| Commit message (Collapse) | Author | Age | Files | Lines |
|\ |
|
| |
| |
| |
| | |
Reviewed-by: trustme
|
| |
| |
| |
| |
| |
| |
| | |
QPanGesture was changed to use QPointF instead of QSizeF, also need to
change all usages of the pan gesture.
Reviewed-by: trustme
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
setReadOnly.
Should set Qt::WA_InputMethodEnabled after set the flags, just because shouldEnableInputMethod() will read the flags.
Task-number: QTBUG-4917
Reviewed-by: Shane Kearns
|
| |
| |
| |
| | |
Also fix grammer and avoid using combined words.
|
| | |
|
| | |
|
| |
| |
| |
| | |
Reviewed-by: Thomas Zander
|
| |
| |
| |
| | |
Reviewed-by: Alexis Menard
|
| | |
|
| |
| |
| |
| | |
Reviewed-by: trustme
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When delivering GestureOverride events by default both the event and
individual gestures will be ignored. We also store the acceptance state
of individual gesture in the event and not in the gesture object, along
with its target.
Reviewed-by: Thomas Zander
|
| |
| |
| |
| | |
Reviewed-by: trustme
|
| |
| |
| |
| |
| |
| |
| |
| | |
When there are two different gestures are being triggered and they are
supposed to be sent to different widgets, don't stop event "propagation"
when the first event is successfully delivered.
Reviewed-by: trustme
|
| |
| |
| |
| | |
Reviewed-by: trustme
|
| |\ |
|
| | |
| | |
| | |
| | | |
Reviewed-by: thartman
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The projection to 2D needs to be done when the rotation is applied,
not after all transformations have been applied.
Reviewed-by: trustme
|
| | |
| | |
| | |
| | | |
Reviewed-by: trust-me
|
| | |
| | |
| | |
| | | |
Reviewed-by: Trust Me
|
| | |
| | |
| | |
| | |
| | |
| | | |
Use an internal class to reset the activationRecursionGuard when the scope is closed.
Reviewed-by: Leo Cunha
|
| |\ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This avoid warnings when using easing curves that progress outside the
range [0.0, 1.0]. Patch proposed by warwick.
Reviewed-by: Warwick Allison
|
| |\ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | | |
Improve the documentation for size policy.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
No need to traverse the tree again to access the topLevelItem in case
there is no common ancestor.
Reviewed-by: bnilsen
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This function is moved to graphicsitem private because it is needed by
multi-touch event handling and is not specific to bsptreeindex.
Reviewed-by: bnilsen
|
| |\ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Auto-test updated.
As a bonus, stabilized tst_QListView::task262152_setModelColumnNavigate.
Reviewed-by: Thierry
|
| |\ \ \ \ \
| | |/ / / /
| |/| | | | |
|
| | |/ / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
The original report was that the customer did resize(main.sizeHint())
instead of the main.adjustSize(); Note that resize(main.sizeHint() still does
not work. However, calling main.adjustSize() should now do what the original
reporter wanted.
The problem was that the resize did not work, because at the point of the resize
the minimumHeight of main was still 22 (8+6+8), and we tried to resize it with
a new height of 10. The resize would bound the height up to 22, and the main
widget would then get a size of (200x22).
The reason why it still had a minimumHeight of 22 was that it was the
minimumSize of the previous layout configuration. Unfortunately the new
minimumSize of the widget hadn't been updated yet because there was a
LayoutRequest event in the queue that hadn't been processed yet. (This
LayoutRequest was triggered by that we called invalidate() from hide()).
Thus, processing the event queue immediately after the hide() could also have
been a workaround for this issue.
There is no really good fix for this issue (and it does not seem to be a
common problem) without introducing a risk for regressions.
Due to that we therefore decided to provide a fix in QWidget::adjustSize().
Reviewed-by: paul
|
| |\ \ \ \ |
|
| | | | | | |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
We now check that the item has children before animating.
Reviewed-by: Alexis
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
On embedded dialog pressing tab stop changing the focus when the focus
was given to QFontComboBox. It's because QFontComboBox embed a
QLineEdit in order to allow editing. But this QLineEdit is a focus proxy
so we need to special case that. The logic is the same in QApplication.
Be careful when changing one of them.
Task-number:QTBUG-4818
Reviewed-by:jan-arve
Reviewed-by:ogoffart
|
| | |\ \ \ \ |
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
The flow positions in ScrollPerItem mode did not take the hidden rows into
account when configuring the vertical scroll bar.
A mapping between the scroll bar value and the flow position has been
added. Auto-test included.
Task-number: QTBUG-2233
Reviewed-by: Thierry
|
| | |\ \ \ \ \ |
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
QGraphicsView.
The main problem here is that QWidget assume that they are in the screen
somewhere, which means inside the available geometry provided by
QDesktopWidget. But in QGraphicsView the button can be in a position
that is way bigger than the screen resolution. Lot of widgets make
this assumption when positionning subpopups or submenus. Instead of
applying the same code on tons of QWidgets, it's better to have an
helper function in desktop widget which catch this case. It's not
pretty (since it has nothing to do with QDesktopWidget) but we don't
have better solution.
Task-number:QTBUG-3822
Reviewed-by:brad
|
| | |_|_|/ / /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
This was reported for QGraphicsAnchorLayout, but affects all
QGraphicsLayout subclasses. However, since it is not even handled
properly in QGraphicsItem::setParentItem() it should be fine that we
don't try to deal with this issue. Thus, the layouts are "consistent"
with graphics items when it comes to detection of cycles in the tree.
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
If an effect is applied repeatedly on the same source, just with
varying parameters, we can save a lot by caching the source pixmaps.
Reviewed-by: Bjørn Erik Nilsen
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
This extra whitespace was introduced by mistake in a previous commit.
Remove it again.
Reviewed-by: Trust me
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Oops.
Reviewed-by: Trust me
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
When exposing or resizing a previously hidden window on Cocoa, we would
go through the CoreGraphicsPaintEngine even when a different graphics
system was set. This was because there were two different paths from
the windowing system into our paint event. The one going through the
virtual drawRect function in QCocoaView did not honor the graphics
system. This patch makes sure the backing store is used for these types
of events as well.
Done with: Gunnar
Reviewed-by: MortenS
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
When setting the TranslucentBackground flag after the window has been
created, we should check whether the widget has a native window id, and
not the WA_NativeWindow attribute which doesn't seem to be set by
default for top-level widgets.
Reviewed-by: Trond
|
| | | | | | | |
|
|\ \ \ \ \ \ \
| |_|_|_|_|_|/
|/| | | | | | |
|
| |\ \ \ \ \ \ |
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
In non-GUI applications on X11, QFont and QFontInfo return different
point size because for QFontInfo, the point size is converted to pixel
size and back, but with different dpis. This commit improves the
situation for the case where font config is used, but the bug still
needs to be fixed properly by using the same dpi for all point<->pixel
size conversions.
Reviewed-by: Trond
|
| | | | | | | | |
|