| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
Wayland 1.4 introduces wl_subsurface which is quite an important
addition.
Change-Id: I48375f60adce556c9989872319f4d073e4a7b13b
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
|
|
|
|
|
|
| |
Remove qwaylandcompositor_p.h leftover.
Change-Id: I2ede92db92894f14301c89cc28c896cfdc6d3563
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
There currently is a QWaylandInputDevice class both in the wayland QPA
plugin and in the QtCompositor API. This causes the qwindow-compositor
example to crash when running nested in a wayland session due to a
mismatch between the two classes.
By namespacing all the plugin code we make sure that name clashes
will not happen anymore.
Change-Id: I17497cff697599200bea68bf01dfde474526390f
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
|
|\
| |
| |
| |
| |
| |
| | |
Conflicts:
.qmake.conf
Change-Id: I65c8503cf632699129ebb6b4f01721b74be78def
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When the scenegraph is invalidated we delete the surfaces texture.
Set the update flag to true so that when the scene graph is again
initialized and starts to draw a frame the textures will be recreated.
Change-Id: I2fb171e4bed517e48fa2311c22651150f3e605f7
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
|
| |
| |
| |
| |
| |
| | |
Task-number: QTBUG-43346
Change-Id: I9bc6d804ddca59a7a0173b8f0d3ec2f268fe9a59
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Clang found two problems with this from QWaylandDataDeviceManager. First,
it's a class, not a struct. Second, nothing is using this private
member.
qwaylanddatadevicemanager_p.h:66:5: warning: struct 'wl_data_device_manager' was previously declared as a class [-Wmismatched-tags]
qwaylanddatadevicemanager_p.h:66:36: warning: private field 'm_data_device_manager' is not used [-Wunused-private-field]
And also this one:
qwaylandextendedoutput_p.h:61:21: warning: private field 'm_screen' is not used [-Wunused-private-field]
Change-Id: Ifc8f28f92190547dcb0a5319aa6fc14277894f15
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
The wl_pointer.event carries the surface local position of the pointer.
Notify Qt of it, without pretending it to be a motion event.
Change-Id: Ibbe1d125a93b478f9c350a50bfea25b01f628178
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
This allows a compositor running with the wayland qpa inside another
compositor to provide EGL to its clients.
Change-Id: I308fb909c8168955148be152bf314a53c0e5ca43
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
|
| |
| |
| |
| |
| |
| | |
Task-number: QTBUG-41141
Change-Id: I0bd789cc1090f7e0e7615b47c33a4e4179677398
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
|
|\|
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
.qmake.conf
src/compositor/compositor_api/qwaylandsurface.cpp
Change-Id: I3ee40153c959dbe31ea685571c5a8bb1ae36606e
|
| |
| |
| |
| |
| | |
Change-Id: Iecd5050097abcafa06073d41a31d1ae31fac5b42
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Calling wl_data_source.destroy(); wl_data_offer.receive(), where the
wl_data_offer is referring to the destroyed wl_data_source, was crashing
the compositor.
Change-Id: Iab352d63f8a93aebd0bfbcf1dc26c0e8a48a523a
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
There was a race condition between the gui and the wayland event thread
which could lead to double freeing the QWaylandShmBackingStore's frame
callback. Protect the wl_callback_destroy calls using a mutex.
Change-Id: Ia70ebac208a6d4450328ba5254a850be26d84d6d
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
The code is correct but some compilers may still warn about dragData not
getting initialized. Silence them.
Change-Id: Ib52321667fc5094e22ebbef538b72b5477e6f10b
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
The surface's resource may have been destroyed, return a null client
in that case.
Change-Id: I3d7e0f0e94008e003e43f1987e6560dd73f5c5db
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
| |
| |
| |
| |
| | |
Change-Id: Ic4c998347e14c17693bea461452d6ee8fc41064b
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The texture of QWaylandSurfaceItem's texture provider was updated in
QWaylandSurfaceItem::updatePaintNode(). That is fine as long as the
texture is accessed after the updatePaintNode() call, but if not we
would access an invalid texture.
That could happen when another QQuickItem subclass was accessing the
textire in its updatePaintNode(). We don't have any guarantee on the
order of the updatePaintNode() calls, so the other item could be
updated before the QWaylandSurfaceItem.
Change-Id: I7320d445c72796ce26f8d8483175a35e9c6840e7
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Many inprovements on windows's decorations,
Also handle compositor events
(test: Super+Shift+F , Super+Shift+M)
The protocol file is a raw copy of
Source: http://cgit.freedesktop.org/wayland/weston/plain/protocol/xdg-shell.xml?id=1.6.0
Task-number: QTBUG-38633/related
Change-Id: I667ec52c8a7e34d74b60174a671c89671f841d6b
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Wayland 1.2 added support for display scaling, where wl_output
has a "scale" event informing the the client that the compositor
will scale the output surfaces by the given factor. The client
then has the option of providing large surfaces to match the
target pixel densety and bypass the compositor's scaling. This
is done by calling wl_surface::set_buffer_scale.
Re-use the current high-dpi support in Qt by implementing
devicePixelRatio() for QWaylandScreen and QWaylandWindow.
Provide high resolution buffers both for raster and OpenGL
graphics.
Introduce a new coordinate system: buffer coordinates, which
is related to the window coordinate system via an "int scale"
scale factor. This scale factor corresponds to Qts qreal
devicePixelRatio, but we keep the name and the type in the
QtWayland implementation.
Change-Id: Ie10d7e5b4833480a9a25d96a26ad02150eb6e83f
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The native resource getters that are supported by eglfs will have to be
supported by the wayland platform plugin too. (on wayland-egl at least)
Change-Id: Ibbab649c04785dbde177342c45b9bc6f1edd954d
Reviewed-by: Andy Nichols <andy.nichols@theqtcompany.com>
Reviewed-by: Andras Becsi <andras.becsi@theqtcompany.com>
|
|\ \ |
|
| |\|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Merge done by Giulio Camuffo
Conflicts:
src/client/client.pro
src/client/qwaylanddisplay.cpp
src/client/qwaylandinputdevice.cpp
src/client/qwaylandinputdevice_p.h
src/client/qwaylandwindow.cpp
src/compositor/compositor_api/qwaylandsurface.cpp
src/compositor/compositor_api/qwaylandsurface_p.h
src/compositor/wayland_wrapper/qwlsurface.cpp
Change-Id: I1df878bb54f49d953c51215a2772869e261ebe81
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The compositor must not destroy wl_resources when it wants to, else it
breaks the contract with the client.
Change-Id: Ic0d298072cdf0954d2504c04bff2bcc99733e621
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
xkb context creation can fail when xkeyboard-config files are not
available.
This changes qwlkeyboard to gracefully fallback on the
QT_NO_WAYLAND_XKB path, as suggested by Giulio Camuffo.
Change-Id: Ic57d80b6aa3e9fbd8a393a3ca54dff768a4c3e83
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| | |
| | |
| | |
| | |
| | |
| | | |
Change-Id: Id7d6582d4e139763634c7022548b558d4a2f8ec2
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Also add a debug message about theme loading failure.
This is not necessarily fatal, some embedded applications could still
function properly without a cursor.
Change-Id: Ib3e2ee43b00630ee77997c7472a95ac5a22c2cc2
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| | |
| | |
| | |
| | |
| | |
| | | |
Task-number: QTBUG-41677
Change-Id: I039ec581e960eb70eb04aed6e65add70c563ba77
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When binding globals take care of the version the compositor provides
and don't call requests that the compositor doesn't implement.
Change-Id: Iea0c534fb9e005720857e2778e380b6a9c22a6f3
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This fixes copying in Weston.
Change-Id: Icbdb81078e8df4575bbe440321773131ba27ebbf
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We cannot just assume all clients implement the same version of the
various interfaces as the compositor does. Keep track of them, and
react accordingly when creating a resource or sending an event.
Change-Id: I9792433a14d49c5c4df0c892fc1349ce0dfb0d43
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
|
| | |
| | |
| | |
| | |
| | |
| | | |
Change-Id: I9700c15066c030436dbe2f9a09b0851b7f4350ae
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
QWaylandShmBackingStore installs a frame callback on flush, and subsequent
flushes will not attach a new buffer until the callback is fired.
If the window is hidden while we're waiting for the callback, we attach
a NULL buffer, so the compositor will not redraw the surface and will not
fire the frame callback. When showing the window again the backing store's
flush() will wait indefinitely for the frame callback to attach its buffer.
To fix it destroy the frame callback when the window is hidden.
This was easily noticeable when fast switching between popup menus.
Change-Id: Ic0c71ed79e2fab9faf452f63b05bc4576ea9a3ba
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
::exit() is not thread safe, so make sure to not call it more than one
time, once from the gui thread and once from the wayland event thread.
Change-Id: I80905c6d996cb827a5101ae6d6c9bc12a267ba71
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Use the EGL_KHR_create_context extension if available to create modern gl
contexts. We disable window decorations on core profiles because that
mandates a VAO to be used.
Change-Id: Id6044e64f6736244ae9d593af4d68c4000a31de6
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We were sending only key press events for the key autorepeat, going
out of sync with the release events. Xorg sends both relase and press
auto repeat events, do so here too.
Change-Id: I41a9bfaff50afea779a7572220f6a01b507e95ac
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We were using the screen physical size to determine if a screen is
portrait or landscape, but we may not always get a valid size.
Instead use the screen resolution.
The orientation is then calculated in the wl_output.done handler,
since we don't know the resolution yet when receiving the first
wl_output.geometry event.
Change-Id: I3554f916e54db829f49fa3d1ea24f7ce1ff24e7c
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
QWaylandAbstractDecoration.
This makes absolutely no sense. Just do things inline.
Change-Id: Ic6d4e62bc9b3c0a4533f63740f2608a0be201047
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Convert to a D-pointer, split between abstract base class and an implementation.
Also move implementation of the current built-in decoration to the "bradient"
plugin, named in glorious memory of the programmer-designed blue gradient that
will forever sear our eyeballs.
The decoration plugin may be specified using the environment variable
QT_WAYLAND_DECORATION.
Change-Id: Idc99ab06ae138ad299bad2b62b9595379bd007ab
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This is in preparation for decoration plugins.
Change-Id: Idb322a7a5cbc2eb5bf2cce019073f9f4fb46297f
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
QWaylandInputDevice recently gained a virtual function, making it a
polymorphic class. It should thus have a virtual destructor as well.
Discovered through a GCC diagnostic:
warning: deleting object of polymorphic class type 'QWaylandInputDevice' which has non-virtual destructor might cause undefined behaviour [-Wdelete-non-virtual-dtor]
Change-Id: Ie816caabea56cde320768dcc8aba6b5ad50bf4c1
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| | |
Change-Id: Ia3ee8b13f1eacc435152149a1ee65fb62658506e
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Allows the registration of multiple input devices for the compositor via
private APIs. Since the Qt stack does not support separate input devices
via the QPA, the identification of each device (wl_seat) is left up to
the implementor.
The compositor will identify input event via the QWaylandInputDevice::isOwner
method. Usually this will happen when an item on the UI has received an
event and would like to send it to the client surface. See
QWaylandSurfaceItem for more details.
Includes basic unit tests
Change-Id: I7ee1db49388713bf3076c23cf8f8a165aefc2fe0
Reviewed-by: Mikko Levonmaa <mikko.levonmaa@lge.com>
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Introduces QWaylandInputDeviceIntegration plugins to allow customization
of input device related behavior. The plugin can be activated via the
environment variable QT_WAYLAND_INPUTDEVICE_INTEGRATION
Change-Id: If5629737752afacb29161f51c1b7c6e171fb2758
Reviewed-by: Mikko Levonmaa <mikko.levonmaa@lge.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|\|
| |
| |
| |
| |
| |
| | |
Conflicts:
src/compositor/wayland_wrapper/qwlcompositor.cpp
Change-Id: Ica26e33eb2b77ed5be0e84e8f4256d2e85abfaf5
|
| |
| |
| |
| |
| |
| |
| |
| | |
When the surface is mapped add or remove the popup grab
as expected.
Change-Id: I807753393ef2753bc96c698f766523186ee88fd2
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| | |
A 0x0 size means the compositor doesn't know what the real size is.
Change-Id: I79cac0482839399d4a53e9556bb999592af555a5
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Weston can send in some cases a negative physical size in the wl_output.geometry
event. While it is not clear whether that is legal go the extra mile and catch
it.
Change-Id: Ie2fbab84e653ad77732a72b6ca61509eb7849895
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Installing a roudtrip on the custom event queue in the wl_data_offer.offer
handler is broken because that triggers a wl_data_device.selection event,
which emits the QClipboard changed signal, so code listening to it may end
up trying to retrieve the clipboard data before the roundtrip ends.
Additionally, we're calling wl_data_offer.receive for each mime type, even
if then we never read from the fd, making the source client do work for no
reason. Instead, call wl_data_offer.receive retrieveData_sys, that is when
actually retreiving the data.
We don't need to install a roundtrip after that, just flushing out the
requests is enough, because we wait up to one second for the source client
to write into the fd.
Change-Id: I180779e375ebd5a22af7084458505a41107fab19
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| |
| |
| |
| |
| |
| |
| |
| | |
If we don't have a valid physical size for a screen, i.e. (0x0),
return the default value from the base class.
Change-Id: Ia7b6f90ee73e07014ab752ceb165426812c49415
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|