| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
Optimizations in change 8e447e8a did not handle the case when the target
width is less than the width of a tile.
Task-number: 260759
Reviewed-by: Samuel
|
|
|
|
| |
Reviewed-by: trustme
|
|\ |
|
| |
| |
| |
| |
| | |
The unicode logic is beyond me, leave these for fixing by someone
else with more unicode knowledge.
|
| |
| |
| |
| |
| |
| | |
Makes the behaviour consistent with the ODBC driver behaviour
Reviewed-by: Justin McPherson
|
|/
|
|
| |
Reviewed-by: trustme
|
|
|
|
| |
Reviewed-by: trustme
|
|
|
|
| |
Reviewed-by: trustme
|
|
|
|
| |
Reviewed-by: trustme
|
|\ |
|
| |\ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The `test' builtin in Solaris' /bin/sh does not understand `-e', use
`-f' instead.
Fixes: configure: test: argument expected
Note, using the `-nokia-developer' configure option hides this
breakage.
Reviewed-by: Rhys Weatherley
|
|/ /
| |
| |
| | |
Reviewed-by: trustme
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
QGLFormat was being deep-copied many times per frame because of
code like this:
if (context()->format().doubleBuffer()) { ...
This change modifies QGLFormat to use implicit sharing to reduce
the overhead of the above type of checks.
Reviewed-by: Sarah Smith
|
| |
| |
| |
| | |
Reviewed-by: trustme
|
| |
| |
| |
| |
| |
| |
| | |
The documentation says "Returns true if all the options of the
two QGLFormats are equal", but that's not what it was doing.
Reviewed-by: Sarah Smith
|
| |
| |
| |
| | |
Reviewed-by: Alex
|
| |
| |
| |
| |
| |
| |
| | |
The custom shader code in the OpenGL2 paint engine needs time to
mature before we make this official public API.
Reviewed-by: trustme
|
| |
| |
| |
| | |
projects where two separate compilation units include `qtestmouse.h'.
|
| |
| |
| |
| | |
qtestmouse.h.
|
|\ \
| |/
| |
| |
| | |
Conflicts:
src/gui/itemviews/qlistview.cpp
|
| |
| |
| |
| |
| | |
Commit 2e0d78836becf24c7f27c982316cf1b4492f27aa fixed this for i386
but omitted the fix for x86_64.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
On 10.5 and below the pixmap data bytes happened to be initialized by
malloc, on 10.6 this is no longer the case.
Revby: Gunnar Sletta <gunnar@trolltech.com>
Similar to 28f94e1ef94f.
|
| |
| |
| |
| |
| |
| |
| | |
We need to clear the QPixmap before the image is drawn using the
CGContextDrawImage().
Reviewed-by: Norwegian Rock Cat
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When setRowHidden() was called after a root index was set then it would
not actually hide the row, if a root index is not set then it worked
fine.
Task-number: 260879
Reviewed-by: Jan-Arve
|
| |
| |
| |
| | |
Q_DECLARE_METATYPE must appear before qRegisterMetaType.
|
| |
| |
| |
| | |
Reviewed-by: TrustMe
|
| |
| |
| |
| |
| |
| |
| |
| | |
QRasterizer, QClipData and QRasterBuffer only need to exported when
QRasterPaintEnginePrivate is subclassed which only is the case for
DirectFB.
Reviewed-by: Noam Rosenthal <noam.rosenthal@nokia.com>
|
| |
| |
| |
| |
| | |
Reviewed-by: Trust Me
As-seen-on: qt-interest
|
| |
| |
| |
| |
| | |
Task-number: 258991
Reviewed-by: Trust Me
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If we make sure to erase the background to background color in connect()
we don't have to do it in setGeometry.
Also clean up the code in QDBWindowSurface::flush
Reviewed-by: Donald Carr <donald.carr@nokia.com>
|
| |
| |
| |
| |
| |
| |
| | |
No need to carry those extra bytes around when QT_DIRECTFB_WM is
defined.
Reviewed-by: Donald Carr <donald.carr@nokia.com>
|
| |
| |
| |
| | |
Reviewed-by: Donald Carr <donald.carr@nokia.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
This code is roughly the same as in other screen drivers. We shouldn't
flush widgets that are in a graphicsscene as a proxy widget. They will
be handled by the graphicsview.
Reviewed-by: Donald Carr <donald.carr@nokia.com>
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When Math3D was introduced, the boxes demo was changed to use the
classes in the new module, but the change caused some bugs with the
object rotations. This commit should fix them.
Reviewed-by: Samuel
|
| | | |
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
If platform headers have been included the Interlocked names may be
macros and we end up with double underscores in our function names. By
prefixing "Interlocked" and then the actual function names we work
around that.
Let's hope nobody redefines Interlocked or Increment and friends... :-)
Reviewed-by: Prasanth Ullattil
|
| | |
| | |
| | |
| | |
| | |
| | | |
the test may not have any file on it.
Reviewed-by: Olivier
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
These functions are implemented directly in assembly, so they need the
proper directives to enable/disable visibility. On ELF systems, it's
.hidden, whereas on Mach-O systems (Mac) it's .private_extern. On
Windows, it's not necessary since you have to explicitly export. I
also implemented the AIX idiom, though it's unlikely anyone will
implement AIX/POWER JIT.
That leaves only HP-UX on PA-RISC unimplemented, from the platforms
that Qt supports. It's also unlikely that we'll imlpement JIT for it.
Reviewed-by: Kent Hansen
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This lets us stream a single QPaintBuffer instead of one QPaintBuffer
per frame in the trace graphicssystem, which leads to not streaming
pixmaps / images once per frame. Performance when doing a trace is also
a lot better for painting heavy applications.
Reviewed-by: Trond
|
| | |
| | |
| | |
| | |
| | |
| | | |
This reduces the size of the trace files significantly in certain cases.
Reviewed-by: Trond
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We shouldn't fully unregister timers when the event dispatcher is stopped
when a thread exits, since this releases the timerId back to the pool.
Instead, only free the OS resources.
Auto-test included.
Reviewed-by: ogoffart
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Second try.
This is a fixed version of the one reverted in commit 25014061b8de
The difference is that it does
text[offset] = QChar::LineSeparator;
instead of
chr = QChar::LineSeparator;
Reviewed-by: Eskil
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This fix prevents crashes in
QGraphicsItem caused by calls to pure virtual functions as an item
tries to access its dying parent. This happens when an item is moved
by an ancestor during this ancestor's destructor.
Autotests included.
Reviewed-by: bnilsen
|
| | |
| | |
| | |
| | |
| | |
| | | |
This also needs to be reviewed by andreas.
Reviewed-by: Michael Brasser
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Note that the constraint is only enforced when resizing the window
interactively. Calling setGeometry() will not try to enforce the
constraint.
See the graphicsview/flowlayout example for a "manual test".
Task-number: 257455
|
| | |
| | |
| | |
| | |
| | | |
Adjust the constraint with the widget margin before passing it on to
d->layout->effectiveSizeHint().
|