QPlatformSurface relates to the platform window, not the wl_surface.
The events are already emitted by QPlatformWindow on create/destroy.
To preserve compatibility for a previous KDE version it was faked to
emit the events when the wl_surface is created/hidden to keep behavior.
This is no longer necessary, and it has caused multiple errors, the latest
being a crash when switching between sub-menus with the Sway compositor.
[ChangeLog][QPA plugin] QWaylandWindow no longer sends fake SurfaceCreated/Destroyed events.
Use expose events to be notified when a Wayland surface appears.
Task-number: QTBUG-76324
Fixes: QTBUG-81952
Pick-to: 5.15
Change-Id: I2f003bc9da85f032a0053677fd281152099fc9eb
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: David Edmundson <davidedmundson@kde.org>
The Qt version was added in 5.14 "for use as eventual replacement for
QString::SplitBehavior." Move another step closer to that goal.
Change-Id: Ie19f7580d75694982ac26c08ad5aabb73d830494
Reviewed-by: Johan Helsing <johanhelsing@gmail.com>
Fixes a crash.
Otherwise we created two shell surface objects in some cases.
Change-Id: Ieded56261061da6acf27ccabba81c8f578ca1596
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@liri.io>
In a319f4a5, we temporarily disabled the override to let qt5.git update, now we
add it back with the new signature. This means QWindow::startSystemMove now
works for Wayland.
Task-number: QTBUG-81751
Task-number: QTBUG-73011
Change-Id: I1459300d6f7fb77a9ae07bb4f5c2b80722f9b02c
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@liri.io>
The new version requires the button that caused the event, as well as the event
type. Instead of adding a conversion functions to and from
QWaylandInputDevice::Type to QEvent::Type, just use QEvent::Type directly, as
there is one-to-one mappings for all the event types we need.
Fixes: QTBUG-80044
Change-Id: I0f6e0e8e7f2e026a1d601f86e819affe0570a1b2
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@liri.io>
Qt window coordinates start inside the decorations, while wl_surface
coordinates start at the first pixel of the buffer. Potentially, that would be
in the shadow, although we don't have those. So for now, it's the first pixel
of the decorations.
Change-Id: Idccf8a359035f5477d6bc9e2e03a0e9fafe16971
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@liri.io>
It's not just popups that may have popups open on them. Always close open
popups when hiding a window to prevent dangling pointers.
[ChangeLog][QPA plugin] Fixed a crash when re-showing a popup after hiding its
parent.
Fixes: QTBUG-80562
Change-Id: I7cdac5c7a30e0add5ebf00259401e4d74626ce96
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Reviewed-by: David Edmundson <davidedmundson@kde.org>
[ChangeLog][QPA plugin] The QPA plugin now supports running with no screens
attached. This is handled by adding a fake screen when the last screen is
disconnected, similarly to what the other QPA plugins do.
Fixes: QTBUG-79111
Change-Id: I4a0e023ae784217dd030f0c62f12487fdff4825c
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
[ChangeLog][QPA plugin] Fixed a crash when showing a window with a hidden
parent.
Now we just avoid creating the subsurface, so nothing is shown. Seems to be
the same behavior as on xcb.
Fixes: QTBUG-79674
Change-Id: Ia46fcd9a0da5aad4704816a41515cb1e128ac65f
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
[ChangeLog][QPA plugin] Fixed a 100 ms freeze that would occur if applications
did not draw after receiving a deliverUpdateRequest().
QtQuick does this at the start of animations. This should get rid of those
backingstore warnings (and also remove a 100ms freeze before animations start
in those instances).
Fixes: QTBUG-76813
Change-Id: Id366bf4a14f402fa44530ae46e7b66d9988c14f6
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Reviewed-by: John Brooks <john.brooks@qt.io>
Previously, we set the max length in QString character length, which means
UTF-16 characters (of potentially three bytes) counts as one character. The max
limit of libwayland, however, is in bytes (and the string itself is converted
to UTF-8).
Fix it by dividing the character limit by three because in the worst case each
UTF-16 character will use three bytes when UTF-8 encoded.
Fixes: QTBUG-78478
Change-Id: Idf4721894e0fe6f3cd92bdc6ada7b0ea4199ea63
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Calling the QOpenGLContext::swapBuffers from
QGuiApplicationPrivate::processExposeEvent in some cases leads to
recursive calls of QWaylandWindow::waitForFrameSync. Since the
mWaitingForFrameCallback check in WaylandWindow::waitForFrameSync
is performed after the mutex is locked, the QMutexLocker tries to lock the
mFrameSyncMutex mutex in every recursive call, that leads to a deadlock.
This patch moves the performing of the mWaitingForFrameCallback check
before locking the mutex.
Change-Id: Ia2d834b7dd03fcd91bbe29a3a897b4db2d155527
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
The wl_surface can be destroyed whilst a render is happening. Calling
wl_surface::frame after the window is reset can crash as wl_surface is
null.
Fixes: QTBUG-77747
Change-Id: I139a9b234cb6acba81d6c1d5fa58629904a25053
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Recent changes in QtBase means QtWayland will have to follow the convention of
the rest of the QPA plugins and have QPlatformSurfaceEvent::SurfaceCreated and
SurfaceAboutToBeDestroyed follow the QPlatformWindow (QWaylandWindow) lifetime
and not the lifetime of wl_surface.
Some users were depending on those events to get notified about wl_surface
changes and used QPlatformNativeInterface in order to get the window's
underlying wl_surfaces in responses to the events.
The good news is that QPlatformNativeInterface is private (QPA) API, so we can
provide an alternative by exposing new private API, which is what this patch
does.
The QWaylandWindow::wlSurfaceDestroyed signal already exists in the dev branch
(introduced in 04be9cf380), so this is a backport of that signal as well as an
addition of a new QWaylandWindow::wlSurfaceCreated signal.
Task-number: QTBUG-76324
Change-Id: Ibc5748474cd52f5b9461fd1ad6cef973491174b1
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@liri.io>
For shells that do not send configure events when exposed:
If setGeometry() was called before the window was initialized, Qt would
not mark the window as mapped, and the window would never become
visible.
Change-Id: Ic933cbbff20702424129c11264215181330cc7d1
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
Implicit capture of 'this' in [=] is deprecated in C++20.
Fix by using explicit captures.
Change-Id: Ie3a94ec60d7c56b2856d201fa3d68d0670bdd7b9
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
If we hide a window whilst a compositor has a pending frame to show,
it's possible the compositor will not render the frame and not return
the callback.
If this happens on the next window expose we can be left with
mFrameCallbackTimedOut still true causing isExposed() to remain false
and us to not send the next buffer when we later show the window again.
Change-Id: I507410415d1a930fd5fa736412055e68fdf6c1d3
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
When we did invokeMethod for handling the frame callbacks, we had to wait for
the GUI thread to finish whatever it's doing before we would stop blocking.
Fix it by clearing the frame callback timer and stop blocking immediately,
while delaying the rest of the work until it can be run on the other thread.
Fixes: QTBUG-76397
Change-Id: I343e4feac4838926b4fa2ccac2948988bc6c3bb7
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Neither Qt::ToolTip nor Qt::Popup are single bits in Qt::WindowFlags, and do in
fact include Qt::Window. This meant that when we or'ed them and did a bitwise
and with QWindow::type(), we would match more types than just Qt::Popup and
Qt::ToolTip. We would for instance get any Qt::Window as well, which meant the
main window would be added to activePopups, leading to strange things
happening, such as crashes and the main window closing unexpectedly.
[ChangeLog][QPA plugin] Fixed a crash when closing multiple popups at once.
Fixes: QTBUG-76124
Change-Id: I1a6a59e161a436604a7ac8ab824396481dc99a20
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
In this patch, we port Q_FOREACH loops to C++11 ranged-for loops. All cases are
trivial in the sense that either the argument is already const or is trivially
marked as const, either by qAsConst(), or, in the case of rvalues, by storing
to a const auto temporary first.
In addition, all loop bodies are clear enough to confirm that the container we
iterate over is not changed under iteration. This does not exclude cases where
a loop is prematurely exited just after calling a modifier on the container, as
that is safe, if not especially elegant.
Change-Id: I87a63f07797437d421567d60e52305391a3c4f21
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
The Wayland plugin now takes full control over delivering update request and
implement frame callbacks for both egl and shm.
[ChangeLog][QPA plugin] The non-blocking version of eglSwapBuffers is now used, if
supported. This fixed a bug where minimized windows would block the event loop.
[ChangeLog][QPA plugin] Windows that don't get frame callbacks from the
compositor within 100 ms are now set as not exposed. This should stop most
clients from rendering unnecessary frames to minimized or hidden windows.
Also, when we relied on the QPA version of requestUpdate, we would sometimes
deliver one update request while we were waiting for a frame callback. When we
implement the fallback timer ourselves we can make sure we only deliver the
fallback if there are no pending frame callbacks.
QtQuick and other applications often depend on blocking swapBuffers to throttle
animations. If the context's surface format has a non-zero swapInterval, try to
emulate a blocking swap.
Fixes: QTBUG-69077
Change-Id: I3c6964f31a16e9aff70b8ec3c5340e640a30fef2
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
windowGeometry is an existing method in QPlatformWindow with an entirely
different meaning to the newly added
QWaylandWindow::windowGeometry.
This led to the existing call in resizeFromApplyConfigure calling the
wrong method.
Change-Id: Ib344cf09178e77e5b4d5427325d840346663ec9d
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
...and introduce a client-side QWaylandSurface.
The goal of this is to remove the assumption that a QWaylandWindow maps 1:1
with a wl_surface and also to be able to share implementation with the cursor
implementation and other parts of the plugin that needs to use surfaces, but
are not QWaylandWindows.
QWaylandWindow and QWaylandSurface are still tightly coupled and a lot of the
implementation remains as part of QWaylandWindow, but I'm hoping to fix this in
later patch sets.
Task-number: QTBUG-74373
Change-Id: Ia5a192c8d133847336d266f63ec216fb3639b8c5
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Because some shells don't allow attaching buffers before configure, we need to
not be exposed until we know that we don't want a shell surface.
Change-Id: Ida7101a99f953d02cf6401e4ea8d28cfabd6e102
Reviewed-by: Giulio Camuffo <giulio.camuffo@kdab.com>
Reviewed-by: David Edmundson <davidedmundson@kde.org>
Use qCWarning instead of qWarning if logging into a category. This
fixes compilation with QT_NO_WARNING_OUPUT, but in addition also
avoids unnecessary string formatting if the category is disabled.
Fixes: QTBUG-74359
Change-Id: I0e316a833f3f7d4a9122377a1154e739bcb080b1
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
Version 5 adds frame events, which groups multiple pointer events together,
enabling diagonal scrolling, leave and enter without an intermediate state, and
together with the new axis_source, axis_discrete and axis_stop also adds what
needed to differentiate between wheel events and touchpad scrolling.
This patch adds scroll phases and pixel deltas to QWaylandInputDevice, and
makes sure handleWheelEvent is called accordingly.
[ChangeLog][QPA plugin] Pixel delta is now set for mouse scrolling events if
originating from an appropriate device such as a touch pad (requires compositor
support for wl_seat version 5 or later).
Fixes: QTBUG-69876
Fixes: QTBUG-63720
Change-Id: I094a1ef0365893bee135cae7e6df89fafdafa2f2
Reviewed-by: Giulio Camuffo <giulio.camuffo@kdab.com>
