QToolBarAreaLayout{Item,Line} are larger than a void*, so holding
them in QLists is needlessly inefficient. Worse, the code could
come to depend on the fragile property of (inefficient) QLists
that references to elements therein never are invalidated.
Fix by marking the types primitive and movable, resp., and
holding them in QVector instead.
Change-Id: I4e68d4bee41040bf84302b8ce8295a11debded70
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
* Transfer propertiesUpdated and updated signals from submenus to parent
menus. Without this, the adaptor only receives this signal from top-level
menu items, and doesn't receive it from items of submenus. Connect to
these signals when a menu item is added or synced, and disconnect when it
is removed.
* Make QDBusPlatformMenus use IDs of items containing them, not their own
IDs (own IDs do not make any sense since they are not exported over D-Bus).
* Store toplevel menus per-adaptor, to make it possible to export multiple
menus (for example a menubar and a tray icon menu).
* Adjust the QDBusMenuLayoutItem::populate methods to always get the menu
via its containing item and to populate the menus recursively.
* Map D-Bus menu AboutToShow method to platform menu aboutToShow method,
and map hovered and closed events to hovered and aboutToHide signals.
(QTBUG-46293)
* Always set the visible property on item. Otherwise, when an item becomes
visible, the D-Bus menu still thinks it's invisible because that property
was not changed back to true. (QTBUG-48647)
* Call emitUpdated from insertMenuItem and removeMenuItem methods, as they
really update layout. Do not call it from syncMenuItem, it changes only
properties but not the layout.
* Start revision numbering with 1, because libdbusmenu-based hosts ignore
updated signal with revision=1.
Task-number: QTBUG-46293
Task-number: QTBUG-48647
Change-Id: Icf713405db0443e25462c1a19046df7689fe5e78
Reviewed-by: Shawn Rutledge <shawn.rutledge@theqtcompany.com>
Reviewed-by: Błażej Szczygieł <spaz16@wp.pl>
As we can load plugins without QLibrary now, we don't have to #ifdef
out the code that does so anymore.
Change-Id: I1dc20216830a882dbd5a1b431183407e6b19c837
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This makes implementing class methods that enable or disable a feature
stored in a QFlags<> member easier.
[ChangeLog][QtCore][QFlags] Added setFlag method to set or unset a flag
Task-number: QTBUG-27100
Change-Id: Ic602abbbf3100df78f21b7918573744d1bbd18c1
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Move instead of copy the bezierCurves and tcbPoints
containers when updating 'config' in setType_helper().
In Qt 5, we still have the unsharable container
issue which causes code for deep copies to be
emitted for most copy operations.
A move is always just re-seating a pointer.
Change-Id: Icff7415dd0ce44df0602273ff42370b26d831b85
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
- don't re-evaluate QMap::end() all the time
- don't copy QVariant more than needed
- pass temporary to QVector::append (enabling moves)
- swap instead of copy-assign a vector into place
Change-Id: I7549812dfbb2dbc9a919fa9565397d50141fc2ca
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
... by replacing them with C++11 range-for loops, or, in one case,
qDeleteAll().
Change-Id: If6980f6c12ac1d359a1f0e062dcc2a86e9685d1e
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
It's used as a global static but according to library policy it's required
to have a trivial dtor and a trivial or constexpr ctor, otherwise it
runs code at library load/unload.
Change-Id: I1fcea3ce388b2a198e3851e6d0aeb91ae0a8e4b1
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
It is very rare that operator+= is used as operator=.
Most of the time, one op+= is followed by more op+=s.
The old code checked whether *this was empty, and simply
shallow-copied the RHS into the LHS in that case. This
has one severe drawback: a reserve() on the LHS is lost.
Fix by not checking for isEmpty() but for d == &shared_null.
Change-Id: Iff28e496cf24cc93f248449d74012c4f3a87253e
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Consider a window which was hidden and shown with hide() and show()
methods and mouse pointer was in window when hide() was called.
At first, window got focusOutEvent and then Qt library sends X server
a message to unmap the window.
Then X server will send client two messages:
1) FocusOut(10) detail=Nonlinear(0x03)
2) FocusIn(9) detail=Pointer(0x05)
QXcbWindow has a logic for not seting active window to 0 if there is
a FocusIn coming (see QXcbWindow::doFocusOut).
So QGuiApplicationPrivate::focus_window still points to the current
window.
Then when show() is called, qt compares previous focus with new focus
and, since they are equal, doesn't do anything. Event focusInEvent
isn't delivered to the window.
Here are two links why X server sends FocusIn just after FocusOut:
http://lists.freedesktop.org/archives/xorg/2008-December/041684.htmlhttps://tronche.com/gui/x/xlib/events/input-focus/normal-and-grabbed.html
Proposed fix ignores FocusIn events with detail==Pointer.
The text of explaining comment is taken from the Chromium project:
https://chromium.googlesource.com/chromium/src/+/master/ui/views/widget/desktop_aura/x11_desktop_handler.cc
from X11DesktopHandler::ProcessXEvent function.
[ChangeLog][module][Linux/XCB] Fix not delivering focusIn event on
hide/show with XCB
Task-number: QTBUG-49071
Change-Id: I433c8b638834c25f113cc134ee4185778c44f540
Reviewed-by: André Hartmann <aha_1980@gmx.de>
Reviewed-by: Lisandro Damián Nicanor Pérez Meyer <perezmeyer@gmail.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@theqtcompany.com>
This patch fixes cursor shape when mouse leaves the window and enters
the window again with pressed mouse button - ignore the mouse enter
and leave event when any of mouse buttons is pressed.
Task-number: QTBUG-46576
Change-Id: Id6ce50cd0d66da51a251d4811bc42cd31606de29
Reviewed-by: Shawn Rutledge <shawn.rutledge@theqtcompany.com>
Added enter/leave event handling in XInput2 to avoid problems with
those events when the mouse is grabbed.
This commit amends: 53d289ec4c
This commit amends: ed2e157803
Task-number: QTBUG-50340
Change-Id: I7a120b46daa4f8fa4c218346273ae90b6abfa156
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@theqtcompany.com>
This is necessary for combined device and simulator builds on Apple
platforms (iOS, tvOS, watchOS) to link properly.
Change-Id: I21e70806643b10f429945d3020995dc94fa5c612
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
This makes the XML parser 100% independent of having regexp
support enabled.
Change-Id: I73004b0fb71e8086618995c71a985a86c292df3d
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
-ldl option was used unconditionally while libdl is not supported
when libc is static.
Add build test to configure which checks if libdl is supported.
QMAKE_LIBS_DYNLOAD in "src/corelib/plugin/plugin.pri" is now used only
if libdl is available.
qt_linux_find_symbol_sys from qlibrary_unix is now used only if
QT_NO_DYNAMIC_LIBRARY is not defined.
Initially reported by Buildroot autobuilder here:
http://autobuild.buildroot.net/results/a85/a85a1839a45fb6102e53131ecc8f6dadf92bcdc2
Change-Id: I0397472456efdc4f3ab5f24d01253bee8048a9d1
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
In a case of having a non English system locale, then the face and name of
the font may be localized but will not match up what is in the registry.
Therefore we need to try with the English name if is possible that there
is one, so in the case of when registerAlias is false we should try to get
the English name if necessary.
Change-Id: Id1064b6f5263677320374dbddfad852f30f041c5
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Reviewed-by: Liang Qi <liang.qi@theqtcompany.com>
This fixes the crash introduced by
75b705fec8. The drag only needs to be ended
on the hide if it is actually in the middle of dragging. This enables the
case of dragging the dock widget out with the mouse to continue to work.
Task-number: QTBUG-50890
Change-Id: I72309dd40ee670319f2ff607ae201c46f8de4652
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
These functions appear to have been dead code always - even in the
earliest Qt4 releases!
Change-Id: Ie2a458092b45113782284cf8ed3aa11a018b60d4
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
While investigating an unrelated rendering bug I noticed that the call
to drawRect here is essentially dead code. However, the original intent
seems to have been to draw a very subtle shadow within the scrollbar
groove, so let's re-enable it. Subjectively, it does add a nice touch,
especially with higher dpi screens.
Change-Id: Iebffe69ecf9c26be617f7e85bd95f85ed7b0fafe
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
Reviewed-by: Jens Bache-Wiig <jensbw@gmail.com>
This is a convenient function which waits (timoutMS) for the runnable to be executed.
Change-Id: I2c0f262d3795c29961e8182b7ffc35a2c6f8a1f6
Reviewed-by: Christian Stromme <christian.stromme@theqtcompany.com>
Regression introduced in e3689949ba
Change-Id: I9d3e37733d933e037f04d3a6efaf05607e16d9c5
Reviewed-by: Christian Stromme <christian.stromme@theqtcompany.com>
Iterate only once over QJsonObject, create key list by
existing loop instead of create by QJsonObject::keys(),
which contains internal loop.
In common case if loop's statement is lightweight,
then effect of optimization is significant, and vice versa.
Also make addJsonArray() and addJsonObject() functions
more homogeneous.
Use reserve to optimize memory allocation.
Change-Id: Id122cd1becfd34bb06640876b1c79e1d396d2a6b
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Set the right sub role (NSAccessibilitySecureTextFieldSubrole) and
return the bullet point character for the text contents. This alignes
the behavior with native widgets.
Change-Id: I7305e08dca61097dd8c050aed64c792c06de0a4d
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
accessible\iaccessible2.cpp(1708,40) : warning: comparison of two values with different enumeration types ('IA2Role' and 'QAccessible::Role') [-Wenum-compare]
Q_STATIC_ASSERT(IA2_ROLE_COLOR_CHOOSER == QAccessible::ColorChooser);
Task-number: QTBUG-50804
Change-Id: I25a1f00fdbd687ff23f875542a6aa0cc0cab3d23
Reviewed-by: Jan Arve Sæther <jan-arve.saether@theqtcompany.com>
The enum needed a name to make that possible, of course.
The one overt int return -1 needed to be made explicitly Neither, too.
Change-Id: I3930bf03a7ee5e1619a1c74f9ca54faf6a6c5b2f
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
If argumentTypes is NULL then it is indeed not equal to the address of
a local static variable, so there is no need to follow the NULL-check
with a check that it's not equal to that non-NULL address.
Change-Id: I62362db747c0620b2195f7997368f026f535d57c
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
When the window flags are changed for the widget then it is possible to
minimize the dockwidget via the titlebar. This will cause it to be ready
to start a drag internally but since the dockwidget never gets a mouse
release event it doesn't end it. Therefore it ends up being in an invalid
state later on when restored, so the endDrag() needs to happen in the hide
event to ensure this is not an issue later on.
Change-Id: Ia84bee96b9eed49896869e6a15d4de6d01964264
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Normally setting the mode again and again on each application startup
is not desirable as it takes time and considerably affects the application
startup experience.
However, in some cases (like shutting down the X server) the output
gets powered down and there seems to be no way to detect this. Here
setting the mode again becomes necessary. Until there is a better
solution, add an environment variable QT_QPA_EGLFS_ALWAYS_SET_MODE
which, upon a non-zero value, leads to always calling drmModeSetCrtc.
The issue is not specific to Qt, the same is visible with the NV examples
as they try to minimize redundant mode set calls in the same manner.
Change-Id: If2b5e3fda74e077a747dd03cee65c936560a5026
Reviewed-by: Louai Al-Khanji <louai.al-khanji@theqtcompany.com>
Add qstrntod as an overload for qstrtod, which takes a len parameter,
so one can pass character arrays which are not null-terminated.
Change-Id: I8c1c6c3627043c1d6ec6eb712efa3abc9e5e9e00
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Milian Wolff <milian.wolff@kdab.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
