The QNX toolchain can use Neon on ARM and SSE<X> on x86/x86_64.
Change-Id: I36c61fa12b65d806b3cc60a0aefcb63964f9ab7e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The generic input plugins do not set the window parameter. So,
use the same technique we use with mouse event handling to determine
the window to deliver the event to.
Done-with: Johannes Zellner
Change-Id: I950c0ad2f330dccfdcc41b8d01f62cd39902bc9c
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
gitorious.org's IP no longer resolves back to gitorious.org.
This fix is temporary, again.
Change-Id: I85b5fe1c5e603d23dd3226b843ef42165d4c417b
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
The current cursor implementation can be a bit hard to read
without hints about which methods are overriden.
Change-Id: I3376890a13be46e1ece03d1442dd5a15ccd61382
Reviewed-by: Johannes Zellner <johannes.zellner@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
qdatetime.h uses std::min/max and on Windows windows.h (or some subsequent
header file) may under certain circumstances define min/max as macros.
The easiest way to prevent the windows header files from doing that is to
define NOMINMAX in the place right before windows.h is included. The other
way is to define min and max to min/max themselves to prevent windows.h
from doing its evil thing.
If a user of Qt (WebKit in this case) chooses the approach of defining
min/max to themselves and then includes qdatetime.h, then a subsequent
inclusion of windows.h doesn't work because qdatetime.h undefines min/max.
We should not enforce the type of workaround needed, therefore this patch
removes the workaround from qdatetime.h and requires user code that
happens to include windows header files before qdatetime.h (seldom case)
to choose either workaround.
Change-Id: I7347eec7369491a065e894cff557004e069453d5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Frequent calls to editorForIndex/indexForEditor are very slow because of an implicit
conversion from QModelIndex to QPersistentModelIndex.
This fix allows to avoid unnecessary conversions when there are no open
editors (most common case)
Change-Id: Ic072880c9f33a43a20b2a61a42c3ba215c5c33cb
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Since local thread storage is used we need to turn this on for the xlc
compiler with the -qtls flag.
Change-Id: Ib40ec87edada56a062b0c72b7d47b38a6d0b5b13
Reviewed-by: Teemu Katajisto <teemu.katajisto@digia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
This extension doesn't work for e.g. default arguments
in function declarations.
Change-Id: I32b7afa6e01b6af55fb2409179b4fd94cb04cd8d
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
If we have it in different functions, then different out-of-line
implementations could be selected for each object file, resulting in
invalid states. The error I caught was when wrapper.cpp was compiled
without tracking and, therefore, did not place a call to
internalSafetyCheckAdd. However, it called an out-of-line copy of
QtSharedPointer::ExternalRefCountWithCustomDeleter::create, which did
set the deleter to remove the safety check.
Therefore, keep everything in one function.
Change-Id: Ib2c6a606699db49d102704bccdd331ec22a8bd78
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Let the constructor initialise the "value" member.
In the case of create(), which already initialised "value", simply
merge the two functions for more readability.
Change-Id: I5638b3d42af3d0f5988f815e0f91d591fa1897a8
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Completing the work of the previous commit: we don't need separate
classes. Merge into the main class's body.
Change-Id: I2f89b34cb6b7f5f9e8d8b809bebd86656f458644
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
The basic class existed for legacy only, when internal reference
counting was a goal. Since it isn't anymore, we can remove the
distinction and simply merge the two classes.
Change-Id: Ib7a1c4158a8d71e71fa6afa447938b8b85ddae87
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
The two classes are QtSharedPointer::ExternalRefCountData and
ExternalRefCountWithDestroyFn. The split existed because of what Qt
4.5 did before custom deleters existed: the ExternalRefCountData class
was a virtual class that contained a destroy() virtual, which was in
charge of deleting the data or returning false if it didn't.
Turns out that virtual classes was a mistake. This commit
de-virtualises them -- we couldn't do it in Qt 4 because of binary
compatibility. This saves us one pointer-size in the size of the
private, plus the fact that fewer symbols are required (there is no
virtual table to be initialised).
Additionaly, since a deleter is always stored with the reference
count, we don't need the split between the two classes anymore.
Change-Id: I1cd9400561dcee089a406a57bd856b1730f18afc
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
This allows a QSharedPointer to be used in contexts where the class in
question is still forward-declared. This produced a warning in Qt 4 due
to the expansion of the template, even if there was no chance of the
pointer being deleted there (because the reference count could not drop
to zero).
Now, not only is the warning removed, but you can actually have the
reference count drop to zero in a forward-declared class and it will
do the right thing. That's because the deleter function is always
recorded from the point of construction and we're sure that it wasn't
forward-declared.
The unit test for forward-declarations had to be rewritten. The
previous version was passing only because the QSharedPointer object
was created under the "tracking pointers" mode, which causes a custom
deleter to be used in all cases.
Task-number: QTBUG-25819
Change-Id: Ife37a4cea4551d94084b49ee03504dd39b8802c1
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
$VAL just happens to work in resolveDeviceMkspec because it is
set in the parent shell environment when the function is called.
Change-Id: I67350f2a9e790cc7eca2a73ef6a4a0d7f09b8d3c
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Follow-up to 6a51062e99 which did this
for win32-g++.
Change-Id: I3ba0dd8ffca46853844b55b16dc92270fa8a623a
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Separate dsp and dspr2 handling. The configure script allows to disable
them separately and with this patch it is possible to compile a dspr2
only libQtGui.so.
Change-Id: Ifca583c9b46a25c93751967a31ac77eafc5d51e4
Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
Follow-up to 0074cc5d34.
The configure script can be used for cross-building for Windows on
unix.
Change-Id: Ie7f9d0ff308ad5763cdf7b2664fa255e89bd5013
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
Since we only support XCode 3 and later, then all of the legacy code can
be safely removed.
Change-Id: I9be8555aaa62c716b2277c2b97f41aa02d27ef13
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Commit d839564c94 was incomplete. It
added the Q_CORE_EXPORT macro to qmalloc.cpp, but the qMemSet and
qMemCopy function bodies are in qglobal.cpp.
Change-Id: I24ee44f04365d8dbdf3f1c0f22b6a72cae9f96bb
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
SSSE3 was first available on the original Intel Core 2 processors, so
add the "Merom" codename. SSE4.1 was available on the 45 nm shrink of
those processors, codename "Penryn", not on the next architecture.
Change-Id: I5fd92db62aa409b7f4e46f9b24d960519177f811
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Modify configure.exe to run some configure-time tests and check if
the SSE and AVX compiler features are supported.
The tests themselves required a bit of changes to compile with
MSVC. The include in sse4_2.cpp was wrong. And for whatever reason, it
didn't like the volatile variables, which GCC, Clang and ICC have been
happy with. This should produce no effect in compilation, though: even
dead code must be syntactically correct. We're not running the output.
Change-Id: Ibe5d0904a378a7efed853c7215f88a2ddcefb1b3
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
The old macro was leading to symbol clashes.
Change-Id: I090c511d4090bc96fc6c88537fae7bbe7f143b6c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
The getUrl: and appleEventQuit: handlers are only called if we register
them with the NSAppleEventManager. The Cocoa documentation says the best
place to do this is in the applicationWillFinishLaunching: delegate
method, so add this method and move the code from
qcocoaeventdispatcher.mm to there. Since QCocoaApplicationDelegate is
only used when AA_MacPluginApplication is not set, we do not need to
check again in the delegate code. Be sure to remove these event handlers
when shutting down the application.
For the getUrl: handler, send file open events when receiving this
event. This restores Qt 4 behavior.
Remove the qDebug() from the appleEventQuit: handler.
Change-Id: Ibcbdd541695176e3d236366d4d541e4811882d6c
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
To avoid a QtWidget dependency, we should use QGuiApplicationPrivate
isntead, but instead of storing, we can use
QGuiAppliationPrivate::instance() instead.
Change-Id: If3f63fee804b7ad32fe8d612bf70c051b70f54c8
Reviewed-by: James Turner <james.turner@kdab.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
This is a cherry-pick of 0ba850c7a2dbccb8dd6aa1664679bda6cce95065
When the mouse button is released at the end of a drag, Cocoa
may simulate an extra mouse moved event. However, the state of
the buttons when this event is generated is already 'no button'.
This leads to some failsafe code canceling out of the drag state
and when the actual mouse release event is finally processed, the
header drag state has already been exited and the header drag
fails.
This patch disables the failsafe code on Cocoa and makes header
dragging work when the mouse goes outside the bounds of the header
view.
Task-number: QTBUG-14179
Change-Id: Ia9fd1ac79f9e7b4b90d3e160298c53d65fb171d3
Reviewed-by: Thorbjørn Lund Martsum <tmartsum@gmail.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
QSharedPointer doesn't work like the other automatic template metatype
declarations because in some cases T* is declared as a metatype, but we
are interested in QSharedPointer<T> (eg QObject*). In other cases, T is
declared as a metatype and we are interested
in QSharedPointer<T> (eg char).
In particular the macro used before this patch was attempting to get the
metatype id of the element_type using for example qMetaTypeId<QObject>()
instead of qMetaTypeId<QObject*>(), which did not work.
Similarly, the variadic macro driven test is no good, because it was
testing QSharedPointer<QObject*> instead of QSharedPointer<QObject>,
so that is removed.
In the end, the only thing we can sensibly automatically declare as
metatypes are QSharedPointers to QObject derived types. That is also
the type that makes the most sense in a QML context anyway.
Change-Id: I13dd40147e2e6bedf38661f898102abaaaa96208
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Some test functions that only test QDate and QTime were in
tst_qdatetime.cpp. Upon moving these into tst_qdate.cpp and
tst_qtime.cpp, there were already some similar tests so I
consolidated them.
Change-Id: I5f8758bf8b4804ae9d3a482f49d21de9f7a1dc03
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Clicking the area just below the close button when Aero is enabled
reveals duplicate caption buttons. DwmDefWindowProc returns hit results
for DWM caption buttons but DefWindowProc may also return hit results
for non-DWM caption buttons.
To resolve this issue, if DefWindowProc returns a window button hit
result then we just use HTCLIENT (the client area) as the hit result
instead.
Change-Id: Ia741ce4f9aa944109d8de54c2f84009f5ea1883f
Reviewed-by: Miikka Heikkinen <miikka.heikkinen@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
The distance field generator was always returning a 64x64 pixels image.
It now returns an image with a variable width (width of the represented
glyph) and always a height of 64px.
Change-Id: Id5f11a50a8031ebca10cd4803adf179ccde6db26
Reviewed-by: Kim M. Kalland <kim.kalland@nokia.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Most fixes are simple and quite obvious. The ones more involved are
the ones to QArrayData, which had probably not been compiled with
strict iterators thus far.
Change-Id: Ic4ff84c34fd9a04fd686fecaa98149b1c47c9346
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
We need to let the QGuiApplication determine whether quitting is appropriate
based on whether there are visible top level QWindows after the last top-level
QWidget was closed.
This solves the issue raised here: http://thread.gmane.org/gmane.comp.lib.qt.user/1880
The transientParent is the QWindow equivalent of parentWidget on QWidget, so the test
in QGuiApplication::shouldQuit is similar to the one in QApplication::shouldQuit.
Change-Id: I500eff8d5887f24415180134b3a4be3c630a896f
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
This also changes the qtdeclarative-specific QMetaObject::activate()
overload to not take a methodOffset argument, since it's no longer
needed.
Change-Id: I4f7ece9f43339f3327419598c032e48fb37b97f0
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Michael Brasser <michael.brasser@nokia.com>
Reviewed-by: Aaron Kennedy <aaron.kennedy@nokia.com>
The extension doesn't work outside of function scopes, so a
function declaration such as
void foo(const QString &str = QStringLiteral("bar"));
would fail on certain gcc versions.
Change-Id: I2971301f2859edd3fc81b95dfa5a7c15f29e395c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
The OpenmaxIL video_render component uses the dispmanx layer 0, so EGLFS
should use at least z index 1. Otherwise the video_render would conflict
with the UI and thus overpaints it.
Change-Id: I3bed23567fa8c4399207289c6ef952c5a5e0d503
Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
Reviewed-by: Donald Carr <donald.carr@nokia.com>
See https://github.com/raspberrypi/firmware/pull/32 for more information.
Change-Id: I51bb532336ed069cde938540cd962721b1a72adb
Reviewed-by: Johannes Zellner <johannes.zellner@nokia.com>
Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
The QSet<int> is a more expensive container to use and create, so
it should be avoided.
This is source incompatible compared to earlier Qt 5 for
QAbstractItemView subclasses which reimplement dataChanged, but this
patch changes nothing compared to already-present SiC compared
to Qt 4.
Change-Id: Id95391dfd62a0a7f487a8765790b007badefb937
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Containers are auto-registered and use normalized names.
Change-Id: Id65c3940401f69436929220e1f6a971135e147ed
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
set WS_EX_LAYERED | WS_EX_TRANSPARENT style if the window has
WindowTransparentForInput flag.
Change-Id: I6b0dcf35abb5fc63c800439e9b81ace1070dcad4
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Added three new methods, which are meant to be used internally to
configureapp.cpp:
- int platform(): returns an integer representing a platform
- QString platformName(): returns the platform name string to be used when
displaying the license agreement.
- QString qpaPlatformName(): returns the value to be defined as
QT_QPA_DEFAULT_PLATFORM_NAME.
Currently supported names are Windows, Windows CE, QNX and Blackberry. Default
is "Windows".
Change-Id: Ifa4d1b9c02cda956be9becdf8db195d3d494f1d5
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
When trying to detect the target architecture, configure.exe will look for a
file called 'arch.exe'. However, in some situations such as when
cross-compiling on a Windows host to a Unix host, and depending on the
toolchain being used, the output file generated by config.tests/arch may
simply be called "arch" instead of "arch.exe", causing configure.exe to fail.
This patches configure.exe to handle both naming schemes.
Change-Id: I5798f716d732388c707564d4d45c4887ab3d3d9f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Re-enable application termination as it was in Qt 4.
QApplication::exec() must return to main() so that the destructors in
main() are run.
The QApplicationPrivate::canQuit() function from Qt 4's
qapplication_mac.mm is gone, so bring back this function in
QCocoaApplicationDelegate instead.
Change-Id: I1c21894d59061687c36ab49bcb2e4e3ae0752fa4
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
