Add override to addPaths and removePaths in kqueue filesystem watcher
Fixes build break in developer-build with gcc 10 on *BSD
Pick-to: 6.1
Change-Id: I7d91251d6bcb1614bb9f21c319abc9c2df681cda
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
DEFAULT_IF needs to be true so that in a static build the
built plugin is automatically linked into the test executable
using the special static plugin per-repo behavior in
qt_internal_add_executable.
The QtPostProcess routines are not executed for this test project
because we don't use qt_build_repo. This means that no
QtFooPluginCMakeConfig.cmake file is generated and thus there's no
point in using qt6_import_plugins because the pulic plugin genexes
won't know about this target anyway.
Explicitly set the CLASS_NAME so that the name expected by the
Q_IMPORT_PLUGIN macro matches the name of the plugin instance that moc
generates in QT_MOC_EXPORT_PLUGIN.
Amends 22e967c304
Task-number: QTBUG-87580
Task-number: QTBUG-90341
Pick-to: 6.1 6.0
Change-Id: I5ef361e7e2cebc46b35310c679f15c84cd61b4a5
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Previously, in a top-level build we always generated the final prl
file somewhere under QT_BUILD_DIR (which is qtbase_build_dir). After
each repo was processed by QtPostProcess.cmake, we installed the prl
files found in PROJECT_BINARY_DIR.
For qtquickcontrols2 this meant that qml plugin prl files were placed
under qtbase/qml, but we tried installing the prl files from
qtquickcontrols2/qml, which didn't have any prl files.
In a static Qt build, qmake's qt.prf calls qmlimportscanner to
identify which plugins should be linked to the executable. This worked
fine because the plugin .pri files were installed correctly.
None of the qml plugin library dependencies were linked in though.
This is supposed to happen in qmake's C++ code where it tries to
find the associated prl file of a linked library in order to extract
all its dependencies. Because no prl file was found, linking failed
with multiple undefined symbols.
Fix this by installing the prl files from QT_BUILD_DIR rather than
PROJECT_BINARY_DIR.
Note that this will create multiple install rules for certain files,
but it's harmless. An example is imageformats.
We process qtbase plugins, see qjpeg, issue an install rule from under
the qtbase/plugins/imageformats folder. We then process
qtimageformats plugins, see webp, issue another install rule from
under qtbase/plugins/imageformats.
The first install rule will install both qjpeg and qwebp, the second
install rule will merely say all plugins are up-to-date.
Pick-to: 6.1 6.0
Change-Id: I8a4bb67bfafc1d016eab62f4fe66b6ba378ceeb2
Fixes: QTBUG-93021
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
The six properties duration, updateInterval, currentTime,
direction, loopCount and easingCurve have been ported to the
new property system and are now bindable.
Drive-by renamed a local variable to avoid shadowing.
Task-number: QTBUG-85520
Change-Id: Ibabf106f5200d2dd4329a1e1f96112eccc29d6b1
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
For items that are children of other items, removeRows calls
beginRemoveRows directly and then once again inside takeChild()
The signal blocker that dates back to the monolitic import from Nokia
prevents the model from emitting extra signals
but the persistent indexes are corrupted nonetheless.
Fixes: QTBUG-90030
Pick-to: 6.1 6.0 5.15
Change-Id: I5bc4b2598bf13247683b113faeec22471f1f04a4
Reviewed-by: David Faure <david.faure@kdab.com>
When inserting rows to a branch with no columns
the tester should not complain about indexes being invalid
Pick-to: 6.1 6.0 5.15
Change-Id: I466f4e5140b10f6dcf65a71f109c2d3be7336507
Reviewed-by: David Faure <david.faure@kdab.com>
The qt_ForceTlsSecurityLevel symbol is not defined if SSL is not configured,
so don't use it in the QNetworkReply test if it's not.
Change-Id: Id3110a1997fea1648b48bfc411e90e7efe306678
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
In a cross built Qt, for example Qt for Android, calling "qmake -spec
android-clang" led to an error message:
"Could not find qmake spec '-qtconf'."
This happened, because:
- the qmake in Qt for Android is a wrapper script that calls
"qmake -qtconf qt_target.conf -spec android-clang"
- the first stage of command line argument handling in qmake garbled the
call to "qmake -spec -qtconf qt_target.conf android-clang"
We do not modify the order of arguments anymore.
Instead, we skip the "-qtconf <file>" arguments in the first argument
handling stage that is supposed to determine qmake's modus
operandi (like -project or -query).
In addition, we need to fix the assignment of
QLibraryInfoPrivate::qtconfManualPath which was only done if
QMakeGlobals::addCommandLineArguments returned ArgumentsOk. However,
this function returns ArgumentUnknown, if it encounters an argument it
cannot handle - like the project name.
Now, we assign QLibraryInfoPrivate::qtconfManualPath if there was no
error detected.
Document the return values of addCommandLineArguments.
This amends commit 661b586a69.
Pick-to: 6.1 6.0
Fixes: QTBUG-93079
Task-number: QTBUG-85136
Change-Id: I12ec25b17d64c00be2a3904b7c4a975b781500a0
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Routine update with minor changes to locale data, no new languages,
territories or scripts. Various Spanish locales change m_grouping_top
from 1 to 2, reversing a change to a test of Costa Rica's currency
formatting made in commit bb6a73260e.
Includes updates to time-zone IDs.
Fixes: QTBUG-91478
Pick-to: 6.1 6.0 5.15
Change-Id: I78ee161275b3c456c5800a7317a96947c932cf8e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The kernel appends that to the symlink targets in /proc for files that
have been deleted.
Change-Id: I7a386ad4f0cb4e2ba629fffd16789b5a52491627
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
There's no need to find the local time zone and do expensive conversions
to compare regular, old seconds-since-1970 to the time now.
Task-number: QTBUG-93069
Change-Id: I6cdea00671e8479b9c50fffd167897b91a04ce03
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
A c&p mistake manage to slip through in 5fabad9a61, amend it.
Change-Id: I706a6643d0745f336afca7bda0d8b0359354a0d6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Qt defines some integral datatypes (qsizetype, qintptr, quintptr,
qptrdiff) not in terms of the corresponding language datatypes (resp.
make_signed_t<size_t>, intptr_t, uintptr_t, ptrdiff_t) but as "integer
types with the same bit size of the corresponding language type" (and of
course the corret correct signedness for the target type).
This makes the Qt datatypes not printable via printf-like formatted
output, incl. qDebug, qWarning, QString::asprintf and so on; that's
because there isn't a format modifier that would universally work
with the Qt definitions.
For instance, on a 32 bit platform, ptrdiff_t may be a typedef for long,
while qptrdiff is a typedef for _int_ instead. Both long and int would
indeed be 32 bits, but they still are different types, and this means
that the ptrdiff_t-specific 't' length modifier would be wrong for
qptrdiff:
qptrdiff p;
printf("%td", p); // WARNING: -Wformat: wanted long, got int
Similarly, not using 't' breaks on 64 bits, and so on and so forth.
There isn't a way out, short of inserting casts on every print
statement.
So, let's adopt the same solution C/C++ use for their own integer
typedefs: the PRIx macros. This allows one to always use the correct
formatting specifier without the need of a cast.
I'm not adding the macros for the qintXX datatypes, as they already
exist in the Standard Library.
[ChangeLog][QtCore][QtGlobal] A series of PRIxQTDATATYPE macros have
been added. They make it possible to print some Qt type aliases
(qsizetype, qintptr, etc.) via a formatted output facility such as
printf() or qDebug() without raising formatting warnings and without
the need of a type cast.
Change-Id: I473226a661868aed9514d793c8e6e4d391ab5055
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Use QBENCHMARK_ONCE rather than QBENCHMARK to avoid skewing the results:
when the QBENCHMARK block is repeated multiple times after the setup
code above runs once, only the first setFilterRegularExpression() call
unfilters some rows, while the subsequent calls simply check that there
is nothing to do.
The added benchmark is sensitive to the inefficiency - quadratic rather
than linear time complexity - fixed by
7d92ef63d7. The following two tables
contain the benchmark results on my GNU/Linux system. The numbers denote
milliseconds per iteration.
1. Qt 5.15.2 without the performance fix:
10K 25K 50K 100K 250K 500K
no match 0 1 2 5 14 28
all 0 0 0 1 3 7
first 0 1 2 5 14 28
1000th 2 6 12 25 68 302
middle 3 34 132 518 3300 13665
1000th from end 1 4 9 19 50 103
last 0 1 2 5 14 30
each 10'000th 0 39 211 937 6326 41050
each 100'000th 0 1 2 5 4226 34780
Without the fix the benchmark times out and aborts at 1000K and 2000K
data rows.
2. Qt 5.15.2 with the performance fix:
10K 25K 50K 100K 250K 500K 1000K 2000K
no match 0 1 2 4 12 26 56 136
all 0 0 0 1 3 7 14 28
first 0 1 2 4 12 26 56 136
1000th 0 1 2 4 13 28 62 145
middle 0 1 2 4 13 27 59 142
1000th from end 0 1 2 4 13 28 60 145
last 0 1 2 4 13 27 59 141
each 10'000th 0 1 2 6 22 69 290 1413
each 100'000th 0 1 2 4 13 30 81 261
Change-Id: I419a5521dd0be7676fbb09b34b4069d4a76423b1
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
On all the platforms we support, we require these sizes in bits:
* char: 8
* short: 16
* int: 32
* long long: 64
We can get rid of the MSVC-specific type aliases (MSVC guarantees
__intXX to be aliases anyhow, not extended integer types) and just
use the builtin types.
Also, we require C++11 and C99, so "LL" and "ULL" are the correct
standard suffixes for (unsigned) long long. Remove the non-standard
suffixes from the Q_(U)INT64_C macros.
Change-Id: If007cd88d74064a163b5e910ca1983acd1dd1d10
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Let N = proxy_to_source.size() before the code modified in this commit.
Let M = (N - proxy_start). Let K = source_items.size(). The algorithmic
complexity of the removed loop is O(N+K+K*M), assuming the number of
O(N+K) reallocations is a constant. The complexity of the QList::insert
and std::copy implementation is O(N+K). This is much faster in practice
when K and M are of the same order of magnitude as N.
For example, this quadratic complexity issue results in noticeable
slowdown in the following scenario:
* a QSortFilterProxyModel is used only for filtering, not sorting;
* first set a filter that matches a single item in the middle of a
huge number of items (about one million) - this is reasonably fast
(takes about a second);
* then clear the filter (i.e. set an empty filter so that no item is
filtered out) and watch your application's UI freeze for a minute.
The "Add QSortFilterProxyModel clear-filter benchmark" commit (with
Change-Id I419a5521dd0be7676fbb09b34b4069d4a76423b1) adds a benchmark
that runs much faster with this performance fix.
Pick-to: 6.0 6.1
Change-Id: Ieaec173e6910f5d21eaee49402087f7711abbedf
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Even if the user (usually accidentally) sets a thread count of zero or
negative. The reporter in the bug report did
QThread::idealThreadCount() - 1 on a 1 CPU system...
Drive-by add to the documentation and the missing #include.
Pick-to: 6.1 6.0
Fixes: QTBUG-93007
Change-Id: I6cdea00671e8479b9c50fffd167807d14e030154
Reviewed-by: Samuel Gaist <samuel.gaist@idiap.ch>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Linking the object resource library to the static library still
causes an issue if the static library contains symbols that are
used in the resource library. The proposed approach creates the
following linker chain:
executable <- resource objects <- static library <- Qt::Core
For the static qml plugins this means that the qmlcache_loader
object file will state in the linker command line before the
plugin library that implements '::qmlData' and
'::aotBuiltFunctions' functions.
We also need to keep the INTERFACE linking of the resource library
to the targets that it belongs too, to collect all the dependencies
that resouce library supposed to propagate to top-level targets.
Pick-to: 6.1
Fixes: QTBUG-93002
Change-Id: Ice0aabb6817317724abeb3db3bb8a954905cfad1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
"Just in case", together with all the others platform checks we
have in qglobal.cpp, check that we're not accidentally breaking
our type aliases for fixed size integers.
Change-Id: Iaa9438741b3a7e89f3c458763c7d222ef2abe961
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Don't call pre routine function in qAddPreRoutine if
the qt_call_pre_routines is not called
Pick-to: 6.1 6.0
Task-number: QTBUG-90341
Change-Id: I0ee70561dc57b857f8b3b1cf42c9dfe0cf45bd49
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Passing Android-related variables to the initial CMake call would have
no effect if no (or an empty) value of CMAKE_TOOLCHAIN_FILE was
specified. To alleviate user confusion, yield an error if an Android
build was apparantly requested, but the toolchain file is missing.
Change-Id: I28ec94eabe436d4b9b410b48c7cad1b48d3e1bec
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Before, we relied on the deprecated "ndk-bundle" subdirectory being
present in the SDK root. Now, we choose the NDK with the latest version
in the SDK's ndk subdirectory.
Pick-to: 6.1
Fixes: QTBUG-87579
Change-Id: I37a9e03184f4518b4074d55375af08209591de94
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
When ANDROID_NDK_ROOT is explicitly set by the user, there's no point in
detecting the NDK. The detection had no effect, because the already set
ANDROID_NDK_ROOT variable would not be overwritten by our set call.
Fix the condition of the check: There is no CMAKE_ANDROID_NDK_ROOT
variable.
Pick-to: 6.1
Change-Id: I8f0d4f2a1a67445f67a79c7d2d0932099828b05e
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
It's needed in qtconnectivity, but currently the
qt_disable_apple_app_extension_api_only function is defined in
QtInternalTargets.cmake which is loaded only in qtbase.
Move the function to cmake/QtTargetHelpers.cmake.
Amends e189126f1a
Pick-to: 6.1
Change-Id: Ia2470e48a91385239394368780f5bbb223000113
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
QDateTimeParser::findTimeZoneName()'s invalidZoneNameCharacter() check
was using QLatin1String::contains(QChar), which converts the Latin-1
string to UTF-16 on each call, despite having pre-checked that the
QChar is ASCII. So use memchr() instead.
Change-Id: I011e2b4ba3be20711fc5005f62e4f9f6a392dd16
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
- Start in root folder so that large directories (/bin/, Windows)
are easily reachable
- Remove the line edit and navigate by double clicking instead
since this is more in line with expectations
- Use a QPlainTextEdit for logging
- Make the log message more informative
- Add icons
Change-Id: Ia3cd7fc143efef80772923291f0b711913aa47be
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Define the template calls in the header to allow for better type
handling and checking with constexpr and avoid overuse of macros.
Depending on the type provided in the QJniObject's call, the
signatures and the correct JNI function variant is used.
If a type is not supported a static_assert throws a compiler error.
Pick-to: 6.1 6.1.0
Change-Id: I8a4d3ce85e1ff76ef385633f2a68511fffd12e55
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Error: qtbase/src/corelib/io/qstandardpaths_android.cpp:49:17: error: no
namespace named 'QNativeInterface'; did you mean
'XXXXX::QNativeInterface'?
Pick-to: 6.1
Change-Id: I0f9fbf1879f72683d31ee0e8603b9bb26c31ae3c
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
When cross compiling from a case sensitive file system, casing
matters, and mingw headers and import libraries consistently
use lowercase.
This was uncovered by d385158d5213ef568b7629e2aa4a818016bbffac;
prior to that, the schannel TLS plugin didn't end up built (at
least when cross compiling).
Fix other similar cases that can be found by grepping the repo.
Change-Id: Ia696e17b7aaa979d7b7f5b0801383f338a8b585b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Currently in order to compile a Qt6 test standalone one needs to use
the qt-cmake-standalone-test script which will load the
Qt6BuildInternals/QtStandaloneTestTemplateProject/CMakeLists.txt project
with the current test source directory to create a complete CMake project.
This commit will allow a test to have these lines at top:
cmake_minimum_required(VERSION 3.16)
if (NOT QT_BUILD_STANDALONE_TESTS AND NOT QT_BUILDING_QT)
project(a_qt_test LANGUAGES C CXX ASM)
find_package(Qt6BuildInternals COMPONENTS STANDALONE_TEST)
endif()
and be standalone and work with any IDE capable of loading CMake projects.
Fixes: QTCREATORBUG-25389
Pick-to: 6.1
Change-Id: If3f878b7e560a8bfbb604a8f1aa30b7962742e66
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Those overloads have been documented as obsolete, but never been deprecated.
Add the deprecation macros to trigger warnings as of Qt 6.2.
The overloads taking a single StandardButton should not be deprecated until
Qt 7, as otherwise porting from old to new API will require an unnecessary
cast to StandardButtons for calls with only a single enum value.
The unit test explicitly tests the deprecated members, so disable warnings
there.
Fixes: QTBUG-92483
Change-Id: I283ddce4681eafda2378607f999946e56bbb777e
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
When we can't get the icon for the first time, we would use the data
in the cache later, and we will never get the new icon when the system
theme is updated while the application is running.
Pick-to: 6.1
Change-Id: I839ad9983918561a1dc6bc842f85477bba53f64a
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
It's allowed to change asynchronously by another thread.
Pick-to: 6.1 6.0 5.15
Change-Id: I6cdea00671e8479b9c50fffd167836a08a42cc1d
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
All TLS (and non-TLS) backends that QSsl classes rely
on are now in plugins/tls (as openssl, securetransport,
schannel and certonly plugins).
For now, I have to disable some tests that were using OpenSSL
calls - this to be refactored/re-thought. These include:
qsslsocket auto-test (test-case where we work with private keys),
qsslkey auto-test (similar to qsslsocket - test-case working with
keys using OpenSSL calls).
qasn1element moved to plugins too, so its auto-test have to
be re-thought.
Since now we can have more than one working TLS-backend on a given
platform, the presence of OpenSSL also means I force this backend
as active before running tests, to make sure features implemented
only in OpenSSL-backend are tested.
OCSP auto test is disabled for now, since it heavily relies on
OpenSSL symbols (to be refactored).
[ChangeLog][QtNetwork][QSslSocket] QSslSocket by default prefers 'openssl' backend
if it is available.
[ChangeLog][QtNetwork][QSslSocket] TLS-backends are not mutually exclusive anymore,
depending on a platform, more than one TLS backend can be built. E.g., configuring
Qt with -openssl does not prevent SecureTransport or Schannel plugin from being
built.
Fixes: QTBUG-91928
Change-Id: I4c05e32f10179066bee3a518bdfdd6c4b15320c3
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
If you have a property myprop of type float, testReadWritePropertyBasics
would happily accept
testReadWritePropertyBasics(myObject, 1, 2, "myProp")
The test would then fail when setting bindings, as we would try to
install a binding of type int on a float property, which gets rejected
at runtime.
To prevent unexpected failures, verify that the types match before doing
any further checks.
Change-Id: I3893563fce0e11f9e20afa7c6a1e1fe0385382ab
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Other manuals have a 'Getting started' section too. Better to qualify
it further, so that searching in the Index of Qt Creator Help mode or
Qt Assistant allows to differentiate between the different products.
Pick-to: 6.1
Change-Id: I7cfdba3c3567bdf49d3fe34307576f88c9f20126
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
If the signal takes a value, we pass the current value of the property
to it.
As we now use eager evaluation, accessing the current value is now
possible.
Change-Id: I5e6947a6575bfa8ca5143f56620c645d4750a686
Reviewed-by: Andreas Buhr <andreas.buhr@qt.io>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This mirrors the functionality of QObjectCompatProperty::notify, and can
be useful to delay notifications until a class invariant has been
restored.
Change-Id: I1c16a0b1537a1b53d144c8abe48e546553edf877
Reviewed-by: Andreas Buhr <andreas.buhr@qt.io>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Certain repositories like qtsvg, qtimageformats and qtvirtualkyboard
build plugins associated with Qt modules from other repositories
(qtsvg's QSvgPlugin associated to qtbase's QtGui).
When configuring in-tree tests in the same build folder as the
repository, the test executables would not automatically link to these
plugins.
Fix this by recording the existence of such plugins in a separate
property of the associated Qt module and only link them when both the
test executable and plugin are from the same project (their
PROJECT_NAME coincides).
This is in addition to linking the plugins associated with the
module where both are built in the same repository.
The logic is a bit tricky and ensures that plugins are not
accidentally initialized twice, so that in-tree tests work for both
top-level and per-repo builds.
As a drive-by, added a TODO explaining why in-tree tests that need to
link to static QML plugins won't work (somewhat unrelated to this
change).
Amends 734d2cdbc4ff6db6b3df8fffbb23dbbb565c076b
Amends b1fcdad9c9b9ad2bddd00f7301c8dd1159d523c2
Pick-to: 6.1
Task-number: QTBUG-87580
Change-Id: I3e1ff8166864f92dea931ec2ea34b6f56b4eec60
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
