| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
While converting the mime to native CF_ formats, we need to check the
vaildity of the list.
Task-number: QTBUG-8446
Reviewed-by: Thierry
|
|
|
|
|
|
|
| |
Grabbing layered windows doesn't work on Vista and above, so instead we
need to grab the corresponding area of the desktop widget.
Reviewed-by: Samuel
|
|\
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
Remove unwanted code after c027f0ae1967ec1d64cb2c9679c8b57f18faf7f5
ActiveQt Internet Explorer component causes Desktop Icons to flicker.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This happens only if IE is embedded in an frameless window. The repaint
caused by calling EnableModeless() is making the desktop icons flicker.
This repaint generated via ActiveQt will be ignored by QtGui. This is
done by checking if the window style is already in the required state
or not.
Task-number: QTBUG-8355
Reviewed-by: Denis
|
| |
| |
| |
| |
| |
| |
| |
| | |
RVCT does not like static inline, and variables should not be defined
in case statement without braces. In this case the temp variable was
actually unnecessary.
Reviewed-By: TrustMe
|
|\ \
| |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Removed an export that shouldn't be exported.
Reduced the code and memory footprint of the keymap.
Enabled Qt key events to work also when native key code is missing.
Crash in QGraphicsScenePrivate::setFocusItemHelper
Improvements to itemview keypad navigation in S60.
Fix SymbianMakefileGenerator::absolutizePath for clean builds
Misc fixes to FLM files
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
This was done by using a static const data structure instead of a
list.
RevBy: Alessandro Portale
|
| |
| |
| |
| |
| |
| |
| | |
If it is missing we base the event on the scan code instead.
RevBy: Alessandro Portale
Signed-off-by: axis <qt-info@nokia.com>
|
| |\ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The logic used in this commit is partially taken from sliders.
This commit makes it possible to interact (change row or column)
in itemview even itemview does not have editFocus. Interacting without
editFocus is enabled when it is not possible to keypad navigate to
reuqested direction. In addition if keypad navigation to any direction is
not possible (i.e there is only one listwidget on screen), there is no
sense to add "done" softkey to get out of edit focus.
Task-number: QTBUG-4802
Reviewed-by: Alessandro Portale
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
QGraphicsScenePrivate::setFocusItemHelper accesses
views.at(i)->inputContext() without checking if it is null.
This leads to a crash if it really is null.
Fixed by adding a null pointer check.
Task-number: QT-3008
Reviewed-by: axis
|
|\ \ \
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
Revert changes made in scroll_sys().
Add benchmarks for accessing meta-object properties via QScriptValue
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The optimization created some glitches. It now
works properly but it could be improved
performance-wise.
Task-number: QTBUG-5926
Task-number: QTBUG-5994
Reviewed-by: Morten Sorvig
|
| |_|/
|/| |
| | |
| | |
| | |
| | |
| | | |
This simplifies the use of model view with qml and proxies.
Merge-request: 2315
Reviewed-by: Olivier Goffart <ogoffart@trolltech.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Don't assume format is the same as systemFormat() in the QNativeImage
constructor, for translucent windows it's always ARGB32_Premultiplied,
and the depth of the window is always 32 bit.
Reviewed-by: Trond
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Problem was that the cached view bounding rect was never updated when
overriding QGraphicsScene::drawItems or QGraphicsView::drawItems, without
calling the base class implementation. The same for 'updateAll' boolean.
We also have to make sure there are no unpolished items before we draw.
Regression against 4.5.
Auto-test included.
Task-number: QTBUG-7880
Reviewed-by: yoann
|
|\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
QLineEdit shows leftovers of edit cursor after clear()
doc: Added indication of default modality value for panels.
doc: Added QScopedArrayPointer and corrected QScopedPointer snippet.
Improve test coverage of QScriptString::toArrayIndex()
QColorDialog::open() freezes the app the *second* time it is used on Mac
doc: Corrected typo.
doc: Removed bad grammar.
doc: Added \obsolete.
doc: Added \since 4.6 to the properties.
doc: Removed some erroneous text.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The commit 0568fb9f428a84a344baaa5c53395db4b99f082c introduced this
regression. Make sure that we have text in the lineedit before showing
the cursor.
Task-number: QTBUG-7826
Reviewed-by: Olivier Goffart
|
| | |
| | |
| | |
| | | |
Task: QTBUG-7970
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The problem here is caused by a boolean flag used to prevent recursion.
The flag was set and never reset. This patch resets the flag the moment
open is called.
Task-number: QTBUG-7825
Reviewed-by: Richard Moe Gustavsen
|
| |\ \
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
doc: Removed some erroneous text.
|
| | | |
| | | |
| | | |
| | | | |
Task: QTBUG-7965
|
|\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/berlin-staging-1:
Fix documentmode tab label text on Mac OSX
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
The problem was that the text highlight is
drawn two pixels too low when document mode
tabs are used.
Task-number: QTBUG-8461
Reviewed-by: prasanth
|
|\ \ \ \ \
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
Regression: QGraphicsScene::render fails to render the entire scene correctly.
|
| | |/ / /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This only happened with items that had either negative width or height
(boundingRect().width()|height()). Problem was that in case of not having
an exposed region (drawing items from QGraphicsScene::render), we simply
checked whether the bounding rect was empty or not. This is fine, however
we have to normalize the rect first.
(Note that QRegion::intersects(rect) always normalizes the rect, so
that's why this use case broke only when calling QGraphicsScene::render).
Auto-test included.
Task-number: QTBUG-7775
Reviewed-by: yoann
|
|\ \ \ \ \
| |/ / / /
|/| | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Always define Q_WS_S60 on Symbian unless configured with -no-s60.
Enable window size caching on Symbian NGA variants.
Use the SYMBIAN_BUILD_GCE macro to check if the GCE variant can be used
enable bytepair compression for S60 3.2 and newer
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Enabling this flag saves us the round-trip to WSERV whenever
RWindow::Size() is called because the size is cached on the client
side. This can improve performance because functions like
eglSwapBuffers() call Size() to see if the window size has changed and
without the cache this introduces an extra IPC call for every frame.
Task-number: QT-2849
Reviewed-by: Iain
|
|\ \ \ \
| |_|/ /
|/| | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
Improve performance of VGPath creation by reusing the same path
Stabilize QWidget benchmarks.
Stylesheet: Fix size specified in QToolbar::handle
fix crash on Windows CE on WM_SETTINGCHANGE
Stabilize style sheet benchmarks.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Regression since f0243e70e05a3368582fd0478d840096d6b60c3f
We added a new rule (SE_ToolBarHandle) and the baseStyle was calling
pixelMetric(PM_ToolBarHandleExtent) on itself, bypassing the QStyleSheetStyle
Reviewed-by: Thierry
Task-number: QTBUG-8348
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
On Windows CE, lParam parameter is a constant, not a char pointer.
The only valid value is INI_INTL.
Task-number: QTBUG-7943
Reviewed-by: ninerider
|
| | |
| | |
| | |
| | |
| | |
| | | |
Forgot to add some #ifdef's.
Reviewed-by: Richard Moe Gustavsen
|
|\ \ \
| |/ /
|/| |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
QFontDialog::exec() never returns on OSX
QNetworkRequest: Doc enhancement
Port http example to QNetworkAccessManager
loopback network example: Make use of bytesToWrite()
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The problem is the fact that this dialog is never meant to be used this
way. Instead it should be called through the static function ::getFont(...).
I reimplemented this code path and made sure that this works.
Task-number: QTBUG-7769
Reviewed-by: Richard Moe Gustavsen
|
|\ \ \
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
Fixed off-by-one blending errors in the NEON drawhelper code.
Cetest extensions for Windows Mobile device power operations.
Remote lib extensions for Windows Mobile device power operations.
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
For example blending alpha 0xff with alpha 0xff and a 0.5 opacity gave
a result of 0xfe instead of the correct 0xff. This caused some
autotests to fail on ARM/NEON.
Reviewed-by: TrustMe
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We are currently getting non premultiplied surfaces
on symbian since we do not forward the egl properties
that we'd like to have to the driver.
Reviewed-By: Jason Barron
|
| |/
|/|
| |
| |
| |
| |
| |
| |
| |
| | |
miutset.h is coming from the app layer that we
are not supposed to use. Since we still need the
constant we're better off working around the
dependency and to forward declare the
one constant we need from the file.
Reviewed-By: Iain
|
| |
| |
| |
| |
| |
| |
| |
| | |
QS60Style should have smaller button margins when icon height is
larger than text height within a button.
Task-number: QTBUG-7586
Reviewed-by: Alessandro Portale
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
QColorDialog creates own instance of S60Data and fetches touch support
value from that. Unfortunately, newly created S60Data is populated
with default values that claim no support for touch.
As a fix, QColorDialog uses the global variable "S60" to fetch touch
support information.
Task-number: QTBUG-8322
Reviewed-by: Shane Kearns
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
On Symbian^3, the virtual keyboard cannot be closed anymore
when the editor area where the text is displayed is tapped and
autocompletion is suggesting a word currently.
This is caused by the fragile native FEP manager which assumes that
client must report certain event (EAknCursorPositionChanged), after
user interaction with the editor area. Otherwise the native FEP
manager is in dead-lock, until somebody reports this event to it.
Task-number: QTBUG-7828
Reviewed-by: axis
|
|\ \
| |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Pass QAudioDeviceInfo when creating audio input/output in examples
Symbian backend for QtMultimedia audio
Modified configure to detect SDK support for Symbian audio backend
Added implementation of epocRoot() function to configure
Removed duplicated implementation of readRegistryKeys from configure
Factored epocRoot implementation out of qmake
Factored readRegistryKey implementation out of qmake
QWebView scrolling doesn't clear old elements before painting again
|
| |\
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Pass QAudioDeviceInfo when creating audio input/output in examples
Symbian backend for QtMultimedia audio
Modified configure to detect SDK support for Symbian audio backend
Added implementation of epocRoot() function to configure
Removed duplicated implementation of readRegistryKeys from configure
Factored epocRoot implementation out of qmake
Factored readRegistryKey implementation out of qmake
QWebView scrolling doesn't clear old elements before painting again
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Author: Petri <petri.ojala@digia.com>
Date: Wed Feb 10 12:27:19 2010 +0200
Patch to QTBUG-7286
QWebView scrolling doesn't clear old elements before painting on top of them.
The bug exists only if web page does not set its backgroung color.
Solution is to set White color as a default color for web page backgrounds.
White color was causing some visibility problem (at least with lineedit) and
it was solved by editing ::canDrawThemeBackground
Task-number: QTBUG-7286
Reviewed-by: Sami Merila
|
|\ \ \
| |/ /
|/| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
Moving QDir benchmarks; making way for new test cases
Fixing QDir benchmarks...
QTreeView: Fix premature pessimization
Simplify code, fix a compiler warning... profit!
tst_qnetworkreply: Add another testcase
QNAM HTTP: Clean code a bit
QNAM HTTP: Always set channel.reply to 0 when done
Revert "Added note to make QList destructor virtual in version 5"
Fix s390(x) atomic ops related crashes
Fix undefined sequence point compiler warning
|
| | |
| | |
| | |
| | | |
Reviewed-by: Olivier Goffart
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Well, it was a compiler warning on master... :-/
Reviewed-by: Janne Anttila
Reviewed-by: axis
|
| |\ \
| | |/
| |/|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
QNAM HTTP: Always set channel.reply to 0 when done
Revert "Added note to make QList destructor virtual in version 5"
Fix s390(x) atomic ops related crashes
Fix undefined sequence point compiler warning
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
level = ++level is not guaranteed to work, as the increment
is not required to happen before the assignment. Change this
code to be less obscure and use standard defined behavior.
Merge-request: 471
Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
|