Directly writing to the underlying property storage has the potential of
breaking all kinds of internal invariants. As we return QBindable in
the public interface, we should not grant callers access to the
internals of the object.
Pick-to: 6.2 6.1
Change-Id: I737ff293b9d921b7de861da5ae23356c17690b78
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Andreas Buhr <andreas.buhr@qt.io>
This does not fix all data races that we have in the system yet.
One major issue is the virtual disconnectNotify(), that can be
called from any thread and thus is inherently problematic, as it
can collide with the object getting destroyed at the same time
in another thread.
Pick-to: 6.2 6.1 5.15
Task-number: QTBUG-88248
Change-Id: I9d841eb363b7e4f0de1657aeb8f5340d0fd55190
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It was suppressed unconditionally (albeit in different places and with
different levels of obviousness) almost everywhere, due to inability
to set the system locale in use by the implementation. Several
test-cases used ISO-8859-1 encoding on Q_OS_MAC, where the tests were
all suppressed anyway. A block of no_NO tests was #if 0'd out; and
should, in any case, have been nb_NO. Tests of any locale but en_US
were skipped on Q_OS_WIN because we can't set the locale (but
including the tests of en_US tacitly assumed that's the system
locale). If setlocale() failed, for ICU or DARWIN, the test was
skipped; but we might as well check for this in the _data() to save
repetition.
The test was laboriously going through the sign cases; relocate the
QString::compare() test's sign() function so that we can use it here,
too, and simply QCOMPARE() signs. Introduce a TransientLocale class,
copied from tst_QLocale, to take care of setting and restoring the
locale using setlocale(). Change the locale name to a QByteArray so
that we save having to convert it to one in order to pass it to
setlocale(). Since that changed every _data() row, reformat those rows
in the process - most of them were long lines.
On the systems where we can't set the locale used by the function
being tested, condition each block of tests in the _data() on whether
LC_COLLATE looks like the locale to be tested, and report how a
determined developer at least can (by repeatedly running the test with
different locales set) test all the cases; and we'll attempt the ones
that we can, when one of the relevant locales is in use. If that
leaves us with no tests we can do, QSKIP() in the _data() to avoid an
assert failure for "Test data requested, but no testdata available."
Change-Id: I75709fda8827dcbe74f80c4136042054da6fcb13
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
For user projects we run the static link order check once
'find_package(Qt6 ...)' is called.
If linker can resolve circular dependencies between static libraries
and object files we set the _qt_link_order_matters property of the
Qt::Platform target. This indicates the use of finalizers is not
required and we may rely on CMake-base propagation of resource
libraries and resource object files.
If linker could not resolve circular dependencies depending on
the _qt_resource_objects_finalizer_mode value:
- Finalizer will be called and collected resource objects will be
linked to the target directly.
- Finalizer will be omitted and resource objects will be linked
using the target_sources function implicitly. This only
propagates resource one level up if consumer links the static
library PUBLICly, but all symbols will be resolved correctly
since object files are placed in the beginning of the linker line.
In the CMake version 3.21 we expect that CMake will take care about
the order of the resource object files in a linker line, it's
expected that all object files are located at the beginning of the
linker line.
TODO: Need to confirm that the CMake 3.21 meets the expectations.
Amends 4e901a2f99
Pick-to: 6.2
Task-number: QTBUG-93002
Task-number: QTBUG-94528
Change-Id: Ia68976df8182d3d3007b90c475c1e3928a305339
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
System V semaphores are not supported in sandboxed applications,
so when Qt is configured with App Store compliance, or the user
requests POSIX IPC explicitly, we use that instead.
https://developer.apple.com/library/archive/documentation/Security/Conceptual/AppSandboxDesignGuide/AppSandboxInDepth/AppSandboxInDepth.html#//apple_ref/doc/uid/TP40011183-CH3-SW24
As the shared memory name limit on Apple platforms is very low,
we have to skip the existing logic for naming, and instead use
a truncated hash of the key. This should still be fine for
avoiding any collisions in practice.
An explicit check for the ENAMETOOLONG error has been added to
catch any cases where they key goes beyond the allowed length.
Sandboxed applications also have an extra requirement that the
key must include an application group identifier. This requirement
has been pushed up to the user and documented, as we don't have
enough information in Qt to know which identifier to use.
Both tst_QSystemSemaphore and tst_QSharedMemory work as before
with both sandboxed and non-sandboxed applications, after removing
some assumptions in tst_QSharedMemory about System V behavior.
Fixes: QTBUG-91130
Change-Id: Iaf1edb36a5d84d69e42ec31471a48d112faa8c6a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Amends 16f927a4f1.
At the time the original change was written, QStringTokenizer
had not been integrated, yet.
Change-Id: I83c31d816199bc48c4baea855d13cbf9eda9aaa2
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
User projects can set the QT_PROMOTE_TO_GLOBAL_TARGETS variable to
true so that the various imported targets created by find_package(Qt6)
are promoted to global targets.
This would allow a project to find Qt packages in a subdirectory scope
while using those Qt targets from a different scope.
E.g. it fixes errors like
CMake Error at CMakeLists.txt:5 (target_link_libraries):
Error evaluating generator expression:
$<TARGET_OBJECTS:Qt6::Widgets_resources_1>
Objects of target "Qt6::Widgets_resources_1" referenced but no such
target exists.
when trying to use a static Qt from a sibling scope.
Various 3rd party dependency targets (like Atomic or ZLIB) are not
made global due to limitations in CMake, but as long as those targets
are not mentioned directly, it shouldn't cause issues.
The targets are made global in the generated
QtFooAdditionalTargetInfo.cmake file.
To ensure that resource object libraries promoted, the generation
of the file has to be done at the end of the defining scope
where qt_internal_export_additional_targets_file is called,
which is achieved with a deferred finalizer.
Replaced all occurrences of target promotion with a helper function
which allows tracing of all promoted targets by specifying
--log-level=debug to CMake.
Pick-to: 6.2
Fixes: QTBUG-92878
Change-Id: Ic4ec03b0bc383d7e591a58c520c3974fbea746d2
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Makes it able to catch the last bug of saturating color values above a
certain value.
Change-Id: Ib2a3918623a1defe2981efe61cf8118e019e9d4b
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Previously, it went direct to QTestResults::addFailure() without going
via the checking for expected failure. Add QTestResults::fail() to
take care of this checking, as for verify() and compare().
Tidied up the code implementing expected failure and QFAIL(), while I
was about it. Adjusted an existing test to verify that expecting a
QFAIL() works, by using QFAIL() instead of QVERIFY(false).
Remove the QVERIFY(false) whose comment brought this to my attention.
[ChangeLog][QtTestLib][QFAIL] QEXPECT_FAIL() now correctly anticipates
a subsequent QFAIL(). Previously QFAIL() counted as a fail regardless.
Change-Id: Icc28cf70e5ff3006363791ea03aa01f2f591eb71
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
In commit 287ace562e, part of this test
was suppressed without filing a Jira ticket (or, at least, without
recording it in the QSKIP message). Since it's a known failure, it
should at least be a QEXPECT_FAIL, not a QSKIP. Since only some of the
subsequent parts of the test fail, I used QEXPECT_FAIL(,,Continue) on
each of the failing tests.
Task-number: QTBUG-94450
Change-Id: Iebc6801210c289b4502e59116e71d5901b71aa46
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Its collation (and everything locale-related) is only supported during
the lifetime of QApplication.
Change-Id: Ide97795664d45768e66248d47c3b1da83935b0d6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Various QDateTime tests relating to transitions
* used a nomenclature that made them confusing to thing about; and
* expected identically-initiallized variables to behave differently.
The latter, naturally, lead to "expected fail" tests.
Rewrote the tests to get the date-times they want to test at by means
that avoid the ambiguities inherent in QDateTime's lack of a way to
distinguish the two passes through the repeated hour in a fall-back
(QTBUG-79923) and added commented-out tests indicating what should be
true once that ambiguity is resolved. Verified the DST status is as
expected in the cases where that's the correct distinction between
date-times with the same date and time. Renamed various things to
(hopefully) make them more intelligible.
In the process, purged some leading 0s from numbers in code.
Fixes: QTBUG-68936
Change-Id: Id7a348995238b70dcb81a96edb8a3fa5315f86fa
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This test failed when checking for a pre-condition of the tested scenario,
so skip the test if that condition isn't met, as the test won't test
anything.
Amends b1fdcc8c0f.
Pick-to: 6.2 6.1 5.15
Task-number: QTBUG-64543
Change-Id: I135cd5b45efcae111305b9be338eb5429d3b97d5
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
The test uses a 64-unit array and deliberately clears various portions
of it, provoking a gcc warning:
warning: ‘memset’ used with length equal to number of elements without multiplication by element size [-Wmemset-elt-size]
The calls to memset() do, in fact, have a sizeof(T) factor in their
size. Suppress this warning for the duration of that test.
Change-Id: I7d144d655a75f5ef4449fa3b956f80bcc509a83b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
A mouse press that transfers focus from an open editor back to the view
will close the editor. To prevent that the corresponding release then
opens the same editor again we need to know that the closeEditor call
was caused by the mouse press. Since Qt first generates the focusOut
event, and then delivers the mouse press, we have to start a zero-timer
to check whether we are in the same event delivery process. If so, ignore
the corresponding release.
Add test case that simulates that chain of events.
Fixes: QTBUG-20456
Pick-to: 6.2 6.1
Change-Id: I28fa32bfbc776db207c594c329961f575ae58ea9
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
After 79fd1cb2c6 the methods for running
QtConcurrent algorithms in the blocking mode aren't used anymore. Since
ThreadEngineBase and ThreadEngineStarter classes aren't meant to be used
externally, it should be fine to remove startBlocking() methods now.
Removed the unused code and adjusted the tests accordingly.
Change-Id: Ifb13820ce207869d6f720bcb5be8d35bb355fe33
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Rather than when the data is received. Source compatibility is
achieved through double-decompressing the data. This lets us know
how many bytes are available just as before but without having the
uncompressed data left in memory.
Fixes: QTBUG-83269
Change-Id: I352bd09581614c582e4628243e2a0e895ba4946b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Small change needed to make QString_char16 and QString_QChar return -1
in this case, but other combinations already returns -1.
[ChangeLog][QtCore][Behavior Change] QString::indexOf(QChar) and
QString::indexOf(char16_t) now treat a negative start-position, from,
bigger than the string's size as invalid. It previously
clipped such start-positions to the start of the string, inconsistently
with other QString indexOf overloads.
Change-Id: Ic56c8a558bf40a94845c649647db569892d4df02
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
tst_QStringList::sort() wants to use the C locale; but, if it failed,
it left that in force, since it only restored the prior locale on
success. It should also use C.UTF-8, since Qt now wants UTF-8.
Change-Id: If62e3d8da682081bf969075a719d03caebf09233
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
We need those events to trigger palette color group changes in QQuickItem
without having to connect every item to yet another QWindow signal.
Task-number: QTBUG-93752
Pick-to: 6.2
Change-Id: I8534808cdaab828e5876f8fda31567aeb1b4272a
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
We use UTF-8 locales by default since Qt 6; and relatively few systems
have the encoding-unspecified locales we were trying to use, with the
result that the setlocale() calls all failed.
Task-number: COIN-689
Change-Id: Id791ba269bf4abac29da3daa4fd01684ca9caa7a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Because they were followed by a QVERIFY(false), no-one noticed until
now. Dates from 2011, when ICU support was first added. I guess
someone fixed the problem in the intervening decade.
Change-Id: I847816c297156e65397c652767f286bc4de193a2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Without the overloads using QLatin1Char in QL1S member functions results
in the QL1Char being converted to QChar and QL1String being converted to
QString.
Change-Id: Ic19545539a207f025a6293f0b2d929de475dc166
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The relevant test is conducted by tst_Moc::warnOnVirtualSignal(),
added in 2009, <quote>
Fix Warning saying that signal cannot be made virtual
The test for virtual signal did not work.
But we cannot make an error right now or it might break existing code
(exemple in task 210879)
</quote> but without removing the #include, so we do still get a
warning about that when building tst_moc.
Pick-to: 6.2 6.1 5.15
Change-Id: I5ea18058f22bbd483d1ff18b0ca7b360f6674ed9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In my prior attempt to handle the last second of 1969, I forgot that
the QTime we're describing is a local time, so whether *it* thinks
we're at the last second of the day is beside the point. Fortunately,
preceding second should get -2 as return if mktime()'s initial -1
actually meant the last second of 1969, so we can test via that, after
a cheap pre-test to save doing this too often (albeit we only even
attempt the check if mktime() returned -1 in any case).
Restructured qt_mktime() in the process to deal with the error case's
early return promptly instead of doing it in an else clause. Also
repackage the calls to mktime to isolate various quirks and simplify
the logic in qt_mktime(). This also prepares for setting tm_isdst as a
hint when we know when we came from, in massageAdjustedDateTime().
Refined one test, added two more test cases. These didn't fail before
this fix, but a judiciously-placed qDebug() in testing revealed that
localMSecsToEpochMSecs() resorted to its fall-back handling - as if
the date-time were outside the time_t range - due to qt_mktime()
failing, for these test-cases (and several others). This fix evades
that fall-back behavior; a judiciously-placed qDebug() shows none of
our test-cases now fail callMkTime().
Change-Id: I11aa5015191dc4a565c28482307f7bc341c207e7
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In setMSecsSinceEpoch(), the CET end-of-time is invalid anyway, so a
whole block (rather than just one sub-test of it) was irrelevant for
the max-qint64 test case (aside from verifying cet *is* invalid).
Split out to a separate test the part of a data-driven test that was
the same for all data rows. Reworked several ill-advised ways to use
QSKIP().
Change-Id: If757d3e722c81fc42a87256125ceef605b6bfb64
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Before Unicode 4.1.0 there was an error in the example code for Hangul
normalization that used <= on the ends of some ranges of values, where
they should have used < tests. This was faithfully copied but the need
for correction has only lately come to light.
Thanks to Ma Lin for pointing this out and providing the fix and
test-cases.
Fixes: QTBUG-71894
Pick-to: 6.2 6.1 5.15
Change-Id: I5c7fec1f9fac1f7a25b2d5e9c3109a90a7ff49e1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QT_NO_INTERNAL_COMPATIBILITY_FUNCTIONS has been removed
Change-Id: Ic77cdae4e5151e3b0fe05cc6dd5b43bffc24768c
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
When the listview setWordWrap is true and ScrollBarPolicy is
ScrollBarAsNeeded, if QStyle::PM_ScrollView_ScrollBarOverlap
returns true, the text displayed an empty line.
Fix this by not reserving the width of the vertical scrollbar
if the flow is TopToBottom and the vertical scrollbar, and QStyle
returns true for PM_ScrollView_ScrollBarOverlap. Amends
aeef92c3c3
Pick-to: 6.1 6.2
Fixes: QTBUG-94248
Done-with: Volker Hilsheimer <volker.hilsheimer@qt.io>
Change-Id: I4d47c7e86bbb86474cb1a99bb26d8b67f0e8a7e2
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
'ld' only capable to resolve circular dependencies by wrapping the
suspected static libraries and objects using --start/end-group
arguments. We want to detect if linker is 'ld' at configure time to
decide how to link the resource objects if finalizers are not enabled.
The qt_config_compile_test function is extended with an extra argument
since it's required to pass custom cmake flags to the ld-related test.
Pick-to: 6.2
Change-Id: I484fcc99e2886952d8b0232f37e4e6a35d072931
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
For auto-tests that were temporarily disabled. Similar to
network-settings.h, header-only stuff.
Fixes: QTBUG-92866
Fixes: QTBUG-92877
Change-Id: I15b5c0b41f0d8bfe59b09c844884ff6d99e6d41a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
When pressing an index in order to start a flick gesture, QAIV sets
the current index. When QScroller changes state to Dragging, then
QAIV restores the current index to what it was before the press, as
the user is clearly scrolling the view. With autoScroll enabled, this
will produce an ugly jump if the old current index is no longer in
the viewport.
To prevent this, disable autoScroll before restoring the currentIndex.
Fixes: QTBUG-64543
Pick-to: 6.2 6.1 5.15
Change-Id: I3e0a18a6a179d80b9d810fce5aa658f0cfff9a29
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
When finding the index we need to scroll to, use the one where both start
and end of the tab rect are outside the currently visible section.
Otherwise we wouldn't scroll when the left-most index ends outside the
visible section.
Add test, which requires that the scroll buttons have object names.
Fixes: QTBUG-70498
Pick-to: 6.2 6.1 5.15
Change-Id: Id153c77dd5fca146612375e0ff39bd1f3e0536b1
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
The error that the actual runtime will encounter, and not the one that compile-time
ifdefs will (potentially) erroneously select.
Change-Id: I8ef4c34bcb8b3e568bc39f8c8ea6bfb7732f9e27
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Do a quick check whether the type name contains an anonymous type. If
so, do not try to use optimized version. The simple check should still
be faster than calling normalizeType unconditionally.
Also only apply the faster version for clang and gcc, instead of all
non-MSVC compilers. Applying it to other compilers would require further
testing to handle anonymous structs.
Moreover, remove space before '(', which is necessary for function
pointers.
Fixes: QTBUG-94213
Change-Id: I795d1964f7a68daa6f9a5f262816d51ee7728788
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Make it return bool since the TriState was really only used signify that
the property was unsupported but there is already a separate way to
check if it's supported. More importantly there is no different set of
actions available to a user if they're in the Unknown or False state.
Because of the change to bool, we also rename the property to have an
'is'-prefix.
Change-Id: Iaaaad5ac31e663c36e00223bf5b0e719f412fc69
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Since 0f1008a593, views record if they
moved the item in the model, and prevent the deletion of the source item
in QAbstractItemView by setting the dropEventMoved private data member.
However, QListView in icon mode is special: it doesn't rearrange the
model, it repositions the icons in the view. While the dropEventMoved
logic was applied to the drag event filter to prevent deletion, the
variable was never set in the filterDropEvent handler. The drop event got
ignored, breaking rearranging of icons.
Fix this by setting the dropEventMoved member in filterDropEvent.
Fixes: QTBUG-94226
Pick-to: 5.15 6.1
Change-Id: I963f5db0f81bcd0d25eef05d9a265be00a5871f6
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
In MultiSelection mode, items are by default toggled on press, which
follows the example of standard Windows controls. However, when dragging
is enabled, then the press might be the beginning of a drag'n'drop
operation, and deselecting the item on press breaks the selection and user
experience.
Don't toggle the selection for presses on an already selected item that
might get dragged; instead, wait for the release event.
Extend the test case slightly to cover the special case. Dragging a
selection in a drag-enabled and MultiSelection item view wasn't possible
before either.
Fixes: QTBUG-59888
Change-Id: Ibd3e95a71ea63dd1e9bc3c8a723eafa9a1c21afa
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: David Skoland <david.skoland@qt.io>
In ExtendedSelection mode, a Ctrl+Press might be both the start of
a selection toggle, or the start of a Ctrl+Drag operation.
If we already toggle on the press, then it's impossible to drag the
existing selection while the Control key is pressed. Ignore Ctrl+Press
events and let the corresponding release event toggle the selection.
Adjust the relevant test cases accordingly. The QItemDelegate test
case used a click+control event incorrectly, such an event doesn't
change the clicked state and should not be eaten, and now it does
change the selection, so fix the test.
Task-number: QTBUG-59888
Change-Id: Ia76126e31c28bc97d3e93e54965bdb1d0b8ac6a4
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Previously, for locales other than the system locale, no entry was
added to the list for the actual locale whose variants - with and
without likely sub-tags - were being appended. In most cases the
standard name will in fact coincide with the variant without likely
sub-tags, so this was unlikely to cause a problem, but it should be
present regardless.
At the same time, turn tst_QLocale::uiLanguages() into a data-driven
test and add another row to its table.
Change-Id: I5cb2d805d78fc3415d82b169caa6154b0f284708
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Previously, the augmentation wasn't being applied to the system
locale, due to a mistaken test claiming the locale's name didn't
resemble the string from which it was constructed. The test dates from
before various fixes to likely sub-tag processing that should make it
redundant now. This makes QLocalePrivate::rawName() also redundant
(and its conversion of QLatin1String to QByteArray relied on '\0'
termination which wasn't actually present in the various codes).
Expanded the test of systemLocale() to also test uiLanguages() turns a
single entry into the list we expect; and add two new test-cases.
(The test uses a mock system locale class, making this independent of
the platform backend.)
Fixes: QTBUG-92234
Pick-to: 6.1 5.15
Change-Id: I0cdf6eae152a42dc377f4ea3e62c282ff4be1764
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Once we're in ~QObject, only methods of QObject are still valid.
Notably, no setter of any derived class is still valid. Thus, to be safe
we must no longer react to binding changes of those properties. To
ensure that this happens for QObjectCompatProperty properties, we
explicitly clear the binding storage.
Fixes a particles3d example crash.
Change-Id: I10d2bfa5e96621ce039d751cffaf3ac41893623e
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
The change in 004e3e0dc2 introduces
Windows junction awareness, though users were still unable to resolve
the junction target. This change adds the ability to solve this.
Fixes: QTBUG-93869
Change-Id: I9f4d4ed87b92e757f7b6d8739e2a61b58c096f63
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Split the code out of QDir::fromNativeSeparator into a separate
reusable function to remove the above-mentioned prefixes. Fixes
and unifies behavior if the prefix was given with slashes instead
of backslashes. Add a couple more test cases.
Fixes: QTBUG-93868
Pick-to: 5.15 6.0 6.1
Change-Id: Ibd94ae283e2fb113f9c2db97475fbc7d89522bbf
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
All repositories define QT_NO_INTERNAL_COMPATIBILITY_FUNCTIONS by now,
and we can remove QtCompatibilityHelpers.cmake altogether.
Change-Id: I4d8104246e96a4514d5651c104607d651d208d95
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Instead of relying on some string comparisons and the current knowledge
of which backend supports DTLS, use the proper API we already have in
place to test if a particular class is supported by the active backend.
Change-Id: I58ca0f7b7fcef68ec375cd64b83e51d4335817da
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Covers all selection modes and various combinations of mouse press, click,
drag, with or without modifiers.
Task-number: QTBUG-59888
Change-Id: Ib8ddc319a89649338d55fbc47e5a7ccfed338af9
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
After d6551fe125 it was no longer possible to start a drag with a
double click (where the first click selects an item, and the second
press+move starts the drag). Resetting the pressedItem variable to block
the emission of the clicked() signal had this unwanted side effect.
Instead, use an explicit boolean to store that the next release event
will be the result of a double click, so that the clicked() signal is not
emitted again (preventing the double-emission was the purpose of change
d6551fe125).
Task-number: QTBUG-77771
Fixes: QTBUG-94087
Pick-to: 6.1 5.15
Change-Id: I082c5169d89eb980dcd7985ef3d302b6ff060fb9
Reviewed-by: Samuel Gaist <samuel.gaist@idiap.ch>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: Olivier BARTHELEMY <perso.olivier.barthelemy@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Test would previously have leaked on failure.
While checking for other uses of delete-after-checks, simplify one
place that tested deleting a calendar widget lead to recreation.
Pick-to: 5.15 6.1
Change-Id: I9524902a02dd42793bcf71113c6799dc925ac4a0
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
Calling create_mapping in setSourceModel as introduced by 8455bfee76
can lead to an early call to filterAcceptsRow, and some existing applications may crash.
It is also an incomplete solution since it was only done for the
toplevel index but not for child indexes.
Instead, go back to creating mappings on demand.
This means coming up with a different fix for QTBUG-87781 (dataChanged
not emitted for indexes that haven't been mapped yet, i.e. not queried
or shown anywhere).
When this happens, we can't know if the index was previously filtered
out or not (for lack of a dataAboutToBeChanged signal...). Creating
the mapping with the new data only gives us the new state of affairs,
there's no reference state to compare to. Therefore, when the mapping
is missing (during dataChanged handling), create it, but skip all the
logic about row insertion/removal, just forward the dataChanged signal
if the row isn't filtered out.
Creating the mapping might require creating first mappings for parents,
recursively, which wasn't done anywhere in QSFPM yet, hence the new
create_mapping_recursive() method.
In addition to all this, the handling of removed items was incorrect,
remove_source_items did nothing if the parent was gone, and then
source_items_removed was trying to adjust indexes in an incorrect list.
If the parent is gone, clear the proxy_to_source list, so there's
nothing to adjust afterwards. This bug actually doesn't happen anymore
in this version of the patch, but the change still seems right and might
prevent repeating a long debugging session in the future.
Thanks to ChunLin Wang for the unittest in this commit.
Done-with: ChunLin Wang
Pick-to: 6.1 6.0 5.15
Change-Id: Id543d0cc98f1a03b5852bda01d2f49b980e06be7
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Handy when one has the data buffer to be read in a QBAV.
This also fixes an issue and compiler warning about passing a
qsizetype data length value as an int, and makes it possible to pass
a qsizetype-size length without going through QByteArray.
Makes the QByteArray overload redundant.
Change-Id: Iba8825cf0fd8003fb2eac5b1d30a61ec91b85ceb
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Both functions now return a result object. Eliminates the need
to pass the errorMessage out-parameter. Adapt auto-tests.
Change-Id: I110b68fedc67b01f76796c44fa55383b2cc03460
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Needed for QML compiler to know property index of C++ type at
compile time without going through QMetaObject::indexOfProperty
Change-Id: I404e71d6071d36812661df17d12b879a8dcbd146
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
The test connects finished to the groups clear method, which in turn
deletes the animation instance. Thus, no member must be accessed after
calling stop, unless we use a (costly) QPointer to guard against
deletion.
Notify earlier that totalCurrentTime changed to avoid the issue.
As a drive-by, modernize the connect in the test.
Fixes: QTBUG-94143
Change-Id: I923101107b7f79115be69a58c8e8d5177a98d48f
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
The JNI interface expects a const JNINativeMethod[] and our wrapper
takes a non-const. Also, this was causing refactoring of exisisting code
with a const JNINativeMethod[] to fail because the call expects a
non-const.
Change-Id: If790c401650cb33fe31f93bafe41aab7714488e9
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
QPropertyChangeHandler is a templated class and it's argument is
a functor. That makes it inherently cumbersome to use the class
in any context where the change handler needs to be stored.
Introduce a QPropertyNotifier class that stores the functor
in a std::function<void()>, and add a QProperty::addNotifier()
method that can be used instead of onValueChanged().
Also make QPropertyNotifier default constructible.
This significantly simplifies the code that needs to be written
and makes it possible to store notifications as class members
without major hassle.
Fixes: QTBUG-92980
Change-Id: Id5b7baec093b9ac0467946cded943d92ad21030b
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Use class WriteIncludesBase and store classes encountered in a
per-module hash (Qt/custom widgets). Write out only the required
classes.
Add --star-import as a fallback should the change cause issues.
Task-number: PYSIDE-1404
Change-Id: Ic50e26758ddd0f2f8aebbce470d32a36fb09a2c4
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Cristian Maureira-Fredes <cristian.maureira-fredes@qt.io>
When rotating M_PI_2 based on x-axis, quaternion to euler conversion
makes NaN for the x-rotation value. This patch fixes this corner case.
Fixes: QTBUG-93600
Pick-to: 6.1 6.0 5.15
Change-Id: Ice321a80ad90dba9cf3ee3a14ec7d3d047c21bd3
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
This patch adds the API, with no supporting backends
Task-number: QTBUG-93848
Change-Id: I50454717f928819e1b990df91872675e842f9987
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Supported on OpenGL (and ES) 3.0+ and everywhere else.
Can also be a render target, targeting a single slice at a time.
Can be mipmapped, cannot be multisample.
Reading back a given slice from a 3D texture is left as a future
exercise, for now it is documented to be not supported.
Upload is going to be limited to one slice in one upload entry,
just like we specify one face or one miplevel for cubemap and
mipmapped textures.
This also involves some welcome hardening of how texture subresources
are described internally: as we no longer can count on a layer index
between 0..5 (as is the case with cubemaps), simply arrays with
MAX_LAYER==6 are no longer sufficient. Switch to sufficiently dynamic
data structures where applicable.
On Vulkan rendering to a slice needs Vulkan 1.1 (and 1.1 enabled on the
VkInstance).
Task-number: QTBUG-89703
Change-Id: Ide6c20124ec9201d94ffc339dd479cd1ece777b0
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
When the PageDown is pressed, the current's rect cannot be
used to match the item, because we don't know the size of
the rect. Move the rect by the height of the viewport, and
then move the rect upwards until it matches the
button <= viewport's bottom of the item
Fixes: QTBUG-92583
Pick-to: 6.1
Change-Id: I210edc0e8b942984f3fc20e7752c6e1315152ea1
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
When changing the style of the combobox, the change will not propagate
to the internal container widget, so the changeEvent handler won't be
called. This is correct (as per QWidget::setStyle documentation).
QComboBoxPrivateContainer asks the combobox style for relevant settings,
such as the frame style, which is then used for sizing and positioning.
If the combobox's and container's settings become inconsistent, then the
combobox popup will not get the correct size and/or position.
Move some of the style-dependent changes into a separate function and
call it when the QComboBox::changeEvent handles the style change so that
both widgets have a consistent set of settings.
Add a test case that verifies that the style is asked for the relevant
setting when the style changes.
Note: QComboBox does a lot of style-dependent setup work in different
places, which is quite messy and complex. Trying to consolidate that
further breaks tests though, so this change is doing the minimum
necessary to fix the reported issue.
Pick-to: 6.1 5.15
Fixes: QTBUG-92488
Change-Id: Ia957d504b2d800add26fc0565be727b5c08a5358
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
This proposal collects all the resource objects to the qt-specific
property of the static libraries. This is done to avoid littering
of other static libraries and put resource object files to the
source part of the linker line when linking the end-point
executable.
The way we link object resource libraries is changed back to the
target_link_libraries approach as we may omit using finalizers
with linkers other than ld. Users may enforce finalizers by calling
the qt6_enable_resource_objects_finalizer_mode function if need.
Refactor tests related to the static resources.
Amends ddaa7150d8
Task-number: QTBUG-93002
Change-Id: I74135e291cd82fb54d1b284b4b4a1e002b1fef98
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Only use plugin finalizer mode if qt_finalize_target is called at the
end of the user project (which we can't really check, the user has to
ensure that) or when qt_finalize_target is automatically defer-called
by CMake 3.19+ (which is done by qt_add_executable).
This removes the previous behavior of using the finalizer mode if
qt_import_plugins is called. Instead the old regular mode is used
if the above preconditions are not met.
The removed behavior had ordering issues if qt_import_plugins was called
before target_link_libraries. The dependency walking would be done
before Qt dependencies were added and thus no plugins would be
linked.
Amends 6fcc272ac9
Task-number: QTBUG-80863
Task-number: QTBUG-92933
Task-number: QTBUG-94030
Change-Id: I78702b653a35596f5581c2f4282b2336f0124e60
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
It's possible that Qt is built with OpenSSL and some other backend.
The ifdefs generated are not mutually exclusive anymore, as it was
in the past. So tests should rely on the actual backend they have
working at runtime.
In the process of fixing the auto-tests, modernized feature #if-ery and
some associated compilation issues.
Fixes: QTBUG-92875
Change-Id: I2be1b924d506fd36dc4e2c6c9158b5023ff87f32
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
This replaces QtAndroidPrivate::runOnAndroidThread{Sync} calls.
This also now allows passing std::function<> that can return values,
and not only an std::function<void()>.
This adds some tests for this calls as well.
Fixes: QTBUG-90501
Change-Id: I138d2aae64be17347f7ff712d8a86edb49ea8350
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Add NEON for RGB32 and RGBA64 writeback, and SSE2 for
RGBA64 writeback.
Change-Id: Id9ee803267a78f5bdff5beaa719e7a59c1dbb9fb
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Useful for some HDR representations and HDR rendering.
Change-Id: If6e8a661faa3d2afdf17b6ed4d8ff5c5b2aeb30e
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
When adding an offset from UTC, arithmetic may overflow. Likewise
when combining a date and time (that have been offset for UTC). Also
check the return from epochMSecsToLocalTime(), as it can fail; and pay
attention to the status stored by setDateTime(), to notice when it
hits an overflow. Fixed some tests that only passed because we
neglected these checks. Extended a test to check we detect overflow in
a couple of cases close to the extremes.
Change-Id: I127a670302f94a07bb9b087b1b9c608b7c08785c
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Øystein Heskestad <oystein.heskestad@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QDateTime has long followed a convention of ignoring what it knows
about time-zone transitions before the epoch. This produces unhelpful
artefacts (such as an ahistorical spring-forward skipping the first
hour of 1970 in Europe/London, which was in permanent DST at the time)
and complicates the code. It documented that DST transitions were
ignored, but in fact ignored all transitions prior to 1970 and simply
assumed that the current time-zone properties (half a century later)
applied to all times before 1970.
This appears to be based on the fact that the MS APIs using time_t all
limit their range to after 1970. Given that we have to resort to
"other means" to deal with times after the end of time_t, when it's
only 32-bit (and after year 3000, on MS systems), we have the means in
place to handle times outside the range supported by the system APIs,
so have no need to mimic this restriction. (Those means are not as
robust as we might want, but they are less bad than assuming that the
present zone properites were always in effect prior to 1970.) On
macOS, the time_t functions only reach back to the start of 1900; it
reaches to the end of its time_t range and Linux covers the whole
range. Given this variety, the range is now auto-detected the first
time it is needed (based on some quick and dirty heuristics).
Various CET-specific tests now need adjustments in tests of times
before the introduction of time-zones (when they are in fact on LMT,
not CET). The systemZone() test of QTimeZone can now restore its
pre-zone test cases. Various comments on tests needed updates.
[ChangeLog][QtCore][QDateTime] Available time-zone information is now
used to its full extent, where previously QDateTime used LocalTime's
current standard time for all dates before 1970. Where we have
time-zone information, it is considered reliable, so we use it. This
changes the "best efforts" used for times outside the range supported
by the system APIs, in most cases giving less misleading results.
Fixes: QTBUG-80421
Change-Id: I7b1df7622dd9be244b0238ed9c08845fb5b32215
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Output object name and class in QCOMPARE(). This should help
to debug flaky QWidget tests that for example check on focusWidget().
[ChangeLog][QtTestLib] QCOMPARE() now reports QObject * values by class and objectName().
Task-number: QTBUG-64446
Change-Id: Ife04e89bba04fc78d077c8f0f07af17a17c9cf8c
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
QNetworkInterface reports as up and running an interface,
seemingly having a valid IP address which actually has
not IPv4/IPv6 configured at all (this is what getifaddrs
gives us).
Change-Id: I5808d5dfdbc54004a1e02b791917e9b2ba62905f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
There were two problems here: first, qHash(std::pair) must be declared
before qHashMulti that might call back to qHash(std::pair) (i.e., a pair
with one element that is also a pair). But moving the declaration above
causes the second problem: the noexcept expression can't refer to qHash
functions that aren't declared yet. So we forward-declare a constexpr
function for that result, but implement it far below.
Fixes: QTBUG-92910
Change-Id: Ia8e48103a54446509e3bfffd16767ed2e29b026c
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Commit 37e0953613 added a to-do, but we
can actually change the type, since we've documented since Qt 5.10 that
setting a non-zero value (aside from -1) with qSetGlobalQHashSeed was
not allowed. Storing a value to be reset later is simply not supported.
Change-Id: Id2983978ad544ff79911fffd1671f7b5de284bab
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
It is now possible to call qt_add_big_resources with multiple .qrc files
that are in different directories but have the same file name.
Before, this resulted in target name conflict errors.
Pick-to: 6.1
Fixes: QTBUG-93230
Change-Id: Ic3d9d6974a107163aeb7ada0b6f1fa113fb18de0
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
For some reason the test case was (no longer?) detected by moc
Change-Id: I20f20de76479f7d9db43fae985cefe0869a7e203
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The existing conversions weren't handling gamma correctly and used
an ad-hoc definition of gray instead of based on true luminance.
[ChangeLog][QtGui] RGB conversions to grayscale formats are now
gamma-corrected and produce color-space luminance values
Change-Id: I88ab870c8f5e502ddb053e6a14a75102239a26f2
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The check is over-arching. is_move_constructible is sufficient;
we don't have to support "ridiculous" types that are copiable but
have deleted move operations (such types are fundamentally broken).
This is in line with the Move* (legacy) named requirements.
Change-Id: Idc7116b39013501b9be39628a4e7afd35fe15530
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
To see if a certain method is supported. To be used in an upcoming patch.
Change-Id: I1a5c2f655585331820701bb54f6991b4aba38273
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The qt6_add_big_resources command was dysfunctional whenever AUTOMOC was
enabled for a target.
We passed multiple object files with the --temp argument of the second
pass of rcc. We must pass exactly one. The spurious object file was the
one that's created by AUTOMOC. Turn off all of AUTOGEN for the C++
source file that's generated in the first pass of rcc.
Enable AUTOMOC for tests/auto/cmake/test_add_big_resource to cover this
situation.
Pick-to: 6.1
Fixes: QTBUG-85051
Change-Id: I6d2ce4953297db0751913456db0a4a126fe33f14
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
It's very likely that we'll have to run certain finalizer code for
targets other than executables.
Rename qt_finalize_executable to _qt_internal_finalize_executable so
it's internal.
Introduce a new function qt_finalize_target which will call the above
internal one when the target is an executable.
This should future proof the API so we have a hook to call code for
any user CMake project that intends to use Qt CMake API.
Change-Id: I03f6f4dcba22461351c247a20241ca7de1a59c1d
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
In QFileSystemModel, in some cases the hostname in a UNC path is
converted to lower case and stored in the root node's visibleChildren.
When QFileSystemModel sets the UNC path as the root path, it tries to
get the row number for the host, but it didn't convert the hostname to
lower case before getting the row number, which resulted in the host
not found in the root node's visible children. As a result, it returns
-1, an invalid row number. Change the behavior to find the node for the
host using the host name case-insensitive and then get the row number.
Fixes: QTBUG-71701
Pick-to: 5.15 6.0 6.1
Change-Id: Ib95c7b6d2bc22fd82f2789b7004b6fc82dfcb13b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
There was previously no way to compare QByteArrayView to with another
QByteArrayView case-insensitively without allocating memory.
[ChangeLog][QtCore][QByteArrayView] Added compare(), enabling case
sensitive and insensitive comparison with other QByteArrayViews.
Change-Id: I7582cc414563ddbde26da35a568421edcc649f93
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Allow linking all plugin initializer object libraries directly
into the final target (executable or shared library).
The finalizer mode is triggered when the project adds a call
to qt_import_plugins, as well when the project has an explicit
call to qt_finalize_executable or when it is defer called by
CMake 3.19+.
Otherwise the old non-finalizer mode is used, where each plugin
initializer object library is propagated via the usage
requirements of its associated module.
A user can explicitly opt in or out of the new mode by calling
qt_enable_import_plugins_finalizer_mode(target TRUE/FALSE)
The implementation, at configure time, recursively collects all
dependencies of the target to extract a list of used Qt modules.
From each module we extract its list of associated plugins and
their genex conditions. These genexes are used to conditionally
link the plugins and the initializers.
Renamed QT_PLUGINS property to _qt_plugins, so we can safely query the
property even on INTERFACE libraries with lower CMake versions.
QT_PLUGINS is kept for backwards compatibility with projects already
using it, but should be removed in Qt 7.
The upside of the finalizer mode is that it avoids creating link
cycles (e.g. Gui -> SvgPlugin -> Gui case) which causes CMake to
duplicate the library on the link line, slowing down link time as well
as possibly breaking link order dependencies.
The downside is that finalizer mode can't cope with generator
expressions at the moment. So if a Qt module target is wrapped in a
generator expression, it's plugins will not be detected and thus
linked.
Task-number: QTBUG-80863
Task-number: QTBUG-92933
Change-Id: Ic40c8ae5807a154ed18fcac18b25f00864c8f143
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
