Suppress message dialog of the test helper as does QTestlib.
Task-number: QTBUG-52714
Change-Id: I5efd7d72f77c7689500ecaccf46f1f9dfb312140
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
Allow sockets in bound state to wait for notifications.
Task-number: QTBUG-52449
Change-Id: I5c02bd61db68abca652312d4d59023963b05b3c5
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
Make the destructor of QDBusArgumentPrivate virtual as it may store
other types (QDBusDemarshaller). Suggested by Thiago.
Fixes a "new-delete-type-mismatch" error reported by AddressSanitizer.
Task-number: QTBUG-52176
Change-Id: I9ac19050840530ca9cae893b10093185d31e0448
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: David Faure <david.faure@kdab.com>
a != A || a != B looks suspiciously wrong.
Found, unfortunately, only by Coverity (CIDs 152247, 152249).
Change-Id: Ia7ec810dde2a3a33ae6fc5644a4eebf279dcffe8
Reviewed-by: Edward Welbourne <edward.welbourne@theqtcompany.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
Use the correct parent path for the renamed file when creating
QFileInfo. It must be a path to the parent directory, not to the
root directory of the model.
Modify tst_QFileSystemModel::setData() to test renames in subdirs
of the root directory of the model.
Change-Id: I69d9e3a937616857a81617791ed118af3f6eea05
Task-number: QTBUG-52561
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Coverity's CID 157687: QCache::insert, indeed, can delete (immediately)
the object we're trying to insert. While this never happens actually in
qlockfile_unix since we have max cost 10 and insert with cost 1, the
code does not look good and Coverity is not happy.
Change-Id: I16a428017bf86e151afe5256906e4cab1ef4044a
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
When IM hints changes (e.g as a result of transferring focus
between Qt objects), we sometimes need to reconfigure the keyboard.
And the way we do that is to create a new QIOSTextResponder that
matches the new configuration and tell it to become first responder.
And in that process we need to ensure that we don't clear the
focus object when the old text responder resigns. After all, it was
the one requesting the new IM configuration in the first place.
Since we set FirstResponderCandidate to point to the new text responder
just before telling the old one to resign, we can use that variable
to check if were in a first responder transition before clearing the
focus object.
Change-Id: I05bfb8180700a92a8f14f496044457583bcd38d3
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Take alignment into account in QFusionStyle::subControlRect().
Task-number: QTBUG-49068
Change-Id: Ia8f2d06c46b24761bff537247bbadd3323e41fa9
Reviewed-by: Alessandro Portale <alessandro.portale@theqtcompany.com>
QWindowsStaticOpenGLContext::doCreate(): Invoke
QWindowsOpenGLTester::supportedRenderers() only when needed; specifically,
do not unnecessarily call it when GLES/SW renderer is specified.
Amends change b4c8e15174.
Task-number: QTBUG-49541
Task-number: QTBUG-52693
Change-Id: I58d1c584d194f8e7fee35ee16b77575f86d3c22e
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Return pixmaps with device pixel ratio similar to QScreen::grabWindow(),
cf c0963486ce.
Adapt kernel tests accordingly.
Task-number: QTBUG-52137
Change-Id: I9ce276d5e2d87ae0d39c8639267d1ac283fed854
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
The old code incorrectly calculated visual region for a selection (specified by
QItemSelection object) in case of reordered / swapped rows or columns. It used
the leftmost and rightmost (upmost and downmost for vertical mode) logical
indexes directly. However some middle logical index (in case of reorder) may be
the leftmost or rightmost visual index. In such cases the repainting didn't
work properly. This fix first checks whether reordering / swapping is in use.
If it isn't (ie visual=logical) we use code similar to the old code. Otherwise
the new code scans all selected logical indexes, translates them into visual
ones and gets the correct leftmost and rightmost indexes.
[ChangeLog][QtWidgets][QHeaderView] Fixed a repainting issue when items had been reordered.
Task-number: QTBUG-50171
Change-Id: If6afabebf445cf32a8e0afe262b6ee149e7c4b2b
Reviewed-by: Thorbjørn Lund Martsum <tmartsum@gmail.com>
Avoid incurring the cost of converting the latin1 data in that case.
Several existing QString unit tests excercise the new code path.
Task-number: QTBUG-52617
Change-Id: I27256d9e7db34f09543e244a79d754ff7932f0d0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Mark the overloads taking an additional QVariant as \internal as (cf
QTextEdit) and the overridden functions as \reimp.
Amends change ab67a5b7b6, fixes warnings:
qtbase/src/widgets/widgets/qlineedit.cpp:1755: warning: Cannot find base function for '\reimp' in inputMethodQuery() /
Change-Id: I9949bf3850ab4200ee500cbe33a44faacecc30b0
Reviewed-by: Jan Arve Sæther <jan-arve.saether@theqtcompany.com>
Removes spacing in QGridLayout when the QWidgets inside of the QGridLayout
are hidden, by checking if the sibling is empty, thus duplicate
spacing can be avoided.
Task-number: QTBUG-52357
Change-Id: I45475e7b264f94ef3bec5f9a4b8cbaa1d53ec6dd
Reviewed-by: Karim Pinter <karim.pinter@theqtcompany.com>
Reviewed-by: Andy Shaw <andy.shaw@theqtcompany.com>
This modification enables to use notification feature of SQLite with Qt
SQL driver, enables to subscribe for notifications and also to remove
notifications. close() is added to destructor to unregister
notifications in case it is used in multiple threads.
[ChangeLog][QtSql][SQLite] Adding notification feature to SQLite driver
Change-Id: I8b98787f5214a406357646a98711a8ff6045a0dd
Reviewed-by: Andy Shaw <andy.shaw@theqtcompany.com>
Xcode does not recognize "<Group>" with a capital G, only "<group>" with
a lowercase g. As a result, paths of files within these groups are
calculated incorrectly. For example, dragging any external file into the
Xcode project would result in its leading slash being removed (while
still treated as an absolute path - broken reference). Furthermore, the
dropdown in Xcode displayed the Group location as an invalid string
instead of the correct "Relative to Group". This patch restores correct
behavior.
This fixes a regression introduced in
f09ec09c208c75a16abe05b6bb505a1fc58775a6.
Task-number: QTBUG-52701
Change-Id: I9af5360049a79e7958301e4090a9a542bab0af8c
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
The startDetached(const QString &, OpenMode) overload and the
QT_NO_PROCESS_COMBINED_ARGUMENT_START macro must point to the
start(const QString &, OpenMode) overload.
Change-Id: I7607fcb92b9f1ef3547a4a1aadc950532024225a
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Set totalScaleFactor on a pinch gesture (OS X) so that it works the same way
as in 'standard' gesture recognizer and corresponds to what we document.
[ChangeLog][QtWidgets] : QPinchGesture on OS X now behaves like on other
platforms: totalScaleFactor is the magnitude of the pinch and scaleFactor
is the delta for the current event.
Task-number: QTBUG-48138
Change-Id: I66b9a1df05cbe106b76aed8f15d900bcdd41fdb7
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
Adding strings in a single expression gets mediated
efficiently by QStringBuilder, where using += repeatedly
forces premature consolidation.
Change-Id: I6c1abb9c9801a016ce6c151cffdf8c7ac9502f4e
Reviewed-by: Edward Welbourne <edward.welbourne@theqtcompany.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
The given size is mostly ignored and the resulting pixmaps typically end up
in a QIcon, which will clobber the DPR anyways when returning a pixmap
for a given size.
Moreover, returning a message box icon with a DPR > 1 when
high DPI scaling is active and Qt::AA_UseHighDpiPixmaps is not set
causes the pixmap to become too small due to some scaling error.
Task-number: QTBUG-52622
Change-Id: I8aaaa97667d6c168040e19b7edad9dfb7517f70f
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
Try to find a window handle so that the pixmap matching the
device pixel ratio of the screen can be found.
Task-number: QTBUG-52622
Change-Id: Iccf3cea82065af5e055d3cd932cd0808b29b15dc
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
For CoW types, prefer const methods to avoid needless detach()ing.
Change-Id: I211ae74434bf3705914764e1e9d02ae823a1be3a
Reviewed-by: Edward Welbourne <edward.welbourne@theqtcompany.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
They were duplicating other enumeration values
(WindowTransparentForInput, WindowOverridesSystemGestures) and were
only supported on Windows CE.
Task-number: QTBUG-42897
Task-number: QTBUG-51673
Change-Id: I993a933ee2a3dc8a0ebd7ba6aef80b456f50a657
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Fixes C4838 warnings:
rfc6234/sha384-512.c(326): warning C4838: conversion from '__int64' to 'quint64' requires a narrowing conversion
rfc6234/sha384-512.c(331): warning C4838: conversion from '__int64' to 'quint64' requires a narrowing conversion
rfc6234/sha384-512.c(796): warning C4838: conversion from '__int64' to 'const quint64' requires a narrowing conversion
Change-Id: Ib531d376d3fd49785e227455f280fc4072bba8c5
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
There's no reason not to allow compiling Qt without any QPA
plugins. It's enough to clearly warn about it at the end
of configure instead of aborting the build.
Change-Id: I52ff681b68fcbfd9c5e28e555224812e6c441dc6
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
In QWindowsXPStyle::sizeFromContents(), scale the margins using the
devicePixelRatio and round the height up. Typically increases
the height by 2 pixels when the device pixel ratio equals 2.
Change-Id: I446ad518c3808a6454a08db7b66a942d37c1703f
Task-number: QTBUG-49374
Reviewed-by: Alessandro Portale <alessandro.portale@theqtcompany.com>
Call QTabBarPrivate::moveTabFinished() from QTabBar::removeTab()
to cancel any drag in progress to prevent crashes due
to invalidated indexes.
Task-number: QTBUG-52527
Change-Id: I5bd8cc6f55e5aea1f1a4710494ba5b92939a42ee
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
QEventDispatcherWin32 on WinCE uses a separate low-priority thread to
monitor sockets activity, so changing the state of notifiers occurs
asynchronously to the main thread. This makes a message-based socket
activation mechanism ineffective.
To avoid timeouts in the helper thread, update the thread's pool
directly from the (un)registerSocketNotifier() functions.
Change-Id: I702c32d69dce09323ca5f65dc2ee1407842e41ef
Reviewed-by: Tobias Koenig <tobias.koenig@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Makes it easier to add stuff to it.
Change-Id: Id75834dab9ed466e94c7ffff1445727a2ed975cc
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
When I tested this, it compiled, but either I wasn't using the same
compiler or the problem happens during linking: we can't use .quad
(64-bit) with a relocation on x32. So instead, let's use .long (32-bit).
Task-number: QTBUG-52658
Change-Id: Ifea6e497f11a461db432ffff14468d1a16f49c67
Reviewed-by: Shawn Rutledge <shawn.rutledge@theqtcompany.com>
Found by Coverity, CID 11351. If we check 'if (backend)' at one line,
not clear why we do 'if (bakcend->canResume())' two lines above without any test.
Also, 'delete backend' is noop if backend is nullptr, so ...
even this 'if (backend)' is ... not needed.
Change-Id: I7a46ce04baeb9251debb7b246954911df58880ca
Reviewed-by: Edward Welbourne <edward.welbourne@theqtcompany.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
This functionality will get replaced by a new and more flexible system
to configure Qt.
Change-Id: I04cf694ab1671eeed39b79a660566595a22f54a7
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
With 5.7, we can rely on c++11 atomics (or the msvc ones), so
the note doesn't make sense anymore.
Change-Id: I10cc67f3dd840f3272ca975c2c9120a8c871a2fb
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Coverity is not happy with us testing a pointer (not null) in a couple of
places but not in initializeInflateStream. In fact we _always_ test this
pointer _before_ calling initializeInflateStream. Anyway, let's make
this pre-condition clear by Q_ASSERT()ing it.
Change-Id: Ibce99ef6ff1d4e09e3a446a672fa28fc93ab832c
Reviewed-by: Edward Welbourne <edward.welbourne@theqtcompany.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
The synthesis of mouse events for unhandled tablet events is now in
cross-platform code, so the platform plugins don't need to do it.
xi2HandleTabletEvent's signature included an eventListener which is
no longer needed; removing it makes it look needlessly inconsistent
with the signature of xi2ReportTabletEvent; and while we're at it,
might as well make the events const and deal with the repercussions.
Task-number: QTBUG-47007
Change-Id: I0f76c8ed2d2b0baed0652bc68286f1734b8b72ff
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
so it turns out that the 'unix' conditional in moc.prf meant that the
feature was not enabled *anywhere* by default, as the unix configure
disabled it everywhere.
amends b3fcaea5.
Change-Id: Ie41ed2ebc338e560b8d6bfbb0bb18888e31b6d13
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
... with STL-style iterators or with algorithms.
Java-style iterators have overhead.
Introduce local template separate_if algorithm from kleopatra
project to simplify current code.
http://api.kde.org/4.3-api/kdepim-apidocs/kleopatra/html
Done-with: Marc Mutz <marc.mutz@kdab.com>
Change-Id: Ib154f80f46f8041d9cafd81bed0e1982b21541cf
Reviewed-by: Edward Welbourne <edward.welbourne@theqtcompany.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Found by Coverity, CIDs: [159014-159383]. Indeed, Q_ASSERTS can
go into the block controlled by the corresponding if-statements,
otherwise - why 'if' at all?
Change-Id: Ibc99cd48b97287ec132e082d14f3c39350d547f9
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
The QTextBlockFormat::FixedHeight overrides the line height
regardless of its calculated height. If the line contains
objects or text which is higher than the specified line height,
using FixedHeight will cause them to overlap with the previous
line. This is not what happens in normal web browsers. The
expected behavior is that the line height given in CSS is the
minimum height, but that we still reserve space needed to display
everything without overlaps.
To make it possible for people to retain the old behavior, we
introduce the -qt-line-height-type property, which allows them
to override the default.
This also fixes output from toHtml() to use the new property
rather than set the minimum height of the paragraph or the
"line-spacing" property, which does not exist in either CSS nor
in Qt.
[ChangeLog][QtGui][Important Behavior Changes] When line height
is specified in pixels, this is now interpreted as the minimum
line height rather than an absolute line height to avoid overlaps.
To get the old behavior, use the -qt-line-height-type property in
CSS and set it to "fixed".
Task-number: QTBUG-51962
Change-Id: Ic2dde649b69209672170dad4c2de1e1c432a1078
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
std::shared_ptr supports it. To resolve an ambiguous overload when a
literal 0 is passed as a parameter, the normal constructors needed to be
made a template, like std::shared_ptr.
Task-number: QTBUG-52569
Change-Id: Id75834dab9ed466e94c7ffff14451417892d2148
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Give the Windows configureapp the same -pch/-no-pch arguments found in
the Linux configure script. Using command line arguments to
enable/disable precompiled header use is preferable to mkspec changes.
Make -pch the default for all toolchains. In particular, this makes
-pch the default for QNX on Windows. Previously, QNX on Windows had
an implied default of -no-pch. Precompiled headers are the default
for QNX on Linux; they should also be the default for QNX on Windows.
A 'dictionary["PCH"] = "no"' will need to be added in
Configure::applySpecSpecifics for any toolchain that should default to
-no-pch (none known at this time).
-no-pch is implemented by putting a "CONFIG -= precompile_header"
in qmodule.pri. This ensures that Qt is built without precompiled
headers (as requested) even if allowing precompiled header use is the
default for the toolchain.
Task-number: QTBUG-52578
Task-number: QTBUG-11545
Change-Id: I1b59bc2d416c5ba169161c5b3cc13accd76eeac8
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
QWindowsXPStylePrivate maintains a list of theme handles for various
style items. The derived class QWindowsVistaStylePrivate had logic
to use a special helper window with the "explorer" window theme set to
obtain the correct treeview arrow branch indicator (arrow shape for Vista as
opposed to '+'/'-' on Windows XP) when creating the "TREEVIEW" theme.
This required calling the helper function
QWindowsVistaStylePrivate::initTreeViewTheming() before
QWindowsXPStylePrivate::createTheme(), which is prone to errors and
initialization order issues.
This could be solved by making QWindowsXPStylePrivate::createTheme()
virtual or similar, but since it the fate of QWindowsXPStylePrivate is most likely
to be merged into QWindowsVistaStylePrivate; it was decided to move
the entire special handling of the Vista treeviews into
QWindowsXPStylePrivate. The existing enumeration value
QWindowsXPStylePrivate::TreeViewTheme is renamed to
QWindowsXPStylePrivate::XpTreeViewTheme and a new value
QWindowsXPStylePrivate::VistaTreeViewTheme is added for which
QWindowsXPStylePrivate::createTheme() invokes the special handling.
This also removes the need to destroy the helper window in unpolish(),
which should save some initializations.
Task-number: QTBUG-52230
Change-Id: I0492ecf38fb3e5eabc4ecbdef70f0bf05e82e104
Reviewed-by: Adam Light <aclight@gmail.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
All internal settings are stored as strings, if the requested setting
value doesn't exist, or the type conversion fails, then an exception
is thrown. Instead of making use of the convenience functions we can
just get the original string an convert it ourself.
Task-number: QTBUG-52574
Change-Id: Ic410244a373318e167437eea4381335a907ccbf9
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
