On systems with weak memory ordering, it was possible for the
storeRelaxed(Initialized) to be observed by another thread performing a
loadRelaxed() without observing the contents of the object itself. The
mutex *does* release the contents of the object to memory, but without
the corresponding mutex acquisition, we couldn't guarantee that the
object's contents would be observed. Now we can.
We don't need to fix the load inside the mutex because the mutex will
have acquired everything from either a previous call to pointer() or to
reset(). The store inside reset() need not be storeRelease() either
because the effect of observing the Uninitialized state will be to lock
the mutex.
None of this is used to protect the data as it is being mutated by the
user in multiple threads, or their access simultaneously with reset()
(which is why the load outside the mutex was removed).
Thanks to litb on Slack for noticing this and bringing to my attention.
Pick-to: 6.5
Change-Id: Idd5e1bb52be047d7b4fffffd1752df5b4d9b2e3f
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Functions in Qt that take a callback need to support callables with or
without context objects, and member functions of an object. The
implementation of those overloads follows a pattern that ultimately
results in a QSlotObjectBase implementation being created and
passed to an implementation helper that takes care of the logic.
Factor that common pattern into a new helper template in QtPrivate
that returns a suitable QSlotObjectBase after checking that the
functor is compatible with the specified argument types.
Use that new helper template in the implementation of
QCoreApplication::requestPermission and QHostInfo::lookupHost.
The only disadvantage of centralizing this logic is that we cannot print
a more detailed error message indicating which argument types the
caller expects. However, that information is visible from the detailed
compiler errors anyway.
Change-Id: I24cf0b2442217857b96ffc4d2d6c997c4fae34e0
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Add the move ctor and move operator for QSqlField
Task-number: QTBUG-109938
Change-Id: Ib66eff76c3a920de9cfb3288f4219555005e7ae5
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Use QSharedData for the private class instead a home-brew version.
Change-Id: Id3625bb0eb8f81c9caa672e2453dab3d44b15ea9
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
When I broke them out from functions in the QLocalTime namespace I
didn't notice that's where I was putting them; they don't belong
there.
Change-Id: If4c9d996b3e46b3b46a29a97d0bcc2cac72c91ab
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Problem: "ninja clean" does not fully clean the build directories of
configure-time executables (e.g. syncqt). This can lead to problems when
building with compiler and linker launchers: on configure time, the
launchers are not used (compare CMake issue #20762). After a "ninja
clean", the executable might be removed but the object files are still
there. This leads to a situation where the object files have been
created without the compiler launcher but are linked with the linker
launcher. We encountered a situation however, where the linker launcher
requires the usage of the compiler launcher.
The configure-time executable has a ${target}_build custom command that
runs "cmake --build" and creates a timestamp file to track when to build
the target. To circumvent the problem of stale object files we add the
"--clean-first" argument to that target to fully rebuild configure-time
executables if the timestamp file is out of date.
The performance this imposes is negligible, because
1. Those configure-time executables are seldom out of date.
2. They are supposed to be "tiny executables with system dependencies
only" anyway.
Pick-to: 6.5
Change-Id: I701f9089f5ad941ffdf235aeccc3119b68c4e3e3
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
If syncqt process crashed due to the incompatibility with the standard
library CMake doesn't produce any useful output. This adds the message
that will give short explanation and possible solutions to fix this
behavior.
Pick-to: 6.5
Task-number: QTBUG-112747
Change-Id: Ib51aec19e3fcacf07515a3d20d72f89753bcdc33
Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Creating a continuation with QtFuture::Launch::Async policy does not
work well with the example, because it still needs to update the UI
once the async continuation is finished. If the user decides to
close the application while the async continuation is executed,
the next continuation will be accessing data from the destroyed
Images object.
Fix it by using QtConcurrent::run() to do the "heavy" work in a
separate thread, and use a QFutureWatcher to handle the results of
the async execution. Update the example documentation accordingly.
After this patch the example still shows the usage of continuations
and onCanceled()/onFailed() handlers. However, it now does not
illustrate the usage of different launch policies and continuation
contexts. It might not be a big issue, because the QFuture
documentation describes these topics rather extensively.
Fixes: QTBUG-103514
Pick-to: 6.5
Change-Id: I8142535064ff7a4e8007a5c0a8fe7709d6d942ec
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Changes code snippets that are visible in the docs.
Task-number: QTBUG-113116
Pick-to: 6.5
Change-Id: If743234bfe6947acf02307bf1144daad4fba5d73
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Use SPDX-License-Identifier also for shell scripts. Also change
license to LicenseRef-Qt-Commercial OR GPL-3.0-only, which makes
arguably more sense than LGPL.
Pick-to: 6.5
Change-Id: I3a46468f264747916aacbf206f5ccba28b6c6593
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
SPDX/reuse does not feature any generic 'public domain' license
identifier. So far we claimed CC0, which is however not entirely correct.
Now that we have LICENSES directory, let's correctly claim this as
a specific license.
Also, remove the LGPL in-code claim. The adaptations for Qt code is
minor so it doesn't make much sense to claim a different license.
Pick-to: 6.5
Change-Id: I4e943a45baae192b54c38184e8eb85fc6d4035e5
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
After changing android target SDK version to 31, some missing values has
to be defined in AndroidManifest.xml.
AndroidManifest.xml template was updated in qtbase in:
56dee3de5e commit. In case when example
uses its own AndroidManifest.xml file, it need to be updated separately.
* android:exported="true": because the manifest sets an intent-filter,
and it then has to explicitly to avoid the warning [1].
* android:allowBackup="true": this has to be explicitly set, we set it
to the default value here [2].
* android:fullBackupOnly="false": SDK 23+ use this to deteremine to
user auto backup or not, we set it to the default value here [3].
[1] https://developer.android.com/guide/topics/manifest/activity-
element#exported
[2] https://developer.android.com/guide/topics/manifest/application-
element#allowbackup
[3] https://developer.android.com/guide/topics/manifest/application-
element#fullBackupOnly
Fixes: QTBUG-112816
Pick-to: 6.5 6.2 5.15
Change-Id: Ia48007a84009901be508d6cc087790d574227ad7
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
QTabBar lays out the tabs when the bar's size or content changes,
often lazily. This should not change the scroll offset of the tabbar,
which is controlled by the user, or at most when a tab needs to be made
visible (e.g. when it becomes the current tab).
Move the logic of updating the scoll offset into the makeVisible
function, so that the scroll is only adjusted to either leave no gap
between the last tab and the right edge of the widget of there is still
a scroll; or to reset it to 0 if there is enough space for the entire
tab bar. Since layoutTabs does show and hide the scroll buttons, we
cannot skip this when the buttons are invisible. However, the
normalizedScrollRect helper now needs to return the entire widget rect
if there are no visible scroll buttons.
Add a test case that simulates the previously broken behavior where
the scroll got unnecessarily reset to 0 when resizing.
Fixes: QTBUG-113140
Pick-to: 6.5
Change-Id: Ic19fbb82821ea09cc5e7646dcbce3aa7607909c2
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Santhosh Kumar <santhosh.kumar.selvaraj@qt.io>
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
This commit introduces fine-grained Bluetooth permissions control to
the QBluetoothPermission class.
For now the fine-tuning of the permissions is only supported on
Android. On Apple enabling any of the permissions is equivalent to
requesting full Bluetooth control.
Task-number: QTBUG-109964
Change-Id: Ie7ac6577cf6a21419b73b33f8cf7e87bc3f8cf43
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The crash is caused by the cleanup sending trace messages when the
plugin has already been destroyed. Add shutdown callback to the plugin
to indicate this has happened. We can't use signals since that also
generetes trace event.
Pick-to: 6.5
Change-Id: I2e490fc51c2aaa97c240c1496a528a6ff6077bd0
Reviewed-by: Hatem ElKharashy <hatem.elkharashy@qt.io>
Reviewed-by: Janne Koskinen <janne.p.koskinen@qt.io>
Add common prologue and add error message when tracing is used in a
module that doesn't have Q_TRACEPOINT enabled, but tracing is enabled.
Pick-to: 6.5
Change-Id: I64ca074942f6e89b4f5b5e3b6048b2b713c06df8
Reviewed-by: Tomi Korpipää <tomi.korpipaa@qt.io>
Reviewed-by: Janne Koskinen <janne.p.koskinen@qt.io>
This patch implements a public getter for the hasStandalone attribute.
It returns true, if standalone has been explicitly declared in an XML
header and false otherwise.
As this is no longer necessary it removes accessing QXmlStreamPrivate
from QDomParser.
[ChangeLog][QtCore][QXmlStreamReader] added hasStandaloneDeclaration()
Change-Id: Iaaa0a728a6f7186e40637186077f7b49c112f7a9
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Yuhang Zhao <yuhangzhao@deepin.org>
These classes depend only on Core, not Gui.
This allows dropping the dependency of tst_qxmlstream and tst_qzip on
QtGui, and prevents a tst_qxmlstream FTBFS when building with
QT_NO_TEXTODFWRITER.
Symbols move from QtGui to QtCore, but the classes are private API, so
not under BC constraints.
The classes are not used outside qtbase, so no other in-tree users
need porting.
Task-number: QTBUG-3897
Change-Id: Ifa148f43ec139d7f9ac1f3893e2fcf4640e3c60c
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Just call itself after copying the buffer, thus avoiding instantiating
insert_helper() for QVarLengthArray.
The other two cases of QtPrivate::q_points_into_range + QVLA in
qstring.cpp are not worth changing.
Change-Id: I9671dee8ceb64aa9b9cafffd1742fa7bb4cbebd8
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
The expected states are:
- nothing sets theMainThread before main()
- theMainThread is reset when the last QObject (the QCoreApplication in
the test) is destroyed
The GUI version of this test appears to leak a lot of QObjects. By the
time this function runs, theMainThread's QThreadData still has a
refcount of 66 on Linux/XCB. The Windows non-GUI version also
failed. Neither situation was investigated to see why objects are
getting leaked.
Pick-to: 6.5
Change-Id: Idd5e1bb52be047d7b4fffffd17507d9e6ef08743
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
MSVC binaries can also be stripped and according to my
experiments they still work normally. So we should not
limit this check to MinGW only.
Change-Id: I026d75a38b94f309ad695cf8f700ed3ac160dd87
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Timothée Keller <timothee.keller@qt.io>
For our purposes, the sem_t code is better, so we've preferred it since
commit c7ab816af1 (5.7). Olivier wrote in
that commit's message, "It makes tst_QMutex::contendedQMutex with no
msleep 8 times faster".
That's true because QMutex didn't lock the underlying pthread_mutex_t in
QMutex::lock(), as we used the inlined atomic code for the uncontended
case.
It is probably possible to merge the qmutex_mac.cpp and qmutex_unix.cpp
code now (both are based on semaphores), but I won't do that for two
reasons:
1) At best, the PThread functions are going to be thin-wrappers around
the code we already have, like FreeBSD's are around usem (see [1])
2) Darwin has a private API that resembles futexes so we may want to
go that way eventually (see [2])
[1] https://github.com/freebsd/freebsd-src/blob/main/lib/libc/gen/sem_new.c
[2] https://codereview.qt-project.org/335849
Change-Id: Idd5e1bb52be047d7b4fffffd175369b13ba47bed
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Instead of having one member signaling !openssl
let's just add booleans for openssl, schannel and securetransport.
The latter two of which are not currently used but may be in the future.
As a drive-by, make a compile-time check into a runtime one.
Change-Id: Id2f51f5396383a3f5836ac708996bfce8ae35c91
Reviewed-by: Konrad Kujawa <konrad.kujawa@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Populate a subset of the font families at startup if the local fonts
access API is supported, and the access permission has been given.
Since this code runs at app startup there is no opportunity to request
font access. That should be done in response to user action, for
example by having a "load local fonts" button in the application.
Pick-to: 6.5
Change-Id: Ib6826deeec06ee3def0e793dd1462977710462be
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Reviewed-by: Aleksandr Reviakin <aleksandr.reviakin@qt.io>
lttng ctf_array does not support float types which
causes compilation error when a float type is passed
to the function. A solution is to pass the array
elements one by one to TP_FIELDS.
Fixes: QTBUG-112761
Pick-to: 6.5
Change-Id: I30e7049d9eda1141298145897df372213145c1b4
Reviewed-by: Antti Määttä <antti.maatta@qt.io>
Reviewed-by: Janne Koskinen <janne.p.koskinen@qt.io>
Start tracking the window geometry before a mouse drag, so that we can
revert back to that geometry when we restore from maximised. Previously,
when dragging from maximised to maximised, the restore geometry would
end up being the final drag place before snapping to maximised, instead
of where the window was before the first maximised.
Fixes: QTBUG-112814
Pick-to: 6.5 6.2
Change-Id: Ic2ddf29d6c4abdc9e8b0c5161b17aa6ee9474ea3
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
We should be able to just pass `0L` and avoid defining a None macro.
Pick-to: 6.5
Change-Id: I513d726120454523627a1e66515a5a533c0238b1
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
This fixes the assert raised on overwriting event handlers
Fixes: QTBUG-113041
Change-Id: Ie2afe09f4111ea542297b82a51382f1eb04ec960
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Aleksandr Reviakin <aleksandr.reviakin@qt.io>
This reverts commit e0cec08480.
Reason for revert: this can't possibly be the correct solution. It
set properties in headers (not .cpp sources) and headers can't cause
a build issue because they aren't built. Moreover, the problems
were seen in another module, so the properties set in those files
shouldn't even be visible to CMake.
Change-Id: I8473bb819e768bd203f89034d18132186da0371c
Pick-to: 6.5
Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
wasm heap uses a shared array buffer, which cannot be fed to
FileSystemWritableFileStream.write due to security limitations.
Heap memory has to be copied to a temporary buffer for the operation to
succeed.
This is only done if __EMSCRIPTEN_SHARED_MEMORY__ is on to optimize the
non-threading path.
Fixes: QTBUG-112881
Pick-to: 6.5
Change-Id: I0d117a8703caf4c17abc67b30df5248a53406d5f
Reviewed-by: Aleksandr Reviakin <aleksandr.reviakin@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Old examples inherited from Qt 4 tend to set some state, such as
enabling the depth test or culling, in initializeGL(). Newer examples
tend not to do this; they rather set the necessary state in paintGL().
This mattered little (or not at all) in the past, but with WebAssembly
and WebGL there are limitations in the GL context management in the
wasm platform plugin. Under certain conditions, esp. when
QOffscreenSurface is involved, it looks like the same native context
gets reused, which means there is a chance of unexpected changes to
the current state between calls to initializeGL() and paintGL(). (and
also between paintGL() calls) See QWasmOpenGLContext for details.
Update the textures example the same way we did for the cube one.
Add a note to the QOpenGLWidget docs about this problem.
Task-number: QTBUG-111304
Pick-to: 6.5 6.4
Change-Id: I29d2b2cdeb07bcecc5dc915d79c12b4323ca9ab3
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Inho Lee <inho.lee@qt.io>
- qt_finalize_target missed a reference to qt_add_plugin()
- qt_generate_deploy_app_script is also supported on Linux
- Use simplers examples for QT_DEPLOY_SUPPORT and
qt_standard_project_setup()
Task-number: QTBUG-113116
Pick-to: 6.5
Change-Id: If6bfd01b9615a73ab1bafddee350e49f6902a6c0
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Make a manual call to the geometry change handling function after a
WM_DPICHANGED event if the window is frameless, since WM_SIZE and
WM_MOVE will not be called.
Fixes: QTBUG-109429
Pick-to: 6.5
Change-Id: I79b9f386fe120ee3d06d6490d3f31a7a5d7121b0
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
QtXml is only DOM and SAX, not QXmlStreamReader/Writer (those are in
QtCore).
Pick-to: 6.5 6.2
Change-Id: I8454d7db90303d347d5b4be94c9f21401d1e273f
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Øystein Heskestad <oystein.heskestad@qt.io>
The \include command includes the source in its entirety when the
second parameter is omitted. This pulled in also the license header
which was visible in the generated documentation.
Add snippet tags and use them to extract only the content we need.
Pick-to: 6.5
Fixes: QTBUG-113138
Change-Id: Ie3fe2fede1e81d08201ec4353352ef069aebc388
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Rounding distances up can result in coordinates outside of
the clip rect. So stick to always round distances down when we multiply
a scaling.
Fixes: QTBUG-109640
Pick-to: 6.5
Change-Id: I784b7c90da9b6e7f5a925d4275eb67497616001d
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
By using an iterator-based for loop.
Change-Id: I9ae12f16bc2a5c2d74c8557a0324438102fec5b1
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Drive-by change: use QByteArrayView instead of allocating a QByteArray.
Change-Id: Iaf7acbbdb4efbb101b73b30061ce38dd1fa99ca3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
"notifier->socket()" returns qint64, but sockfd should be int, because
pollfd.fd is a gint (aka int).
Change-Id: If6618aa5e652d4284b989352d61a28b605106d09
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
b1802a164b added handling for a parent to
become focus proxy of a child. The respective 'else if' branch didn't
check whether setFocusProxy() was called with a nullptr argument.
This patch adds the missing nullptr check.
It also adds functionality to tst_QWidget::tabOrderComboBox() to test
the removal of a focus proxy, as well as the complete removal of an
element from the focus chain.
Change-Id: I4cb865b9ac4497fc5e2595910738fb77694f5837
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
