Change-Id: I2d30d6079d3982e5ca2b8d9346d0d511279792ad
Reviewed-by: Nicolas Arnaud-Cormos <nicolas@kdab.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
This makes it clear that the m_childWindows size must be 0 at this point,
indicating otherwise an inconsistency in the logic of the child windows
management.
Change-Id: I04e418bc6e1d23681bd96f4d619cde9645dc6a22
Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com>
It was introduced in the early days of the docs'
modularization and appears not to be used anywhere
anymore.
Change-Id: I5b0c60d92828624af2129153fce96ad01aec861c
Reviewed-by: Martin Smith <martin.smith@digia.com>
MinGW-64 defines _CrtSetReportMode, but only as a noop, resulting
in a warning "statement has no effect [-Wunused-value]"
Change-Id: I6e509246134dd5929b23b07abdb6d705db603793
Reviewed-by: Jonathan Liu <net147@gmail.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Screen readers do random things with the dbus interface,
we should be careful when handling what they send.
Change-Id: Ie8fc515c8d399dae9b5e892a135b54a1820d09e2
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Task-number: QTBUG-27336
Don't check 'widget != m_widget' in
QWidgetWindow::handleDragEnterMoveEvent() since the current window's
parent widget may be the actual drop target. I replace it with a check
'!widget->isWindow()' to prevent we pass through a top level window.
I also change 'widget->mapFrom(m_widget, event->pos())' to
'widget->mapFromGlobal(m_widget->mapToGlobal(event->pos()))' since m_widget
may not be widget's parent.
Change-Id: Ia4f10f85ccdf1e27223ddc51afabd98b5d16f2fb
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
This is exactly the type of pointer arithmetics that GCC likes to
miscompile.
Just use offset >= size() (=subtract begin() from both sides).
Change-Id: Ifb13652d7b96bf4c06727d9c965516e95d16ab5c
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
These indexes are later used as the boundary points of a QItemSelectionRange
anyway, which means that they're going to become QPersistentModelIndexes again
soon.
Because QPersistentModelIndex::row and ::column API are not inline, we cache
the resulting values in the loop.
Change-Id: Ib5099148269a8ccbb7ff2d8819a347e429c55dd1
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
QPersistentModelIndex is not so big that it causes very bad QList performance,
but considering how we use the container, QVector is fine. We never prepend to
the container, we only ever append, clear and iterate over it.
Change-Id: If1e53db9252c22340f1e0179b14a460495b5b35d
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
The implementation of QAIM can implement an efficient version of sibling.
Task-number: QTBUG-17732
Change-Id: I474dbc11e52b3ccc42e2165bc9336882fab13d03
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
This model apparently pre-dates the API required to do resetting properly.
Change-Id: I0ea903cef87077ca4a3d8ca2b53eba213e77cc26
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
This is compared later with the internalId() of other indexes, which
produces a warning.
Change-Id: I57ea248c792db3b49947cc6a8fe989c68b67ff0f
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
- Also get rid of all deprecated calls
Task-number: QTBUG-21609
Change-Id: I284587517537c7be4229d99637c6907d7e7652a5
Reviewed-by: Christoph Schleifenbaum <christoph.schleifenbaum@kdab.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Instead of having QGLPixelBuffer be a unusable stub implementation we
deprecate it and implement it in terms of QOpenGLFramebufferObject.
Framebuffer objects are anyway the recommended replacement for
pixelbuffers in modern OpenGL, as the context switching overhead is
avoided.
Change-Id: Ia220c358ee92813e87981d297c51d84525010322
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
On Mac, highlighted text remains the same color as normal text. That's
not the case for menu and combobox items.
Change-Id: I3efe2547413f77921524b833b4b10cf2d533e110
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
We resort to custom Qt rendering when selected since it looks better
than using Carbon's HITheme engine.
Change-Id: Iefcc2bebb1ba58d0d99f078c4960c5126542f637
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
These function were replaced by QMetaTypeFunctionHelper class that is
a bit more flexible.
Change-Id: I5aee2089f543b6e05c4d65a688fa8ca721f8abe0
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
MSVC2008 tries to see if a container is fully ordered.
This causes a build failure when operator< is not defined
between the elements in the container (but is instead defined
between the element type and the item passed to std::lower_bound).
Change-Id: I093d6f67f64a247f0aae763acb35ee73614f4208
Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
QPostEventList is kept sorted by priority; std::upper_bound
is used to insert a QPostEvent in the right place in the list.
Turns out that MSVC2008 is a bit too picky and tries to see if the
list is actually ordered. This causes a build failure as there is
no operator< defined between two QPostEvents (in fact,
an integer -- the priority -- is passed to std::upper_bound).
Work around this issue by defining operator< between two QPostEvents.
Change-Id: Ie3562dd0cc7253e25fc988b25d566d9d9e9fe62b
Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
1) It is not strictly necessary since the method argument type is already
designed for clickable here. Thereby, it is just a duplication.
2) It is not done inside the removeNativeEventFilter, so it is not consistent
now.
3) Similarly, the following documentation in 4.8 did not have another explicit
reference to the documentation of the EventFilter typedef:
EventFilter QCoreApplication::setEventFilter ( EventFilter filter )
Change-Id: I45fd42a03b16fd173fca9d734a7771e804d577d8
Reviewed-by: Jerome Pasion <jerome.pasion@digia.com>
Reviewed-by: David Faure <faure@kde.org>
Cause Qt5's source file use "cpp" as suffix
and automoc use "cpp" as suffix too.
Keep it with same feel better.
Change-Id: Iba3f8ffb5c3dbf003a9e1aacf3706089c1b6765c
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Before we would draw dock widget handles on windows classic
and no handles on XP, Vista and Windows 8. This would make it
very hard to make a consistent look and feel across those styles.
I also added a workaround to avoid the redundant toolbar borders
when they are used outside of the toolbar areas.
Change-Id: Ib703453677fcac8d51b2180abe45527297af0a80
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
It is time to clean up some of our legacy code. These styles have
not been actively maintained for a long time and I think it is safe
to say that they should no longer belong as part of the default
distribution of Qt. We dont support any platforms based on CDE with
our source packages.
Note that even if we are removing these styles from the default
distribution of Qt, applications that depend on them
will still be able to bundle the existing (and unmodified) styles
along with their own source code as we are not breaking compatibility.
Change-Id: I1709630c20ba8e8088cd01628628d86856db57a4
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
The format of the PPS object has changed, it is now prefixed with
[n], so be more leninent in parsing to support both formats.
Change-Id: I54eb1502a6231acf2dd121531539435c59d7d668
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Fix some typos. Fix some sentences to make them clearer to understand.
Since both of them are widgets, there's no Qt4-centric info.
Change-Id: I5c3c4a23621505d47b00cf466e4daad9763c3076
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Fix warnings:
Change-Id: Ia68607f72087c0085e528fee0e6270b80692e389
warning: enumeration value 'SynthesizeMouseFromTouchEvents' not handled in switch
warning: suggest parentheses around assignment used as truth value
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
When a QLabel was displaying rich text, the raw html was being returned as
accessible name. Now the plain text is returned.
Task-number: QTBUG-27302
Change-Id: I169d5eff527a0aef810af11d7712a362148974a5
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
There is no outline function in Windows CE,
so add a failure to inform the user that, he
needs to switch the font rendering.
Change-Id: Ieceb68bec49a88c7f7eee99f1c62cf1701da6aec
Reviewed-by: Miikka Heikkinen <miikka.heikkinen@digia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Set fd so that it gets closed in QPdfPrintEnginePrivate::closePrintDevice()
Task-number: QTBUG-14724
Change-Id: Ic9c02c92b07ece92911d7bc3eac79689b4864253
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
This seems to fix the ability to run a Qt app across a forwarded X
connection (e.g. ssh -X).
Change-Id: I2b7a26985cf11107e69b303337a5fbb369e38c9e
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
Reviewed-by: Uli Schlachter <psychon@znc.in>
Although not widely used, the raster operations from Qt 3 were useful
and some of them were already implemented, this brings the rest of them
back for those who need them.
Change-Id: Id538611eaaba9be3d39bf2dd33b6c532f5d4aba8
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
QMetaType has a way to "switch off" some types from the build.
QtMetaTypePrivate::TypeDefinition<T>::IsAvailable is defined as false
for all unaccessible types. Sadly that information was never used by
gui and widget handlers. The patch implements it.
Change-Id: Ie5835be4c88cfbbca8a4e9199e31ddfc20cae190
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Reviewed-by: Jing Bai <jing.bai@digia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There is another iterator called it at the top of this function.
Change-Id: I8addf47d66f098c5d811eb54674525eddb4d4a1f
Reviewed-by: David Faure <faure@kde.org>
is_signed/is_unsigned check whether the type argument is signed
or unsigned, resp., and will be used in QFlags to select the
correct underlying integer type.
not_ is used in the implementation of is_signed
not_c is for completeness (version of not_ that takes a bool
instead of an integral_constant).
Change-Id: I77cc445e8c6cf3181336505c9c13478fba3e7890
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
An int is not enough to hold all values of a 64-bit enum, so
don't allow to compile QFlags for such enums.
Change-Id: Ia291c4a5e2160dfd6d5a35e61a7f71ab571b9d9a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The back-end might report screen orientation changes at any point and we
record it in screen.d->orientation. However QScreen::orientation()
returns the orientation filtered according to the mask.
Changing the mask sends a notification to the back-end, which might send
another update as a result of a possible subscription to system services
(accelerometer). However on platforms where no subscription is required, where
the platform plugin ignores the mask and always sends the latest orientation,
we should "simulate" the update by updating the filtered orientation according
to the new mask. The function is cheap to call as it won't emit any signals
unless the orientation actually changes.
This patch also adds missing flush() calls after handleScreenOrientationChange
calls in the tests to ensure that the (synthetic) window system events are
actually delivered to QScreen/QGuiApplication.
Change-Id: Iebdd050f947e658ff5bc388629aa4cb31ab497fe
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
Reviewed-by: Laszlo Papp <lpapp@kde.org>
The window orientation changes are handled through a generic QPA plugin and
sends an initial window orientation change event on startup for the initial
orientation and subsequent updates when the device orientation changes.
Unfortunately the plugin was never loaded after the introduction of the new
plugin loading mechanism, because the factory key ("MeeGoIntegration") was
not provided to the plugin registry.
Change-Id: Ib7324561993abfbac1dccb0839aab0a8e870d231
Reviewed-by: Jocelyn Turcotte <jocelyn.turcotte@digia.com>
Reviewed-by: Laszlo Papp <lpapp@kde.org>
Thomas McGuire
Rafael Roquetto
Pierre Rossi
Kai Koehne
Geir Vattekar
Frederik Gladhorn
jian liang
Marc Mutz
Gabriel de Dietrich
Stephen Kelly
Pasi Matilainen
Samuel Rødal
Joerg Bornemann
Jędrzej Nowacki
Giuseppe D'Angelo
Miikka Heikkinen
Laszlo Papp
Yuchen Deng
Jens Bache-Wiig
David Gil
José Millán Soto
Andreas Holzammer
Teemu Katajisto
Shawn Rutledge
Andy Shaw
Marcel Krems
Thiago Macieira
Simon Hausmann