listeners would be removed twice, which crashes.
This fixes a crash in wlroots listener checks. See #8509.
sway_input_method_relay can be destroyed from two sources, either the
seat is destroyed or the manager protocol objects are destroyed due
compositor exit.
This fixes a crash in wlroots listener checks. See #8509.
While we we cannot easily check for true visibility of layer surfaces as
easily as for views, we can check at least check that the output
associated with the surface is enabled.
We have historically considered surfaces without a view visible. This
made sense in case of layer surfaces which do not have a view, but it
also allows unmapped surfaces to act as global inhibitors irrespective
of the current view state, which is not the intention fo the protocol.
As we now explicitly handle layer surfaces, assume that views are only
visible if they can be found and their visibility checked.
When a session is locked, no views are visible and there is no way to
interact with the user session. This means that all inhibitors based on
visibility - with the exception being inhibitors on the session lock
surfaces themselves - become inert, allowing the session to go idle.
The only inhibitor type on normal views that one could argue should
remain active is INHIBIT_IDLE_OPEN, but for now we disable all view
inhibitors regardless of type.
`swaymsg -t get_inputs --raw` calls it a pointer but `--pretty` calls it
a Mouse. Previous commit 6737b90cb that set this to pointer probably
forgo to update the pretty one.
closes#8584
Commit c2d6aff added a bounds check on `height - title_bar_height`,
repurposing the local variable `height` in an attempt to DRY out the
expression.
However, because re-assignment occurs inside the loop body, its result
would leak across loop iterations, compounding its effect and leading
to the artifact reported in issue #8625, where each child except the
first in a tabbed container would acquire a visible waterline.
Introduce a second variable and reset it in each loop iteration to get
rid of the waterline.
Fixes#8625.
Replace the XWayland-only class attribute in the examples:
- The first example given should work for Wayland-native windows.
- The example 'Kill all windows with the title "Emacs"' should use
title, not class. Also, it's a substring (regex) match.
There are many different implementations of regular expressions with
incompatible syntax. For example, GNU grep alone provides three
different ones. Clarify the use of PCRE2 by sway criteria.
If far too many containers are created, they can become so small that
their size calculations come out negative, leading to crashes on
asserts.
Instead, set a lower bound for sizes and disable the container entirely
if it goes below it, giving whatever space it used to the last
container.
The splits are not recalculated, so currently the effect is that if all
containers have the same width fraction, they keep getting narrower
until at some point they all round to zero and the last container will
be given all the available space.
A better behavior would have been if the additional container did not
contribute to size and fraction calculations at all, but it's an extreme
edge-case, anything is better than crashing, and this is easier to
implement.
has_prefix() expects the prefix to be the 2nd argument, not the first.
The config parsing was broken when using `--input-device=`.
Introduced by: 0c60d1581f "Use has_prefix()
instead of strncmp() throughout"
in the origin text_input.c, we only check the sway_view and layershell,
but now we have the third shell named sessionlock, so we need to modify
both text_input.c and view.c to handle the new type of shell
When the repaint timer fires, we check if the sway_output is disabled,
and if so, skip the output commit after having reset frame_pending.
The sway_output enable flag is only updated if the output is disabled
and removed from the layout, not if the power is disabled for e.g. idle.
This can lead to situations where a commit is attempted on a disabled
output, which will lead to an attempted and failed primary swapchain
allocation.
Use the wlr_output.enabled state to check if the output is active.
Call wlr_scene_output_set_position when in global fullscreen to
correctly set output positions when repositioning outputs (using
swaymsg output or similar).
swaybar and the exec command reset signal masks, signal handlers and
NOFILE limit before exec, but swaybg was missing all that.
Reset it for swaybg as well.
The behavior of handlers registered with signal(3p) is not well-defined
for signals delivered more than once, as laid out in the man page.
We should replace our use of signal with sigaction, but for SIGCHLD
specifically we can also just skip the signals altogether by setting the
handler to SIG_IGN which causes child reaping to not be required.
Fixes: https://github.com/swaywm/sway/pull/8567
To use something like:
[con_id=__focused__] mark --add --toggle foo
The container must currently have a view. However, it is possible to
focus parent containers that do not have a view. For example, via:
focus parent
Since containers without views can be the focused (meaning the container
is marked "focused": true in the output of: swaymsg -t get_tree), it
seems reasonable that a view is not required to target a container via
__focused__.
cmd_exec_process is used whenever sway is meant to execute a child
process on behalf of the user, and had a lot of complexity.
In order to avoid having to wait on the user's process, a double-fork
was used, which in turn required us to wait on the outer process. In
order to track the child PID for launcher purposes, a pipe was used to
transmit the PID back to sway.
This resulted in sway blocking for 5-6 ms per exec on my system, which
is quite significant. The error handling was also quite lacking - the
read loop did not handle errors at all for example.
Instead, teach sway to handle SIGCHLD and do away with the double-fork.
This in turn allows us to get rid of the pipe as we can record the
child's PID directly. This reduces the time we block to just 1.5 ms on
my system. We'd be able to get down to just 150 µs if we could use
posix_spawn(3), but posix_spawn(3) cannot reset NOFILE. clone(2) or
vfork(2) would be alternatives, but that presents portability issues.
This change is replicated for swaybar, swaybg and swaynag handling,
which had similar albeit less complicated implementations.
max_width was applied to the source box, but not to the cairo surface.
The cairo surface would therefore take on arbitrarily large dimensions
according to the required dimensions to fit the text input, which if
large enough would cause failures during output rendering and leave a
black hole in the titlebar.
Calling container_update() wasn't enough: If there is no visible window
decorations (title bar, borders) container_update would basically no-op
and the scene wouldn't repaint with the update alpha. By also calling
output_configure_scene() we force a call to
wlr_scene_buffer_set_opacity() thus ensuring we update the scene.
Closes: #8580
currently, the output background command handler prematurely
returns with an error if the background file cannot be accessed.
It should only error if user did not provide fallback color.
closes#8556
Changes
- Introduce variables to avoid uneccessary writing on output members
- Log a debug message when fallback is being used over inaccessible
file
- Always parse the background color and swaynag warn if it is incorrect
- when updating output member variables, free previous values
- add cleanup label and goto it if `strdup` fails
- Move output->member initializations to before parsing fallback, Also
free and init output->background as well
