| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Reviewed-by: Markus Goetz
Task-number: QTBUG-10106
|
|\ |
|
| |
| |
| |
| |
| | |
Merge-request: 644
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
|
| |
| |
| |
| |
| | |
Merge-request: 2395
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
|
| |\
| | |
| | |
| | |
| | | |
Merge-request: 649
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
|
| | | |
|
| | |
| | |
| | |
| | |
| | | |
Merge-request: 647
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
|
| | |
| | |
| | |
| | |
| | | |
Merge-request: 648
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
|
| | | |
|
| |\ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
In change 817469519a784b1cc84f89cb3cb84f7560874f8e, there was a
behavioral change which can cause regressions, as the bounding rect of
the QTextLayout would previously return the set line width when this
was greater than the calculated natural text width. We revert to this
behavior to avoid regressions and add an autotest for it. When the
line width is not set (and si.width is equal to QFIXED_MAX), then we
will still return the natural text width.
Reviewed-by: Lars
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
When no lineWidth is set on the QTextLine, the QScriptLine::width will
be the maximum value that can be represented by QFixed, thus always
overriding the width of the bounding rect. The lineWidth of the
QTextLine should never be used as the width of the bounding rect, since
it's not a calculated value. The real bounding rect is found by querying
the natural text width.
Task-number: QTBUG-11104
Reviewed-by: Rhys Weatherley
|
| |/ /
| | |
| | |
| | |
| | |
| | | |
Adding note to the QDrag::exec docs alerting the user about Windows speciffic behaviour.
Task-number: QT-640
|
| |\ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Base class is type-dependent. One must not use unqualified-id to access base member.
Reveiwed-by: Roberto Raggi
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
Merge-request: 652
Merge-request: 652
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
|
| |/ /
| | |
| | |
| | | |
Task-number: QT-1520
|
| | |
| | |
| | |
| | | |
This reverts commit 9b4ff3deb28b3d642dc4480207f2f23841cf26e9.
|
| | |
| | |
| | |
| | | |
This reverts commit 26f3dbf0e20dbaf131bbbb35440b1ecb40b3f405.
|
| | |
| | |
| | |
| | |
| | | |
Merge-request: 655
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Methods setNativeKey() and nativeKey() were added to
QSharedMemory API. Shared memory's native key is returned
by nativeKey() and it is set with either setKey() or
setNativeKey(). setKey() leads to a native key that is
platform independent while setNativeKey() directly sets
the native key without any mangling.
When using setNativeKey(), key() returns a null string
and shared memory's system semaphore is not set. This
means that is up to the user to define a such protection
mechanism (i.e. lock() can't be used on native keys).
QSharedMemory tests were updated.
Merge-request: 1497
Reviewed-by: Benjamin Poulain <benjamin.poulain@nokia.com>
Reviewed-by: Andreas Aardal Hanssen <andreas.aardal.hanssen@nokia.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Broken by 0cdf33e9acb00b8f3654e8268253a3fb7c5db92c, which assumes the
binary and sources are in the same directory. The fix reverts the code
back to how it was in 4.5 (where it still works with shadow builds).
Reviewed-by: Denis Dzyubenko
|
| |\ \
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
Fixed bug where testlib would not respect the -iterations option.
Improve precision of testlib benchmarking.
QMetaMethod::invoke: Document that the function does not check the arguments
add test for QMetaMethod::invoke
Make QGLBuffer copiable
|
| |\ \ \
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
qdoc: Fixed the model/view programming page; tried a style change.
fix visibility detection for g++ with a qualified binary name
|
| | |\ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
it may be called g++-4.4 (versioned) or x86_64-pc-linux-gnu-g++
(cross-compiler), so make the pattern just match anything with g++.
it may also be aliased to "c++", so catch that as well.
otoh, "gcc" would be a poor choice for QMAKE_CXX, so don't recognize
that.
Reviewed-by: thiago
Task-number: QTBUG-11065
|
| | | | | | |
|
| |\ \ \ \ \
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
* '4.7' of scm.dev.nokia.troll.no:qt/qt-qml:
Active window focus gain should not open virtual keypad
Use QElapsedTimer rather than QTime::elapsed()
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Task-number:
Reviewed-by: Martin Jones
|
| | |\ \ \ \ \ |
|
| | | | | | | | |
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Reviewed-by: Trust Me
Task-number: QTBUG-10854
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Reviewed-by: Trust Me
Task-number: QTBUG-10479
|
|\ \ \ \ \ \ \ \
| | |_|_|_|_|/ /
| |/| | | | | | |
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Reviewed-by: Morten Sørvig
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
20 ms is a bit too to get a precise measurement due to limited timer
resolution. We increase the limit to 50 ms and introduce a warm-up iteration
to prevent any state or cache-related issues which can be hard to account
for in the tests.
On the N900 this seems to decrease the variance of test results by 30 %,
and fixes some cases where the results were way off due to hidden state.
Task-number: QT-3390
Reviewed-by: Morten Sørvig
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
And added an assert to ease debugging
Task-number: QTBUG-10945
|
| | | | | | | | |
|
| |/ / / / / /
| | | | | | |
| | | | | | |
| | | | | | | |
Reviewed-by: Gunnar
|
| |\ \ \ \ \ \
| | |_|_|_|/ /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
Wrong QGraphicsItem::childrenBoundingRect() when applying effects.
Fix a bug in QDirectFBPixmapData::fromImage
Replace memmove with memcpy
Fixed compilation of QtOpenGL.
Optimize initialization of QStaticText
|
| | |\ \ \ \ \
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
Wrong QGraphicsItem::childrenBoundingRect() when applying effects.
Fix a bug in QDirectFBPixmapData::fromImage
Replace memmove with memcpy
Fixed compilation of QtOpenGL.
Optimize initialization of QStaticText
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Problem was that we used the children's raw bounding rect instead of
using their effective bounding rect when calculating the bounds.
Auto test included.
Task-number: QTBUG-10756
|
| |\ \ \ \ \ \ \
| | |_|_|_|/ / /
| |/| | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
fix escaping - by not using eval
escape backslashes in QMAKE_PRL_LIBS
fix & unify path separator escaping
Removed unused variable
Upgrade harfbuzz to the latest version
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
the left-hand-side of qmake assignments is expanded, so there is
no need to use eval.
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
they may contain (absolute) windows paths, which need escaping now.
Reviewed-by: mauricek
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Reviewed-by: mauricek
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Reviewed-by: Olivier Goffart
|
| | |/ / / / /
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Merge the folling changes from harfbuzz:
commit 85ad0ddd092522b4cff251f324128662f100991f
Author: Lars Knoll <lars.knoll@nokia.com>
Date: Thu May 27 16:05:00 2010 +0200
use unsigned char instead of unsigned in the bitfield.
This saves significant memory, as the HB_CharAttributes
and HB_GlyphAttributes structures are not 4 byte aligned
anymore, by aligned to 1 resp. 2 bytes.
The change is not compliant with ISO C, but accepted by
almost all compilers.
commit 30f0e6c3d0aa67ab9ff95c60903fe28c4df80fbc
Author: Behdad Esfahbod <behdad@behdad.org>
Date: Mon May 17 15:43:46 2010 -0400
[hangul] Fix typo
Reviewed-By: Thiago Macieira
|
| |\ \ \ \ \ \
| | |/ / / / /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
* '4.7' of scm.dev.nokia.troll.no:qt/qt-qml:
Tweak aesthetics of QML viewer inside QtDemo
Add key event handling to the declarative web view
Improve gradient docs
|