This fixes the recently added QEXPECT_FAIL about glob-deleteall
in a local directory (with a binary cache). Before adding a glob match
we ask the more-local (higher-precedence) directories if they have
a glob-deleteall for that mimetype, and skip it then. This "asking"
is a virtual method, implemented for both XML and binary providers.
Change-Id: I6e4baf0120749f3331fd2d9254bea750a322b72d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The old name Asia/Calcutta is being phased out. We can't assign
QTzTZP, so select between new name and old using a reference variable.
In the process, fix a QCOMPARE() against bool to a QVERIFY().
Pick-to: 6.6 6.5
Change-Id: I7cd8a813f8a88c8ae4ba07213f04f4ad0860cec0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
A QMimeTypePrivate used to belong to a single provider, which would
provide the complete data for it.
But since the redesign in commit 7a5644d648, each provider
represents is a single mime directory, and the merging happens at the
QMimeDatabase level. So we need a QMimeType[Private] to be just a name
(a "request" for information about this mimetype) and the information
for that mimetype is retrieved on demand by querying the providers
and either stopping at the first one (e.g. for icons) or merging
the data from all of them (e.g. for glob patterns).
The XML provider was using QMimeTypePrivate as data storage,
give it its own struct QMimeTypeXMLData for that purpose instead.
Task-number: QTBUG-116905
Change-Id: Ia0e0d94aa899720dc0b908f40c25317473005af4
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When running tst_qmimedatabase with the full shared-mime-info test suite
(which unfortunately requires local setup so this is easy to overlook),
we need *.webm to still be associated with video/webm.
So to test glob-deleteall, do that in installNewLocalMimeType(), with
other similar tests.
This however unearthed the following bug: the handling of glob-deleteall
is only correct when the local dir has no binary cache. It's broken
when using a binary cache. Added a QEXPECT_FAIL for now because this is
going to be fixed as part of a major redesign, coming up.
I also found out that neither xdgmime nor gio do this correctly...
Change-Id: Ib075fcdb792f60a859f23db8c2d7e1c6524f9050
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
For historic reasons, the test had a single list of override xml files
to copy first into the global dir, and then into the local dir.
But glob-deleteall only makes sense in the local dir (as per the MIME
spec). Having two definitions for the same mimetype in the same dir
is undefined behavior, so the test was working by chance only, and
my upcoming refactoring/fixes caught that.
Change-Id: I4717683b4b3f9ba69f1fd815669460789700e877
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Since compilers don't provide such macros, do it ourselves.
In order to test these macros, add ad-hoc specializations of
QTest::toString() for qint128 and quint128 locally to the test. Turns
out it's not too hard to write them, so we might move them to a public
header, yet.
Pick-to: 6.6
Change-Id: I1483f3af2ccec6038e1c780649f9ffe413bb59ef
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Check that QIntegerForSize<16> and std::numeric_limits<quint128> work
and that q(u)int128 are available in C mode, too.
Pick-to: 6.6
Change-Id: I44af8282399c78f6e74a8268af53bad64407ca34
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There were two problems:
- On platforms where QFLOAT16_IS_NATIVE == true, a qHash(qfloat16{})
call has become ambiguous between the three FP qHash() overloads
(float, double, long double), where it was unambiguously calling the
float one in Qt 6.4. This SiC was caused by the replacement of
operator float() by operator __fp16() in
99c7f0419e, which is in Qt 6.5.
- On platforms where QFLOAT16_IS_NATIVE == false, qHash(qfloat16{})
would produce a different value from qHash(float{}), and therefore
Qt 6.4, when the seed was != 0, because the former would go via the
one-arg-to-two-arg qHash adapter while the latter one would
not. Since participating functions are inline, this causes old and
new code to produce different hash values for the same qfloat16,
leading to a BiC possibly corrupting QHash etc.
Fix both by adding an explicit qHash(qfloat16). This function is
inline, so it doesn't add a new symbol to 6.5.x.
[ChangeLog][QtCore] Fixed qHash(qfloat16) which was broken from 6.5.0
to 6.5.3, inclusive. If you compiled against one of the affected Qt
versions, you need to recompile against either Qt 6.4 or earlier or
6.5.4 or later, because the problematic code is inline.
Pick-to: 6.6 6.5
Fixes: QTBUG-116064
Fixes: QTBUG-116076
Change-Id: Id02bc29a6c3ec463352f4bef314c040369081e9b
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QWindow::setMask() is not guaranteed to turn the masked out areas
transparent, and it's up to the client to ensure this during
painting.
Pick-to: 6.6
Change-Id: I1155b3ad095152a993532f2290cacb670e20daa7
Reviewed-by: Liang Qi <liang.qi@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The event might come from a partial expose, but we want to
fill the entire bounds of the window, as we're filling with
a gradient.
Pick-to: 6.6
Change-Id: I66cedb160fb0ed06935c06ba2fe5dec9ed468833
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Liang Qi <liang.qi@qt.io>
Visualize what's going on with tst_QTextCursor::insertMarkdown() at least.
But this could be expanded for other purposes.
It's also interesting to test drag-and-drop with this. And you can save
the result to any supported text format.
Change-Id: I363c32ff9b1bd7c53c562b08c58de95a69be0aa9
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
0b421fa58b9a73d657bf17834788fd1175c4767e ensured a correct focus chain,
when buttons in a QDialogButtonBox were hidden.
The implementation did not check, if a hidden button was added via
setStandardButtons(). In consequence, it was removed from the
standardButtonHash and never added again.
QDialogButtonBox::button() returned nullptr for a standard button,
once it had been hidden. That introduced a regression.
This follow-up patch makes sure, a standard button is not removed
from standardButtonHash, when hidden. By no longer removing it from
standardButtonHash, it makes showQDialogButtonBox::button() always
return the pointer to the standard button, even if it is hidden.
The function handleButtonDestroyed() used the argument
QDialogButtonBoxPrivate::RemoveRule::KeepConnections, in order to leave
signal/slot connections untouched. It expected the the destroyed button
to be removed from standardButtonHash. In order to retain that
functionality, the enum class RemoveRule is renamed to RemoveReason,
and one value was added. QDialogButtonBoxPrivate now handles all
necessary cases of removing a button:
ManualRemove (previously Disconnect):
- remove button from roles
- remove button from standardButtonHash
- disconnect all signals
LeaveEvent (previously KeepConnections):
- remove button from roles
- do not remove button form standardButtonHash
- do not disconnect signals
Destroyed (new):
- remove button from roles
- remove button from standardButtonHash
- do not disconnect signals (QObject will do that)
An autotest is added to tst_QDialogButtonBox.
Task-number: QTBUG-114377
Pick-to: 6.6 6.5
Change-Id: Ib28625d44fa89c3d06f181f64875c2e456cebbfa
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
By first checking if the list has any matches before potentially making
it detach.
Change-Id: I7a42c2910ef6efc45033e562573414a3a9ef972e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
91dcc76fc1 might have fixed the underlying issue, so we no longer
need the XFAIL codepath at all.
Fixes: QTBUG-114720
Change-Id: I67ccbed67a0536b679c50c26eb0b3e51c93dceeb
Pick-to: 6.6
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
rowsAboutToBeMoved() and rowsMoved() signals aren't emitted for
QSortFilterProxyModel, which meant the two connections in the
ObservingObject's constructor didn't trigger the slots, which let the
test pass (the store/checkPersistentFailureCount stayed at 0).
- Instead connect to layoutAboutToBeChanged() and layoutChanged()
respectively, these two are emitted for QSFPM
- Use PMF syntax
- Verify m_persistent{Proxy,Source}Indexes aren't empty
Change-Id: I8b83989de02c2bfb22bde9b230cb5b68814f74b6
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Because the local `seed` variable shadowed the member one, this test
was run for each QFETCH_GLOBAL with the same data and seed. That
doesn't make sense, so make the test use the member variable `seed`,
as all other tests already do.
Since zero is one of the seeds coming from QFETCH_GLOBAL, drop the
seedless calls to qHash(), too.
Amends 64bfc927b0.
Pick-to: 6.6 6.5 6.2
Change-Id: I1e22ec0b38341264bcf2d5c26146cbbcab6e0749
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The old code only tested with seed = 0 and seed = 1045982819, the
latter being a "random number", which, however, fits into
32-bits. Since Qt 6.0 increased the seed from uint to size_t, amend
the test to actually test a seed value with some of the upper half of
bits set, too, also in 64-bit mode.
While we're at it, also test with each seed's bits flipped for extra
coverage.
Remove a static assertion that prevented testing seeds with the MSB
set.
Pick-to: 6.6 6.5 6.2
Change-Id: I5ed6ffb5cabaaead0eb9c01f994d15dcbc622509
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
This commit reverts 2d77051f9d.
When requesting an allocation of size 0, we will actually get
a nullptr.
qarraydata.cpp:
~~~
if (capacity == 0) {
*dptr = nullptr;
return nullptr;
}
This will let the Q_CHECK_PTR trigger falsely. Such an occurrence was
initially detected during the cmake_automoc_parser build-step.
Found-by: Marc Mutz <marc.mutz@qt.io>
Task-number: QTBUG-106196
Pick-to: 6.6
Change-Id: Icb68c5dd518c9623119a61d5c4fdcff43dc4ac5d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QStaticLatin1StringMatcher is a static templated Latin-1 Boyer-Moore
string matcher which can be case sensitive or not. It should be used
when the needle is known at compile time so there is no run-time
overhead when generating the skip table.
The convenience functions qMakeStaticCaseSensitiveLatin1StringMatcher
and qMakeStaticCaseInsensitiveLatin1StringMatcher should be used to
construct the matcher objects.
Green Hills Optimizing Compilers are currently not supported.
[ChangeLog][QtCore] Added QStaticLatin1StringMatcher, which can be used
to create a static constexpr string matcher for Latin-1 content.
Task-number: QTBUG-100236
Change-Id: I8b8eed1e88e152f29cbf8d36d83e410fafc5ca2c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The latest XDG spec (0.8) defines XDG_STATE_HOME that does not exist
in QStandardPaths::StandardLocation.
Some Linux distributions clean XDG_CACHE_HOME on restart which makes
XDG_STATE_HOME useful as a path for saving application state.
This commit adds StateLocation and GenericStateLocation to serve as a
StandardLocation for XDG_STATE_HOME for all platforms.
This commit also updates docs and tests to fit the new changes.
[ChangeLog][QStandardPaths] Added StateLocation &
GenericStateLocation to StandardLocation
Change-Id: I470602466c37f085062cc64d15ea243711728fa5
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QWidget::setVisible is virtual, and called via hide() by both the
QDialog and the QWidget destructor. A dialog that becomes invisible
when getting destroyed will at most execute the QDialog override.
Subclassing QDialog and overriding setVisible() to update the state
of the native platform dialog will not work, unless we explicitly
call hide() in the respective subclass's destructor.
Since e0bb9e81ab, QDialogPrivate::setVisible is
also virtual, and gets called by QDialog::setVisible. So the clean
solution is to move the implementation of the native dialog status
update into an override of QDialogPrivate::setVisible.
Add test that verifies that the transient parent of the dialog
becomes inactive when the (native) dialog shows (and skip if that
fails), and then becomes active again when the (native) dialog is
closed through the destructor of the Q*Dialog class. The test of
QFileDialog has to be skipped on Android for the same reason as the
widgetlessNativeDialog.
Fixes: QTBUG-116277
Pick-to: 6.6 6.5
Change-Id: Ie3f93980d8653b8d933bf70aac3ef90de606f0ef
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
This allows us to control whether to run this particular test in ASan
mode or with specific loggers or not. Adding the expected log output for
tst_silent for other loggers is left as an exercise to the reader.
Change-Id: Ifa1111900d6945ea8e05fffd177f1548c8c8e714
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
This reverts commit f20adcde30. The
implementation had the right idea, but this is not expected to work
reliably in C++. It's jumping out of several frames without cleaning
them out properly and our ASan-based memory leak-checker has started
complaining (the next commit will move this test elsewhere).
==19313==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 258 byte(s) in 1 object(s) allocated from:
#0 0x7ffa505c8e48 in __interceptor_malloc (/usr/lib64/libasan.so.5+0x109e48)
#1 0x7ffa4f2d7ff9 (/home/qt/work/install/lib/libQt6Core.so.6+0x896ff9)
#2 0x7ffa4f2d834d in QArrayData::allocate(QArrayData**, long long, long long, long long, QArrayData::AllocationOption) (/home/qt/work/install/lib/libQt6Core.so.6+0x89734d)
#3 0x7ffa4f23b700 (/home/qt/work/install/lib/libQt6Core.so.6+0x7fa700)
#4 0x7ffa4f1f6cc8 in QString::reallocData(long long, QArrayData::AllocationOption) (/home/qt/work/install/lib/libQt6Core.so.6+0x7b5cc8)
#5 0x7ffa4f1f68a7 in QString::resize(long long) (/home/qt/work/install/lib/libQt6Core.so.6+0x7b58a7)
#6 0x7ffa4f2092ff (/home/qt/work/install/lib/libQt6Core.so.6+0x7c82ff)
#7 0x7ffa4f209e09 in QString::vasprintf(char const*, __va_list_tag*) (/home/qt/work/install/lib/libQt6Core.so.6+0x7c8e09)
#8 0x7ffa4ed0d83d (/home/qt/work/install/lib/libQt6Core.so.6+0x2cc83d)
#9 0x7ffa4ed114a9 in QMessageLogger::fatal(char const*, ...) const (/home/qt/work/install/lib/libQt6Core.so.6+0x2d04a9)
#10 0x5641d2604c40 in tst_Silent::messages() /home/qt/work/qt/qtbase/tests/auto/testlib/selftests/silent/tst_silent.cpp:77
#11 0x5641d26050fb in tst_Silent::qt_static_metacall(QObject*, QMetaObject::Call, int, void**) tests/auto/testlib/selftests/silent/silent_autogen/include/tst_silent.moc:118
The restoration of the signal handler (which QtTest now has) is also
wrong: this needed to use sigaction() instead.
Change-Id: Ifa1111900d6945ea8e05fffd177f14fbc09a1d7d
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Nothing in those files uses QPair; and a local build finished fine without them.
Task-number: QTBUG-115841
Change-Id: I669cfecaa9129bce6b31e464826287f138b159db
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In order to test the impact of migration to QASV.
Task-number: QTBUG-101707
Pick-to: 6.6 6.5 6.2
Change-Id: I17f84ca98fc87d89bb4cd6ad98c8a12aecd315ee
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Add a change-of-zone manual test, based on a bug-report by Felix
Kälberer. This failed: debugging revealed that MS's mktime() doesn't
pick up a change to system zone.
Fixes: QTBUG-83881
Change-Id: I9b86398ef870627a059e269f85a0f5d9d9de284b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The test case is an incomplete version of the test that will be added to
verify the fix for the referenced bug report. The test crashes already
when showing the dialog without this fix.
Task-number: QTBUG-116277
Pick-to: 6.6 6.5 6.2
Change-Id: I969a723157f6453b78bafae5cb24a6b37b1eea50
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
... to replace the old stateChanged(int) one, which a) had the wrong
name and b) the wrong argument type.
Mark the old one as \obsolete, so new users don't see it
anymore. Prepare for deprecation in the test, but don't actually
deprecate, yet (this author does not know how to deprecate signals).
Found in API-review.
Amends 37b47ebf94.
As a drive-by, replace explicit qWait() calls with QTRY_COMPARE().
[ChangeLog][QtWidgets][QCheckBox] Added new
checkStateChanged(Qt::CheckState) signal, obsoleting
stateChanged(int).
Fixes: QTBUG-104688
Change-Id: I01791fd003b752c47d99bea65151202be9175c21
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The palette set by windows vista style during polish doesn't allow style-
sheet style to override it.
This patch reset resolve mask for the palette set by windows vista style
and thus it can be overridden.
Fixes: QTBUG-115511
Pick-to: 6.6 6.5
Change-Id: Ifcaf441f806cfa0273599b3dce83fdfaec3f5a66
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
There was an extra `q` before dnslookup.
Found while trying to build tst_qdnslookup, the target wasn't seen by
CMake/Ninja.
Pick-to: 6.6
Change-Id: Id594aab30dc9081fc269541561e0f2db5e615657
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
See code review 496440 on Gerrit for the details.
Change-Id: Ibd32a44cf7e2e07f36687cc2f0eeaf3008f64e73
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
By saying what's special about some of them
Pick-to: 6.6 6.5
Change-Id: I17bf2e12a27bf55f621020ddf3819ee9e606847d
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Instead of adding it after the block size was calculated. This makes no
difference for non-growing (exact) blocks. For growing blocks, this
means we take that extra element into account before rounding to the
next power of two, instead of after. That results in a change of the
thresholds of when a block grows and also what capacity it will
contain.
For example, for a QString growing to 22-25 elements:
Request | Previously | Now |
elements | bytes | malloc()ed | capacity() | malloc()ed | capacity() |
22 | 44 | 66 | 24 | 64 | 23 |
23 | 46 | 66 | 24 | 64 | 23 |
24 | 48 | 66 | 24 | 128 | 55 |
25 | 50 | 130 | 56 | 128 | 55 |
To avoid wasting elementSize - 2 bytes in this footer, we only include
this footer if elementSize <= 2. Thus, for a QList<int> growing to 11-13
elements:
Request | Previously | Now |
elements | bytes | malloc()ed | capacity() | malloc()ed | capacity() |
11 | 44 | 66 | 12 | 64 | 12 |
12 | 48 | 66 | 12 | 128 | 28 |
13 | 52 | 130 | 28 | 128 | 28 |
In both cases, we now only allocate powers of two while growing, which
may be beneficial to some allocators.
Pick-to: 6.6
Change-Id: Ifa1111900d6945ea8e05fffd177dcb96e251d0a1
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
The SizeShift was not taken into account when constructing QASV from
QL1SV. This is not an issue in normal Qt builds, because SizeShift == 0
there.
But in bootstrapped case (and in future Qt 7) SizeShift changes to 2,
and the bug becomes visible.
The added test-cases do not really reveal the issue, because we do
not run tests in bootstrapped builds, but at least they will help
to prevent the issues in Qt 7.
Pick-to: 6.6 6.5 6.2
Change-Id: I337b37b5230323a5357f48fd1c9bf799ca507d52
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The pre-existing test for QLocalServer was testing only a minor subset
of cases, so replace it with the call to
QTestPrivate::testReadWritePropertyBasics.
The test for QLocalSocket's bindable properties was missing, so add
it.
The new tests didn't reveal any problems.
Task-number: QTBUG-116346
Pick-to: 6.6 6.5
Change-Id: I695bb050d39eeae9ffb84c097c36601a4ca89af6
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
The pre-existing tests were not using the QTestPrivate helpers, so
extend them with the call to QTestPrivate::testReadWritePropertyBasics.
The updated test didn't reveal any problems with binding loops, so no
other action is required for now.
Task-number: QTBUG-116346
Pick-to: 6.6 6.5
Change-Id: I51a17974a7f5bec3c969fcb55b6f28e3e9218eb5
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The pre-existing tests were not using the QTestPrivate helpers, so
extend them with the call to QTestPrivate::testReadWritePropertyBasics.
The updated test didn't reveal any problems with binding loops, so no
other action is required for now.
Task-number: QTBUG-116346
Pick-to: 6.6 6.5
Change-Id: I360614a40fe2bacb796051607ed67e7e666b4f22
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
The bindable property tests were not using the QTestPrivate helpers, so
add a new test which uses them.
The new tests revealed a binding loop for the interval property.
Fix it in a usual way by explicitly removing the binding and using
{set}ValueBypassingBindings() in the setter.
Task-number: QTBUG-116346
Pick-to: 6.6 6.5
Change-Id: If94f57938da449a68e3527aead5ebd55ba410adb
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
The old tests were not using the test methods from QTestPrivate, so
add another test case.
Task-number: QTBUG-116346
Pick-to: 6.6 6.5
Change-Id: I291ede26461e79a615630f1decad2ad7549b4dd8
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
... by using valueBypassingBindings() when accessing the properties
from the setters.
This commit is mostly trivial.
Had to change the template parameters in the unit-test, because the
updated QTestPrivate::testReadWritePropertyBasics() creates an instance
of the TestedClass, and QAbstractProxyModel cannot be instantiated,
since it has pure virtual methods.
Task-number: QTBUG-116346
Pick-to: 6.6 6.5
Change-Id: I0cae29263ea9bb92c9de06891b0ba8633fb9fd72
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
The bindable property tests should use the helper functions from
QTestPrivate.
Task-number: QTBUG-116346
Pick-to: 6.6 6.5
Change-Id: Ie1a61ab80e6f737eac02246214c2c93129a1cf94
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Extend the unit-tests for bindable properties and fix the discovered
binding loop by using {set}ValueBypassingBindings() in the setter of
the duration property.
The code refactoring does not modify the setter logic, because
previously the binding was anyway implicitly removed when calling the
assignment operator. The updated code just does it explicitly.
Task-number: QTBUG-116346
Pick-to: 6.6 6.5
Change-Id: I0f339d182efb60500ee7f12e407f200d739da312
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
libstdc++'s std::filesystem::path implementation incorrectly assumes
that any 8-bit char input is UTF-8, when it patently isn't on Windows.
Reported at https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111244
Pick-to: 6.5 6.6
Fixes: QTBUG-116609
Change-Id: I2b24e1d3cad44897906efffd17803f2862935c9b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
It's very hard to debug a macro.
Pick-to: 6.5 6.6
Change-Id: I2b24e1d3cad44897906efffd17803b8eac9bd844
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Accessibility implementations rely on correct information about the
model dimensions when operating on item views. An item view that has a
root index set needs to report it's size based on the root index, rather
than for the view's model directly.
Pass the rootIndex to all calls to QAbstractItemModel::column/rowCount.
Refactor the code to avoid excessive dereferencing of a QPointer, apply
const and fix/improve coding style in touched lines.
Emit a ModelReset notification when the root index changes, or (in the
case of QListView) when the model column changes.
Split long Q_ASSERTs into multiple lines to be able to better trace the
exact reason for an assertion, and replace the assert with an early
return of nil when it's plausible that a cached cell is no longer part
of the view (i.e. because the root index changed).
Add a test case that verifies that changing the root index changes the
dimension of the view as reported through the accessibility interface.
Pick-to: 6.6 6.5
Fixes: QTBUG-114423
Change-Id: I7897b79b2e1d10c789cc866b7f5c5dabdabe6770
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
Installing a second mimetype with *.txt as glob had a different
effect depending on whether it was installed into the same prefix
or a different prefix as the one where text/plain is installed.
Pick-to: 6.6
Change-Id: I7f54b8efe22f620eb57257745c48fe5402c87626
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QTextMarkdownImporter::import() took a QTD* as if it would be ok to
reuse one instance of QTextMarkdownImporter for repeated importing into
different documents; but in practice, we never do that: in fact it's
usually a short-lived, stack-allocated object, as in
QTextMarkdownImporter(&doc, QTMI::DialectGitHub).import(input);
So it's less clumsy internally to require the document be provided to
the constructor: that way a QTextCursor can be constructed immediately
too, as part of the importer object rather than separately on the heap.
This is private API, unused outside qtbase.
Change-Id: I8041ceb33cb7e7608df55dc5a963292c585afb90
Reviewed-by: Oliver Eftevaag <oliver.eftevaag@qt.io>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
The 6.5 versions of the overload not taking a context/receiver object
were constrained by requiring a functor to be free function or lambda.
207aae5560 removed that constraint, which
might be source incomaptible if wrapper functions in user code forward
the constraint using Expression SFINAE. Those wrappers would no longer
be removed from the overload set based on the same criteria as the
function they wrap.
We can't constrain the new functions based on the same predicate as
before, as after the simplification we have only one overload with, and
one without context object. But we can still remove overloads for
incompatible functors.
Add the respective scenario to the QPermission test as a compile-time
test.
Found during 6.6 header review.
Pick-to: 6.6
Change-Id: Id21391b4a6b78a29de2f8fa04374f4262e5fafa7
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
...to set m_o to nullptr and prevent an existing QSignalBlocker from
touching the QObject, which it was created for.
Add documentation and implement an autotest.
Change-Id: Ic18e80af5a57df1928f9d36aa0ab7ad79b6525fd
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Private libraries were linked conditional to QT_FEATURE_private_tests in
the CMake file. qstatictext_p.h was included conditional to developer
build. A developer build fails, with test enabled and private tests
disabled.
=> Change the CMake condition to QT_FEATURE_developer_build, to resolve
mismatch between CMake and cpp file.
Pick-to: 6.6 6.5 6.2
Change-Id: I79213e7d3c38851b8b80cb8ab248d7bff750c227
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
QLabel *l is declared uninitialized, assigned in a for loop. The last
object is deleted for testing purposes.
This leads to a false compiler warning about deleting a potentially
unintialized pointer.
=> initialize with nullptr to silence the warning.
Pick-to: 6.5 6.6
Change-Id: I1422b04fc1fdbfc7248de577884aabfb539f3f4b
Reviewed-by: Doris Verria <doris.verria@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Private libraries and WrapOpenSSL were linked conditional to
QT_FEATURE_private_tests in the CMake file.
qsslkey_p.h and open ssl symbols were included conditional to developer
build. A developer build fails, with test enabled and private tests
disabled.
=> Change the CMake condition to QT_FEATURE_developer_build, to resolve
mismatch between CMake and cpp file.
Pick-to: 6.6 6.5
Change-Id: I3ac93b02701e467a0b548c35d441d35a45c4568b
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
QItemSelectionModelPrivate::initModel() uses string based connections,
to connect/disconnet its QAbstractItemModel.
The QObject::destroyed signal is connected to modelDestroyed(), which
does not disconnect other signals.
QQuickTableView's selection model binds to its QAbstractItemModel.
The binding also reacts to QObject::destroyed
Eventually, QItemSelectionModel::setModel(nullptr) is called.
At this point, only a QOBject is left from the QAbstractItemModel.
That leads to warnings about disconnecting string based signals, which
belong to QAbstractItemModel.
This patch changes the connect syntax to the QObjectPrivate::connect
API. Instead of keeping a list of string based connections around, the
connections themselves are kept in a list member. Disconnecting happens
based on that list.
Connections are also disconnected in
QAbstractItemModelPrivate::modelDestroyed.
An auto test is added in tst_QItemSelectionModel.
Fixes: QTBUG-116056
Pick-to: 6.6 6.5 6.2
Change-Id: I57e5c0f0a574f154eb312a282003774dd0613dd6
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
This functionality was lost when we switched to Catch2.
Change-Id: I2b24e1d3cad44897906efffd177fb4cff641546f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Amends 118f2210c6. That commit added
#ifdef Q_OS_UNIX
if (test == "assert"
|| test == "crashes"
|| test == "failfetchtype"
|| test == "faildatatype")
return; // Outputs "Received signal 6 (SIGABRT)"
#endif
Which duplicated 4 out of the 5 tests in the block:
#ifdef Q_OS_LINUX
// QEMU outputs to stderr about uncaught signals
if (QTestPrivate::isRunningArmOnX86() &&
(test == "assert"
|| test == "crashes"
|| test == "faildatatype"
|| test == "failfetchtype"
|| test == "silent"
))
return;
#endif
But as Linux is Unix, we never got to that second block for those 4
tests.
Pick-to: 6.6
Change-Id: I2b24e1d3cad44897906efffd177fb4b5507d190a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
QString::fromJsString -> QString::fromEcmaString()
QString::toJsString() -> QString::toEcmaString()
For API naming compatibility with QByteArray::fromEcmaUin8Array()
Pick-to: 6.6
Change-Id: If6e2121e31e630d6728ed24e41d14b763f395aaa
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
QWidget already handles this, but it might be useful for non-Widget
object hierarchies as well, such as in Qt Quick.
The flag is opt in, and as QWidget already handles these events by
itself (without checking any flags), we assert that we don't end up
in this code path, instead of enabling it for QWidget. The latter
would mean refactoring the QWidget code, with possible regressions.
Docs and header comments have been updated to reflect that this
event is not widget specific. (This is an issue with other events
as well, that are documented to say "widget", since they came
from a time when there was only QWidget, but nowadays apply to
e.g. QWindow as well. That's something for another fix though).
Change-Id: Ib71962131d6011c17dcce8c01bd8adcdaa58d798
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
OpenSSL 3.1.2 can be configured with no-deprecated option, in this
case test fails to build.
Pick-to: 6.6 6.5 6.2
Change-Id: Icaf457f55fb001b632922856dbe4bbb5bdba220e
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
It's currently conditional on two features used in QNetworkInterface.
The fact that QNetworkInterface misses those two features is probably
correlated to the failures observed on QNX, but is not the cause, so
this code was actually wrong and was possibly disabling the execution of
similar content on other OSes.
Therefore, this commit removes them and changes the conditional to
exclude the OS that is failing (QNX).
I find this situation unacceptable. IPv6 support is mandatory for any
application after 2011-01-31, the date when IANA delegated its last IP
block, and definitely after 2019-11-25, when RIPE NCC ran completely
out. But since there's no SDK available for it, I'll grudgingly accept a
grandfathered exception because there's nothing I can do about it (I
tried to fix it; look at the change history of this patch set). I will
block any new OSes in that situation, though.
Task-number: QTBUG-116503
Change-Id: Ifa1111900d6945ea8e05fffd177ed6979c3e5916
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Let's do it in initTestCase(), where we're already searching for IPv6
addresses, instead of storing the information in a local static.
Pick-to: 6.6
Change-Id: Ifa1111900d6945ea8e05fffd177efb6a055aaa58
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This was disabled in d0d1d74033, I guess
accidentally, by a too-wide conditional. The change the same commit
applied to QtNetworkSettings didn't make the same mistake.
I am also opportunistically updating the conditional to QT_CONFIG (I
missed this in 9d4579c1cd) and adding the
Linux-specific check, as the AF_NETLINK implementation does not rely on
getifaddrs() or if_nametoindex().
Drive-by fix indentation.
Pick-to: 6.6
Change-Id: Ifa1111900d6945ea8e05fffd177ef8fcb11b4e1e
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This involves reimplementing QWidgetPrivate::grabFramebuffer().
Widgets call this function whenever a texture-based widget is
encountered.
This implies however that we rename QRhiWidget's own, lightweight
grab function, grab(), because it kind of shadows QWidget's grab().
Switch back to grabFramebuffer() which is what QQuickWidget and
QOpenGLWidget both use.
Supporting QWidget::grab() is particularly important when grabbing
an ancestor of the QRhiWidget, because that has no alternative.
Right now, due to not reimplementing the QWidgetPrivate function,
the place of the QRhiWidget is left empty.
In addition, grabFramebuffer() is now const. This is consistent
with QQuickWidget, but not with QOpenGLWidget and QOpenGLWindow.
Change-Id: I646bd920dab7ba50415dd7ee6b63a209f5673e8f
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
localStorage is unavailable on workers. Operations will now get proxied
to the main thread instead.
This - among other benefits - makes tst_QSettings::testThreadSafety
pass.
Task-number: QTBUG-115509
Change-Id: Iebbe5e9f9069948f8728e0a82628cc082b30de12
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
We know that all lines from /proc/<PID>/maps end in a newline, so we
trim exactly that one byte, then we put it all back together with
.join('\n') to check if we've read the entire file.
Linux virtual files are usually served in 4 kB increments; tst_qfile's
maps file is about 16000 bytes for me, just short of the QIODevice buffer:
[pid 414315] read(5, "55c6afe04000-55c6afe11000 r--p 0"..., 16384) = 4049
[pid 414315] read(5, "7f215fd25000-7f215fd26000 r--p 0"..., 12335) = 4038
[pid 414315] read(5, "7f2160800000-7f21608c7000 r--p 0"..., 8297) = 4072
[pid 414315] read(5, "7f216119f000-7f21611a0000 rw-p 0"..., 4225) = 3994
It is not a coincidence that the reads are at line boundaries, though
it's not a guarantee from the kernel.
We appear to have accidentally fixed the QEMU emulation bug by reading
another process' /proc/<PID>/maps (hypothesis: QMU emulates the target
system in /proc/self, hiding itself in maps, but makes no translation
for other process map files.
Change-Id: Ifbf974a4d10745b099b1fffd1777ac919b12dc90
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
The test changes couple variables to emulate the user project
environment. These variables also affect the policy handling.
The test will build and work properly only if tests are built
standalone. So add this limitation.
Amends 2e340cea88
Pick-to: 6.6
Change-Id: I0cc49bf55bf7763e4c3ecdfa5333fb0453f06794
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
We are BLACKLISTing on macos/arm - the test is flaky.
Pick-to: 6.6 6.5
Task-number: QTBUG-115945
Change-Id: I3be28c895d46ce5ba86e00d597016334bdb16021
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
QObject::children() returns a const QList&, and nothing in the loop body
changes the objects children, so straightforward port to ranged-for.
Task-number: QTBUG-115839
Change-Id: I78827fd986d6ff2607cc2616ff23580c9d830f1b
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
It's very flakey in CI: http://testresults.qt.io/grafana/goto/XZQAAPRSg
What we want to test is whether Qt issues paint events in response
to enabling an already enabled effect. Doing so via qWait will process
both window system events and posted Qt events, and the former might
include spontaneous paint events from the system that we can't control.
Task-number: QTBUG-115945
Pick-to: 6.5 6.6
Change-Id: I65e5c6a4458e77b3bd2ad700c5caf6d441f4ca53
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Neither fork() nor execvp exist on Vxworks so they should not be used
Task-number: QTBUG-115777
Change-Id: I6de4e9ec67741466de1b1f4bd89d9c962e539bb8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Due to the pre-existing overload of invokeMethod that accepts a void-pointer there is
concern that the new overloads would not be preferred.
Here we add a test for this to verify all supported platforms work.
Change-Id: Ie5ac7bf16643599006ac57e0145feb6aace3fa87
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This was missing for a while, and there is nothing fundamentally
missing for it to work.
With the more recent work around slot objects and invokeMethod in
general, it is a good time to add support for this.
In this patch, when connecting to a functor, it automatically deduces
the overload to call based on the arguments passed to invokeMethod.
Sharing code with QObject::connect could be done, but they have a
key difference that makes it harder:
With signal emissions we throw away trailing arguments that are not
used: i.e. `signal(int, int)` can be connected to `slot(int)` or
`slot()`. With invokeMethod that's not a thing. So we will need a way
to toggle that behavior during resolution.
[ChangeLog][QtCore][QMetaObject] Added support for passing parameters
to the overload of QMetaObject::invokeMethod that takes a functor. These
new overloads must have the return-value passed through qReturnArg().
Change-Id: If4fcbb75515b19e72fab80115c109efa37e6626e
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Tests in non-qtbase modules could not find_package their own Qt6*Tools
packages, because add_subdirectory(tests) was called before the config
files for Qt6*Tools were created.
The creation of tools config files is done in QtPostProcess.cmake, which
was included in qt_build_repo_end(). Move that include into its own
macro, qt_build_repo_post_process() and remove it from
qt_build_repo_end(). Call qt_build_repo_post_process() before the
'tests' directory is added in qt_build_repo().
Every call site of qt_build_repo_end() must now be adjusted and call
qt_build_repo_post_process().
Task-number: QTBUG-88264
Change-Id: I80d60a1b5c0e9b715c298ef4934b562f815432d1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
qt6_android_generate_deployment_settings is implicitly used by
qt6_add_executable in user projects. This test makes sure that the
function behaves as expected for user projects specificly, since in Qt
tests it behaves differently because of Qt-specific conditions inside.
Task-number: QTBUG-116037
Pick-to: 6.6
Change-Id: Iea10eca7a780ebaff0c05b91ebe47b821b9ec956
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Remove all "#undef QT_NO_FOREACH" white-listing from source files.
Previous commits have removed all remaining Q_FOREACH/foreach uses in
this sub-tree.
Also remove one source file from NO_PCH_SOURCES in CMakeLists.txt.
Task-number: QTBUG-115839
Change-Id: I02cf994eda720c028e613407342fbd6658fa62b1
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The QHash "cache" isn't modified in the loop body (method is const), in
general cacheSize() shouldn't be susceptible to data races.
Task-number: QTBUG-115839
Change-Id: I122a03ddd5e648d16736c16fa9289eafc886979d
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The toString() method's parameter is a const&, the loop body doesn't
change the container; and the container the parameter refers to isn't
changed during iteration.
Drive-by, remove braces from single-line if blocks.
Task-number: QTBUG-115839
Change-Id: I363e1ed37c0f75fa6a9f8eac3393a6c10d756c1b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Straightforward ports where the container could be made const.
Use C arrays instead of QList if the data is known at compile time.
Drive-by, where appropriate make the for-loop variable a const& (e.g.
QString) instead of copying it for no reason.
Task-number: QTBUG-115839
Change-Id: I273a386e414e5923e750072f0407226efcd4531e
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
In both cases, the container is a member of the unittest, initialized in
initTestCase(), then not changed after that. So use std::as_const.
Task-number: QTBUG-115839
Change-Id: I3b66127e10ac94137260d99f354de9f66a74bec7
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The containers are created locally in the top level test functions, they
can't be made const due to the way they are filled, however the loop
body don't change them; even if the code in a loop would cause
re-entrance du due to signal emittance or events processing, those
containers aren't affected and aren't changed during iteration because
the top-level test functions themselves aren't re-entered, hence
use std::as_const.
Drive-by change: take QHostAddress by const& when it's used as a
for-loop variable (it has a QExplicitlySharedDataPointer d-pointer).
Task-number: QTBUG-115839
Change-Id: I443169e10d973aba2f62854aba200fc2dc2c80aa
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The loops were iterating over a temporary, so use a local const auto
variable to hold it, and use ranged-for.
Drive-by, make the for-loop variable const& instead of copying it,
for any object that has a d-pointer (QNetworkAddressEntry, QHostAddress,
QNetworkInterface).
Task-number: QTBUG-115839
Change-Id: If96c0b2a6142fe2fa2ed45ed7e2435cc1f80e005
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The was introduced with the rewrite of QMetaObject::invokeMethod() in
commit 0f76e55bc4, because we have an
optimization for zero timers to avoid creating a temporary
QSingleShotTimer object. The old implementation did attempt to normalize
the target slot name, but did so because it looked metamethods up using
QMetaObject::indexOfMethod:
int idx = meta->indexOfMethod(sig.constData());
if (idx < 0) {
QByteArray norm =
QMetaObject::normalizedSignature(sig.constData());
idx = meta->indexOfMethod(norm.constData());
}
The new implementation does not use this method so it didn't need to
attempt to normalize.
I am fixing this only in QTimer and not in QMetaObject::invokeMethodImpl
(even though it is trivial to do so) because I don't believe spaces in a
pure string to invokeMethod were ever expected to work:
QMetaObject::invokeMethod(obj, "slotName ", Qt::QueuedConnection);
The Q_ARG and Q_RETURN_ARG (for code not recompiled) still does
normalization inside QMetaType::fromName().
Fixes: QTBUG-116060
Pick-to: 6.5 6.6
Change-Id: I964c2b1e6b834feb9710fffd177cac60c83ef413
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The recent change in QCocoaScreen (108d2e4486)
fixed the problem with no widget found at 'hot spot' (while we know
it's there, since we create it and wait for it to get expose event).
Pick-to: 6.6 6.5
Fixes: QTBUG-108402
Task-number: QTBUG-115945
Change-Id: Ibbf6867bb3381b8137d64cdbd15cc467d8fcf348
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
