Previously, a QVariant parameter would be wrapped inside a new
QVariant, and you would have to cast the QSignalSpy's QVariant to
a QVariant to get the actual value. This behavior was unintuitive
and undocumented.
Check if the parameter type is QVariant, and copy it directly if it
is. This makes the QSignalSpy's QVariant directly usable (no need to
"unwrap" the value in user code).
Existing tests that use QSignalSpy together with QVariant parameters
(such as tst_QPropertyAnimation::valueChanged()) and do cast the
QVariant parameter to a QVariant, continue to work after this change;
this is because qvariant_cast<QVariant>() returns its input value
(unchanged) when the type is not QMetaType::QVariant.
Task-number: QTBUG-21645
Change-Id: Ibfb171edd60c0d3f7ca1d5419e5c5f3d0380d5b3
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
The argument has been obsoleted and not documented since 2007. Get rid
of it now before Qt 5.0
Task-number: QTBUG-25089
Change-Id: I91a5508a5e1606f5b5c289501295c67be4abe6a0
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
- Replace SRCDIR define by QFINDTESTDATA, simplify code.
- Introduce a test for stdout mode that verifies the newline
convention.
- Use a temporary directory as not to clobber the
test directory and introduce an environment variable
UIC_KEEP_GENERATED_FILES to keep them for error
analysis.
Task-number: QTBUG-26730
Change-Id: I22e3bb5a9ca92a1977c29b165ea605f1017baa02
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Add new qt_handleXXX functions that forward to the QWindowSystemInterface
functions, and use those in the testlib inline functions. Remove use of
struct QWindowSystemInterface::TouchPoint from the testlib header files
(requiring some slight increase in ugliness in the two tests that use
that struct).
Also remove the qmake hack that adds private headers to all tests
Change-Id: Iec23537e55a44802f6e9cd463f7a0f82007c5250
Reviewed-by: Paul Olav Tvete <paul.tvete@nokia.com>
We need to handle FD_CLOSE separately on Windows as this will be sent
only once. When we get FD_CLOSE we need to check if there is more data
available for reading. It there is this might indicate that there is
another FD_READ that we need to handle after the FD_CLOSE. So in this
case we will manually create another close event.
Task-number: QTBUG-19409
Task-number: QTBUG-25386
Change-Id: Ie19906bc3f64fb6a85a508a5ab12caac5d70ccdb
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
Qt 5.0 beta requires changing the default to the 5.0 API, disabling
the deprecated code. However, tests should test (and often do) the
compatibility API too, so turn it back on.
Task-number: QTBUG-25053
Change-Id: I8129c3ef3cb58541c95a32d083850d9e7f768927
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
The SCXML spec states that entry order should be equivalent to
"document order" and exit order should be "reverse document order".
Since QStateMachine uses child order for the entry order, the exit
order should be reverse child order.
Change-Id: Ia7b05fdd5c9261ccf202f64f8d23f5c88b20a8c3
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Ensure that the parsing mode is cascaded down from setAuthority and
setUrl so that the hostname parsing does not attempt to decode
percent-encoded hostnames when it shouldn't.
Take the opportunity to also remove the "Boolean Trap" from
QUrlPrivate::setHost.
Change-Id: Ia64754c4a4900182700b7af1382aea8410abc7e9
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
The URI RFC defines schemes as containing only a very restricted set
of characters, none of which require encoding, so don't even
try. Testing this behaviour in some web browsers indicate that they do
not accept percent-encoded schemes either.
Change-Id: I692dd20e1aac7e8a1bcb276cb5113b5802393d38
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
If the password is empty (but present), the userinfo component of the
URL should end in a colon (":"). QUrl already supported that and it
was tested (case "password-empty").
If the username is *also* empty but present, the userinfo component is
just the colon (":"). Fix support for that case by checking if we
stored the presence flag instead of checking the size of the
component.
Change-Id: Ie224493a997dbf76b2e44dd6d55fd9674ac83c1c
Reviewed-by: David Faure <faure@kde.org>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
QString::fromUtf8, without an explicit size, (currently) defaults to
stopping at the first NUL. That means we need to pass an explicit
size.
Also take the opportunity to test that QUrl::toPercentEncoding also
works with the same data.
Change-Id: I79362d67afda624b01ca07b0315b611c4aa3fdda
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: David Faure <faure@kde.org>
QWindow::setWindowState is not supposed to set the window active.
The method requestActivateWindow() should be used for that.
When switching from and to fullscreen mode we're always passing
SWP_NOACTIVATE to SetWindowPos to not change the activation state
of the window. This is inverse to the old behaviour, which did not
have an effect.
Change-Id: I339337935cdad76b3ef252202e92177f37543038
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Unloading and reloading a plugin didn't work correctly,
because we didn't reset instance to 0 on unload.
Task-number: QTBUG-26098
Change-Id: Ic3e4497f359b1ca455be949dce9cafa9d67d8039
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QApplication::setActiveWindow doesn't activate the native window but
marks the widget as active inside Qt.
We need to use QWidget::activateWindow instead. See docs.
Also moved the activation call further down because on Windows a
minimized window cannot be activated using the activation-by-focus
fake we're currently using.
Change-Id: I752f6ada1f463931fa9cfb3c35f42dbec0207bfa
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Previously synchronous window system events were
implemented by bypassing the queue and processing
the event immediately. This is not ideal since the
event order is not preserved - there might be "happened
before" events waiting in the queue.
Add QWindowSystemInterface::flushWindowSystemEvents
and change all handleSynchronous* to 1) queue the
event 2) call flushWindowSystemEvents.
flushWindowSystemEvents is almost identical to the
already existing sendWindowSystemEvents with the
exception that it does not call QApp::sendPostedEvents.
Move the common implementation to a new private function.
Task-number: QTBUG-20778
Change-Id: Ie98a83875bc0a14e335e36bed0dd9e0ed4a1dea0
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
The test has a member QWidget *topLevel which it recreates
and shows in init() without waiting for it to be exposed
although it is not used in every test case. This apparently
interferes with some tests that create separate top levels.
Do not show in init(), delete the topLevel.
Add wait to the cases where the topLevel is shown.
Change-Id: Ib428020b36dc82991d41e68478fd583bdfb004c7
Reviewed-by: Sergio Ahumada <sergio.ahumada@nokia.com>
Use QTRY_VERIFY for the mappped attribute check.
Change-Id: I3cbde9122405bf7067f3702193e80636edc8c5c6
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Don't rely on a black list of codecs for the serialization
test. This breaks badly when new codecs get added to Qt
(e.g. through ICU). Instead use a white list of known
codecs that can encode/decode the test data.
Change-Id: I1dc55a25e852198bb935f070a4a21e8369f56268
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Use ICU to do code page conversion instead of the
builtin text codecs. With this QTextCodec simply
becomes a wrapper around ICU's ucnv_* methods.
We only keep our own codecs for UTF-*, ISO-8859-1,
ISO-8859-15 for performance reasons, and for TSCII
and iscii-* because they aren't supported by ICU.
Change-Id: I4fc49eba55cf772b9772c6dac606a47a44346a60
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This is working, but let's just be sure by adding a testcase.
Change-Id: I8c6b5ded0c7b6c90645dbf70a7ce6c1ba447a284
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
contains(QT_CONFIG,opengl) tells whether Qt itself is able to use
OpenGL, which is not the same thing as whether the QtOpenGL API is
available. Make the check correct; fixes compilation when Qt is
configured with -no-widgets (which also disables QtOpenGL).
Change-Id: Iaa296c2b10650971ef4846f8bc6f44761fadcf7c
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
This test was using some QSpinBox instances to generate signals for
testing. Use our own QtTestObject to generate the signals instead.
Change-Id: I3714955ae040d541c3b613a478945c38a18be18d
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
the system path separator and shell are bound to the host system
(system() will use cmd even on mingw with sh.exe in path).
the makefiles otoh may depend on what the qmakespec defines.
consequently, add $$system_path() and $$system_quote() (for use with
system() & $$system()). $$native_path() is renamed to $$shell_path() and
should be used with $$shell_quote() to produce command lines in
makefiles.
$$QMAKE_DIR_SEP needs to be applied to Option::dir_sep right after
parsing the spec, so it is available to $$shell_{path,quote}().
Change-Id: If3db4849e7f96068cf03a32348a24f3a72d6292c
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
- Introduce smart pointers to delete widgets and resources
to ensure tests are not affected by left-over widgets
also in case of failure.
- Replace deprecated QTest::qWaitForWindowShown() by
QTest::qWaitForWindowExposed() and use QVERIFY,
remove some hard-coded timeouts.
- Set some titles and object names.
- Add verbose debug output of event lists in tests
childEvents.
- Set minimum sizes on widgets to avoid Windows warnings.
- Stabilize GDIWidget, trigger on first event only.
Change-Id: I64119a2e7113e4a9f0156d00c72ce0935d03bb81
Reviewed-by: Joerg Bornemann <joerg.bornemann@nokia.com>
This is useful for inserting a string without space-handling, given that
dbg.nospace() followed by dbg.space() inserts a space.
It's also useful for QDebug operators for custom types, so that they
can disable space handling and then restore to whatever it was before
(rather than forcing it to space() mode).
Change-Id: I9d72e9ffbcbc581ed093168752c29af924405b33
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When containing a QObject (or sub-class) pointer and trying to convert
to a QObject pointer canConvert() did dereference the pointer without
checking for it being null.
Change-Id: Ie274e54f2f817f2b6c5df64504f8af6359b8f38d
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
In touch event terminology the global position is the screenPos,
scenePos is the windowPos.
Fixes a tst_qdeclarativepincharea test failure in qtquick1.
Change-Id: Ie98fe12be8cbedc9b019913b066e7c4bce75278d
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Make sure the test subdirectories are actually visible
in the file model before the tests start.
Change-Id: If640456bba4362b19d7ad9d9184736c2eb8d3bde
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
This avoids test instabilities and prevents test directories
from being cluttered with temporary files. Change tests
accordingly. Remove unused createLink() method.
Change-Id: I843c28ab81c8a476c71c5211a7479b22d3d9fc93
Reviewed-by: Sergio Ahumada <sergio.ahumada@nokia.com>
This allows to check whether QMetaTypeId2::MetaType exists, and can help
turn run-time into compile-time expressions, even without constexpr support,
or in situations where constexpr can't be used (because you can't overload
on it). This was designed for the QMetaType::registerConversion feature,
but it's much more widely applicable.
Change-Id: Iafa04add04bcb531b3f7fe3e751c7e91ee6a3bc0
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Using the nullary version has the advantage that multiple calls
during a program run are much more efficient, since an inlined
atomic is used to store the result. It also ensures that
Q_DECLARE_METATYPE(T) has been used, whereas qRegisterMetaType<T>("T")
will happily register anything. So I've added the macro where it
was missing, or moved it to a central place when it existed
hidden.
In tst_qnetworkreply, this became a bit tricky, because a private
header is conditionally included, so moved the Q_DECLARE_METATYPE()
into a conditional section, too.
Change-Id: I71484523e4277f4697b7d4b2ddc3505375162727
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Remove usages from autotests with the exception of
widgets/kernel, widgets/widgets and widgets/graphicsview.
Change-Id: I917b2857ed0cd07a6b3dbcd69244f558086c6586
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Also check the return value of
DefaultValueFactory<QMetaType::Void>::create(), the same way it's
done in testCreateHelper<QMetaType::Void>().
Change-Id: I3e6d7fca4ea74dbe65009f2eb2c64a1b3a370d68
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
This is a simplified port of KDTools' KDAB_SET_OBJECT_NAME.
It simply assigns the variable name as the objectName of
a QObject, uic-style. It uses a small helper function so
that it works on references as well as pointer variables.
QLabel label;
QLabel *pLabel = new QLabel();
Q_SET_OBJECT_NAME(label);
Q_SET_OBJECT_NAME(pLabel);
Change-Id: I25fec0c90f33249a3ea5d2dd622ab708019fd101
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: David Faure <faure@kde.org>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Using qRegisterMetaType<T>() has the advantage that multiple calls
during a program run are much more efficient, since an inlined
atomic is used to store the result. It also ensures that
Q_DECLARE_METATYPE(T) has been used, whereas qRegisterMetaType<T>("T")
will happily register anything.
Had to add Q_DECLARE_METATYPE to QFileInfo, for
QList<QPair<QString,QFileInfo>> of QFileSystemModel to work with
the partial specialisations of Q_DECLARE_METATYPE for QList, QPair.
In order to synchronize this change with other modules that did
their own Q_DECLARE_METATYPE(QFileInfo), a sync macro is defined
that can be tested in other modules, and will later be removed again.
Change-Id: I3004664e07e64cd885d5a03a57ff4e4379804aec
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
While the qWaitForWindowShown(QWidget *) is inherited
from Qt 4.8, the qWaitForWindowShown(QWindow *) was introduced
in Qt 5. As it is identical to qWaitForWindowExposed()
and removed already, it can be deprecated in Qt 5.
Remove its usages in qtbase.
Change-Id: I28788d120ad687a49f02b2b44de6b38a2832fe5c
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
This makes development on the meta type system easier because only
QtCore must be re-built to run most of the tests. The existing
QGuiVariant test needs to be run before pushing anyway, but not
so frequently.
Change-Id: I1fa66edbd790c957e1a232226847dd550227a477
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
The type needs to be large enough that
QVariantIntegrator<T>::CanUseInternalSpace is true.
Change-Id: I311c44bedfebd946e41639975df206c27b6d55ca
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
For consistency with qMetaTypeId, and because we can.
Change-Id: I6882a16ef3c0d84539048c9f2c201c4a2b2ca7ad
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
This patch implement the equivalent of
468626e99a90d6ac21cb311cde05c658ccb3b781 in qtdeclarative but for
QtWidgets.
If a widget doesn't accept a touch event, then QApplication gives it
another try by synthesizing a corresponding mouse event. This way
QtQuick and QtWidget behave in a similar way, removing the need for
platform backends to try to emulate a mouse event from a touch event
unconditionally.
Also add relevant unit tests and adjust old QApplication ones.
Change-Id: Iddbf6d756c4b52931a9d1c314b50d7a31dbcdee9
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
QModelIndex is a build-in type nowadays and doesn't
need to be registered anymore.
Also remove them from the tests.
Change-Id: I47029972651c045c880cee86fb292116a29493d5
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
The behavior of QTEST_MAIN depends on whether QT_GUI_LIB or QT_WIDGETS_LIB
is defined. It could create a QGuiApplication or QApplication which
could cause linking issues if the corresponding library is not linked to.
The failure cases are also tested.
Change-Id: I61ed0bc760564ef42ce1dbd86c83c06348c860ff
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
No need to pass the dispatcher. Get rid of Windows logic to maintain
a stack of dispatcher associated with flags.
Change-Id: Ic2daad4b6762a46fac3274937effc188af436c9a
Reviewed-by: David Faure <faure@kde.org>
This allows building the QVariant tests without the QtWidgets module.
Change-Id: I7cd7e78a60c7bc7614ec16df1abe1e93e45d4923
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
The previous API was hard to use (global function, no type safety,
manual chaining), and confusing (app vs dispatcher split only made
sense on Windows). Installing and removing out of order would have
the risk of setting back a dangling pointer (crash). Meanwhile QPA
added type safety, and this new API models the QObject::installEventFilter
API for ease of use. The virtual method is in a new interface,
QAbstractNativeEventFilter.
QPA was even calling the dispatcher event filter with QPA-private event
classes, which made no sense (refactoring leftover from when the code
was in the dispatcher). Now the QPA plugins trigger the qcoreapp event
filters with the actual native events directly.
Change-Id: Ie35e47c59c862383bcaf857b28d54f7c72547882
Reviewed-by: Marius Storm-Olsen <marius.storm-olsen@nokia.com>
- Implement waitForWindowExposed() for toplevel windows.
- Implement waitForWindowShown(QWidget *) and mark as
deprecated in line with waitForWindowShown(QWindow*).
- Use in tests.
- Simplify tests (collapse waitForExposed, setActive
into setActiveWindow, waitForActive), remove most
hard-coded timeouts.
- Stabilize graphicsview tests by using waitForWindowActive.
Change-Id: Ic7c061e2745b36f71a715ee4e47c0346b11a91e8
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
When turning off fullscreen mode and restoring the widget's geometry
we must inform the QWindow about the geometry change synchronously.
Otherwise QWidget::geometry() will return the old value.
Using the same technique for the state transition to fullscreen mode
without sending a separate resize event.
Autotest: tst_QWidget::saveRestoreGeometry
Task-number: QTBUG-26421
Change-Id: I869e36cd302d9a94e398f48949ab3cb7ee9cdf51
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
QtWidgets and QtGui are currently compiled without exceptions,
which causes a crash with gcc 4.6.3.
Change-Id: I8f872f3bec6266444adf08d51a6678150c5fae8e
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
It should be possible to include both:
* <QtGui/private/qfoo.h>
* <private/qfoo.h>
Change-Id: I83ed5bba633b4a6b9bd38e315c987d78beecfb1b
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
When mixing native and regular widgets in same QMdiArea, some
subwindows didn't properly get set native. This was because
when a native parentless widget was given a parent, it wouldn't
enforce native window on the new parent and its ancestors.
This happened because window flags were adjusted too late in
relation to createWinId() call in setParent_sys().
Fixed by moving the createWinId() call to its proper place.
Also removed some old Q_WS_* ifdeffing in QWidget::setParent() that
masked some native enforcement code.
Additionally removed few QEXPECT_FAILs from QWidget autotest now
that those cases work correctly.
Task-number: QTBUG-26424
Change-Id: Ib6f9d0531e5c7299e2c307734d49c81f1ffa9713
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
WinAPI GetTempPath() sometimes returns short names
for C:/Users/<user>/AppData/Local/Temp.
Change-Id: I33f991acc06e652ccd484d36a5a384eb776f8395
Reviewed-by: Joerg Bornemann <joerg.bornemann@nokia.com>
SQL Server 10 introduced stricter rules for TIMESTAMP validation,
making it necessary to specify the decimal digits.
Other databases might do the same as well, so this patch introduces
a check for the TIMESTAMP column size and adjusts the decimal digits
parameter as needed.
Task-number: QTBUG-2192
Change-Id: If6d798c6c928ebda75bc474e49a07fbbfbe5816c
Reviewed-by: Mark Brand <mabrand@mabrand.nl>
A virtual method was reimplemented to return an always-empty string,
probably a leftover from a refactoring.
This fix showed that tst_qwidget_window was buggy: between Qt4 and Qt5,
a "Before" became "After", which made "Before" unused, and was masking
the fact that the app name was empty by default. In addition, the
earlier Qt5 change that made the app name default to argv[0] now requires
updating this test, now that it's actually working.
Change-Id: I5360026821a9b95bedd0ff09dba3d51a22e542b7
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@nokia.com>
This makes it possible to add API for setting the restore policy
per state, or even per property assignment (QTBUG-17861).
This change is fully source compatible with Qt4.
Change-Id: I53628546b070f6fc84891f86e7ad7bd8ef5ba285
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Back when QStateMachine was changed to inherit QState, this
constructor was conveniently left out because setting the state
machine (root state) to be a parallel state group didn't actually
work. But as of commit d281aa6936,
it does work, so add the missing constructor.
Task-number: QTBUG-15430
Change-Id: I68c599baa0ef1bfc869195140cf5daf645e75b8b
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
under extremely rare circumstances this would have actually failed
Change-Id: I4132d0f82e9f924e92e9e96f6d34451c94a67201
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
This is just for completeness of the understanding of the limitations
of private signals. There are no private signals in Qt which have
overloads.
Change-Id: Ic34c555aea360ee34beec796e597657888573da9
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
When the pushbutton is shown, it will generate both a ShowEvent and
a StateChange with active=1 (because it is a top level window).
This patch relaxes the reqirement in which order events are delivered.
Ideally the order should also relied on, but I'm not sure if that
is feasible due to differences among window managers across all
platforms.
This got provoked by codereview.qt-project.org/#change,26014
Change-Id: I96159fbb1b64f0ca8d13833d8a4c6799c655afc2
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@nokia.com>
Since Qt's connections are thread-safe, QStateMachine's plumbing
around them should be thread-safe too.
Change-Id: I8ae91c2edc2d32ca4ed4258b71e5da22de30ed91
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
By default, QStateMachine lazily registers signal transitions (i.e.,
connects to the signal) when the transition's source state is
entered. The connections are established in Qt::AutoConnection mode,
which means that if the sender object lives in a different thread,
the signal processing will be queued.
But if a sender object's signal is used in an out-going transition
of the target state of the queued transition, it's possible that a
second signal emission on the sender object's thread will be
"missed" by the state machine; before the machine gets around to
processing the first queued emission (and registering the
transitions of the new state), a sender object on the other thread
could have emitted a new signal.
The solution employed here is to eagerly register any signal
transition whose sender object is on a different thread; that is,
register it regardless of whether the transition's source state is
active.
Conversely, when a machine's transitions are unregistered (i.e.,
because the machine finished), signal transitions with sender
objects on other threads should be left as-is, in case the machine
will be run again.
This doesn't solve the case where the sender object is moved to a
different thread _after_ the transition has been initialized.
Theoretically, we could catch that by installing an event filter
on every sender object and handle the ThreadChange events, but
that would be very expensive, and likely useless in most cases.
So let's just say that that case isn't supported for now.
Task-number: QTBUG-19789
Change-Id: Ibc87bfbf2ed83217ac61ae9401fe4f179ef26c24
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Some of the transition constructors didn't call the maybeRegister()
function, causing the transitions to be ignored if they were created
when the state machine was running and the transition's source state
was active.
Added tests that cover all possible cases.
Change-Id: If1b593b127bd719e3be4e5a2e6949a780c4e97c3
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
The originalSignalIndex member was not set if the signature had to be
normalized. This caused the SignalEvent passed to onTransition() to
report a signal index of -1.
Improve the signal transition tests so they check both the event
passed to eventTest() and onTransition().
Change-Id: I5331fd1944d53310b6d11eb2fd8713b80faa53a1
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Use the same trick as used for private signals in the models.
Change-Id: I4235788490cae0e3d554565621d145652dc5b0ca
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
The private signals can not be used as function pointers, as
required by the new syntax, so we introduce a parameter which
can only be created privately.
Change-Id: I3d7bb8a163e764d685e8007cba831fb77e3c6855
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Moc checks for the use of the QPrivateSignal struct, which is part of
the Q_OBJECT macro and is private to each class that uses it. Moc then
generates a name of the signal which does not include the private
struct, and generates code to invoke such signals with an instance of
the private struct.
This way we can mark private signals as such and prevent them from
being emitted from subclasses or from outside of the class entirely.
The drawback to this is that it only works if the private
signal has no default arguments. However, at least in Qt, there are
no such signals.
Change-Id: Id16eadaa8d3c36a2c3b265077877f3e1d8304c84
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Type traits can not be changed durring Qt5 life time.
Change-Id: If69f65ff2113c901580afee91b11ae1b11c13a4f
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Replace a list of QFuture<void>s and a loop that calls waitForFinished()
on each of them with a QFutureSynchronizer<void>, which does exactly that.
Change-Id: I1f2e90169a5b2949bd8cb9d1009a5a7af1500139
Reviewed-by: David Faure <faure@kde.org>
The SCXML spec had a bug that would cause the initial state of a
compound state within a parallel state group to be entered even if
the transition specified another (non-initial) state of the compound
state as its target. This only happened if the transition had
multiple target states.
The bug has been fixed in recent revisions of the SCXML spec. This
commit implements the fix, which is to walk the ancestors of the
transition's target states only after all the target states
themselves have been added, so that the default initial states are
correctly overridden/ignored.
Task-number: QTBUG-25958
Change-Id: Iac532047678c483a4a3996e24dacf30e00f6bbe0
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Remember to register the metatype where we use it, so we don't depend
on another test being run previously.
And skip the setWorkingDirectory test completely on Unix. I don't know
why it needs to be skipped, but if we're not going to verify anything,
don't even try to do anything. This saves us one memory leak at least.
Change-Id: I22e151cc3fa7b4e976972aca8978b88b263d9bee
Reviewed-by: Mitch Curtis <mitch.curtis@nokia.com>
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
This is accomplished by introducing dependencies to catalogs.
This requires one API change:
QTranslator::load(const uchar *, int);
changes to
QTranslator::load(const uchar*, int len,
const QString &directory = QString());
Since now, even the load from memory might need a directory if
the memory block contains a qm file with dependencies.
Change-Id: I781f333d07f53bb431d0a7b5fa1abe282dc4d338
Task-number: QTBUG-26138
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
When you're inserting a column in front of a rowspanned cell
and this cell is not the first in the rowspan, we would get
the wrong logical index of the new cell (putting it in
front of the initial cell with the rowspan). If the cell
does not span all rows, the table will get into a broken state
and trigger asserts in update(). To fix this, we search for
the first cell after the insertion point which has a logical
index higher than the cell directly before the insertion point.
Change-Id: I42e91a20d77b2ba9c5607f6cab23f51ed888cbd3
Reviewed-by: Simon Hausmann <simon.hausmann@nokia.com>
Deprecate qWaitForWindowShown for Qt 6 as it is just a wrapper.
Change-Id: I0f8195679679120bd402e273fed4d331dc926708
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
