| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
This was causing a crash on windows because the buffer and period sizes
were worked out to 0 with an invalid sample size and dividing one by the
other is division by 0.
Task-number: QTMOBILITY-438
Reviewed-by: Justin McPherson
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When inputting only numbers, the symbol menu should not do anything.
However in the old code the resource id of the symbol table was still
being set, so the symbol key on N97 would look up a table that was
not valid for the current input mode and crash.
Fixed by setting the symbol table id to zero under those conditions.
RevBy: Sami Merila
Task: QTBUG-13472
AutoTest: Included
(cherry picked from commit 5cef786a651c675d3428060a19bfd9d9ecee6083)
|
|
|
|
|
|
|
|
|
|
|
|
| |
Problem was reproduced on N97. If the FEP detects that the widget is
full while still editing text, it will try to send those events as
key events instead. Since this screws up the content in the widget,
we stop those events from reaching the widget in the input context.
AutoTest: Passed
Task: QTBUG-12949
RevBy: Miikka Heikkinen
(cherry picked from commit 5ca6264933af60b3cd376b7f08bea008fa69b515)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This fixed a case where using password mode would lead to Qt
believing that the cursor was one step to the left of where it really
was. This would have the effect of replacing the last character
instead of appending to it, and even crashing if the cursor was all
the way to the left.
The code is also much cleaner this way, but it meant that QTBUG-9867
had to be solved differently this time. We do this by assuming that
and empty FEP update means "erase last character", which seems to
work well in practice.
Also added a long overdue autotest for the FEP input methods. Most
tests pass, but some don't, which I will try to fix later.
Task: QTBUG-9867
Task: QTBUG-12949
RevBy: Miikka Heikkinen
AutoTest: Included
(cherry picked from commit 52cf47565e402dc682038ccaf8d725401802b603)
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This has already been fixed in 4.7 (QTBUG-9770), but the change is
too big to backport.
The general idea is the same: Only operate on UTC dates internally
(since that's how JS dates are stored), and let QDateTime take care
of converting from/to local dates as necessary.
The fix itself shouldn't be merged to 4.7, but the autotests should.
Task-number: QTBUG-9770
Reviewed-by: Jedrzej Nowacki
|
|
|
|
| |
Task-number: QTBUG-12908
|
|
|
|
|
|
|
|
|
| |
This fixes Westpoint Security issue with Advisory ID#: wp-10-0001.
Before, we would allow wildcards in IP addresses like *.2.3.4 ; now,
IP addresses must match excatly.
Patch-by: Richard J. Moore <rich@kde.org>
Task-number: QT-3704
|
|
|
|
|
|
| |
Merge-request: 731
Task-number: QTBUG-4455
Reviewed-by: Peter Hartmann
|
|
|
|
|
|
|
|
|
|
| |
Mac Phonon backend does not support opening a resource file via the
MediaSource(const QString &) overload. The MediaSource(QIODevice *)
overload must be used. The Symbian backend, on the other hand,
requires the former overload to be used.
Task-number: QTBUG-6562
Reviewed-by: trustme
|
|
|
|
|
| |
Task-number: QTBUG-6562
Reviewed-by: Justin McPherson
|
|
|
|
|
|
|
|
| |
* Removed Q_SKIP on Symbian
* Divided test into two, to exercise the two possible ways of providing
the resource path to Phonon::MediaSource
Task-number: QTBUG-6562
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Calling QIconModeViewBase::initDynamicLayout() on the second and
successive segments would return QPoint(-1,-1), resulting in a
totally empty area rectangle for all the items while in
QIconModeViewBase::doDynamicLayout(). This rectangle is used to
initialize the BSP tree, and produces an arithmetic exception when
empty.
Furthermore, a rendering bug was also apparent when displaying the
first item of a segment while the last item of the previous segment
was hidden.
Auto-tests included.
Reviewed-by: Olivier
Task-number: QTBUG-12308
|
|
|
|
|
|
| |
Task-number: QTBUG-10643
Task-number: QTBUG-11376
Reviewed-by: Jason Barron
|
| |
|
| |
|
|
|
|
|
|
|
| |
Trivial fix, valid formats have pixel format != Invalid, not ==.
Task-number: QTBUG-12337
Reviewed-by: Andrew den Exter
|
| |
|
| |
|
|
|
|
|
|
| |
This logic for private_tests was the opposite of what it should be,
meaning that network tests which use Q_AUTOTEST_EXPORTs have not been
run for some time :-(
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We used to delete the tracked object before the new tracking was
properly set up in QSharedPointer. That means if the tracked object's
destructor or deletion function recursed into the QSharedPointer, it
would find itself in an inconsistent state.
So instead finish the setup and only then call out to user code. That
ensures the internal state is always valid.
Task-number: QTBUG-11730
Reviewed-by: ossi
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Change 4b709b41f5a7ae8dc6e537b644158f5201ce0a98 tried to make sure that
rows with merged cells would not be completely removed, as this would
cause a crash. However, when removing just a few columns from a merged
cell, the span of the cell should be reduced by the number of columns
removed. The "touched" guard would cause the span to be decreased a
maximum of one time, regardless of how many columns were removed,
leaving the table in an invalid state, as the column count would be
smaller than the span of the cell. This would assert later on.
To avoid this, we only guard against removal, not against decreasing
the span.
Task-number: QTBUG-11646
Reviewed-by: Thomas Zander
|
|
|
|
|
|
| |
8601-2004 standard."
This reverts commit 752d46c90ee0fc5f06f01feedd8e0659178b15d4.
|
|
|
|
|
|
|
|
|
| |
8601-2004 standard.
Task-number: QTBUG-11623
Reviewed-by: Denis Dzyubenko
Reviewed-by: David Boddie
|
|
|
|
|
|
|
|
|
|
| |
If item clips children to shape we should take that into account when
computing the source rect, to avoid creating unnecessary large source
pixmaps which might have a significant performance impact.
Task-number: QT-3551
Reviewed-by: Gunnar
Reviewed-by: Trond
|
|
|
|
| |
Reviewed-by: Trust Me
|
|
|
|
|
|
|
|
|
|
|
| |
The test code used the 'using' keyword to try and change access control
of a base class method from protected to public.
With the RVCT 2.2 compiler, 'using' imports the function(s) from the base
class, but they retain their existing access control.
Used an inline public function to call the base class as a workaround
Reviewed-by: mread
|
|
|
|
|
|
|
|
|
|
|
|
| |
Deployment of the test files to ../xmlpatterns resolves to
c:/private/xmlpatterns, which is an illegal location to deploy to on
symbian.
Changed the deployment and test to put the xmlpatterns directory inside
the app's private directory on symbian
Regression tested on windows build.
Reviewed-by: mread
|
|\
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
Fix possible crash in QTextLayout for glyphless items
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Change e1915815bc5ef86b3844608bba46769da5173363 moved part of the
right bearing check out of the "non-whitespace-or-object" block of the
layout, which could potentially cause crashes for layouts that contained
items that were line separators or tabs etc. because we would access
the logical clusters array based on the position of e.g. the tab even
though it didn't have an entry. This could potentially give us an
arbitrary index which might cause an out of bounds when accessing the
glyphs array.
Task-number: QTBUG-11427
Reviewed-by: Simon Hausmann
|
| |
| |
| |
| | |
Use the "slow" delays in all cases, not just on WinCE.
|
| |
| |
| |
| | |
Increase timeout running lconvert from 1 to 10 seconds.
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Replace this:
QTest::qWait(some_delay); QVERIFY(some_async_condition);
With this:
QTRY_VERIFY(some_async_condition);
|
|/
|
|
| |
Increase timeout when running lupdate from 5 to 30 seconds.
|
|\
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
Bump version number after 4.6.3 release.
|
| |
| |
| |
| | |
Reviewed-by: Trust Me
|
| |
| |
| |
| | |
This bug was introduced with bug fix 558089fb21e7f388f9810c51abbd9bf3872b2178
|
| |
| |
| |
| | |
Reviewed-by: ossi
|
| |
| |
| |
| |
| |
| |
| |
| | |
In waitForReadyRead we didn't check for synchronous connection loss.
Autotest added: tst_QLocalSocket::syncDisconnectNotify
Reviewed-by: ossi
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If we notice a broken pipe via _q_notified, we should call close
in case the internal read buffer is empty.
Auto test added: tst_QLocalSocket::asyncDisconnectNotify
Reviewed-by: ossi
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
QTextBoundaryFinder was not consistent with ICU.
See also:
https://bugs.webkit.org/show_bug.cgi?id=31076
The previous definition of a line break was that the index
in the string after which the line break should occur.
Now it is the index of the boundary at which the break should
occur (hence one more).
Task-number: QT-3495
Reviewed-by: Simon Hausmann
Reviewed-by: Lars Knoll
|
|
|
|
| |
Reviewed-by: ossi
|
|
|
|
|
|
|
|
|
| |
Reading from a socket with a broken connection didn't work, even if
there were still bytes to read in the internal read buffer.
Autotest: tst_QLocalSocket::writeToClientAndDisconnect
Task-number: QTBUG-10921
Reviewed-by: ossi
|
|
|
|
|
|
|
| |
Fixed race condition on connection.
Done-with: ossi
Reviewed-by: ossi
|
|
|
|
| |
Reviewed-by: ossi
|
|
|
|
|
|
|
|
|
| |
add the containing directories of all files specified on the command
line to the list of project roots. otherwise, the strings from headers
which are included before being encountered on the command line will be
omitted.
Task-number: QTBUG-10345
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Removed compiler warnings
Window visibility changes update TLW backing store reference count
Enable visibility change events on all Symbian native windows
Added reference counting to QWidgetBackingStore
Added tst_QWidget::destroyBackingStoreWhenHidden
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* When a native window becomes completely invisible, the reference count
of the top-level widget's backing store is decremented.
* When a previously invisible native window becomes either partially or
fully visible:
- If the top-level widget has a backing store, its reference count
is incremented.
- Otherwise, the top-level widget creates a backing store with an
initial reference count of 1.
Task-number: QTBUG-8697
Reviewed-by: Jason Barron
|