| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
qwaylandinputdeviceintegration_p.h:64:85: error: ‘uint32_t’ has not been declared
Change-Id: Ic747cc2ab45e4dc6bb70ffff143843d1ed4a341b
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I3ebcb49a4acf230b9515c0317aab2307dcb9270b
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@gmail.com>
|
|
|
|
|
|
|
|
|
|
| |
Let Ctrl+A cause a key event with text Ctrl+A (0x01), like we do on
the other platforms. Previously we would use 'A', which would confuse
some applications.
Task-number: QTBUG-50398
Change-Id: Ie3def4bd9c96e538b2572b88fae029d347e08814
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
This was too aggressive, if there is a queued buffer the current one
needs to be invalidated or the compositor will hold references to
both buffers while the scene graph is stopped which will block
clients in eglSwapBuffers.
This reverts commit 438063af85d0999f9dead6f04a3285edb175e24c.
Change-Id: I3247cfc94e12322fcccc07b5dd64146b055ab431
Reviewed-by: Giulio Camuffo <giuliocamuffo@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
contains(QT_CONFIG, egl) and CONFIG += egl is the only sane way to test
for and pull in EGL headers and libs. This is particularly important when
trying to be robust and guard against half-broken sysroots on embedded
where a naive PKGCONFIG += egl breaks.
Also add an EGL_WAYLAND_BUFFER_WL define to keep wayland-egl compiling.
We are not testing for that in any config tests may cause a failure in sysroots
that have parts of Mesa thrown in but pick ip an older EGL header from the
vendor's driver.
Change-Id: I7b7e6a7a91e78dbda5b6954ad08761298c538efc
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Directly calling eglCreateGlobalImageBRCM and friends is not ideal due to the
messy RPi distro setups: shipping Mesa (sw rasterizer only) in standard locations
may cause picking up those EGL/GLES libs instead of the Broadcom ones from /opt/vc
when building Qt. While this is something that should be fixed in the sysroot (since
it is a problem at runtime anyway), let's make life easier by resolving via
eglGetProcAddress, which is the right approach in any case.
Change-Id: I73f7698b638691d97743a6f819bc7c8ee6ebd245
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|
|
|
|
|
|
|
| |
qwaylandinputcontext.cpp:57:16: warning: unused parameter 'sym' [-Wunused-parameter]
qwaylanddatadevice.cpp:173:135: warning: 'dragData' may be used uninitialized in this function [-Wmaybe-uninitialized]
Change-Id: Id654360fd9b7fdb572565ad39b664af3355b5e79
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If the compositor sends events to us while the main thread is blocked
the socket notifier in the events thread would keep sending out the
activated() signal, but no events would actually be read until the
main thread starts to run again. That causes the event thread to keep
queueing new events, and so allocating memory, potentially forever.
This patch fixes the issue in maybe a bit radical way, that is by removing
the event thread. The socket notifier now runs in the main thread so it
will block if the events are not being read.
Nowadays there is no real reason to keep the event thread around, as every
thread that needs to receive wayland events can dispatch them on its own,
we don't need a central dispatcher thread anymore.
Change-Id: Ib7885e4b038b82719d78d193f465618a72cbe6af
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@gmail.com>
|
|
|
|
|
|
|
|
|
| |
Also add unit tests
Change-Id: I66de887607f73b318884e4a35f18510b90cf0315
Task-number: QTBUG-47357
Reviewed-by: Jan Arne Petersen <jan.petersen@kdab.com>
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|
|
|
|
|
|
| |
Allows to get the wl_shell_surface for a QWindow, if it exists.
Change-Id: I16b1c578a1c605e58c96e94ae55a3331ecfa353d
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|
|
|
|
|
|
|
|
|
| |
With commit c55a36cb9015cf1eebd49eaa5b1b4f4ec9b28451 in qtbase the
QSimpleDrag code changed in a way that caused a deadlock due to
drawing a window without a role. However, thanks to that we can
now remove that start/cancel hack and simplify the code.
Change-Id: Icba6e7c9c4927855e48fb21632db1a10332c4ffb
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I47d7f7d2cb589df8087c23eda60137e7b263f11f
Reviewed-by: Andy Nichols <andy.nichols@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For Wayland on NVIDIA. Tested with a Jetson TK1 Pro and Vibrante Linux.
With just the hw integration no compositors would work out of the box
since EGL_KHR_stream_consumer_gltexture only allows connecting to the
texture bound to GL_TEXTURE_EXTERNAL_OES, meaning that assumptions about
the target always being GL_TEXTURE_2D break horribly both in C++ and in
shader code.
In addition, buffers have to be extended with an additional updateTexture()
operation as EGLStream requires to call ConsumerAcquire on every
frame. Previously there was no concept of this as calling
createTexture() on attach() was sufficient.
Qt Quick bits are omitted since the refactored compositor API is pretty
different. This means that QML compositors will not currently function
in this environment.
The qwindow-compositor example is enhanced to support the external
texture target, but this won't apply for the refactored branch
either. It is provided for testing purposes for the time being, and to
show how C++ compositors can support different texture targets and
correct operation with EGLStreams.
Done-with: Louai Al-Khanji <louai.al-khanji@theqtcompany.com>
Change-Id: I0e209fc0cbcf435cca83528d938eb50e4bdceb82
Reviewed-by: Louai Al-Khanji <louai.al-khanji@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If a QtWayland compositor gives focus to a surface right after its
creation during the client startup, the keyboard resource may not be
bound yet by the client.
In this case, the surface is correctly marked as focused, but the
keyboard resource is never marked as focused, and thus no keys are
ever sent to the client.
To fix this, the focusResource is updated if needed after wl_keyboard
is bound.
This can be reproduced with weston-simple-im (patched to use wl_shell
instead of xdg_shell) and qml-compositor, modified to enable
TextInputExtension and call takeFocus() at the end of windowAdded().
Change-Id: I551cb5bc56c05a1e5187b23108f4ef80468782dc
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|
|
|
|
|
|
|
|
| |
When no keyboard resource is bound, the destruction handler for the
focused surface was not called, and the pointers to focused objects
were not reset. Thus the m_focus surface pointer could become invalid.
Change-Id: Iee404219304ae7b2bae87131ab140ab134e98118
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|
|
|
|
| |
Change-Id: I9217c633bcbecd330715b6dbc654245b3fb739a8
Reviewed-by: Louai Al-Khanji <louai.al-khanji@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I2a7b82bd2705919a91492edfccac02f4d8fc2c9b
Reviewed-by: Louai Al-Khanji <louai.al-khanji@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I70c01453b3939d1d645d626ae84c21ab4c9d267a
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|\ |
|
| |\
| | |
| | |
| | | |
Change-Id: Ib3c5bec85a09d92e37a3af8e1c749a80cb40c056
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: I4f83fdb59a60bb1f1720047f0b442d70192be5da
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: I75eda7e766182025b54f8df851ff1d5a72eefb17
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The unconditional grab from 498946d6e9c20d4e8af28ee942de5bf4cf5eaf78
did not have a corresponding ungrab. That only worked well as long as
there were only QWaylandSurfaceItems in the scene. Each of them just
stole the grab for itself at the first TouchBegin. Regular QML items
(some people have UI in the compositor process...) do not participate
in that mutual stealing so they'd sometimes not get any touch events
because the events still went to the last QWaylandSurfaceItem to
receive a TouchBegin.
Note: The grab would also go away if touchEvent() didn't accept()
the event. Without the accept(),
QQuickWindowPrivate::translateTouchToMouse()
would be called at TouchEnd and do the ungrab.
Change-Id: Ia7d1b15fff366542d382eb97494b1e43dedb65a4
Reviewed-by: Andrew den Exter <andrew.den.exter@qinetic.com.au>
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When calling ::exit() the socket notifier in the events thread may still
fire before the process actually exits, using objects that are being
destroyed and resulting in a segfault.
Stop the events thread before calling ::exit().
Change-Id: I187762da2a7efa83db1e62b0e28dfab89f478c7d
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@gmail.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When hiding a window we attach a NULL buffer, which means that any
frame callback that is still waiting will not be sent. In the SHM
case we just destroy it, but we cannot do that in the EGL case, since
we don't have a hold of it. So destroy the EGLSurface, which will
in turn destroy the callback.
Task-number: QTBUG-46921
Change-Id: I20e0d1ca7b53fb88096101d46979996dd944d429
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
|
|/ /
| |
| |
| |
| | |
Change-Id: I19fd0bc7a56b4baa5b6d40836d5f0401a3b11c49
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Handle transition from Normal to Maximize then Fullscreen and back to Maximized.
Avoid to maximize if no requested size
The protocol file is a raw copy of
Source: http://cgit.freedesktop.org/wayland/weston/plain/protocol/xdg-shell.xml?id=1.8.0
Minor Nitpick fixes (arrays of bytes, conditionnal or test, c++ function call)
Task-number: QTBUG-47327
Change-Id: Ib508e2166cc1337fd93454f30814136839cffa29
(cherry picked from commit 6906a6445c0cbf9d11f8d5d32b181f558a2292c9)
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
initialization order.
qwaylandxdgsurface.cpp: In constructor 'QtWaylandClient::QWaylandXdgSurface::QWaylandXdgSurface(xdg_surface*, QtWaylandClient::QWaylandWindow*)':
qwaylandxdgsurface.cpp:55:31: warning: base 'QtWayland::xdg_surface' will be initialized after [-Wreorder]
qwaylandxdgsurface.cpp:55:31: warning: base 'QtWaylandClient::QWaylandShellSurface' [-Wreorder]
qwaylandxdgsurface.cpp:48:1: warning: when initialized here [-Wreorder]
qwaylandwlshellsurface.cpp: In constructor 'QtWaylandClient::QWaylandWlShellSurface::QWaylandWlShellSurface(wl_shell_surface*, QtWaylandClient::QWaylandWindow*)':
qwaylandwlshellsurface.cpp:55:31: warning: base 'QtWayland::wl_shell_surface' will be initialized after [-Wreorder]
qwaylandwlshellsurface.cpp:55:31: warning: base 'QtWaylandClient::QWaylandShellSurface' [-Wreorder]
qwaylandwlshellsurface.cpp:49:1: warning: when initialized here [-Wreorder]
Change-Id: If7ed49fa6f788db7407b5ee82aa252e19e2d4747
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The texture often represents the last reference to the last buffer of
a destroyed surface and invalidating releases that reference which
shortcuts any attempt to keep the buffer alive with QWaylandUnmapLock.
Change-Id: I4641a87a32f8639c3ffb6c3b5dafd21c77bbb2c1
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| | |
Change-Id: I55961a32a17dbf35b380f8a4b75433a6b694dd4a
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@gmail.com>
|
|\|
| |
| |
| |
| |
| |
| | |
Conflicts:
src/client/qwaylandwindow.cpp
Change-Id: I80ead87f3e1f36299b1bf9cc9c7db6cd48c0e86f
|
| |
| |
| |
| |
| |
| |
| | |
Instead throw a warning and keep using the old one.
Change-Id: I1533b3930d4d8c884db922b0d6dd567ef7e4f67c
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@gmail.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
If a transient parent of a surface is destroyed before the surface
itself, reset all references to that parent surface.
Change-Id: I5b748099d8a5502632d8846c5de5381b5f28b665
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| | |
Change-Id: I51da06261acfb193a59db34bc2dd25e3452fce8d
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
|
| |
| |
| |
| |
| |
| |
| | |
LGPLv3 refers to it but does not include it in its body.
Change-Id: Ib056b47dde3341ef9a52ffff13eed18cf3504738
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
It is a bit dangerous to call wl_display_dispatch() in the event thread,
since it may race with the dispatch called e.g. in QWaylandDisplay's
blockingReadEvents() and lead to a dead lock. Instead, use wl_display_prepare_read()
and wl_display_read_events() in the event thread, which doesn't block, and only
dispatch in QWaylandDisplay.
As a result we don't need the additional wayland queue anymore, so remove it.
Change-Id: I9fbbe5d2f38d06773beb7847df1a0212cca92c37
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
QToolTip has the same problems as QMenu, that is you can create one
without a parent and with a global position. Use the same trick and
use the window that last received an input event as the parent.
Change-Id: I093c8da0d54110903f35670b01dea6fa96abecf4
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
It was not possible to use a different input context than the platform
context. This has been unified across major platforms, depending on the
environment variable "QT_IM_MODULE", the following context is selected:
- null: default platform context (Wayland in this case)
- empty: no context
- set: set one, if it exists and is valid (otherwise no context)
Change-Id: I07c6fb339c434f99fc6e092a2e18f00600daa3bc
Reviewed-by: Nedim Hadzic <nedim.hadzic@pelagicore.com>
Reviewed-by: Jan Arne Petersen <jan.petersen@kdab.com>
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
QMouseEvent -> QPoint in QSimpleDrag::move and drop
Task-number: QTBUG-46615
Change-Id: I9cda0f039ee8f5a70219b320abbb65f8649747e1
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
To manage the keyboard state and send key events to clients we use
the xkbcommon library (unless disabled). That means that we currently
have a separation between key events outgoing to clients and key events
being used internally in the compositor, resulting in different keysyms
being produced in the two cases for the same hardware key.
Filter the key events coming from the underlying platform and modify
them according to the xkb state instead.
Change-Id: I4c6678be68e5b214b2b9cd0222dc9f425c580085
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@gmail.com>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| |
| |
| |
| |
| | |
Change-Id: Ibe1811ee0b7057d082a55c4ae62e5932f7e10b84
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| |
| | |
Change-Id: I52628f87dbea5383db06468f9748a9bacdec3179
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When a client attachs a NULL buffer on a surface we may want to hang on
to the old buffer for a bit, e.g. to do a hide animation.
Add the QWaylandUnmapLock class to keep a surface mapped after an attach(0)
request.
Change-Id: I5d5be0b36410a76a58d2b68c1b2ac5e3b875a801
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
| |
| |
| |
| |
| | |
Change-Id: Iac40929aebc1e0f507b729b0c01a2ccb18516c37
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
|\|
| |
| |
| |
| |
| |
| | |
Conflicts:
.qmake.conf
Change-Id: Ic75157b11eee188608c3ac7ed6fb1a033bb72750
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
With a queue of just two buffers BufferAttacher can end up holding
references to both after the renderer is stopped starving the client
application and causing it to block in the glSwapBuffers until the
renderer restarts or the surface is destroyed. Release the current
buffer to the client when the renderer is stopped so it can continue.
Change-Id: Ica0e13ef78f7e6058e273c26b517a88d07f958c7
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| |
| |
| |
| | |
Change-Id: I7000630632373cf7d81c6c0b307f1ab41c2e133f
|
| |\
| | |
| | |
| | | |
Change-Id: Iaf9eaa171c141d79c52441de8a697c87e2e42407
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: Ia8a927c5bcfb45bc7a8c16d1c8ff75487fe71cdd
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This occurs for example when a Qt Quick2 application calls
showFullScreen() while a rendering is already in progress.
The GUI thread was calling destroyEglSurfaces() from setGeometry() and
thus destroying the buffers, while the render thread was trying to use
the buffers from swapBuffers().
There was a division by zero in swapBuffers() with the modulo m_count
operation, because EGL surfaces were destroyed and m_count was 0.
This fix is quite fragile, proper locking on buffers would be better.
Change-Id: I41fcb2fcebec84c0dfc82ce56c6e323d3722b30a
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
|