GitLab

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>

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#overviewSigned-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>

Clients must assume that every key, including modifiers, are lifted
after getting wl_keyboard.leave event, not doing so could lead to key repeat
keep going after the user changes focus, or stuck modifiers, when the user
interacts with client without keyboard focus.
Signed-off-by: Kirill Chibisov <contact@kchibisov.com>

Consider that we set the modifier for the modifier key as well, and that
you have the option to change your layout when both shifts pressed. This
realized by making the shift keys produce 'Group_switch' when you press
the shift key and the shift modifier is already active. So sending modifier
before the key event will result in 'Group_switch' each time you press
shift.

That being said, the order of modifiers should be updated after the
key/enter event, so it'll only affect future keypresses, not the current one.

See: https://lists.x.org/archives/xorg-devel/2014-July/043110.htmlSigned-off-by: Kirill Chibisov <contact@kchibisov.com>

Signed-off-by: Emmanuel Gil Peyrot <linkmauve@linkmauve.fr>

Signed-off-by: Emmanuel Gil Peyrot <linkmauve@linkmauve.fr>

Signed-off-by: Emmanuel Gil Peyrot <linkmauve@linkmauve.fr>

Signed-off-by: Emmanuel Gil Peyrot <linkmauve@linkmauve.fr>

The protocol says:

> It is a protocol violation to issue this request on a seat that
> has never had the pointer capability.

But never defines an error code.
Signed-off-by: Simon Ser <contact@emersion.fr>
Closes: #165

Instead of building a new debian image every time we build the repo, let's build
it once and re-use it. This way we're more  likely to spot actual bugs vs issues
with the distribution (or mirrors). This is the same approach mesa, gstreamer,
libinput, etc. already use.

The pipeline consists of two parts: container-build and distribution-image. Both
use the FDO_DISTRIBUTION_... variables to do their thing, the former to build an
image, the latter to use that image. The container-build step is a noop if an
image with that tag already exists in the registry, the templates take care of
all that.

The .debian.buster section groups the variables needed to easily extend all jobs
requiring buster. Not all variables are used by all jobs but meh.

The grouping is slightly odd because some debian-specific variables are in
the global variables and others in the .debian.buster section. This grouping
will make things easier if we extend to build on other distributions - then we
have all packages and tags in one place.

Because buster doesn't have a recent-enough version of meson, we install that
from pip.

Fixes #79Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.com>

Different indentation and we don't want a fixed line length here, wrapping yaml
is prone to introduce bugs.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.com>

Signed-off-by: Yann Dirson <ydirson@free.fr>

This is the style used in wayland.xml which is the only file we really
care about for git blame information. So let's adjust all others to that
style for consistency and fix editorconfig to avoid messing this up in
the future.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>

This allows to include client and server headers in the same file
fixing warnings like

In file included from ../subprojects/wlroots/include/wlr/types/wlr_layer_shell_v1.h:16,
                 from ../src/desktop.h:16,
                 from ../src/server.h:13,
                 from ../tests/testlib.c:8:
tests/59830eb@@footest@sta/wlr-layer-shell-unstable-v1-protocol.h:80:34: warning: redundant redeclaration of ‘zwlr_layer_shell_v1_interface’ [-Wredundant-decls]
   80 | extern const struct wl_interface zwlr_layer_shell_v1_interface;
      |                                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ../tests/testlib.h:8,
                 from ../tests/testlib.c:7:
tests/59830eb@@footest@sta/wlr-layer-shell-unstable-v1-client-protocol.h:77:34: note: previous declaration of ‘zwlr_layer_shell_v1_interface’ was here
   77 | extern const struct wl_interface zwlr_layer_shell_v1_interface;
      |                                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ../subprojects/wlroots/include/wlr/types/wlr_layer_shell_v1.h:16,
                 from ../src/desktop.h:16,
                 from ../src/server.h:13,
                 from ../tests/testlib.c:8:
tests/59830eb@@footest@sta/wlr-layer-shell-unstable-v1-protocol.h:106:34: warning: redundant redeclaration of ‘zwlr_layer_surface_v1_interface’ [-Wredundant-decls]
  106 | extern const struct wl_interface zwlr_layer_surface_v1_interface;
      |                                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ../tests/testlib.h:8,
                 from ../tests/testlib.c:7:
tests/59830eb@@footest@sta/wlr-layer-shell-unstable-v1-client-protocol.h:103:34: note: previous declaration of ‘zwlr_layer_surface_v1_interface’ was here
  103 | extern const struct wl_interface zwlr_layer_surface_v1_interface;
Signed-off-by: Guido Günther <agx@sigxcpu.org>
Closes: #158

There is only one page written. Having manually-written man pages duplicates
information with doc comments. Besides, man pages are already generated by
Doxygen.
Signed-off-by: Simon Ser <contact@emersion.fr>
Closes: #156

Add a paragraph about WAYLAND_SOCKET and describe what happens when the display
name is a relative path.
Signed-off-by: Simon Ser <contact@emersion.fr>

This commit updates the README build & install instructions.
It replaces that obsolete "autogen && make" with "meson && ninja"
Signed-off-by: Ricardo Quesada <ricardoquesada@gmail.com>

Signed-off-by: Matt Turner <mattst88@gmail.com>

Many new and valuable features were added between Meson 0.49 and 0.52.1.
We would like to use some of them.
Signed-off-by: Matt Turner <mattst88@gmail.com>

Stretch is old-stable and will reach end of life this year.

buster-backports has newer Meson available, so switching to Buster will
allow us to bump the Meson requirements.
Signed-off-by: Matt Turner <mattst88@gmail.com>

Wayland requires a binary, wayland-scanner, to be run during the build
process. For any configuration other than native builds (including
cross compiling and even 32-bit x86 builds on an x86-64 build machine)
Wayland's build process builds and uses its own wayland-scanner.

For any builds using a cross file, wayland-scanner is built for the host
machine and therefore cannot be executed during the build of the Wayland
libraries. Instead builds using a cross file must execute the build
machine's wayland-scanner (typically /usr/bin/wayland-scanner).

As such, to build Wayland's libraries for a non-native ABI a package
manager must build and install /usr/bin/wayland-scanner first. But then
the build for the native ABI then rebuilds wayland-scanner itself and
doesn't use the system's, and worse, wants to install its own, which
conflicts with the /usr/bin/wayland-scanner already installed!

So, add the -Dscanner=... option to control whether to install
wayland-scanner.
Signed-off-by: Matt Turner <mattst88@gmail.com>

We have always built libwayland with the scanner from the same build so
that the generated code and installed headers are exactly up-to-date
with the libwayland version. If libwayland was to use a scanner later
than itself, the scanner might do things that are not available in the
libwayland at hand, leading to a broken build or a broken library
(headers).
Signed-off-by: Matt Turner <mattst88@gmail.com>

This allows the compositor to send an error when the client submits a buffer
whose size is not divisible by the buffer scale. Previously, the protocol said
it was a client error but didn't specify any error code.
Signed-off-by: Simon Ser <contact@emersion.fr>
Closes: #145

conversion to 'wl_fixed_t {aka int}' from 'int64_t {aka long int}'
may alter its value [-Werror=conversion]

Signed-off-by: Jan Beich <jbeich@FreeBSD.org>

Signed-off-by: Jan Beich <jbeich@FreeBSD.org>

ZFS on FreeBSD >= 12.0 returns EINVAL, see
https://svnweb.freebsd.org/changeset/base/325320Signed-off-by: Jan Beich <jbeich@FreeBSD.org>

In file included from ../tests/connection-test.c:43:
In file included from ../tests/test-compositor.h:30:
../src/wayland-client.h:40:10: fatal error: 'wayland-client-protocol.h' file not found
 #include "wayland-client-protocol.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ../tests/display-test.c:45:
In file included from ../src/wayland-server.h:104:
src/wayland-server-protocol.h:4454:2: error: unterminated /* comment
        /**
        ^
In file included from ../tests/cpp-compile-test.cpp:2:
In file included from src/wayland-server-protocol.h:8:
In file included from ../src/wayland-server.h:104:
src/wayland-server-protocol.h:3:2: error: unterminated conditional directive
 #ifndef WAYLAND_SERVER_PROTOCOL_H
  ^
../tests/headers-protocol-test.c:33:2: error: including wayland-server-protocol.h did not include wayland-server.h!
 #error including wayland-server-protocol.h did not include wayland-server.h!
  ^
In file included from ../tests/headers-protocol-test.c:26:
In file included from src/wayland-client-protocol.h:8:
In file included from ../src/wayland-client.h:40:
src/wayland-client-protocol.h:1358:2: error: unterminated conditional directive
 #ifndef WL_SHM_FORMAT_ENUM
  ^
In file included from ../tests/protocol-logger-test.c:34:
In file included from ../src/wayland-client.h:40:
src/wayland-client-protocol.h:2613:1: error: unterminated /* comment
/**
^
../tests/resources-test.c:49:36: error: use of undeclared identifier 'wl_seat_interface'
        res = wl_resource_create(client, &wl_seat_interface, 4, 0);
                                          ^

This allows users to build Wayland with Meson from an autotools-generated
release tarball.
Signed-off-by: Simon Ser <contact@emersion.fr>
Closes: #141

When running tests with ASan, proxy-test fails at the proxy_tag test:

    ==27843==ERROR: LeakSanitizer: detected memory leaks

    Direct leak of 32 byte(s) in 1 object(s) allocated from:
        #0 0x7f65a732dada in __interceptor_malloc /build/gcc/src/gcc/libsanitizer/asan/asan_malloc_linux.cc:144
        #1 0x7f65a71cb3ea in wl_display_add_protocol_logger src/wayland-server.c:1813
        #2 0x557c640c0980 in proxy_tag tests/proxy-test.c:104
        #3 0x557c640c1159 in run_test tests/test-runner.c:153
        #4 0x557c640c1e2e in main tests/test-runner.c:337
        #5 0x7f65a6ea0ee2 in __libc_start_main (/usr/lib/libc.so.6+0x26ee2)

    SUMMARY: AddressSanitizer: 32 byte(s) leaked in 1 allocation(s).

Destroying the logger fixes the leak.
Signed-off-by: Simon Ser <contact@emersion.fr>
Fixes: 493ab79b ("proxy: Add API to tag proxy objects")

Some filesystems do not support fallocate and return EOPNOTSUPP.

On musl-based distros libwayland-cursor exits abruptly which causes the
application to crash. Unlike glibc, musl does not provide a fallback
mechanism for handling unsupported fallocate. Instead, musl developers
argue that application should handle the case of unsupported system
call.

This commit allows falback to ftruncate in case when EOPNOTSUPP
was recieved.
Signed-off-by: Ihor Antonov <ihor@antonovs.family>

Signed-off-by: Emmanuel Gil Peyrot <linkmauve@linkmauve.fr>