Rename screenshooter protocol to weston_screenshooter

Due to the effort of moving a way from non-prefixed protocols, rename
the weston specific screenshooter protocol to weston_screenshooter.

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

desktop-shell: Rename protocol weston_desktop_shell

In the effort of going away from generic names of protocols only
relevant for weston, rename the weston desktop shell
weston_desktop_shell.

This also resets the version to 1, as there will be no prior versions
to weston_desktop_shell.

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

Use xdg_shell protocol from wayland-protocols

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Acked-by: Jasper St. Pierre <jstpierre@mecheye.net>

Makefile.am: Make the external xml scanning rule version generic

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

Use input method protocol from wayland-protocols

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Reviewed-by: Jan Arne Petersen <janarne@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

Use text input protocol from wayland-protocols

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Reviewed-by: Jan Arne Petersen <janarne@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

Use linux-dmabuf protocol from wayland-protocols

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

Use fullscreen-shell.xml from wayland-protocols

Use the fullscreen-shell protocol XML from the wayland-protocols
installation, and remove the one we provide ourself.

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

compositor-wayland: Change focus to a bool instead of an int

Also rename focus to has_focus to be slightly less confusing.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

desktop-shell: use weston_pointer_send_axis

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>

shared: work around EGL header idiosyncrasy

When no X11 headers are present eglplatform.h will break unless
certain defines are set prior to its inclusion.

including wayland-egl.h defines WL_EGL_PLATFORM which will stop
the attempted inclusion of the X11 headers.

Maybe this isn't the best solution to the problem, but it's harmless
and gets the job done.

Closes bug: https://bugs.freedesktop.org/show_bug.cgi?id=92104

Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
(patch by Ahmet Acar, commit log by Derek Foreman)

compositor-drm: pass NULL to mmap() instead of 0 as the address

mmap() function expects to be passed a void pointer as the address
here. Passing NULL is technically more correct than passing 0.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

compositor-drm: fall back to EGL visual_id 0 if no others work

commit e76f185 stopped using visuals with "native visual id" == 0

This broke some systems, so we now try 0 as a worst case fallback if
everything else has failed.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

input: Pass axis events through pointer grab interfaces

Don't only send motions and buttons but also axis events through the
pointer grab interface.

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-By: Derek Foreman <derekf@osg.samsung.com>

desktop-shell: Add surface_keyboard_focus_lost helper

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-By: Derek Foreman <derekf@osg.samsung.com>

Provide damage region for screen recording on Raspberry Pi

When trying to do sceen recording using the screenshooter, no screen data
was ever processed because the rpi-renderer never set the previous frames
damage area.

Update the rpi-renderer to copy the necessary data.

Signed-off-by: Shinya Saito <ssaito@igel.co.jp>
Signed-off-by: Tomohito Esaki <etom@igel.co.jp>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

ivi-shell: rename ivi_controller_interface

The name of ivi_controller_interface is changed to ivi_layout_interface
with this patch.

This name is better suited to the interface, because it is implemented
in ivi-layout.c and its methods are linked to ivi_layout* functions.

Furthermore, the controller modules (e.g. hmi-controller) are the users
of this interface and they have their own interfaces,
which are called *_controller_interface,
e.g.: ivi_hmi_controller_interface.

This causes confusion about the software architecture.

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
Tested-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>
Reviewed-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>

compositor: Disallow negative geometries in backend output configs

Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Giulio Camuffo <giuliocamuffo@gmail.com>

desktop-shell: Make activate_binding take a view instead of surface

In preparation for further refactoring. This patch also removes a
redundant NULL check. Since we pass views, and views will always have an
associated surface, there is no point of checking if it has.

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

main: stub the functions to load the backends with the new config system

Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

compositor: pass the backend config struct to the backends init function

Add new configuration argument to the backend_init() function, which
will replace the current argc, argv, and config arguments.
After each backend is converted individually the unused parameters
will be removed.

Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

compositor: introduce structs to handle backends configuration

This commit introduces the structs weston_backend_config and
weston_backend_output_config, to prepare for the new config
system for the backends.

Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

zunitc: made name of test fixture parameter explicit.

Instead of using the implicit name 'data', changed the test
with fixture macro ZUC_TEST_F() to use an additional value
to explicitly set the name to use for test data from the
fixture.

Signed-off-by: Jon A. Cruz <jonc@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

zunitc: Clarify documentation on return behavior.

* Clarify documentation on ZUC_ASSERT_* behavior in regards to return
   vs. abort()
* Added overview section on return behavior.
* Fixed spelling
* Removed outdated reference to tap function.

Changes since v1:

* Incorporated grammatical feedback.

Signed-off-by: Jon A. Cruz <jonc@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

clients: don't discard motion if we have a grab

We discard motion outside the window on the assumption it's from before
some event that caused the window to shrink. However, if we have a grab
it's likely that this motion is actually from dragging from the inside
of the window out.

This fixes a problem where drag selecting in weston terminal behaves
oddly - it doesn't update the select region while the drag is happening
outside the window.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Tested-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

clients: try harder to discard motion after resize

If we're going to ignore motion below and to the right when coming
out of maximize, we should probably also ignore it above and to
the left.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Tested-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

weston-launch: Remove unused event source

vt_source is never added to an event loop in this launcher, but
it's potentially removed.

Let's just remove the variable entirely.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

clients: Correct drawing glitch in stacking demo

Fix a graphics glitch in the stacking demo in which a transient
window's drop shadow is visibile within the interior of the window.

Signed-off-by: Ben Hummon <benjamin.hummon@gmail.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

compositor/clients: Protect CLOCK_BOOTTIME with ifdefs

CLOCK_BOOTTIME is a relatively new* feature that may not actually be
present everywhere (I'm looking at you wheezy).  Since our use of it
is actually only cosmetic, I've just ifdef'd if.

* No it isn't.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Giulio Camuffo <giuliocamuffo@gmail.com>

build: Add -lrt for programs that use clock_gettime()

glibc requires this prior to 2.17, and we already do it in a few other
places.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Giulio Camuffo <giuliocamuffo@gmail.com>

devices-test: Fix typo 'destroyes' in a comment

clients: track seat_version per seat, not per display

Apparently it's possible for a compositor to advertise seats with
different versions on the same connection, so this makes us more robust
against that dubious behaviour.

This also tracks the seat version we requested instead of the advertised
maximum.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: David FORT <contact@hardening-consulting.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

libinput: remove unused #define

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

configure: Up the libwayland version requirement

weston commit f7bb9352 requires recent libwayland changes for providing
‘WL_POINTER_RELEASE_SINCE_VERSION’.  Increase the version requirement to
indicate that current weston git requires development version of
wayland.

NOTE: At release we should probably increase the wayland requirement for
weston to 1.10.0.

Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

ivi-shell: install ivi-layout-export.h

IVI-Shell is designed to be used with other controller modules
than hmi-controller.These controller modules require
the ivi-layout-export header file to properly integrate
with the ivi-shell. The header file should be installed
when ivi-shell is enabled, because these controller modules
are not a part of the weston repository.

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

Add commented-out option to weston.ini to set Xwayland path

User can remove the comment marks to make xwayland.so use a local
installed Xwayland server.

compositor-drm: fix cursor size test in drm_assign_planes

In commit 70d337dfd we changed one cursor size test from a hard coded 64,64
to the actual device provided width, height.

The test in drm_assign_planes remained fixed at 64, 64.

The simple-shm test ended up being small enough to fit into a cursor plane
by one test, but too large by the test in drm_assign_planes.  We'd assign
to the cursor plane but not keep a reference.

weston-simple-shm would disappear and be replaced with the previous
cursor image.

This partially "fixes" T3361.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>

screenshooter: correct output ordering of screencast recording when Y-flip is off

The screenshooter encoder wrote the output from either top-to-bottom or
bottom-to-top, depending on the Y-flip setting, but wcap-decode only
decodes from bottom-to-top. Make the encoder always output from
bottom-to-top, to match the decoder, and flip the input (source)
according to the Y-flip setting.

Signed-off-by: Tomohito Esaki <etom@igel.co.jp>
Reviewed-by: Giulio Camuffo <giuliocamuffo@gmail.com>

input: Split get_udev_seat() into its own function

We're going to need this on device removal in the future, so pull it out
into a separate function now.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

compositor-x11: Fix Weston running on a secondary X screen

When DISPLAY is anything other than #.0 weston would still use .0

Fixes
https://bugs.freedesktop.org/show_bug.cgi?id=90532
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

shell: fix maximizing windows on a secondary output

commit f814c5dc9 changed get_output_work_area behaviour
which broke the code for positioning maximized window.
The x position was set to 2*output->x instead of to output->x

fix https://bugs.freedesktop.org/show_bug.cgi?id=92357

Signed-off-by: Marek Chalupa <mchqwerty@gmail.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

weston: release keyboard/touch/pointer objects

This patch adds the missing calls to release when the seat has capabilities
changes. It also fixes a missing release of the touch object and a leak with
old clients.

Signed-off-by: David FORT <contact@hardening-consulting.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>

input: add new client to resource_list instead of focus_resource_list

When new client registers touch listener, it was added to focus resource list.
At this point if another client "without" touch listener is in focus then
subsequent touch events are sent to new client with another client's resources
causing new client to stop rendering.
Now new client is added to resource list by default and it'll be added to focus
resource list only if its in focus.

Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

zoom: remove temporary variables

Signed-off-by: Giulio Camuffo <giuliocamuffo@gmail.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

zoom: Remove unneeded usage of wl_fixed_ts in favour of doubles

Signed-off-by: Giulio Camuffo <giuliocamuffo@gmail.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

input-panel: Fix show_input_panel_surface crash when no surface has focus

Signed-off-by: Nicolas Guyomard <nicolas.guyomard@open.eurogiciel.org>
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

doc: Correct absolute path in doxygen graphs.

Adds a config item to prevent doxygen from using absolute paths in
generated diagrams. This trims off the base directory during config to
leave just relative paths.

Signed-off-by: Jon A. Cruz <jonc@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

cosmetic: Remove a few double semicolons

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>

desktop-shell: NULL check whether a popup parent is a shell surface

get_shell_surface(parent) may return NULL if the client passed a
unassigned wl_surface or a wl_surface with a non-shell surface role
(such as cursor role).

https://bugs.freedesktop.org/show_bug.cgi?id=92316

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Tested-by: Olivier Fourdan <ofourdan@redhat.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

weston-input: Reformat weston_log line to add missing space between words

Signed-off-by: Chris Michael <cp.michael@samsung.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

clients: Moving and rotating transient surfaces

Transient surfaces use child/parent surfaces for stacking order. This
change resloves an issue in which attempting to move or rotate a
toplevel transient surface can move or rotate its ancestor.

animation: Fix potential leak of memory pointed to by move

Free move before return if animation is null.

Signed-off-by: Lucas Tanure <tanure@linux.com>
Reviewed-by: Giulio Camuffo <giuliocamuffo@gmail.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

main: fix destroying the compositor on error while starting

Signed-off-by: Giulio Camuffo <giuliocamuffo@gmail.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

compositor: don't crash if destroying a compositor without a backend

Calling weston_compositor_destroy() on a pointer returned by
weston_compositor_create() should be always valid, even if the
compositor does not have yet a backend.

Signed-off-by: Giulio Camuffo <giuliocamuffo@gmail.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

compositor-rdp: Add missing '\n' in weston_log usage

Signed-off-by: Chris Michael <cp.michael@samsung.com>

cms-colord: Add missing '\n' in weston_log usage

Signed-off-by: Chris Michael <cp.michael@samsung.com>

screen-share: Add missing '\n' in weston_log usage

Signed-off-by: Chris Michael <cp.michael@samsung.com>

ivi-shell: Add missing '\n' in weston_log function usage

Signed-off-by: Chris Michael <cp.michael@samsung.com>

ivi-shell: introduce get_weston_view

The internal API "get_weston_view" is introduced, which returns the
weston_view of the given ivi_layout_surface. The API returns a NULL
pointer, if the ivi_layout_surface does not have any weston_view.

The weston_view is required in many places of ivi-shell implementation.
Therefore, this API will reduce lines of code. Furthermore, it will
increase the maintainability of the ivi-shell implementation.

Old way of getting the weston_view was flawed, because the views list of
the weston_surface is read without controlling the existence of the
weston_surface.  New implementation explicitly throws an error if the
weston_surface does not exist.

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
Tested-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>
Reviewed-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>
[Pekka: Line-wrapped the commit message.]

ivi-shell: remove is_surface_in_layer API

This internal API is redundant, because a surface is allowed to be only
on one layer.

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
Tested-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>
Reviewed-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>

ivi-shell: remove is_layer_in_screen API

This internal API is redundant, because a layer is allowed to be only on
one screen.

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
Tested-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>
Reviewed-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>

ivi-shell: remove struct link_layer

link_layer's sole purpose is to link a surface to multiple layers, if
the surface should be shown in multiple layers.  This can be only
achieved, if the surface has multiple weston_views with different
transformation matrices.

Current implementation assumes in many places that a ivi_surface has
only one weston_view.  Therefore, a surface can be only shown on one
layer.

Although this (a surface on multiple layers) is a nice to have feature
for ivi-shell, it is not very crucial.  In any case, it is not an easy
task to implement this feature, because it has lot of corner cases.

I removed with this patch the link_layer data structure, because it does
not have any purpose in current implementation.

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
Tested-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>
Reviewed-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>
[Pekka: Line-wrapped the commit message.]

ivi-shell: remove struct link_screen

link_screen's sole purpose is to link a layer to multiple screens, if
the layer should be shown in multiple screens.  This can be only
achieved, if surfaces of the layer have multiple weston_views dedicated
to the different screens.

Current implementation assumes in many places that a ivi_surface has
only one weston_view.  Therefore, a layer can be only shown on one
screen.

Although this (a layer on multiple screens) is a nice to have feature
for ivi-shell, it is not very crucial.  In any case, it is not an easy
task to implement this feature, because it has lot of corner cases.

I removed with this patch the link_screen data structure, because it
does not have any purpose in current implementation.

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
Tested-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>
Reviewed-by: Nobuhiko Tanibata <NOBUHIKO_TANIBATA@xddp.denso.co.jp>
[Pekka: Line-wrapped commit message.]

xdg-shell: Clarify ack_configure behaviour

Right now many toolkits (toytoolkit, gtk+ and EFL) will send an
ack_configure request immediately in response to a configure event,
even if they're not immediately committing the surface at that time.

This leads to a situation where multiple configures receive ack_configure
before any commit happens.

There's really no reason for that sequence of events to bother a compositor,
so this just clarifies the language to make it ok.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>

cosmetic: add missing line between variable declarations and code

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>

ivi-layout-transition: Standardize the check for layout_transition_register

Check the return from layout_transition_register in order to fix
potential leak of memory pointed to by transition. And don't register a
null transition.

Signed-off-by: Lucas Tanure <tanure@linux.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

xwm: use always a valid 'primary view' for an X window

The xwm gets a primary view for a X window using the get_primary_view
vfunc of the shell_interface struct. Storing it is dangerous though
because it doesn't listen for its destruction so it may end up using the
old stored view pointer after that view was freed, or after the primary
view for that window was changed to another one.

Fetch the primary view just before using it every time and try to not
abuse this 'primary view' concept which may map badly to some shells:
iterate over all the views instead when it makes sense.

Tested-by: Benoit Gschwind <gschwind@gnu-log.net>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

configure.ac: add explicit enable/disable for systemd-login support

Otherwise, auto-enable depending on whether the system has the necessary
libraries.

[Updated help text as per pq suggestion -- bwh]

Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Tested-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

compositor: remove systemd-notify.h

Systemd notifications support was converted into loadable
module, so systemd-notify.h header is not needed.

Signed-off-by: Egor Starkov <egor.starkov@ge.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

compositor-wayland: Terminate weston_log error message

Signed-off-by: Chris Michael <cp.michael@samsung.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
[Pekka: manually applied]
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

configure.ac: add support for new versions of systemd

Starting from systemd version 209, a single libsystemd.pc is provided.
For previous versions, fall back on libsystemd-login.pc.

Signed-off-by: Frederico Cadete <frederico@cadete.eu>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

weston_wm: Implement _NET_ACTIVE_WINDOW

Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

compositor-wayland: Expose zlinux_dmabuf

This mirrors what 5ffb440c3fe19669de73e51b90ea337fa2d5ae78 did for
compositor-x11, and allows testing without having to restart the
compositor-drm everytime.

Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

systemd: Fix distcheck for typo in header path

compositor: systemd notifications support

Add systemd status and watchdog notification support.
Feature is not compiled by default and can be enabled by
"--enable-systemd-notify" configuration flag. It compiles
into module "systemd-notify.so" and can be loaded by
adding it in weston.ini like any other module, i.e.
"modules=systemd-notify.so". Watchdog timeout equals to
half of timeout defined by "WATCHDOG_USEC" environment
variable, which is set by "WatchdogSec=" setting in
service file.

Signed-off-by: Egor Starkov <egor.starkov@ge.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

tests:  Migrate screenshot code from internal test to client helpers

These routines provide test cases an ability to capture screen images
for rendering verification.

This commit is a no-change refactoring, except for making the routines
non-static.  Makefile rules are also updated; most notably, this links
test clients against the cairo libraries now.

v2: Fix pointer code styling, suggested in review

Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
Acked-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

text: Switch to weston_client_start instead of weston_client_launch

We really don't want to deal with sighups and pids.  It's far easier
to just deal with the client destroyed signal to respawn the input method.

Inspiration taken from Pekka's commit 826dc14ec43b92d

This fixes a potential crash on shutdown.  If the input method client
pointer is still set (sighup handler hasn't run) but the client is
already destroyed, a call to text_backend_destroy() will try to destroy
the client a second time resulting in a segfault.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

text: Don't track binding in struct text_backend

The binding is only used for preventing launch, and we've already got the
pid for that.

With multiple seats there are multiple bindings, so if the most recent
binding unbinds it will clear the pointer, so using it like this is risky
anyway.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

text: Launch input method in text_backend_init

Previously we tried to launch the input method every time a seat was
created, and the launch function would notice it was running and not
bother to actually launch.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>

launcher: fix typo

In the launcher split up patch WESTON_LAUNCHER_SOCK was accidentally
changed to WESTON_LAUNCH_SOCK

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

launcher: fix broken distcheck

We were missing launcher-impl.h in Makefile.am

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

compositor: remove incorrect idle timer initialization

weston_compositor_create() should just create idle timer but not arm it,
because idle-time setting is not ready at this point.

Remove idle timer first update in weston_compositor_create() since
idle_time variable is not set at this point. Idle timer is armed
properly later in weston_compositor_wake().

Signed-off-by: Egor Starkov <egor.starkov@ge.com>
[Pekka: tweaked commit message]
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

fbdev: Ensure strncpy results are null terminated.

Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

ivi-layout-transition: Fix potential leak of memory pointed to by transition

Several fixes to handle invalid transition objects:

1. Free transition in ivi_layout_transition_fade_layer if we fail to
   allocate memory for data

2. Check if transition is not null and if layout_transition_register was
   ok before return, if not, free transition

3. Destroy transition if layout_transition_register was not ok, we can't
   just free transition, we need to call layout_transition_destroy in
   order to free private_data from transition

Signed-off-by: Lucas Tanure <tanure@linux.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

launcher-logind: Remove old VT switching code, move to SwitchTo/Activate

In the time since this code was written, logind has gained new APIs to
handle VT switching automatically and activate sessions. Switch to that.

Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

launcher: Split out launcher implementations into three distinct ones

We now have a launcher interface and distinct implementations for
logind, weston-launch, and direct DRM, each in their own files.

This helps up clean up the spaghetti code into something that's
hopefully a bit more understood. There should be no functional
changes here.

Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
update:  Dropped redundant free() in weston_launcher_destroy()

launcher: Rename logind-util to launcher-logind

We're refactoring this to have multiple launcher "implementations".

Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

compositor-drm: Ensure strings from EDID blobs are null terminated

strncpy only adds null terminating bytes if the source string is smaller
than the destination string.  Since this function relies on the string
being null terminated when checking its contents, we better make sure
there is at least a \0 as the last character.

Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>

terminal: allow copy-paste when selection is scrolled out

We could not paste text when its source went outside the
visible part of the buffer ; this is because we were
incorrectly assuming that our iterator should start at
row 0, while it could very well be negative.

Signed-off-by: Manuel Bachmann <manuel.bachmann@iot.bzh>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

xwm: flush xcb connection only when we processed some event

xwayland source is checked, so it dispatches twice on any event.
If the other turn has no events to dispatch, we flush the connection
redundantly

v2. do not flood logs with 'unhandled event' messages

Signed-off-by: Marek Chalupa <mchqwerty@gmail.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

tests: Convert 'xalloc's to 'zalloc's

xalloc terminates the program abruptly if the requested amount of
memory couldn't be allocated. To insure that the errors are handled
cleanly, use zalloc instead.

Signed-off-by: Seedo Eldho Paul <seedoeldhopaul@gmail.com>
Acked-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

clients: ungrab the correct input when menus close

We need to input_ungrab() on the stored input, not the one that caused
the release - otherwise bad things can happen in multi-seat environments
when a seat that didn't open the menu closes it.

To reproduce:
 configure two seats
 launch weston terminal
 open the right click pop up
 select a menu item from the other seat

The next click from the seat that opened the menu will cause a segfault.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

ivi-shell: initialize all fields in create_layout_transition

create_layout_transition does not initialize is_transition_func.

This patch initializes it to NULL.

Signed-off-by: John-John Tedro <johnjohn.tedro@gmail.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

vaapi-recorder: Fix allocator sizeof operand mismatch

Result of 'calloc' is converted to a pointer of type 'unsigned int', which is
incompatible with sizeof operand type 'int'

Signed-off-by: Lucas Tanure <tanure@linux.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

ivi_layout_surface_add_notification: Fix potential leak of memory pointed to by 'notification'

Signed-off-by: Lucas Tanure <tanure@linux.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

ivi-layout-transition: Fix potential leak of memory pointed to by transition

Check if layout_transition_register worked, otherwise destroy transition

Signed-off-by: Lucas Tanure <tanure@linux.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

ivi-layout-transition: Change layout_transition_register return type

layout_transition_register could not work and we need to free trans in
the calling function.

Signed-off-by: Lucas Tanure <tanure@linux.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

ivi-layout-transition: Fix potential leak of memory pointed to by transition

Free transition before return. We don't need to use layout_transition_destroy,
because transition was not registered yet.

v2: consolidate memory leaks fixes

Signed-off-by: Lucas Tanure <tanure@linux.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>

releasing: Update command path to reflect new scripts dir in wayland-web

configure.ac: bump to version 1.9.90 for open development

Master is open for new feature development again.

Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>