| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
| |
not C99-compliant.
C99 says vsnprintf returns >= `size' when not enough memory was
available. In practice, on Windows, glibc < 2.0.6, and probably plenty
of other places it returns -1 instead.
Reviewed-by: Rhys Weatherley
|
|
|
|
|
|
|
|
|
|
|
| |
Most log elements in all loggers can now be of (more or less)
arbitrary length. Selftests pass and are memcheck-clean. The
increase in heap usage seems negligible (it's below the default
minimum threshold of massif).
Task: 253861
Reviewed-by: Michael Goddard
|
|
|
|
|
|
|
|
|
|
| |
This commit contains the bare minimum needed for the API to use dynamic
allocation; some parts of the code still use static buffers and
therefore have constraints on the size of output messages.
Task: 253861
Reviewed-by: Michael Goddard
|
|
|
|
|
|
|
|
|
|
|
| |
This enables very long failure messages (e.g. including an entire
build log in a failure message).
This change modifies only the plain test logger to use this feature.
Task: 253861
Reviewed-by: Michael Goddard
|
| |
|
|
|
|
|
|
|
|
| |
If comparing two variants fails, the failure message will now output
the type and value of the variants (rather than
"Compared values are not the same").
Reviewed-by: Thiago
|
|
|
|
|
| |
Don't use old-style casts in Qt code. And avoid signed/unsigned
comparisons (sizeof returns size_t, which is unsigned).
|
|
|
|
| |
Whoops, don't do sigaction for (nonexistent) signal 0.
|
|
|
|
|
|
| |
Feature define logic was wrong
Reviewed-by: Robert Griebl
|
| |
|
| |
|
|
|
|
|
|
| |
Don't try to write to files with special characters in the name.
Also fix bizarre indent.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
terminated by a signal.
The new XML testloggers in Qt 4.6 only output a testlog after the test
completes. Therefore, if a test crashes, no testlog will be output at
all. This is a functional regression from Qt 4.5 and earlier where the
testlog would be output up to the point where the crash occurred.
This is a Unix-specific fix for this problem.
This change is also useful for hanging tests; if the test is killed
with SIGTERM it will output the test log up to the current test before
it exits.
|
|
|
|
|
| |
Merge-request: 604
Reviewed-by: Marius Storm-Olsen <marius@trolltech.com>
|
|
|
|
| |
Reviewed-by: Trust Me
|
|
|
|
| |
QTest::touchEvent().
|
|
|
|
|
|
|
|
|
| |
This enum indicates what kind of device generated the touch event
(TouchScreen or TouchPad). We use this information to control how touch
events are sent, specifically we restrict touch events to a single
widget/QGraphicsItem on touch-pads, since there is no direct
relationship between the physical touch location on the pad and the on-
using the touch-pad).
|
|\ |
|
| |
| |
| |
| | |
Fixes the very frequently asked question "how do I see the full output?"
|
| | |
|
|\ \
| |/
| |
| |
| | |
Conflicts:
src/gui/graphicsview/qgraphicsscene_p.h
|
| |
| |
| |
| | |
Reviewed-by: Trust Me
|
| |\ |
|
| | |
| | |
| | |
| | | |
Reviewed-by: Trust Me
|
| | | |
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
the API for these 2 classes is identical, the implementation is almost
identical, they share the same data structures, so bite the bullet and
merge them.
this means we go back to using screenPos() instead of globalPos()
again
|
| | |
| | |
| | |
| | | |
how embarassing :/
|
| | |
| | |
| | |
| | | |
Need to move the extern declaration out of the QtTest namespace
|
| | | |
|
|\ \ \
| |/ / |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
seem fixable easily)
Merge-request: 594
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
All Qt headers must have QT_BEGIN_HEADER, QT_END_HEADER and
QT_MODULE. These headers didn't have this.
Reviewed-by: TrustMe
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* Added a context for the touch positions in press(), move(),
release() functions.
* Sending RawTouch events to QApplication object if the target widget
is not specified.
|
|/ /
| |
| |
| | |
Modified it the same time to remove refcount and static variable.
|
| |
| |
| |
| |
| | |
used character operations whenever possible
better usage of QLatin1String
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
xunitxml testlib logger.
In the testlib plain logger, xfail is considered a pass and xpass
considered a fail. xunitxml had the opposite behavior; change it
to be the same.
Autotest: included
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
It is now possible run the benchmarks with "-chart" to generate
charts for benchmarks that produce chartable data. See
examples/qtestlib/tutorial5 for an example.
This patch has minimal inpact on QTestLib itself. Most of the
functionality is implemented in a separate tool which QTestLIb
calls after detecting the chart command line option.
|
| |
| |
| |
| | |
Reviewed-by: Trust Me
|
| |
| |
| |
| | |
Reviewed-by: Trust Me
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
xunitxml testlog.
`error' in xunit causes many xunit processing tools to consider the test
as failing. This is different semantics from native testlib XML, where
a qDebug is considered a message and not a failure.
Change it to put qDebugs under the "system-err" tag when using xunitxml
to retain semantic compatibility with testlib XML.
Autotest: tst_selftests
|
| |
| |
| |
| |
| |
| |
| | |
When failing without a filename, setting file to 0 does not work.
File must be set to "".
Autotest: 2d194009379a4e199ec246ca053ee97ee733482f
|
| |
| |
| |
| |
| |
| |
| | |
Obvious typo in format string: `<DataTag><![CDATA[%s]]></Description>',
closing tag should be `DataTag'.
Autotest: 2d194009379a4e199ec246ca053ee97ee733482f
|
|\ \
| |/
| |
| |
| |
| |
| | |
Conflicts:
src/corelib/global/qfeatures.h
src/gui/painting/qtransform.cpp
util/scripts/make_qfeatures_dot_h
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
`-developer-build'.
This was an undocumented hack to make it easier for our autotest scripts
to detect crashes in unit tests. Going forward, it should not be
required.
Reviewed-by: Shane Bradley
Acked-by: Jesper Thomschütz
|
| |
|
|
|
|
| |
Reviewed-by: Andy Shaw
|