Blocks are likely to have been created in a differnt thread from the one
performing their deletion, so we need an acquire fence.
The rest of the atomics use in the class looks ok, but nevertheless warrants
a deeper analysis.
Change-Id: I1571ded3a06695b0d58b5bf1d80d6283ac21f959
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit 6fa34930c2)
(cherry picked from commit 51bcc7e07e2bb5b42bb200dcd5269e9e9e2fe240)
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The pointer value is not the only data we're interested in, but
instead points to indirect data, so we need a release fence on store
(present) and a corresponding acquire fence on load (was missing).
Change-Id: I51f8251c0c7f4056192880430f2be5e0836dbed6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit 6f84829031)
(cherry picked from commit 4cc6e1419294a729e53d698bace2254903c1429b)
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
As with widgets, items that are disabled should not receive any input
events.
Similar to QGraphicsScene, which ignores disabled items when handling
mouse presses, the view should also ignore them when handling mouse
moves to update the cursor.
Since QGraphicsView only adjusts the cursors on mouse moves, reenabling
an item that is currently under the mouse will not change the cursor.
This is consistent with other changes of item attributes that would
position the item under the mouse (such as moving it). The overhead of
hit-testing items for every such attribute change would be too large,
and applications can generate a mouse move event if they really need
to adjust the cursor in all situations.
[ChangeLog][QtWidgets][QGraphicsView] Ignore disabled items when setting
the mouse cursor.
Fixes: QTBUG-76765
Change-Id: Ifcd31fc0581e8421e58eeb436a55b031909eed7e
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
The tests send QEvent::MouseMove events to the view, but don't fully
construct the event with both local and global position. Consequently,
QMouseEvent will use QCursor::pos as the global position, which is
unreliable, as QTest::mouseMove can not guarantee that the mouse really
moves - when running the tests locally on e.g macOS, it never does.
So instead construct the QMouseEvent with the trivially calculated
global position.
Change-Id: Ic4c914e3af7f15751545080d4743b06d3887cce8
Reviewed-by: Andreas Aardal Hanssen <andreas@hanssen.name>
This patch amends 730cbad882
The issue was that the event reader thread (QXcbEventQueue::run()) can enqueue
events sometime between GUI thread has last time peeked at the queue and before
it has called waitForNewEvents() and hence started waiting for more events (via
QWaitCondition). This scenario is even mentioned in the QWaitCondition documentation:
"[..] if some of the threads are still in do_something() when the key is pressed,
they won't be woken up (since they're not waiting on the condition variable) and
so the task will not be performed for that key press. [..]"
And if there are no more events on the X11 connection, the waitForNewEvents()
in QXcbClipboard::waitForClipboardEvent() would timeout.
Fixes: QTBUG-75319
Change-Id: I8990a2a0c00571dfc334fb57d616dee999042885
Reviewed-by: Igor Kushnir <igorkuo@gmail.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Implicit capture of 'this' in [=] is deprecated in C++20.
Fix by using explicit captures.
Change-Id: I1633446f4670202b0d1aca938d8c27dbc0c1411e
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The plain bool variable wasDeleted is set to true in the
QHostLookupManager dtor before the call to clear(), which calls
waitForDone() on the thread pool performing the lookups. All tasks on
the thread pool start by checking this variable so as to return early
when destruction is in progress. But the check was outside the
mutex-protected area, so as a non-atomic load, without a
happens-before relation to the write, this is a Data Race, thus UB.
Fix by moving the check past the mutex locking into the critical
section. This way, tasks that were waiting for the mutex after seeing
no wasDeleted before get the message reliably.
This does not introduce a dead-lock, since the call to waitForDone()
is outside any mutex protection leaving a window for the tasks to
obtain the mutex and react on wasDeleted.
Change-Id: Ied4b9daa7dc78295b0d36a536839845c4db2fb78
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: David Faure <david.faure@kdab.com>
Factor out the screen area check to a separate helper and apply to
restoredGeometry, too, fixing an oversight of
2f2bfc4e59.
Change-Id: I795d8d5e3cddb5e986c96c08a342d69063d04970
Fixes: QTBUG-76900
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
restoredFrameGeometry is only used for a sanity check in format
version 0, add a comment and remove code.
Task-number: QTBUG-76900
Change-Id: I797b07d069f8568cb39541bcbe9009935a4a79f7
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
This is an enabler for supporting pen color for color fonts in
Qt Quick.
Task-number: QTBUG-74761
Change-Id: I3e605f939e6677cbbd4a650ae7998dea8fd2d7a9
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
A previous change to QLineEdit::sizeHint() has caused the QSpinBox
arrows to break the widget frame on a high-DPI display with some
particular scale values like 150 or 175%. This change updates
QLineEdit::sizeHint() so that it has a minimum height, using the same
values used in 5.11.
Fixes: QTBUG-76047
Change-Id: I21f3e736da473b10fdf52e5a60e5fc5d07f270a1
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Color fonts may also contain regular alphabet characters that
should be rendered with the current pen. In Qt, however, these
characters were drawn into the cache with a default pen color
of black.
Since all characters in a font is currently backed by the same cache,
and it would require a lot of plumbing to get around this, a step
in the right direction is to include the current pen color in the
cache as long as it is an RGB cache. This means that drawing
text with the color font with different pen colors will create
different caches.
There is no API to select font color on Freetype currently, but
this problem has also not been observed there, as the fonts
in question, with both regular and color glyphs, are not being
detected as color fonts (so the text color will be correct).
So Freetype will be left out for now.
[ChangeLog][QtGui][Text] Fixed bug where regular text rendered
with a color font would always display in black.
Task-number: QTBUG-55096
Task-number: QTBUG-74761
Change-Id: Icc7dbf73241db1e7cc6a0de18c2de927aeecf713
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Remove the config.unix condition from the ccache feature. This enables
us to use ccache for non-Unix cross-builds like MinGW on Linux.
Change-Id: I3b108c2288754ad5dd05834e3d5a487c2da4ac00
Fixes: QTBUG-76681
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
This reverts commit 224a60989e.
Turns out that we cannot just untangle the determination of
source root and build root, because this breaks the assumption
that every .qmake.conf results in a separate .qmake.cache in the
build tree. QTBUG-76140 must be fixed differently.
Fixes: QTBUG-76907
Change-Id: I5c0a3719d5e00a0f1cacad51651b47c1f284d22d
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The header clean command must not be prefixed by "@echo ...", because it
now starts with "$(CXX)" which already is prefixed.
This amends commit 6fa5dfdd.
Change-Id: I5c2e0d2c2ed91c7232fce0a4a49db0fccfdc005d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
A blinking white dot could appear on dark backgrounds on Windows 7/8/8.1.
It was caused by a workaround added to trigger the on-screen keyboard
on the Microsoft Surface. This change effectively restricts that
workaround to Windows 10, where the blinking dot was not an issue.
And anyway, impact over the Surface functionality should be minimal,
as all models produced since 2015 only support Windows 10 (and it's
unknown if the workaround was ever necessary for early models running
Windows 8.1).
Fixes: QTBUG-74492
Change-Id: Ic9b9c0f71f23b75212054c56a29796cf0efa109a
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
In some cases, the wheel event coordinates would be incorrect, as the
local coordinates were being determined relative to one window and
the event being sent to another window, possibly incorrect.
Fixes: QTBUG-75820
Change-Id: I4c3c4c6c4688bd9232d67ce4052d24365f6aea3a
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
This reverts commit 9a25d27b9d58316dee5d2305135d2d74ad5d51e7.
The QSKIP is no longer needed as the imap server's certificate
was updated
Task-number: QTBUG-76610
Change-Id: I1007ce50d6f7f6258fdeb8894c66678a660b03ca
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Because it is. It's just QCoreApplication::postEvent(), which is thread-safe.
It also _has_ to be, because we recommend to use deleteLater() to delete
QObjects that live in another thread:
Quoting the ~QObject() docs:
> Warning: Deleting a QObject while pending events are waiting to be delivered
> can cause a crash. You must not delete the QObject directly if it exists in
> a different thread than the one currently executing. Use deleteLater()
> instead, which will cause the event loop to delete the object after all
> pending events have been delivered to it.
If deleteLater() is not thread-safe, it cannot be used for one of its intended
purposes.
Change-Id: I333d506b42bdfcdff00fe6cefa234c21865625a6
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This is the most flaky-pass test currently.
It fails the first time it is run on MacOS_10_12, but
succeeds all the following times.
This happens extremely often, so disable it until the issue
is resolved.
Task-number: QTBUG-76566
Change-Id: I94359eceb91c3b958930424e6c8b5957fb3f1252
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
QWidgetPrivate::updateIsTranslucent sets the surface format of the
window with the alpha based on the translucency attribute, so we need
to call this function when the attribute value changes. The test can
confirm that the window's requested surface format has changed, we
can't rely on what is actually set, and don't have to rely on
hard-coded values like 8bit alpha.
While WA_NoSystemBackground needs to be set for WA_TranslucentBackground
to have an effect, we can't clear the attribute when clearing
translucency (as it might have been set explicitly).
Change-Id: I238d6930b7e0488397467a4e035b5f530566a1ff
Fixes: QTBUG-60822
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
When calling waitFor{ReadyRead,Disconnected} it will wait for data but
if the data is already received and the read notification has been
queued (and there's no more data coming in) it will return false.
By checking if a read notification has been queued and then handling
this we can easily take care of this scenario.
Fixes some flaky tests which missed the read data in waitForDisconnect
and similar.
Fixes: QTBUG-38385
Change-Id: Ic05d59883c1175783e56ff1822b6636c35aec874
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
If we end up in the connected state then we should pass on any
remaining data immediately instead of waiting until the next time we get
a read notification.
The other `case`s in the switch might be able to do something similar,
but I don't want to introduce that logic now in case it breaks something
else, the Connected branch is small and simple to deal with.
Should severely reduce flakiness with socks proxy in CI under pressure.
Task-number: QTBUG-76367
Change-Id: I0965d4c62a29a25ce6b8dd60862a464279aef0b4
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Ryan Chu <ryan.chu@qt.io>
Like in Qt code itself we should use forward slashes consistently.
This enables the vulkan test to run on Linux for MinGW
cross-compilation.
Task-number: QTBUG-76660
Change-Id: Ifc6dec11bac3c3769d5d06e49da529f66f7b5843
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Otherwise, the math will fail badly. Documentation improved to reflect
reality.
Change-Id: I9e3d261ad9bf41cfb2b6fffd159085cd38e3c388
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The make executable that comes with Android's NDK tries to execute the
shell-builtin "move" as a stand-alone executable unless you trick it
to execute "move" through the shell by surrounding one argument with
double quotes or such.
Force the execution of "move" through shell by altering QMAKE_MOVE for
Android on Windows.
Change-Id: I5b1490ad0606960dbd06a4cafb0b0b983e265159
Fixes: QTBUG-35713
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Although QPlatformInputContext::reset() documentation says that reset()
should not send any QInputMethodEvents, implementations on Windows,
macOS and iOS do send a QInputMethodEvent which clears preedit text in
their reimplementations of reset(). Text editing controls depend on that
and may not clear preedit text if such event is not sent.
Change-Id: I75ab73946cb06e93e5fc5e98e0cc503a7de5c2e0
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
There is no need to tell the editor to stop composing if user taps so
close to the cursor position that the cursor will not move anyway. If we
do stop composing in such case, then since there will be no cursor
position change notification, we will never start composing again
(before the cursor is actually moved), and the current composing region
will remain being displayed as normal text instead of being displayed as
composing text.
Change-Id: I4ebe6442e1ba8c365d6754c1a8487235d177c732
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Fixes auto-capitalization in AnySoftKeyboard. It was typing the whole
first word in a sentence in upper case.
Change-Id: I605a1aee39d432a3474c0bf706445d354562285f
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
7a7c722782 caused a regresssion in some
cases.
Change-Id: I1089a79534d811b195de663ff664d9ba5a6ac6c5
Fixes: QTBUG-74110
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
On Windows buildToolsVersion 29.0.0 have problems, therefore,
it's better to use a version that we know it works on all platofrms.
Change-Id: I25cdea4b8101bfe5f022025fcd7cc4cbf358fa03
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
This fixes the usability issue of a modal dialog showing up behind a
splash screen, not visible to the user, but blocking user input and the
application startup sequence until discarded.
[ChangeLog][QtWidgets][QSlashScreen] On macOS, lower the splash screen
when a modal dialog is shown to make sure the user sees the dialog.
Change-Id: Ibae768f76909d930cb25dcf5cee31edc5f15c29a
Fixes: QTBUG-49576
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Don't go through the doScaledRead path (i.e. calling read_image_scaled)
when reading an image that the target size is the image of the file we
are opening.
This makes the loading of the file much faster while keeping the output
correct.
[ChangeLog][QtGui][QImage] Improve loading time when loading png files
that have the same size as the target.
Change-Id: I2a33c49fe1ce52ec296c2175ee542b5bcdec2c4b
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
When using a http proxy (and presumably other proxies) we might have
failed/aborted (aka "finished") the request and _then_ receive a
"proxy authentication required" message from the proxy. In this case
there is no spdy/http2 reply in the queue, so asserting is wrong.
Change-Id: Id9b76b580299f6a6cd6efad62d6aaf63183816fb
Fixes: QTBUG-76426
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
In some scenarios with QNAM we call socket->close, leading to a flush,
leading to an error, leading to another error emission... To work
around this scenario we stop trying to close the socket if the network
channel is already closing.
Change-Id: Id15504f476484ce61f11ba83a5755ceb5f581f9b
Fixes: QTBUG-76567
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Inspired by:
https://codereview.appspot.com/6812076/
Resizing a window larger results in the newly exposed region being invalidated
but the old region is treated as valid.
This can result in the old region no longer updating.
Was added to D3D9.
Improving a fix from Filippo Cucchetto:
https://codereview.qt-project.org/c/qt/qtbase/+/195336
and pushing to D3D11.
ifndef protects against compilation error for WinRT.
Invalidate() should be used only for desktop apps.
Task-number: QTBUG-46074
Change-Id: Ie24b8dffe130b970f2362337ac4f9bee666f82b2
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
The directory must be prepended by = to denote the sysroot.
This amends commit 797f686e.
Change-Id: Ib85c0abc58fce3504ecccef0e223b2618ac9149a
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
We must use the same two-level replacements like for .prl files.
This amends commit d5071a40.
Change-Id: Iea065d01dee61cf2d1ff78640d045c3c76db9ac8
Fixes: QTBUG-76625
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Converts from OpenGL formats to Vulkan formats.
There are commented out lines for the formats in QOpenGLTexture::TextureFormat
for which it was hard to find an unambiguous mapping to vkFormat.
Task-number: QTBUG-75108
Change-Id: I06a7fd8df7d98cef314410ffd79ca9cff6599357
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
(cherry picked from commit b21b07877a)
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
See comment in qnumeric_p.h:convertDoubleTo for details.
Change-Id: Ifcd13f7f67995af6a60e50ccabe843a855be04ae
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Calls like
QHighDpi::fromNativePixels(point, window)
would return device independent coordinates outside any
screen in cases where the window is spanning multiple
screens and the native point was not on the main screen.
Correct this by looking up the correct screen and use
its scale factor and origin when scaling coordinates.
Task-number: QTBUG-73231
Change-Id: I01a3a42f42121b8d9f4ced2bb0fb023d6ae6bfe7
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Using QImage allows creating the style before the application has
been created, and is the more modern API.
No changes to the documentation needed.
Change-Id: Ifa0e5fa1113802fca18fbd45bb3c0a5ba1dbfeab
Reviewed-by: Andreas Aardal Hanssen <andreas@hanssen.name>
Fix most clang warnings about override, nullptr, range-based for loops.
Change-Id: Id47e57adb63a38e2f397a31511b788a2432c97cf
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
