| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|\ |
|
| |
| |
| |
| | |
Reviewed-by: paul
|
|/
|
|
|
|
|
| |
The test passes without these workarounds in 5800 XpressMusic and
emulator
Reviewed-by: TrustMe
|
|
|
|
|
|
| |
See: 82275b4c03a0. This workaround applies also for Symbian.
Reviewed-by: TrustMe
|
|
|
|
| |
Reviewed-by: TrustMe
|
|
|
|
|
|
|
| |
Symbian OS has working networking support now, so we donät need to skip
this test anymore.
Reviewed-by: TrustMe
|
| |
|
|
|
|
| |
Reviewed-by: Bill King
|
|
|
|
|
|
| |
Added more checking in each step of the playback or recording tests.
Reviewed-by:Justin McPherson
|
|
|
|
|
|
|
| |
The SRCDIR entry was added to the project file to correct the deployment
for WindowsCE. Also the process start is now monitored in the test.
Reviewed-by: Joerg
|
|
|
|
|
|
|
| |
The deployment for Windows CE did not work as the processes could not
be located properly on the device.
Reviewed-by: banana joe
|
|
|
|
|
|
|
|
|
|
| |
qsystemsemaphore autotest does not compile for S60 and WinCE without
QtScript and QtScript currently does not compile for S60, that's why
the following statement was added to pro file:
requires(contains(QT_CONFIG,script))
Reviewed-by: TrustMe
|
|
|
|
|
|
|
|
|
|
| |
qsharedmemory autotest does not compile for S60 and WinCE without
QtScript and QtScript currently does not compiled for S60, that's why
the following statement was added:
requires(contains(QT_CONFIG,script))
Reviewed-by: TrustMe
|
|
|
|
|
|
| |
Symbian OS doesn't support environment variables.
Reviewed-by: Miikka Heikkinen
|
|
|
|
|
|
|
|
| |
Specialized test code to set higher stack for this test case is not
needed anymore since default stack for Qt thread in S60 now is 80 KB.
See commit: 9324fdde
Reviewed-by: TrustMe
|
|
|
|
|
|
|
|
|
| |
The polling watcher was generating two events for directory content
deletion on Symbian emulator. 1 for content change, 1 for time change.
Seems within the bounds of reason for a poller, so test relaxed to
allow it.
Reviewed-by: Liang QI
|
|
|
|
|
|
|
| |
The image comparison reference snapshots were not up to date. Also
the comparison ignores the title bar completely.
Reviewed-by: banana joe
|
|
|
|
|
|
|
|
| |
This test would trigger random Q_ASSERT failures but should
work fine after the previous commit.
Signed-off-by: Eduardo M. Fleury <eduardo.fleury@openbossa.org>
Reviewed-by: Jesus Sanchez-Palencia <jesus.palencia@openbossa.org>
|
|
|
|
|
| |
Signed-off-by: Jesus Sanchez-Palencia <jesus.palencia@openbossa.org>
Reviewed-by: Eduardo M. Fleury <eduardo.fleury@openbossa.org>
|
|
|
|
|
| |
Signed-off-by: Jesus Sanchez-Palencia <jesus.palencia@openbossa.org>
Reviewed-by: Eduardo M. Fleury <eduardo.fleury@openbossa.org>
|
|
|
|
|
| |
Signed-off-by: Jesus Sanchez-Palencia <jesus.palencia@openbossa.org>
Reviewed-by: Eduardo M. Fleury <eduardo.fleury@openbossa.org>
|
|
|
|
|
|
|
| |
Test that the variable is stored in the correct activation object.
Currently it's marked as an expected failure because it doesn't work
with the JSC-based back-end. It did however work with the old back-end,
so this is a regression.
|
|\ |
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Some network tests depend on reading file and comparing it with
data fetched from network. On Windows, if not set properly, git will
append CRLF when checking out. This makes tests to fail as sizes and
binary content are not as expected.
Reviewed-by: Janne Anttila
|
| |\ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When we changed the sibling stacking order to be defined (4.5) instead of
undefined (4.2, 4.3, 4.4), the need to control this stacking order
arose. Before we could just say the order was random, but stable,
and the only way people could rely on order was to set Z. Now, when the
default order is defined as "insertion order", people start relying on
this order, and incidentally they want more control.
In QML, the need to have insertion order semantics is very evident as
the order you define the elements in QML more strongly implies a
graphical stacking order than the imperative order they get when added
in C++.
This change adds QGraphicsItem::stackBefore(const QGraphicsItem *), which
works similarily to QWidget::stackUnder(). It moves the item in front of
the sibling item passed as an argument.
While implementing this function, and writing tests for how this
function behaves in combination with Z values, I found that the code
we had for updating siblingIndex was broken in the case where you remove
an item from the middle of the children list. In this case newly added
items would be assigned the same sibling index order as one that's already
in the list.
So in order to get the tests to pass I had to fix this bug as well.. The
approach is to sort the children list by insertion order, so that we can
fix up the sibling indexes.
Performancewise this has little implications. If there are gaps in
the sibling index list, which only occurs if you remove an item from the
middle of the children list, will the sibling index list be adjusted /
corrected before used (for example, by stackBehind()). Multiple calls to
stackBehind will be fast, and the list is flagged for resorting (including
Z order).
Reviewed-by: jasplin
|
| |/
| |
| |
| | |
One item slipped and was now added.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Tests showAndMoveChild and rectOutsideCoordinatesLimit_task144779
were made desktop client area dependent so the screen capturing
targets meaningfull rectangles on Windows CE.
Reviewed-by: Joerg
|
| |
| |
| |
| |
| |
| | |
WindowsCE will not do any tests involving dynamic casts.
Reviewed-by: Joerg
|
| |
| |
| |
| |
| |
| | |
Also added check, whether some processes where actually not started.
Reviewed-by: Joerg
|
| |
| |
| |
| |
| |
| | |
The <sys/types.h> header is unavailable in Windows CE
Reviewed-by: Joerg
|
| |
| |
| |
| |
| |
| | |
Mismatching preprocessor statements cleaned up.
Reviewed-by: Joerg
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The failing (3) values are explicitly excepted here.
The source values for the comparison table should remain untruncated
doubles and the error bound checking function should be made dynamic.
Also the source values should come from a "trusted" source and not
from QEasingCurve itself.
Reviewed-by: Joerg
|
| |
| |
| |
| |
| |
| |
| | |
Changed the absolute size values for the combobox to desktop
dependent sizes.
Reviewed-by: Joerg
|
| |
| |
| |
| |
| |
| | |
Tests made unavailable for Windows CE testing
Reviewed-by: Joerg
|
| |
| |
| |
| |
| |
| |
| | |
QPlastiqueStyle is not compiled for Symbian OS, but default ones seems
to make the job fine since all get passed.
Reviewed-by: Janne Koskinen
|
|/
|
|
|
|
|
| |
Nokia X86 compiler does not resolve QCOMPARE corrctly for derived types,
that's why we cast it explicitly.
Reviewed-by: Janne Koskinen
|
|
|
|
| |
Reviewed-by: TrustMe
|
|
|
|
| |
Reviewed-by: Jason Barron
|
|
|
|
| |
Reviewed-by: TrustMe
|
|
|
|
| |
Reviewed-by: Simon Hausmann
|
|\
| |
| |
| |
| |
| |
| | |
Conflicts:
examples/webkit/formextractor/formextractor.pro
mkspecs/features/qt.prf
src/gui/painting/qpaintengineex.cpp
|
| |
| |
| |
| |
| |
| |
| | |
This reverts commit df47e0d40290f5e40054a9612f75177d9ef8537a.
There is no "private_build" in Qt 4.5
Reviewed-By: thiago
|
| |
| |
| |
| | |
Reviewed-By: TrustMe
|
| |
| |
| |
| |
| |
| |
| |
| | |
QMenu actually doesn't use QFontMetrics::width to calculate its bounding
rect but QFontMetrics::boundingRect. That's why this test failed on
Windows CE.
Reviewed-by: thierry
|
| |
| |
| |
| |
| |
| | |
static_cast from QGraphicsView* to QWidget*. Change QPlastiqueStyle to QWindowsStyle for some platforms like S60.
Reviewed-by: Jason Barron
|
| |
| |
| |
| |
| |
| | |
Test qaudiodeviceid was replaced with qaudiodeviceinfo.
Reviewed-by: Paul
|