This improves iterating over /usr/bin with QDirIterator by more
than half, from 36 to 13 milliseconds.
Change-Id: Ib3a9271c3a6f81c1ea3c21d012c875c7e9bad2ad
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
By depending on a simple variable instead of a more complex expression, we
can more easily use this file from other contexts.
Change-Id: Ib6ce1e2537f3e7ac19d25ad33454f25e85f71040
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
.qmake.conf (and previously .qmake.cache) already does that for us.
Change-Id: I06cc01fa45921d7bd66dda7a0f88729faeff37bd
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
... so we can test those functions with host and cross compilers.
Change-Id: Ifebfdac54580633c797f77b139514cf9d66edd8c
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Native widgets have a native window each, so QPA plugin sends enter and
leave events for associated QWindow whenever mouse cursor moves from
one widget to another. QWidgetWindow had no context to interpret these
events as moves from one widget to another, since they were sent
separately. This resulted in leaves and enters for each widget in
parent chain, when only the bottom child should have gotten them.
Fixed by peeking into window system message queue when handling leave
in QWidgetWindow and retrieving the entered window from queued enter
event.
Also provided a convenience function that QPA plugin can use to
ensure both leave and enter events are in the event queue when
moving from one QWindow to another.
Task-number: QTBUG-27550
Change-Id: I74fec0ac90f6848495c2392c5f7e41624ad8aea2
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
Instead omit the whole test when Q_COMPILER_INITIALIZER_LISTS
is not defined.
Change-Id: I34017484a027b95a2677e1c4cb9231fa2aeb5680
Reviewed-by: Caroline Chao <caroline.chao@digia.com>
Instead omit the whole test when appropriate.
Change-Id: Iebd569676cc7b2f8fe4a1d272107e092410d397b
Reviewed-by: Caroline Chao <caroline.chao@digia.com>
The goal is to cut as many QWidget dependencies as possible and make
stylesheets eventually work for the desktop components.
Change-Id: Ib4aa47af07379fc39fd6df1961e113d03df6df35
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
These containers don't make sense and will just result in no action
being taken (all items added will simply be discarded), but it
shouldn't crash due to a division by zero.
Update the documentation to explain the fact.
Task-number: QTBUG-27339
Change-Id: Ib9acf5c0a9a826e6853e7beaf5e56511fde98dc6
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Ian Walters <ian@walters.id.au>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In the event dispatcher native events will be processed in a tight loop
to drain the queue. IO events and timers will be postponed.
Change-Id: Ic2c06ed182027289eb5e7042fbae99efbd01ea27
Reviewed-by: Fabian Bumberger <fbumberger@rim.com>
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com>
Documentation has been updated to reflect changes in Qt5.
Change-Id: I378858cf61f4bf62375c30c3903818d754dadcf0
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
The IA-64 C++ ABI does not encode the return type for non-template
functions (QVector is the template, not the function), which means that
these two functions have the same signature:
Node *QVector<Node>::begin()
typename class QTypedArrayData<Node>::iterator QVector<Node>::begin()
[both are _ZN7QVectorI4NodeE5beginEv]
When linking compilation units compiled with different
QT_STRICT_ITERATORS settings, only one of the two out-of-line copies
will survive. Depending on the ABI, we may have a problem: the ABI can
say that a function returning a structure takes an implicit first
parameter, which a function returning a regular pointer doesn't.
Task-number: QTBUG-27277
Change-Id: I57a59e5a7c46f55faabfe85c073dca89d2a7bbf3
Reviewed-by: Jan Kundrát <jkt@flaska.net>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Instead omit the whole test when Q_OS_WIN is not defined.
Change-Id: I311538ce839353d4d5e83edfd12c68968fe61215
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
The problem is that isValid() should return true as
soon as we have a view widget.
That sadly means we need to verify that the model is still
valid whenever accessing it.
Change-Id: I9237528abf2f5c75a73382525103307e9ca15f05
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
Having them empty is the common style and is less noisy.
Change-Id: I30c0c4d297a7bdb373a57a636f4fdbc488a69fcd
Reviewed-by: Clinton Stimpson <clinton@elemtech.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Reviewed-by: Alexander Neundorf <neundorf@kde.org>
This is being changed because:
- The OpenGL paint engine in Qt only supports GL2
- QML2 only supports GL2
- QSurfaceFormat has a default value of 2
Applications that want to use GL1 on a QGLWidget will have to
explicitly request this format using QGLFormat::setVersion.
Task-number: QTBUG-27589
Change-Id: Ieb283ef7d6e15a29ec28ce7e4363dbf477decaa7
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Removing the lackey executable that needs QtScript to be built.
This was a qscript bastard that was able to run a client and a server
script. It's replaced by a C++ version with the same functionality.
Nice side effect: the two second wait per test row could be removed.
The client executable is now waiting for the server to be created.
Task-number: QTBUG-24142
Change-Id: I135b75abf6620d3b0af50dc226ea8c81c2bf4149
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
The dbus_get_version function was introduced in 1.2, so we'd need to
detect pre-1.2 by the absence of the function. But if we're going to
detect the presence or absence of any function, we might as well do it
on dbus_connection_can_send_type, which is the function we wanted anyway.
Change-Id: I6e17a3a8f1382c6a489490084f6e3f61aa5a1947
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
tst_QProcess::echoTest_performance() is not an unit test but a
performance test, so moving it from 'tests/auto/corelib/io/qprocess'
to 'tests/benchmarks/corelib/io/qprocess'
Change-Id: I796788534eafc5ca3b8d86c0ec46998285fd4b8f
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Qt 5 is the time to get rid of all the old inconsistencies in the raster
paint engine caused by trying to preserve the old X11 based
coordinate system where (0, 0) is in the center of the top-left pixel
instead of the upper left corner of said pixel. However, this was only
adhered for line drawing and path / rect filling, and not for image or
pixmap drawing and not at all when doing antialiased painting. By
defining the antialiased coordinate system as being the right one and
letting the aliased fill rules follow from that we finally end up with
some consistent behavior that doesn't lead to surprises and workarounds
in application code.
It is still possible for applications to get the old behavior by
setting the QPainter::Qt4CompatiblePainting render hint. This should
make porting easier for the few cases where an application relies on the
aliased fill rules we used to have in Qt 4.
Task-number: QTBUG-27500
Change-Id: If86b95e77d838ec83033d64af86632b9a73c74a9
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Don't add invalid ranges to the result. They will be removed
whenever d->ranges is processed for public consumption
anyway. For example, QItemSelectionModel::selection() calls
merge() with another selection. The merge() method removes
invalid ranges already. But the invalid ranges don't need
to be there in the first place, so this patch removes them.
A longer-term goal is to maintain d->ranges as an always-sorted list.
This method can be called with a vector containing invalid
QPersistentModelIndexes when those persistent indexes are made invalid
in between layoutAboutToBeChanged and layoutChanged. It's a normal
thing to happen and a case that should be handled deliberately.
Change-Id: I741ed9208d8a75644975c9e8d61f0d6d78e20576
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
QListView::currentChanged and QListView::selectionChanged changed to
use 0-based indexes instead of 1-based indexes.
Change-Id: Ie618970c9a37ec76156dbfab2e86685c3c8f374c
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
The linux test would actually fail because the qobjects created
could end up having the same address (create-delete-create...).
After an object is deleted, it's not instantly removed from the
cache of valid objects.
Instead it would stay in the list with it's smart pointer becoming zero.
This patch adds the missing null pointer check so we are always up to date.
Change-Id: Ia7be14741d4798c2b8e75cb7127298c73cf206ef
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
It's actually looking for the mkspecs (so it can read qconfig.pri to
get the Qt version), so give it exactly what it wants.
Change-Id: I2957b2d93a8837b8492d313209d45ff3ec01704c
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Previously recommended fonts sizes in pixels do not work very well
on the new high resolution devices.
Current UX documentation recommends certain point sizes for certain
roles. Those work well on both type of devices.
Change-Id: Ia4faa4cf2d753cab9b10f91d01dfef1450683272
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
It turns out that the calculation of mouse pos over the frame
strut (title bar) was wrong. When outside the content view, the
nsevent coordinates are negative, so to get this correct, we
need to calculate the height of the window above the content view
and use this information to get the mouse pos in positive
coordinates with origin window top left.
This bug was especially apperent with QDockWidget, as it became
almost impossible to dock a window under such circumstances.
Change-Id: I2faf6aab5e2aa0b4e217ea087ceec8c1b1e978bb
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
This is a new non-native style for Qt.
It is intended as a replacement for the now aging
Plastique and Cleanlooks styles.
Change-Id: I30c0518a69e4e3b8b2b05ee7d84c3a5a1f307578
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Adding QCheckBox, QComboBox, QLineEdit etc. into an HBox layout or grid
layout makes the layout look like a snake's trail. Fixed the positioning
of these widgets to make the layouts visually more appealing.
Updated qmacstyle auto-test accordingly.
Task-number: QTBUG-13635
Ammended-by: Gabriel de Dietrich
Change-Id: I89461f9aad68ea8488070ed06257b9b8e7f493c5
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Just like 'make docs' is used to build documentation, you will now have
to run 'make install_docs' to install it.
Change-Id: I57db53160ca91618784f4e39da0a47322c070208
Task-number: QTBUG-27590
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
The qmake function prepareRecursiveTarget can now be used both by the
existing logic in default_post, as well as future recursive targets that
will be needed as part of the modularization of documentation builds.
Change-Id: Ibc72c3e224cb57c9f1796de3b04fda9de663dbb4
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Changed it to check the Frame*Margin property instead of the
Block*Margin property as this was incorrect for a QTextFrameFormat.
Task-number: QTBUG-22173
Change-Id: I2c3066165fb592ed034874b1180593822859f933
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
that way we don't have to auto-generate code for that in the configures.
note that we now load qt_build_config.prf instead of just qmodule.pri,
which means that exceptions_off is set everywhere. we forcibly re-enable
them for testcases to minimize the deviation from default 3rd party usage.
testlib selftests are not qt testcases, so the one that needs exceptions
needs to enable them explicitly.
Change-Id: I1b9360bb11f2e80c92a2b63a7c45991ad17fda1b
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Now that we no longer re-use the animation code in gtk,
it is time to remove this strange dependency.
Change-Id: Ib672a9b110dfba08d79b8654e572994007675957
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
We also made an accessibility auto-test more stable, so that it would
not be influenced by this patch.
Task-number: QTCREATORBUG-7966
Change-Id: I6ce4e2361c8847aee5dd1cf5664c29cd2d83e38d
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
Using dynamic properties on the native interface is deprecated.
Change-Id: Ia3411780dad15af61d4805c0d9fabf00dba92301
Reviewed-by: Miikka Heikkinen <miikka.heikkinen@digia.com>
