| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Replace Nokia contact email address with Qt Project website.
- Remove "All rights reserved" line from license headers.
As in the past, to avoid rewriting various autotests that contain
line-number information, an extra blank line has been inserted at the
end of the license text to ensure that this commit does not change the
total number of lines in the license header.
Change-Id: Ie7ba62011752fcb149b99b26317c54f2a0cfa931
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
On Mac we have to fake the QEvent::Enter event when entering a widget
so we need to ensure the buttons held down are set as part of that event
otherwise it causes problems with graphicsview which will update the
grabbed item based on this event.
Task-number: QTBUG-19353
Change-Id: I89252a46b5edd0d82b6b1a0cf2592f72b472e5cc
Reviewed-by: Nils Jeisecke <jeisecke@saltation.de>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
|
|
|
|
|
|
|
|
|
|
| |
When the language menu was shown while IME composition was underway, it
would cause the composition text to be in an invalid state.
Task-number: QTBUG-25107
Change-Id: Iaf06f330fb055e30c25ccc9874f15132ea000bc2
Reviewed-by: Tasuku Suzuki <tasuku.suzuki@nokia.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
|
|
|
|
|
|
|
|
|
|
| |
In this case these string are not input through a composing process
(no marked text is set) but they should not be handled by Qt key
event handler.
Change-Id: I9ea96c29db69d51abc6821e72a3228158012bebc
Task-number: QTBUG-23216
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
|
|
|
|
|
|
|
|
|
| |
Simple search and replace. This commit doesn't touch 3rd-party files,
nor translations (where the change is not so simple and will be handled
in a separate commit).
Change-Id: I4e48513b8078a44a8cd272326685b25338890148
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
|
|
|
|
|
|
|
|
|
|
| |
Apple changed the API for pixel based scrolling, which results
in runtime warnings on the command line. This patch will make
Qt use the new API. In addition, apple removed suppert for
z-scrolling. Since that was never supported in Qt as well, we
also remote code that handles that.
RevBy: msorvig
|
|\
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
src/plugins/platforms/wayland/qwaylandclipboard.cpp
src/plugins/platforms/wayland/qwaylandclipboard.h
src/plugins/platforms/wayland/qwaylanddisplay.cpp
src/s60installs/eabi/QtOpenGLu.def
|
| |
| |
| |
| |
| |
| |
| | |
Updated version of LGPL and FDL licenseheaders.
Apply release phase licenseheaders for all source files.
Reviewed-by: Trust Me
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
| |
When many NSView are being drawn, calling CGContextFlush()
is too expensive and some flickering becomes apparent.
We now call CGContextSynchronize() instead.
Since this solves the flickering problem, we can now
call setNeedsDisplay: for items in the unified
toolbar. This allows us to smootly trigger many
flushings inside the unified toolbar.
Task-number: QTBUG-19267
Reviewed-by: Jiang Jiang
|
|
|
| |
Reviewed-by: Jiang Jiang
|
|
|
|
|
|
|
|
|
| |
The check on flushRequested was needed since we
were flushing the toolbar with the main window,
but the flushing is now independent so we don't
need it anymore.
It is now making us skip some flushing.
Reviewed-by: Samuel Rødal
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Selected range is the range of the text that has been edited with
input method, but not yet committed, normally the cursor should be
placed that the end of it (or hidden).
Task-number: QTBUG-17923
Reviewed-by: Morten Sørvig
|
| |\
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Conflicts:
src/gui/kernel/qcocoaview_mac.mm
src/s60installs/bwins/QtGuiu.def
src/s60installs/eabi/QtGuiu.def
src/testlib/qtestcase.cpp
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The crash happends when you close down an application that has a
drawer that contains a widget with keyboard focus. The NSView
backing this widget is then the first responder in Cocoa.
This bug was a bit more hard-boiled than usual, and I can't say I truly
understand whats going on. My findings are that both the drawer and the
parent window both points to the same view inside the drawer as its
first responder. And when deleting the drawer (togheter with the focus
widget), the parent window is left pointing to a first responder that
is actually deleted. Is seems that us refusing to release a view as first
responder if we have no attached widget is wrong. So we choose to return
YES instead, which after all makes much more sense.
Task-number: QTBUG-15897
Reviewed-by: msorvig
|
|\ \ \
| |/ / |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
It turns out that we sometimes hit a strange bug with enter/leave
events when a popup is showing. If you righpress to show the popup,
and then move the mouse outside the window, we get a continues series
of leave events. This patch separates more the native vs alien logic
for dispatching enter/leave to accommondate this problem
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
No reason to use the optimization when only scrolling in one
direction. In those cases we are better off scrolling immidiatly
when requested to minimize the shuffling of
scrolls and repaints done by the application
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
On Mac/Cocoa, enabling touch on a widget slows down the scrolling
performance of the whole application. It seems Cocoa spends some
time in the background figuring out what to do with the touch
events, and whether or not it should convert them to scroll/wheel
events. Therefore, it makes sense to no subscribe for touch when
the mouse is not over the widget, This patch implements that
strategy, and the effect is huge when tested agains creator.
Rev-By: brad
|
| | | |
|
| |\ \
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
alien-squached
Conflicts:
src/gui/kernel/qcocoasharedwindowmethods_mac_p.h
src/gui/kernel/qcocoaview_mac.mm
src/gui/kernel/qwidget_mac.mm
|
| | |\ \
| | | |/
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Conflicts:
examples/webkit/imageanalyzer/imageanalyzer.h
examples/webkit/imageanalyzer/mainwindow.h
mkspecs/unsupported/qws/linux-x86-openkode-g++/qplatformdefs.h
src/corelib/io/qfsfileengine_iterator_unix.cpp
src/corelib/io/qfsfileengine_iterator_win.cpp
src/corelib/kernel/qcoreapplication.cpp
src/network/access/qnetworkaccessdatabackend.cpp
src/plugins/bearer/connman/qconnmanservice_linux.cpp
src/plugins/platforms/openvglite/qwindowsurface_vglite.h
src/s60installs/bwins/QtCoreu.def
src/s60installs/eabi/QtCoreu.def
src/s60installs/s60installs.pro
tools/assistant/tools/assistant/helpviewer_qwv.h
tools/qdoc3/test/qt-html-templates.qdocconf
|
| | | |
| | | |
| | | |
| | | | |
Reviewed-by: Trust Me
|
| | | |
| | | |
| | | |
| | | | |
Give Alien on Cocoa a warm welcome.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The over flushing of the toolbar had a huge
performance impact so now we only flush the toolbar
when necessary and moved the rendering out of the
flushing process.
Reviewed-by: Richard Moe Gustavsen
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Instead of flushing by grabbing the Core Graphics
context, we do it directly in the drawRect:
method.
This approach allows us to call flush() many times
and only proceed to flush when the system is
ready.
Reviewed-by: Richard Moe Gustavsen
|
| |/ /
|/| |
| | |
| | | |
Reviewed-by: Eskil Abrahamsen Blomfeldt
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Update the retain/release mechanism for
CGContexts. Some leaks were due to unnecessary
retain calls.
Task-number: QTBUG-15373
Reviewed-by: Samuel Rødal
|
|/ /
| |
| |
| | |
Reviewed-by: Fabien Freling
|
|\ \
| |/
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
demos/declarative/snake/snake.qml
qmake/generators/symbian/symbiancommon.cpp
src/network/access/qnetworkaccessmanager.cpp
src/s60installs/s60installs.pro
tests/auto/qitemselectionmodel/tst_qitemselectionmodel.cpp
tests/auto/qnetworkreply/tst_qnetworkreply.cpp
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
event is sent to two stacked MDI windows
The problem is the fact that we were propagating the event through the
class hierarchy. This means calling the super class implementation,
which does nothing but pass the event to the next guy, which is the
other window, which in turn does the same thing...
The point is, there is no need to propagate this event on the class
hierarchy. If anything, this should be propagated on the widgets
hierarchy, which is already handled by Qt. Removing the code so this
problem does not happen again.
Task-number: QTBUG-12952
Reviewed-by: Richard Moe Gustavsen
|
| |
| |
| |
| | |
Reviewed-by: Olivier Goffart <olivier.goffart@nokia.com>
|
| |
| |
| |
| |
| |
| |
| | |
the raster engine on Mac OS X.
Task-number: QTBUG-12615
Reviewed-by: Samuel Rødal
|
| |
| |
| |
| | |
Reviewed-by: Fabien Freling
|
| |
| |
| |
| |
| |
| |
| | |
This text is delivered through the NSTextInput protocol. Since it was not
initiated from a keyDown message, we were ignoring the insertText: message.
Reviewed-by: Denis
|
| |
| |
| |
| |
| |
| | |
raster engine on Mac OS X.
Reviewed-by: Richard Moe Gustavsen
|
|\ \
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
src/gui/kernel/qapplication.cpp
tests/auto/qfileinfo/tst_qfileinfo.cpp
tools/qdoc3/test/assistant.qdocconf
tools/qdoc3/test/designer.qdocconf
tools/qdoc3/test/linguist.qdocconf
tools/qdoc3/test/qmake.qdocconf
tools/qdoc3/test/qt-build-docs.qdocconf
tools/qdoc3/test/qt.qdocconf
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
While generating the QDragMoveEvent, we were always reusing the last
drop action set by the user. This is not correct, we should copy the action
only if a valid action set before. This is the behavior on Windows.
Task-number: QTBUG-9486
Reviewed-by: Denis
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Previos versions of Mac OS X (< 10.6) didn't have support for such 'Not
Allowed' cursors. 10.6 introduced a new method for NSCursor called
operationNotAllowedCursor. This fix uses the new cusor on available platforms.
Task-number: QTBUG-5186
Reviewed-by: Denis
|
| |
| |
| |
| |
| |
| |
| |
| | |
a bad context, resulting in wrong paintings.
It also unify the raster engine behavior among
platforms.
Reviewed-by: Jason Barron
|
|\ \
| |/
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
bin/syncqt
src/gui/text/qtextlayout.cpp
tools/assistant/tools/assistant/helpviewer_qwv.cpp
tools/assistant/tools/assistant/helpviewer_qwv.h
tools/configure/configureapp.cpp
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
On other platforms then when a drag is finished then it cleans up by
calling setMimeData(0) and deleteLater() on the drag object, this didn't
happen before on the Mac which meant it did not delete the QDrag objects
until the parent of them was deleted thus taking up memory.
Task-number: QTBUG-11613
Reviewed-by: Prasanth
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
raster engine on Mac OS X.
This is mainly done by reverting the commit
04e34fe3aecca482abeeabe2e31778e9102eeb08
Task-number: QTBUG-11518
Reviewed-by: Samuel
|
|\ \
| |/
| |
| |
| | |
Conflicts:
tools/qdoc3/test/qt-html-templates.qdocconf
|
| |
| |
| |
| |
| | |
Merge-request: 535
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
|
|/
|
|
|
|
|
|
|
|
|
|
| |
Up til now, we always used to create a new NSView for a top
level widget, and make it a child of the windows content view.
This is/should be unnecessary, as NSWindow has a method for
setting the content view directly, thereby allowing for reducing
the stacking level with one view. Not the biggest gain if the
year, but it seems effortless to implement. And I do it as
a part of Alien preparation.
Rev-By: msorvig
|
|
|
|
|
|
|
| |
This happens only if a new drag is started while a drag operation is
going on. It seems to be an unsupported operation on Mac.
Reviewed-by: Fabien Freling
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The reason turns out to be that we:
1. Half-way fake mouse events from QCursor::setPos
2. Don't fake mouse move upon mouse enter unless traking is on
The test failed when mouse trackin was on, so that we ended up
faking the same mouse move event twize (in QCursor::setPos and
mouseEnter). We now do a different implementation for setPos on
cocoa (so that native events will be generated), and always
fake a mouse move event from mouse enter to make tooltips (and
other event filters) work
Reviewed-by: MortenS
|
|
|
|
|
|
|
|
|
|
| |
On all platforms, we always send a mouse move event when the
mouse enters a widget (if mouse tracking is on, that is). Then
macnativeevents auto test revealed that we did not do this for
the cocoa port. This patch will fix this.
Reviewed-by: cduclos
Reviewed-by: prasanth
|