The assets file engine is created before main() is called, thus
prepopulate was called before main(). In this function, we created
a QDataStream which internally created a QBuffer which inherits
from QObject. This caused the main thread pointer to be initialized
early, and the old "QApplication is not on the main() thread" warning
and crash returned. The fix is to prepopulate the first time the
file engine is used instead.
Task-number: QTBUG-37444
Change-Id: I2c6e5f1a8ca88b5dc7d8e145fffeb7587dc0e623
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
When we scroll, the keyboard will change position relative
to QScreen, even if it appears to stay put. For that
reason we also need to update the keyboard rect after
doing a scroll.
Change-Id: I9bda2ab5b5e4970f488d3e69e44cf58e273f8fcd
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
The variant taking x, y, z has been broken in all Qt 5.x releases.
This is now corrected by making the GL2 paint engine capable of applying
a Z translation.
Task-number: QTBUG-31156
Change-Id: I119566e9e9577f6cdf7e8bae56cac1e34995e622
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
Implement PdmDevicePixelRatio.
Task-number: QTBUG-36419
Change-Id: I4d2822d01e08c80e9b829d5f524e8e761e694fba
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
The POINTBLOCK struct in qregion.cpp is badly defined. By using ints
as the base storage of the internal array it forces not only bad casting
from int to QPoint, but also allocates an array four times too big.
This patch changes to char, since a char pointer may alias anything and
this gives the right size of the array.
Change-Id: I608eaf39ac7306c71314a139bed6e2352249c0ab
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
As soon as we receive an event not related to dragging
a QDockWidget out of its area (or a similar use case),
we no longer need the forward window.
Task-number: QTBUG-37265
Change-Id: I310e9cc2cf099c76e7a88427826d4b97ca0cd9b9
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
These internal and obsolete hooks are no longer used. The correct way
to provide platform dialogs in Qt 5.x is to implement a platform theme
plugin, as was done for GTK+ 2.x.
Change-Id: I3f1474fbf760130106b3c47173eaedd2f1a919bf
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: David Faure <david.faure@kdab.com>
Show how use QCommandLineParser with additional
parameter checking for custom options and positional
arguments.
Also explain how to display help in GUI applications.
Change-Id: I03513e09b7dd5b150259593da0af2ef2a281cab2
Reviewed-by: David Faure <david.faure@kdab.com>
it's QT_INSTALL_CONFIGURATION, not QT_INSTALL_SETTINGS.
probably the most useless of all properties, but anyway.
Change-Id: Ia1e739ab4ed59e2bcd3199914caed2b3db9070ee
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
It could be that the channel has its reply already reset to 0, while
the protocol handler thinks the reply is still active, which might
lead to weird behavior including hard to reproduce crashes.
Task-number: QTBUG-37424
Change-Id: I89b65d34caaa546a343edc2ee205aa76425de88f
Reviewed-by: Richard J. Moore <rich@kde.org>
The 4.8 compiler in the Android NDK sometimes produces uncompilable
code for armv5. In change b4252802b3
we reintroduced thumb instructions on armv5 with a work-around for
a particular case of this bug. However, the compiler now breaks
for Qt Script. There's no time to try to find a work-around for
this case, so we need to disable thumb again. It actually also
seems that this is completely broken with this compiler, so it would
probably be best to keep this disabled until the compiler is fixed.
Especially since armv5 will become part of the CI builds now, so the
compiler bug can end up blocking critical changes in the future if it's
re-enabled :/
Task-number: QTBUG-37376
Change-Id: I9e1b050ce596717ba1fa7ec2f5e8a3ce3581a3af
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Change-Id: Ia2b34e3ee6954342501a805f9e047cb90078e9d5
Reviewed-by: Martin Smith <martin.smith@digia.com>
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
The old code tried to check the 2nd parent for inheritance from
QAbstractItemView. This also triggers for line edits on a QDialog parented
on the item view. Introduce convenience function that checks for
top levels in the chain.
Task-number: QTBUG-37504
Change-Id: I932f8efdb4764e9b1eea84c802bf7e8718338e1d
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
By placing debug messages in shared memory, the application can share
debug messages with winrtrunner (or any utility which passes -qdevel
to the app and opens the corresponding shared memory area).
Task-number: QTBUG-37308
Change-Id: Id0e04cfd5f0f701d552a448f780788c7cbf9b438
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
__IPHONE_NA was previously defined to 9999, but with
iPhoneOS7.1.sdk it is now undefined. From Availability.h:
"__IPHONE_NA is not defined to a value but is uses as a
token by macros to indicate that the API is unavailable"
This causes Qt to evaluate QT_MAC_PLATFORM_SDK_EQUAL_OR_ABOVE
to true when given __IPHONE_NA as argument. And then the
build will fail.
Change-Id: I11f1d0285329d90c633c00c0c4d446ef5cd8089b
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
This simplifies the compiler control semantics by always using the
packaged compiler if it is available. With no packaged compiler, the
service is assumed to be running if the directory structure is present.
Use of a packaged library can be overridden by setting the environment
variable QT_D3DCOMPILER_DISABLE_DLL to 1.
When the runtime compiler is used, the source will no longer be
logged, and the compilation output will no longer be cached.
Change-Id: Ib07f517e7043d7785bdfa9da55abd34df518eeed
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
Following Apple's documentation, the first step to identify a
bundle is to check if it has a known extension. Currently, it's
done using string comparisons ored in an if statement. The list
is not complete and new types, whether provided by a system update
or other means, can't be detected.
The new approach is to use Uniform Type Identifier which queries
the OS directly to check whether the extension conforms to
kUTTypeBundle. That includes e.g. applications, frameworks etc.
Task-number: QTBUG-31884
Change-Id: Ief73a83904adf27ccb71b0070e67cba081d1fd4a
Reviewed-by: Jake Petroules <jake.petroules@petroules.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Migrating this is out of scope for the time being.
Change-Id: Iac4a98d8db8e132a6ffa28075548fe2af76637fc
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Change-Id: I86341ccbd0251a9c858a5e1a9b7ea291d73cda01
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Reviewed-by: Kai Koehne <kai.koehne@digia.com>
Based on a patch by Marcin Juszkiewicz <marcin@juszkiewicz.com.pl>,
licensed under CC0 (aka Public Domain) or BSD license.
Change-Id: I60815d6893c7a9d2873864ff626b865881ec5ee9
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
testcase.prf cannot be loaded from pro file for various reasons,
see qtbase commit history for details.
Moved runtime testdata logic from pro file to testdata.prf, and
thus made is reusable in other test cases as well.
Change-Id: I500d08dc4951e4eda862071e4ddd3e0f6de8c3d2
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
A key event with no unicode would get a text of "\0" instead of
the empty string. This would lead to various widgets interpreting
the Key_Back event as text input, meaning that it would be grabbed
by the widget instead of being propagated.
Task-number: QTBUG-35784
Change-Id: Ibdb0e491572e41dd1aaf3b03ae1a780731f0559a
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Update the 10 year old FIXME hack to allow Key_Back (which will
close the dialog on Android). Also clean up nearby code that has
been dead for seven years.
Task-number: QTBUG-35784
Change-Id: I609858afb2caefe7025e421406288ae56717fea5
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
In case of a SUBDIR qmake project, it tried to cd into a directory
that did not exist yet. I needed to run qmake twice to get it working.
This fixes it.
Change-Id: I6d322e9a7c96a9d82df77b9ba5f19711a8180ed0
Task-number: QTBUG-37429
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Seen on Windows:
XPASS : tst_QGraphicsAnchorLayout::stability() QCOMPARE(sameAsPreviousArrangement, true) returned TRUE unexpectedly.
tst_qgraphicsanchorlayout.cpp(1399) : failure location
Move check into loop with descriptive message on fail, do not error
out if it actually passes. This now typically prints:
XFAIL : tst_QGraphicsAnchorLayout::stability() The layout has several solutions, but which solution it picks is not stable ( QRectF(30,0 0x10) != QRectF(30,0 10x10) , iteration 3 , item 4 )
Change-Id: Iae8553dbbcedeb70d5d672e3cefbd1f06a63d27d
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
Includes a one-pixel fix for widgets, where the menu item text
would be one pixel too far away from the vertical gutter.
Task-number: QTBUG-37451
Change-Id: I2539233814a427c5bed66a5c07584717eb3df418
Reviewed-by: Jens Bache-Wiig <jens.bache-wiig@digia.com>
Reviewed-by: Mitch Curtis <mitch.curtis@digia.com>
Change the code so that it doesn't force application windows to full
screen when the no-fullscreen parameter is used in the QT_QPA_PLATFORM
environment variable.
Task-number: QTBUG-37456
Change-Id: If9612cc2ca69f829d45e7fd4b83338c3a41cf986
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Fabian Bumberger <fbumberger@rim.com>
Use QPointer to make sure that we are not trying to destroy already
deleted object.
Task-number: QTBUG-35702
Change-Id: Ib746996787488e636f25e6ea5be0571607ee2ded
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
The tasks might not have run yet at the time of the QCOMPARE, so we need
to acquire on the semaphore in order to ensure that this is the case,
just like in the previous testcase.
Change-Id: I1da72bb07c2f53760b3bf912fc26aaf10ed18d48
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
QXcbDrag inherits QBasicDrag which takes care of calling
updateAction() when needed, we don't need to call it explicitly.
Task-number: QTBUG-33057
Change-Id: I754408f74f56de36ace8ffa40a61bd7c64619899
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Changed clip rectangle handling from bounding rect to region. Now also
sub region rectangles are taken in to account.
Task-number: QTBUG-37199
Change-Id: I9e09376e2c6d3fee8f85db753295a6138a03096f
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Janne Koskinen <janne.p.koskinen@digia.com>
Reviewed-by: Louai Al-Khanji <louai.al-khanji@digia.com>
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Add a screenChanged handler slot to QWidgetWindow,
which calls markDirty() on the backing store with
the BufferInvalid and UpdateNow flags set.
Update CocoaBackingStore to create a new buffer on
window devicePixelRatio change. Use the
QCocoaWindow::devicePixelRatio() implementation instead
of a duplicate implementation in the backing store.
The plan is to replace this implementation with one
based on QUpdateWindowRequestEvent for Qt 5.4
Change-Id: I8e521c53df4ac90815613e730fe821996334721f
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
xcb_randr_get_screen_resources() and xcb_randr_get_screen_resources_current()
do basically the same, but for one case: if nothing has queried these
information ever before. So if an application is the very first client ever
to connect to an X server it may just return nothing. A call to
xcb_randr_get_screen_info_reply() will then cause the X server to allocate the
needed information and send an update notification, resulting in a second
QXcbScreen being created, but the other one is still around and probably used.
The behavior I observed with a simple test application was that the application
was not visible on the screen when it was the first client ever on the X
server. Killing the application and starting it again made it work just fine.
Change-Id: Id64f267e8ebcfa5b39d21d98307170a09e5169df
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Gatis Paeglis <gatis.paeglis@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
