| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Commit 25e76959da84fe4c40f98cf32b7b8c69e5087681 changed it for the
interpreter, this commit makes it work with the JIT enabled as well.
|
| | | | |
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Since QScopedPointer can take a custom cleanup handler we can use this
instead of the original QDirectFBPointer.
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This just caused too much problems and must be solved another way.
In qfsfileengine_win.cpp we define FSCTL_GET_REPARSE_POINT and all
of the other stuff that's needed for NTFS symlink support, if
its not defined. This is the case if _WIN32_WINNT is less
than 0x0500.
All other changes in this commit are just reversions of commits
that were done for the infamous qglobal.h change.
Discussed with prasanth, tested by pulse.
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
and y-axis inversion and overall less conversion, making significantly faster
for plain usecases
Reviewed-by: Trond
|
| | | |
| | | |
| | | |
| | | | |
Reviewed-by: Trust Me
|
| |/ /
|/| |
| | |
| | | |
Reviewed-by: Trust Me
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This makes mixing GL and QPainter code safer. We need to be able to
assume default GL state in begin(), and set back whatever we change to
the default state in end() in the GL 2 paint engine.
Reviewed-by: Trond
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Only use bilinear filtering when SmoothPixmapTransform render hint is
used.
Reviewed-by: Kim
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The recommended specifiers are lowp for colors / normal vectors, mediump
for texture coordinates when a limited range is sufficient, and highp
for generic texture coordinates and vertex coordinates / transformation
matrices.
We used to use mediump for texture coordinate in some places, but
since we don't control the texturing scenarios we need to handle the
worst case, which is zooming in on part of a large texture (2048x2048)
with bilinear filtering. To properly handle this case without color
banding mediump is probably not sufficient, so we'll use highp for
texture coordinates.
Reviewed-by: Tom
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Previously we were using QPaintEngine::syncState() which is not ideal
naming-wise, since it actually prepares for native painting instead of
syncing the painter's state to native state.
Reviewed-by: Trond
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This adds the function beginMoveRows, endMoveRows, beginMoveColumns,
endMoveColumns
Reviewed-by: Olivier Goffart <ogoffart@trolltech.com>
Acknowledged-by: Thierry
Merge-request: 972
|
|/ / /
| | |
| | |
| | |
| | | |
Task-number: 201882
Reviewed-by: Trust Me
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
belonged to a deleted script engine
When the engine is deleted, the JSValue is invalidated, but the
QScriptValue's type will still be QScriptValuePrivate::JSC.
Use a new helper function, isObject(), that checks both that the
value is of type JSC _and_ that it is valid, before calling
JSValue::isObject() (JSValue::isObject() assumes that the value
is valid).
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
If you do a selection with the mouse and react to selectionChanged by
changing the selection. Those changes would be overwritten by
QAbstractItemView::mouseReleaseEvent.
It is useless to set the selection on mouse release. We already do
that on mouse press.
Task-number: 250683
Reviewed-by: ogoffart
|
| | |
| | |
| | |
| | | |
Graphics effects are no longer in the Graphics View module.
|
| | |
| | |
| | |
| | | |
Reviewed-by: Kavindra
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
If GetTextMetrics() should fail, the results are undefined. When the
undefined data are used, e.g. when painting text, this can cause a
crash. To avoid the crash and make it clear that the metrics cannot be
retrieved, we zero out the entire structure.
Task-number: 251172
Reviewed-by: gunnar
|
| | |
| | |
| | |
| | |
| | | |
This reverts commit f7ebdd380d16a7be9713930b5ab41c32e996dcdb.
defining WINVER in qglobal.h is about to be reverted.
|
| | | |
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The extra ".data()" call is not needed here because QScopedPointer
overloads operator!() to do the right thing here.
Reviewed-by: Eskil Blomfeldt
|
|/ / /
| | |
| | |
| | |
| | | |
Task-number: 254511
Reviewed-by: Trust Me
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Column number in executed JS source code is correctly passed to
debugger. Few autotest were repaired.
Reviewed-by: Kent Hansen
|
| | |
| | |
| | |
| | | |
Reviewed-by: janarve
|
| |/
|/|
| |
| |
| |
| |
| | |
Previously it wouldn't and the layout could appear to be broken.
Task-number: 248817
Reviewed-by: Gabriel De Dietrich
|
| |
| |
| |
| |
| |
| |
| | |
Calling QScriptEngine::toStringHandle() is dead slow, so don't call it;
use JSC::Identifier directly.
This is the same issue as was fixed for setProperty() in commit
a8574172dd5e6bc11cf6f69b6fad5a063549e88d.
|
| |
| |
| |
| |
| |
| | |
The precision specifiers need to be there on OpenGL ES 2.0.
Reviewed-by: Tom
|
|/
|
|
|
|
| |
Otherwise the property is stored on the wrong object (the proxy).
This fix makes the Qt bindings generated by qtscriptgenerator work
again.
|
|
|
|
|
|
| |
classes
Reviewed-By: TrustMe
|
|
|
|
|
|
| |
The member need to be initialized in the same order as in which they are declared
Reviewed-by: Thierry
|
|
|
|
|
|
|
| |
This fixes the lifetime of the mutex locker, since before it would
already be destroyed on the next line.
Reviewed-by: Olivier Goffart <olivier.goffart@trolltech.com>
|
| |
|
|
|
|
|
|
|
| |
Lots of warnings in the qreal == float case.
Some Q_UNUSED added.
Reviewed-by: thartman
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Don't crash when parsing
"<?xml version=\"1.0\" encoding=\"iso-8859-1\" ?><guid><http://www.foo.dk/artikel/8938</guid>"
(unmatched "< tag followed by "foo:") using QDomDocument::setContent
together with a QXmlSimpleReader with the
"http://xml.org/sax/features/namespaces" feature enabled.
Fixes task tracker issue 254700. See there for a test case.
Merge-request: 1322
Reviewed-by: Peter Hartmann <peter.hartmann@trolltech.com>
|
|
|
|
| |
Reviewed-by: hjk
|
|
|
|
|
|
|
|
|
|
| |
QVectorPath::hints() is not a strict bit field, and thus can not be
anded with RectangleHint. Instead, QVectorPath::shape() should be
directly compared with RectangleHint to check if the vector path is a
rectangle or not. In this case the first four points of a regular
painter path were treated as a rectangle with dire consequences.
Reviewed-by: Tom
|
|
|
|
|
|
|
| |
We now need to connect to the slot in the QLineControl and not to the
slot of the QLineEdit (the QLineEdit slot is now also removed).
Reviewed-by: Alan Alpert
|
|
|
|
| |
13 pairs missing...
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Why this happends is a bit blurry. From before, I know that
cocoa is a bit buggy regarding setting a menu item
hidden or not. The solution back then resulted in the function
syncNSMenuItemEnabled in qmenu_mac.mm. This patch basically
applies the same (silly) trick; disabling the menuitem before
enabling it. This seems to force an update to the menu items
enabled state. For the record: this is not a fix that I
embrace. I hope we can remove it again some day. See
task for how to reproduce.
Task: 259600
Rev-By: alexis
|
|
|
|
|
|
|
|
| |
Had left the cursor flash time as half the QApplication cursor flash
time by mistake. The new function sets the whole period and not just the
time between toggling cursor visibility.
Reviewed-by: Trust Me
|
|
|
|
|
|
|
| |
To prevent using IDirectFBImageProviders one can define
QT_NO_DIRECTFB_IMAGEPROVIDER
Reviewed-by: Donald Carr <donald.carr@nokia.com>
|
|
|
|
| |
Reviewed-by: Donald Carr <donald.carr@nokia.com>
|
|
|
|
|
|
|
|
| |
Some DFB implementations do not support rendering the cursor using the
intended interfaces. In these cases one can define
QT_DIRECTFB_WINDOW_AS_CURSOR and use a window to render the cursor.
Reviewed-by: Donald Carr <donald.carr@nokia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
If you define QT_DIRECTFB_IMAGEPROVIDER_KEEPALIVE we make sure atleast
one IDirectFBImageProvider is alive at all times.
Apparently this is refcounted by DirectFB on atleast one implementation
and there's considerable overhead involved when releasing the
last/creating the first image provider.
Reviewed-by: TrustMe
|