| Commit message (Collapse) | Author | Age | Files | Lines |
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
This also adds QStyleOptionTabWidgetFrameV2 so that we do not have
to do ugly hacks in the style to obtain it.
Task-number: QTBUG-5187
Reviewed-by: ogoffart
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
While drawing the tabbar frame, mac style needs the QTabBar pointer to
calculate the correct size.
Since the QTabWidget also uses the PE_FrameTabBarBase to draw background
of the corner widgets, the mac style has to use position passed with the
style option. This only works with horizontal tabs.
Reviewed-by: Jens Bache-Wiig
|
| |
| |
| |
| |
| |
| | |
This was a platform regression for the cocoa platform
Reviewed-by: Prasanth
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The subcontrol rect offset was correctly reporting
not to work on Vista style. The problem was that we were
getting the size of the arrow by subtracting the xoffset.
But this would mean that the
Reviewed-by: ogoffart
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When a tool button is not in a tool bar on XP and Vista it will get a
slightly different appearance from normal tool buttons. I resolved
this by drawing a normal tool button with a divider line on top
if the autoraise property is not set on the button. (which is by
default enabled only for buttons in a tool bar).
Task-number: QTBUG-5061
Reviewed-by: prasanth
|
| |
| |
| |
| |
| |
| |
| |
| | |
Dragging is handled by a seperate window in document mode. The currently
selected tabbar item is drawn to a pixmap for this purpose. That
paintdevice was not initialized correctly. (e.g. font)
Reviewed-by: Trond
|
| |
| |
| |
| |
| |
| | |
There is no reason to use QMap when the key is a pointer.
Reviewed-by: Thomas Zander
|
| |
| |
| |
| | |
Reviewed-by: David Boddie
|
| |
| |
| |
| |
| |
| |
| | |
When application closes and we haven't deleted the unregistered gestures
and gesture recognizer, we should delete them.
Reviewed-by: Thomas Zander
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When there are several gesture recognizers registered for the same
gesture type, we need to know which recognizer we need to get a state
object for since those QGesture objects are tied to the recognizer that
created them.
Reviewed-by: Thomas Zander
|
| |
| |
| |
| | |
Reviewed-by: Thomas Zander
|
| |
| |
| |
| | |
Reviewed-by: Denis Dzyubenko
|
| |
| |
| |
| |
| | |
The widget() getter is still publicly documented, follow the lead of
other events to make the setter internal.
|
| |
| |
| |
| |
| | |
As QApplication is a singleton this makes usage of these
easier and also in line with many other methods on the class.
|
| |
| |
| |
| | |
Reviewed-By: TrustMe
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
While querying for the text in the pasteboard, it was looking in the
wrong place. The helper function qt_mac_get_pasteboardString() always
searched in generalPasteboard instead of the pasteboard referred by the
QMacPasteboard.
Reviewed-by: MortenS
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
On accepting one gesture Qt can automatically cancel other gestures
that belong to other targets. The policy is normally set to not cancel
any other gestures and can be set to cancel all active gestures in the
context. For example for all child widgets.
Reviewed-By: Denis Dzyubenko
|
| |\
| | |
| | |
| | | |
4.6-WM_NULL-driven
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Emit the QDesktopWidget::workAreaResized() signal when the _NET_WORKAREA
property of the root window changes.
Merge-request: 1111
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Qt up to 4.5.x is missing whole setup of multimedia keys already defined by X
Merge-request: 1742
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
|
| | |
| | |
| | |
| | |
| | |
| | | |
Decided after review by David Boddie.
Reviewed-by: trustme
|
| | |
| | |
| | |
| | | |
Reviewed-By: trustme
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The default height of tab bar is 22 pixels in Snow Leopard. We used to
draw tabs taller than 21 pixels to a pixmap and stretch to the required
size. The limit is now changed to 22 pixels (which is the most common
use case). The stretched drawing is not perfect in Snow Leopard due to
some changes in how HITheme draws tabs.
Reviewed-by: Jens Bache-Wiig
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When a tab is clicked and mouse is moved outside that tab, it should
be drawn as normal (not in sunken state).
Reviewed-by: MortenS
|
| | |
| | |
| | |
| | | |
Reviewed-by: trustme
|
| | |
| | |
| | |
| | | |
Reviewed-by: Denis Dzyubenko
|
| | |
| | |
| | |
| | | |
Reviewed-by: Denis Dzyubenko
|
| | |
| | |
| | |
| | | |
Reviewed-by: Denis Dzyubenko
|
| | |
| | |
| | |
| | | |
Reviewed-by: Denis Dzyubenko
|
| | |
| | |
| | |
| | |
| | |
| | | |
type argument.
Reviewed-by: Thomas Zander
|
| | |
| | |
| | |
| | | |
Reviewed-by: Thomas Zander
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We shouldn't add several graphicsobject contexts for the same gesture
type when looking for the gesture-enabled QGraphicsObject under a
hotspot.
Reviewed-by: Thomas Zander
|
| | |\ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
We have some internal hidden widgets which should not come up in the
QApplication::topLevelWidgets() list. So the known ones are being
removed from the QWidgetPrivate::allWidgets set.
Task-number: QTBUG-739
Reviewed-by: Denis Dzyubenko
Reviewed-by: Bradley T. Hughes
|
| |\ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
We should continue processing as much as we can, and report the
WM_QT_SENDPOSTEDEVENTS at the end of processEvents().
|
| |\ \ \ \ |
|
| |\ \ \ \ \ |
|
| |\ \ \ \ \ \
| | |_|_|_|/ /
| |/| | | | | |
|
| |\ \ \ \ \ \ |
|
| |\ \ \ \ \ \ \ |
|
| | | | | | | | | |
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
This function works more reliably than PeekMessage() with different
flags.
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
This includes the startstop timer delay in QAbstractAnimation, and the
inSizeMove workaround for paint events.
Reviewed-by: Prasanth Ullattil
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
(which is just WM_USER+1)
Delay the next WM_QT_SENDPOSTEDEVENTS iff there is a WM_TIMER
or input event pending
We also need to break out of processEvents() after seeing this
message, to prevent livelocking in the prescence of fast timers.
I also took the liberty of defining WM_QT_SOCKETNOTIFIER (WM_USER)
at the same time (to give clear meaning to what WM_USER and WM_USER+1
are used for).
Reviewed-by: Prasanth Ullattil
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
The state machine algorithm frequently needs to know what type a state
is, e.g. if it is atomic, final or a history state. We were using
qobject_cast() to determine this, but that function is expensive.
This commit introduces an internal StateType to be able to differentiate
between the different types of state. This vastly improves performance.
Reviewed-by: Eskil Abrahamsen Blomfeldt
|
|\ \ \ \ \ \ \ \ \ |
|
| |\ \ \ \ \ \ \ \ \ |
|
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Reviewed-by: trustme
|