To update your local repository to use the new default branch, these
commands may help:

$ git fetch origin
$ git checkout master
$ git branch -m main
$ git branch --set-upstream-to=origin/main

Signed-off-by: Daniel Stone <daniels@collabora.com>

Alexander Dunaev authored Feb 17, 2021 and Simon Ser committed Apr 27, 2021

The user may install cursor themes manually, and the desktop environment
may provide a special directory for storing them.  For instance, GTK puts
those themes into ~/.local/share/icons, and many Linux distributions suggest
using that directory for user-specific themes.  However, users of
libwayland-cursor cannot load these themes using the API provided by the
library because the latter does not look into that directory.

This patch adds ~/.local/share/icons to the search path, so user-specific
themes can be loaded through the API provided by libwayland-cursor.

Signed-off-by: Alexander Dunaev <adunaev@igalia.com>
Reviewed-by: Emmanuel Gil Peyrot <linkmauve@linkmauve.fr>
Reviewed-by: Simon Ser <contact@emersion.fr>

When building for a product, tests are not needed.

Besides, one test requires a C++ compiler, which is not always
available.

So, add an option to configure to disable building tests altogether.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>

Signed-off-by: James Hilliard <james.hilliard1@gmail.com>

Alexander Richardson authored Mar 18, 2021 and Simon Ser committed Apr 15, 2021

The fcntl() argument is defined to be an int and not a long. This does not
matter on most architectures since the value is passed in registers, but
it causes issues on big-endian architectures that pass variadic arguments
on the stack.

Signed-off-by: Alex Richardson <Alexander.Richardson@cl.cam.ac.uk>
Reviewed-by: Simon Ser <contact@emersion.fr>

Alexander Richardson authored Mar 18, 2021 and Simon Ser committed Apr 15, 2021

We can't just unconditionally read the optional arguments (and also read
it as a void* despite actually being an int).
While this happens to work on most architectures because the first few
variadic arguments are passed in registers, this is non-portable and
causes a crash on architectures that set bounds on variadic function
arguments (for example CHERI-enabled architectures). It could also cause
problems on big-endian architectures that pass variadic arguments on the
stack rather than in registers.

For CHERI-MIPS, reading sizeof(void*) causes a read of 16 bytes from the
bounded varargs capability. This always crashes since even calls with the
optional argument only have 4 bytes available.

Signed-off-by: Alex Richardson <Alexander.Richardson@cl.cam.ac.uk>
Reviewed-by: Simon Ser <contact@emersion.fr>

This code is only required for building wayland-scanner so it should be
scoped accordingly. libxml-2.0 will only be required if both "scanner"
and "dtd_validation" are set to true.

Signed-off-by: Michael Weiss <dev.primeos@gmail.com>

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>

While MAP_FAILED is generally defined to that value, we should not be
relying on implementation details of system headers.

Signed-off-by: Alex Richardson <Alexander.Richardson@cl.cam.ac.uk>
Reviewed-by: Simon Ser <contact@emersion.fr>

Meson now replaces autotools.

Signed-off-by: Simon Ser <contact@emersion.fr>

Vlad Zahorodnii authored Feb 24, 2021 and Simon Ser committed Mar 05, 2021

This can be useful if the compositor wants to call wl_global_destroy() with some
delay but it doesn't have the wl_display object associated with the global,
which is needed to get access to the event loop.

Signed-off-by: Vlad Zahorodnii <vlad.zahorodnii@kde.org>

In wl_proxy_set_queue, passing a wl_event_queue from a completely
unrelated wl_display could lead to object IDs mismatches.

Add an assertion to catch this case. It's always a user bug if this
happens.

Signed-off-by: Simon Ser <contact@emersion.fr>

Signed-off-by: August Svensson <a.sve@live.se>

Tadeo Kondrak authored Jun 11, 2020 and Simon Ser committed Jan 27, 2021

It's the only destructor event in the core protocol, and destructor
events were previously unannotated.

Signed-off-by: Tadeo Kondrak <me@tadeo.ca>

Tadeo Kondrak authored Jun 11, 2020 and Simon Ser committed Jan 27, 2021

This allows specifying events to be destructors, which is useful for
non-C language bindings. It is unused in wayland-scanner.

Signed-off-by: Tadeo Kondrak <me@tadeo.ca>

This adds 4 new formats.

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>
Acked-by: Daniel Stone <daniels@collabora.com>

The compositor should handle absolute paths in WAYLAND_DISPLAY like the clients, ie not
adding the XDG_RUNTIME_DIR prefix if it's an absolute path.

This allows to create the wayland socket in a separate directory for system compositors if
desired. Clients could then directly inherit the environment variable.

Signed-off-by: Loïc Yhuel <loic.yhuel@softathome.com>

These TODO items are outdated (either done or not applicable anymore). Let's
migrate items that we want to keep to GitLab issues.

Signed-off-by: Simon Ser <contact@emersion.fr>

This is already done in autotools, and fixes the build with glibc < 2.17.

Signed-off-by: Loïc Yhuel <loic.yhuel@softathome.com>

The current wl_output.mode event has several issues when used to advertise
modes that aren't current:

- It's not possible to remove some modes. This is an issue for virtual
  outputs and when the kernel prunes some modes because of link limitations.
- wl_output.mode fails to carry metadata such as aspect ratio, which results
  in duplicated or missing modes.
- It's not clear, given the current set of protocols, how non-current modes
  are useful to clients. Xwayland ignores non-current modes.

GNOME and wlroots already only advertise the current mode because of these
issues.

If a protocol needs the clients to know about all available modes, it should
advertise these modes itself instead of relying on wl_output.mode.

Signed-off-by: Simon Ser <contact@emersion.fr>
Closes: #92

Some clients rely on wl_surface.enter/leave to start/stop their rendering
loop.

There are cases where this doesn't work:

- Some compositors don't send wl_surface.leave when a toplevel is hidden.
  For instance Sway doesn't send this event when a toplevel is on an inactive
  workspace (but doesn't send wl_surface.frame events).
- Some compositors might still want applications to continue to render even if
  away from outputs. For instance a compositor that allows to screen record
  individual toplevels might not send a wl_surface.enter event for hidden
  toplevels.

Signed-off-by: Simon Ser <contact@emersion.fr>

Alyssa Ross authored Oct 29, 2020 and Simon Ser committed Oct 29, 2020

This setting makes Docbook section IDs consistent, and should allow
Wayland builds that include documentation to be fully reproducible.

Signed-off-by: Alyssa Ross <hi@alyssa.is>

Nick Yamane authored Jul 08, 2020 and Simon Ser committed Oct 29, 2020

Explicitly say that destroying a wl_data_source previously used in a
wl_data_device::start_drag request will cancel the DND session. This
is currently the only way to do it from client side (besides those
"indirect" ones already documented) and all compositors work like that
but it is not clear spec-wise it is the expected behaviour.

Signed-off-by: Nick Diego Yamane <nickdiego@igalia.com>

Without the casts the bytes accesses get converted to int. but int is
not guaranteed to be 4 bytes large. Even when it is 4 bytes large
`bytes[3] << 24` does not fit because int is signed.

Signed-off-by: Valentin Kettner <vakevk@gmail.com>

This type is meant to be 4 bytes large as seen in _XcursorReadUInt which
always reads 4 bytes. An unsigned int is often 4 bytes large but this
isnt' guaranteed so it is cleaner to use the exact type we want.

Signed-off-by: Valentin Kettner <vakevk@gmail.com>

Peter Hutterer authored Jul 09, 2020 and Pekka Paalanen committed Aug 18, 2020

This makes the Bug template available in the new issues form. Ideally a majority
of users will select this template and then realize that they're in the wrong
repo.

Making this a default is a feature of the enterprise gitlab version only, see
https://docs.gitlab.com/ce/user/project/description_templates.html#overview



Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>

Signed-off-by: Christopher James Halse Rogers <christopher.halse.rogers@canonical.com>

Explain that wl_keyboard.key yields platform-specific key codes.
Some compositors use Linux key codes (defined in the
linux/input-event-codes.h header file, e.g. KEY_ESC), however
clients should not assume that this is always the case. The only
reliable way for clients to interpret key codes is to feed them
into a keyboard mapping.

Signed-off-by: Simon Ser <contact@emersion.fr>