| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Reviewed-by: Trust Me
|
|
|
|
| |
Reviewed-by: Trust Me
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The current implementation was based on the GtkExpander widget.
However we do not have such a widget and the primary use case is for
expanders inside item views. Hence it makes sense to base it on
gtktreeview instead. As I can not reproduce the crash with
any recent versions of cleanice I will remove the workaround as well.
omment and edit as applicable ]---|
Task-number: 256146
Reviewed-by: Thorbjørn
|
|
|
|
|
|
|
|
|
|
|
|
| |
It turns out that the fix 7bf4512659113f8cc78e72f1c84158ce4f70a526
caused the QCompleter popup from appearing for a split second on
Windows. This is because the popup is shown as a toplevel window
on this platform. On other platforms than Mac and Windows it doesn't
seem to matter, but be on the safe side, the fix is modified to
apply for Mac only.
Reviewed-by: Daniel Molkentin
Task-number: 255374
|
|
|
|
| |
Reviewed-by: Richard Moe Gustavsen
|
|
|
|
|
|
|
|
|
|
| |
On Mac, a new completer popup must initially have its show()
method called rather than its hide() method. Otherwise the event
handling done by the completer results in a bad state.
On other platforms it doesn't matter.
Reviewed-by: Richard Moe Gustavsen
Task-number: 255374
|
|
|
|
|
|
|
|
| |
Forwarding the ignoreOpacity flag to children in QGraphicsItemPrivate::fullUpdateHelper.
This is a complementary fix to task 252913, partly fixed in commit 2e3a5ea44...
Reviewed-by: bnilsen
BT: yes
|
|
|
|
|
|
|
|
|
|
|
|
| |
We need to subscribe to xfixes selection notify events on all
available screens.
Also implemented delayed subscription to xfixes events since we don't
really need clipboard change notifications unless the application
explicitely asked for by (i.e. created a qclipboard object).
Task-number: 255609
Reviewed-by: Bradley T. Hughes
|
|
|
|
|
|
| |
documentation.
Reviewed-by: TrustMe
|
|
|
|
|
|
|
|
| |
The predefined dash patterns for Mac have always been off, compared to
the ones in the raster engine and the GL engine.
Task-number: 255292
Reviewed-by: Kim
|
|
|
|
|
|
|
|
|
| |
The border was always drawn before the background, and if there were
both a border and background brush set, the background ended up
being drawn ontop of the border.
Task-number: 255868
Reviewed-by: Samuel
|
|
|
|
|
| |
Task-number: 253086
Reviewed-by: Joerg
|
|
|
|
|
| |
Task-number: 255742
Reviewed-by: Kim
|
|
|
|
|
|
|
|
| |
Though this variable always will be initialized in
QX11PaintEngine::begin() valgrind complains about conditional jump or
move depends on uninitialised value(s).
Reviewed-by: Donald <qt-info@nokia.com>
|
|
|
|
|
|
|
| |
If KDEDIRS was set to an empty string, this would unintentionally bring
/share into the icon search path.
Reviewed-by: Olivier Goffart
|
|
|
|
|
|
|
| |
Fixes issue with decimals in the spinbox used for the double
Task-number: 255502
Reviewed-by: mbm
|
|
|
|
|
|
|
|
|
|
| |
We didn't actually check the depth of the target window before
calling the qt_x11_drawImage() fu, that will only work with
depths >= 24.
Task-number: 255311
Reviewed-by: Samuel
BT: yes
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
events."
This reverts commit 7314c07a3e443b1d5349b419a03db8d41ca43f7e.
As reported by Eike, this patch caused several problems for Qt Creator.
Potentially it may cause problems for other (external) applications as well.
An alternative fix (scheduled for 4.5.x) needs to be found for tasks
254456 and 254460.
Reviewed-by: Richard Moe Gustavsen
|
|
|
|
|
|
|
|
|
| |
The reason is that cocoa looses the first responder when
we raise the fake window inside the MDA area. So we need
to re-set the first responder again
Task-number: 255040
Reviewed-by: Trenton Schulz
|
|
|
|
|
|
|
|
|
|
|
| |
There were several problems with antialiased text in 16 bit mode
under Windows. No gamma correction was done, yet we prepared the
cached glyphs for gamma correction. The mask format we rendered
the glyphs into was also set to the desktop depth, which implied
that information was lost and the text looked rather odd.
Reviewed-by: Samuel
BT: yes
|
|
|
|
|
|
| |
We were saying that it connected to accepted it all these cases, but it
actually is doing some nice magic that makes sense assuming you document
it.
|
|
|
|
|
|
|
|
| |
No progress bars on the mac show text and it would be bad if we allowed
it. There's nothing stopping people from connecting the valueChanged()
signal to a slot and have a real label layed out correctly that actually
updates with the amount of time it takes to complete, etc. This is more
what they do on Mac OS X if they decide to show a label.
|
|
|
|
|
|
|
|
|
|
|
| |
QMotifStyle
Same fix as in e9a7e43031d7c1ee712e43be682c4e2c183759c4 but with motif
Reported by https://bugs.kde.org/show_bug.cgi?id=193911
Task-number: 255138
Reviewed-by: jbache
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This was quite a bug and it showed to some issues that I hadn't taken
into account when doing the initial port to Cocoa. The issue was that we
weren't "merging" items into the application menu if an item had already
been associated with it. Which seems OK for applications that create one
window with one menubar, but breaks down horrible when you have multiple
windows with each having their own menubar. The result is that items in
the application menu potentially go to the wrong window (and the
potential crash). Since there can only ever be one "Quit", "About", or
"Preferences" menu item in Cocoa, we need to make sure that we keep
these items in sync whenever we switch the menubar or remove actions
that are being deleted. That's what we do here.
FWIW, QActions with "ApplicationSpecificRole" for their menu role have
potential to cause memory leaks or other bugs if abused. If you are a
happy open source hacker who wants a thankless job, solving them would
get you lots of goodwill in my book.
Task-number: 255038
Reviewed-by: Richard Moe Gustavsen
|
|
|
|
|
|
|
|
|
|
|
| |
The blend functions assume the width / height of the images being
blended to be greater than 0. A width of 0 caused the first iteration of
a duff's device memcpy (QT_MEMCPY_USHORT) to be executed, thus blending
8 pixels instead of none.
BT: yes
Task-number: 255014
Reviewed-by: Trond
|
|
|
|
|
|
|
|
| |
The R and B channels were swapped on little endian machines with
BGR layout.
Task-number: 254934
Reviewed-by: Samuel
|
|
|
|
|
|
|
|
|
|
|
|
| |
both visible and invisible widgets.
This is a quick hack to avoid a crash in Qt when setting a focus on a
visible widget that has invisible parent. Proper fix was committed
into master 1a7da7096bbda17197738061902f4489af234bc0, see it's
description for more details.
Task-number: 254563
Reviewed-by: Thierry Bastian
|
|
|
|
|
|
|
|
|
|
|
| |
QFont has a feature that you can pass a comma-separated list and it will
walk through the list and match the font that it hits first. There's a
nice static function that X11 and Windows uses, but the Mac was using an
older copied version of it. This old version didn't handle quoting which
is what happens in the style sheet. So, using the same code makes
everything work well. As a bonus, Creator looks correct again.
Reviewed-by: Simon Hausmann
|
|
|
|
|
|
|
| |
This won't leak on error case, and will work with arbitrary sizes.
Also changed from int to short as instructed by Samuel.
Reviewed-by: Samuel <qt-info@nokia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On Mac, a widget with a NoFocus policy could still get focus
(if only temporarily) as the result of a native focus event.
In particular, a line edit with a completer should
not lose focus (if only for a brief moment) as a result of the
completer popup being shown. This will for example cause an
item delegate to think that the user has navigated away from
the cell and delete the line edit as a result. This will in turn
cause the completer to access a null pointer.
Reviewed-by: Richard Moe Gustavsen
Task-number: 254456 and 254460
|
|
|
|
|
| |
KDE Bug: https://bugs.kde.org/show_bug.cgi?id=191759
Reviewed-by: Bradley T. Hughes
|
| |
|
|
|
|
|
| |
Task-number: 254549
Reviewed-by: TrustMe
|
|
|
|
|
|
|
|
|
| |
We were adding two times in the QActionPrivate list the entry for
the current QGraphicsWidget if the action was existing before.
Task-number:KDE
Reviewed-by:bnilsen
BT:yes
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
An optimization was made to the sorting of QFileDialog to sort only
the current root (meaning what the user see). This avoided slowness
when the model was big with lots of leafs. The problem here is
for the treeview, the root is always the same, we expands only
nodes. In that case, a recursive sorting is needed to ensure that
all expanded nodes are correctly sorted (and filtered). This will
be slower that's why i use an hidden flag in the d pointer to
deactivate the recursive sort for the QFileDialog.
Task-number:254701
Reviewed-by:olivier
BT:yes
|
|
|
|
|
|
|
|
| |
When QWidget::scroll() is called on a widget with WA_PaintOnScreen,
scroll the dirty region.
Task-number: 254742
Reviewed-by: bnilsen
|
|
|
|
|
|
|
|
|
| |
If pixmaps were added through QIcon::addFile() with different sizes
than the sizes of the pixmaps themselves, streaming the icon in didn't
work properly.
Task-number: 254374
Reviewed-by: Kim
|
|
|
|
|
|
|
| |
Instead, query the mode at startup, store it and restore it on exit.
Reviewed-By: TrustMe
Task-number: 254194
|
|
|
|
|
|
|
|
|
|
| |
Make use of setInformativeText in qmessagebox for aboutQt dialog for
now. Proper fix might be to add scrollable widget to the dialog, or
split the about info into several pieces, though it cannot be done in a
patch release.
Task-number: 254464
Reviewed-by: Trenton Schulz
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This was already fix. But there was still a frame if there was a
stylesheet on the applicaiton.
The reason is that the frame's constructor call the style for some
hints. And later the combobox will query the style again for the frame
hint, before the view get polished.
The problem is that the StyleSheetStyle will fill the css cache with
wrong information on the first call.
This is not visible if there is no stylesheet as in the constructor, the
widget style is still the default application stylesheet if there is no
global applicaiton stylesheet.
Task-number: 254589
Reviewed-by: jbache
BT:
|
|
|
|
|
|
|
|
|
|
|
| |
We now set and unset the GTK_HAS_FOCUS flag on the same
painting call only if focus is set instead of resetting it
on each painting call. This is a tiny optimization but also kills
a possible assert on certain versions of Gtk+ (as reported with
Red Hat Enterprise Linux 5).
Task-number: 254614
Reviewed-by: denis
|
|
|
|
| |
Removed unused line and added task number to autotest.
|
|
|
|
|
|
|
|
|
| |
When setVisible() is called on a QGraphicsItem, if the parent of that
item was hidden, the child shouldn't be actually shown.
Task-number: 197802
Reviewed-by: leo
Reviewed-by: alexis
|
|
|
|
|
|
|
|
| |
The problem was that we don't clear the selection model if the previous
selection was valid.
Task-number:251341
Reviewed-by:jasplin
|
|
|
|
|
|
|
|
|
|
| |
We should use the "statusbar" string when we draw the sizegrip.
Since gtkstatusbar is the only gtk widget using the paint_resize_grip
and it uses the "statusbar" string itself it should be safe to
assume this.
Task-number: 254535
Reviewed-by: rosch
|
|
|
|
|
|
|
|
|
|
|
|
| |
Reworked the 85f98acaa3a38079071bea711e43c9a86edec1f6 fix, since
it broke glyph positioning in the GL engine under Windows.
Instead of changing the glyph cache margin, which impacts where
the glyph is positioned, we just make the image the glyph is drawn
into 4 pixels bigger in width/heigth.
The margin in QImageTextureGlyphCache needs to be reworked..
Task-number: 254450
Reviewed-by: Eskil
|
|
|
|
|
|
|
|
|
|
| |
Seems like no code was written to handle other font engines than
CoreText. Unfortunatly the engine on Carbon is ATSUI. This patch
adds general code for converting a QFont to an NSFont so the dialog
can support other engines than CoreText
Task-number: 251957
Reviewed-by: Trenton Schulz
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
the model initially
This was caused by two different bug:
- In the QSortFilterProxyModel, we need to re-sort when setting the
source model change the sorting column (happen when setting a model
initially)
- In the treeview, we need to activate the sorting even if there is no
column yet (because the initial model is empty
Task-number: 254234
Reviewed-by: Thierry
BT:
|
|
|
|
|
|
|
|
|
| |
Before adding a / to the path one should check if it doesn't end with
one already. Otherwise one might get paths like //My Documents on
WinCE causing the native call to crash the filesystem service on that
system.
Reviewed-by: alexis
|
|
|
|
|
|
|
|
|
| |
If a QCursor with a predefined shape is declared static, it could be
destroyed after the application dtor has already cleaned up QCursor
memory.
Task-number: 254467
Reviewed-by: Rhys Weatherley
|