Long-lived threads started by Qt itself can now receive events even if
QCoreApplication hasn't been created. This is required in all threads we
start that will handle events, unless we're sure that the thread will
exit before the global application object begins destruction.
Otherwise, those threads will have race conditions dealing with the
event delivery system trying to call the QCoreApplication::notify()
virtual while the object is being destroyed.
Change-Id: I27eaacb532114dd188c4ffff13d4ad2a4bb443e6
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
This commit makes QCoreApplicationPrivate::checkReceiverThread,
notify_helper, and sendThroughObjectEventFilters be static functions,
since they only deal with global data or the parameters only.
Making notifyInternal would have been binary incompatible (it's called
from inline functions QCoreApplication::sendSpontaneousEvent and
QCoreApplication::sendEvent), so instead add a new static
notifyInternal2 and mark the older function deprecated and to be removed
in Qt 6.
Change-Id: I27eaacb532114dd188c4ffff13d59fe3b0661489
Reviewed-by: Albert Astals Cid <albert.astals@canonical.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
ButtonInfo is larger than a void*, so holding them in a QList is needlessly
inefficient. Worse, the code could come to depend on the fragile property
of (inefficient) QLists that references to elements therein never are
invalidated.
Fix by holding it in QVector. Also reserve() the vector, even though we
can't tell the size exectly. It's a short-lived vector.
When appending, add an optimistic qMove().
I would have liked to use std::vector instead, but QRenderRule, thus
ButtonInfo, isn't nothrow-move-constructible, because of missing
move constructors on QBrush and QFont, among others.
Change-Id: I89164f4ed5745498093102f022a7ef32186e8045
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
I can't possibly understand the meaning of the note, especially
given the snippet. (However ,it has been there since Qt 3,
so perhaps it refers to some weird QValueList behavior?)
Change-Id: I952e9a7cb687b94cfccb927eb359b635804f4ade
Reviewed-by: Martin Smith <martin.smith@digia.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Prefer -std=gnu++11 unless strict_c++11 is defined. You can enable
strict C++11/C++14 mode by using
CONFIG += strict_c++
That is enabled for Qt's own code, so we we don't accidentally use GNU
extensions in portable code.
There's no support for strict C++98 mode (that is, the -ansi option).
[ChangeLog][qmake] By default, GNU extensions are now enabled with
Clang, GCC and ICC even in C++11 and C++14 modes. To disable the GNU
extensions, add to your .pro file: CONFIG += strict_c++.
Change-Id: Ib056b47dde3341ef9a52ffff13ef14de2169bef5
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
The pipe2/dup3/accept4 functions and SOCK_CLOEXEC are quite old nowadays
on Linux. They were introduced on Linux 2.6.28 and glibc 2.10, all from
2008. They were also picked up by uClibc in 2011 and FreeBSD as of
version 10.0. So we no longer need the runtime detection of whether the
feature is available.
Instead, if the libc has support for it, use it unconditionally and fail
at runtime if the syscall isn't implemented.
Change-Id: Ib056b47dde3341ef9a52ffff13efcc39ef8dff7d
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
FD_CLOEXEC is implicit for us, so no caller will need to set O_CLOEXEC.
Change-Id: Ib056b47dde3341ef9a52ffff13efcb635dea95f0
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
There's really no difference between them, other than -force-pkg-config
skipping the check if the tool is even available and printing a warning.
Change-Id: I04cb83c6649ef73866a84032ea46093c4a00ce00
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
This also finishes the implementation of the static methods which
were being passed a supportedSchemes argument but weren't using it.
Now they can pass it along to the QFileDialog instance, which then
passes it to the helper used by the QPA implementation of the file
dialog. The default implementation only supports local files and
can therefore ignore this, but other implementations can use this
argument to restrict the protocols allowed to the user.
[ChangeLog][Widgets][QFileDialog] Add supportedSchemes property.
Change-Id: I5235f70e785da1c06866a8355ef98f571890c4a2
Reviewed-by: Kevin Ottens <kevin.ottens@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
I show this code during trainings and this explicit template type
just makes the line of code more complex than it has to be.
Change-Id: I294c65f48967e54c67d9c3b8e1de96c3c2495a5b
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
From MSVC 2015 onwards, MSVC formats floats using printf("%g") like
g++, so the fiddling of the expected output needs to restricted
accordingly.
Change-Id: I6e7f78e5e90f70886a8b2ef37c0fb9bf82b5e1a3
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
- Introduce Qt 5 signals & slot syntax.
- Replace QSignalMapper used for the corner/area
actions by a by a functor.
- Improve command line parsing.
- Reorder class declarations.
- Remove commented-out code.
- Use QDialogButtonBox in dialogs.
- Fix minor issues in code, use multi-argument version
of QString::arg(), QDir::toNativeSeparators() to
present file paths to the user, static method
invocations.
Change-Id: I865c56639c74135b59740797e9a9dfbfca2e72b6
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
- Introduce Qt 5 signals & slot syntax.
- Use QCommandLineParser to obtain file arguments, factor
out positioning into tile() function to be able to
use it from command line and open/new slots.
- Merge MainWindow::createMenus()/createQToolBars()
into MainWindow::createActions(), removing the need
to store the actions as member variables.
Use QMenu::addAction() for brevity.
- Use QIcon::fromTheme() to obtain system icons and use
resource icons as fallback.
- Rewrite settings code to use
QWidget::saveGeometry(), Widget::restoreGeometry() since
saving size and position does not work well with multiple
screens. Query the available size when determining
the initial size instead of using hard-coded values
for High-DPI screens.
- Fix minor issues in code, use multi-argument version
of QString::arg(), QDir::toNativeSeparators() to
present file paths to the user and and static method invocation.
- Fix snippet references accordingly.
Change-Id: I3ea0372bc7ff82247192f54620289352fb68d60f
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
When running the test, one gets the impression that it hangs.
Add some debug output including time to show what happens.
Change-Id: Iac6b4f0518ecec62169bf2269a0a8ec9192da570
Reviewed-by: David Faure <david.faure@kdab.com>
This mapping has to be done manually, like we do on Windows
for example.
libinput maps through xkbcommon, like xcb, so it is already correct.
Task-number: QTBUG-46845
Change-Id: I61f3f1160e2581aae2ef43cc260f191f6d344fec
Reviewed-by: Louai Al-Khanji <louai.al-khanji@theqtcompany.com>
On Windows, having read-only files in a directory can cause removal
to fail. When file deletion fails, check on the permissions, set
write permissions and retry.
Split apart code paths by OS in tst_QDir::removeRecursivelyFailure();
deletion of the read-only directory on UNIX should still fail.
Change-Id: I36e54be5229a7b552e90fd5f42722b868fa0b6ee
Reviewed-by: David Faure <david.faure@kdab.com>
The test showed failures on OS X:
FAIL! : tst_QListView::batchedMode() Compared values are not the same
Actual (ba.size()): 2
Expected (3) : 3
Loc: [tst_qlistview.cpp(848)]
Use QTRY_COMPARE() to count the number of visible indexes with
a helper function instead of using hard-coded timeouts. Item 3
appears with a little delay on OS X.
Change-Id: I2fb2ff5ebdf9dbe85bdc79401375ad6f47b7b12b
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
The code samples seem to be using Qt::FindDirectChildOnly. Changed this
to Qt::FindDirectChildrenOnly.
Task-number: QTBUG-46383
Change-Id: I99b05e5e279bff8f799a09b0d7b02512fc73170b
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
Right now the two functions being added are just wrappers, but this will
allow us in the future to support FreeBSD and Linux's system calls that
do the equivalent of forkfd, but have slightly different semantics in
the actual getting of the information from the file descriptor.
See-Also: https://lkml.org/lkml/2015/3/12/1044
See-Also: http://www.freebsd.org/cgi/man.cgi?query=pdfork
Change-Id: Ia0aac2f09e9245339951ffff13c94acb5f4ff204
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Commit 5bf67f5f41 did it for qobject.cpp,
but I missed qwidget.cpp (I hadn't realized the test existed there too).
Change-Id: Ib056b47dde3341ef9a52ffff13eeceafcfc64893
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The Intel compiler's compilervars.bat also puts the MSVC compiler in
PATH, so cl.exe was always being found first.
Change-Id: I72e524da10fb0e221c4530a3e5c1a4a347c3f878
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Otherwise the information is missing from configure's output summary:
EGLFS ................ no
EGLFS i.MX6....... .
EGLFS KMS .......... no
EGLFS Mali .........
EGLFS Raspberry Pi .
EGLFS X11 .......... no
Change-Id: Iee8cbc07c4434ce9b560ffff13cb331778c70261
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
The test checks whether a child window receives
mouse events synthesizes from touch. Enlarge the child
window so that it fully covers the parent and move the
touch point a bit inside so that it is not affected
by window manager positioning issues.
Use QTRY_VERIFY.
FAIL! : tst_QApplication::touchEventPropagation() 'widget.seenMouseEvent' returned FALSE. ()
Loc: [/work/build/qt/qtbase/tests/auto/widgets/kernel/qapplication/tst_qapplication.cpp(2107)]
Change-Id: Ic08e68b1e547cc7148cd8994464fdc2a14ac507b
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
The test sends touch events to the root item at the top left
corner which fails if the views starts to scroll.
Make the view sufficiently large to prevent scrolling and align
at top left.
FAIL! : tst_QTouchEvent::touchBeginWithGraphicsWidget() Compared values are not the same
Actual (((root->touchBeginCounter))): 0
Expected (1) : 1
Loc: [/work/build/qt/qtbase/tests/auto/gui/kernel/qtouchevent/tst_qtouchevent.cpp(1471)]
Change-Id: I357322ccc809ddb5cb587febf3c75cbe497e59d8
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
They can't be ref-qualified if the QVersionNumber object doesn't actually
hold a QVector<int>, as the next commit will make it: for segments(),
there might not be a QVector to be moved; for normalized(), the common
case will be that there's no gain in ref-qualifying.
Change-Id: I4bfb8b8765a502c0de6aed693752217106e575a2
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Example on how to reproduce the bug with examples/widgets/mainwindows/mainwindow:
- Select the option "Groupped Dragging" from the "Main window" menu
- Drag and drop dock widget so there are at least three of them in
the same group of tabs (for example red, black, and green)
- Drag the tab bar of one of them outside to make it floating
(for example black)
- Drag away the group of two remaining tabs to make them floating
- Drack back the floating-alone (black) dock widget somewhere
back inside the main window.
=> The black dock widget is tabbed with the other instead of being
where we put it
That's because it was still inactive (skipped) in the tab group.
Task-number: QTBUG-47209
Change-Id: I84ac33d4ca1df0381c6007486b41830bd7be99a8
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
Even though we expect very few option names, GCC unrolls the
loop calling IsInvalidName::operator() four times. Together
with the lead-in/lead-out for size() % 4 != 0, that makes
for five inlined copies of the function call operator.
I don't know how to prevent the unrolling, but at least I can
stop the inlining.
Change-Id: Ib9d6c33e8193464a73927a92b8b54c3452cb26a3
Reviewed-by: David Faure <david.faure@kdab.com>
We suppose that this function will never be called, yet the
compiler still inlines it in all four places.
Stop it.
Change-Id: If601bc3644cacb77aeab9d615578e4afb349a73e
Reviewed-by: David Faure <david.faure@kdab.com>
There are other literal 4's used in the implementation, so
disambiguate.
Change-Id: I5b0e4ac0018f9d5734e3de8b4f75259e175b23d3
Reviewed-by: David Faure <david.faure@kdab.com>
QBackingstoreTextureInfo is larger than a void*, so holding them in
a QList is needlessly inefficient. Worse, the code could come to
depend on the fragile property of (inefficient) QLists that
references to elements therein never are invalidated.
Change-Id: I5edf846ee8f01ae36c9314147261748270e1fdf6
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Only one of the three following branches actually uses the ButtonInfo
object, so create it only in that branch.
Change-Id: I65c1d45d807258ed3352018c088bb4c4c66f9b84
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
These types are held in QVariant.
This change should be BC, because it's private API, even though
it is located in a public header file. The classes are used as
private data members in the respective public API classes, but
we don't change the size or composition, and holding these
types in a QList would be outside the use of non-private API
for which we promise BC.
Change-Id: I7ef1c1ca57e9d87c6474c97bb2fa8afef170c88f
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
In preparation of marking it movable. For this, it is important to
have a good overview over the data members. Distributing them among
the function members is not helpful for that. Fix.
Change-Id: I29e6e270c698ce32f7e58e083e6205bdba9c1362
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
First, all values in the array fit into 8-bit, so change the type
from ushort to quint8.
Then, instead of appending a zero as end marker, statically
determine the size of the array and use that number.
Also fixes (benign) off-by-one error in the existing reserve()
call.
Change-Id: Id05b6a56b55d2e0769b00496a55808bb7351b084
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Same pattern as used for Q{Sequential,Associative}IterableImpl.
Change-Id: Iacec3266af80eecf491d2bb766c6fddd1365bdaa
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Make sure that the QString is not in scope anymore when the pthread
cleanup happens. C++ destructors in scope are pthread cleanup handlers.
Change-Id: I9a75ad8521ae4e5cbbe5ffff13d1ae1c7e31f6bb
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
These are all error conditions, and should not happen in practice.
Naturally, there's no savings in executable size, because the
compiler just shuffles the layout of the branches around.
Change-Id: I52b98cc696fd808735c7d73c1f21e02478ff4f5a
Reviewed-by: David Faure <david.faure@kdab.com>
