| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
In the create_sys() function where native handles are created, we were
only looking at the "isOpaque" flag to determine if a window should be
translucent or not. This is not the correct thing to do because
transparency should only be granted if the application explicitly sets
the Qt::WA_TranslucentBackground flag.
The fix is to defer the transparency decision to s60UpdateIsOpaque()
which does the right thing.
Reviewed-by: Jani Hautakangas
|
| |
|
|
|
|
|
|
|
|
|
| |
If a widget is initially shown as maximized or fullscreen then the
normal geometry is calculated before the status pane has been created.
When returning to the "Normal" window state the geometry needs to be
tweaked to compensate the fact that there is a status present now. This
should only happen if the application has not been given an explicit
position.
Reviewed-by: Janne Anttila
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Now that the S60 screen furniture construction is delayed, we need to
manually set the status pane observer to get notifications about
changes in the size of the status pane. Using the AppUi would have been
the obvious choice here, but sadly CAknAppUi (base class for
QS60MainAppUi) uses private inheritance when implementing the
MEikStatusPaneObserver interface.
To work around this problem we make QSymbianControl implement the
interface and use the instance of this class that is associated with
QDesktopWidget to recevive the notifications and then call the function
in the AppUi to maintain the same behavior.
Reviewed-by: Janne Anttila
Reviewed-by: mread
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously the S60 screen furniture (Status Pane and CBA) was being
constructed when QApplication was constructed. This meant that if you
never created a widget, you would still get those app panes which is
not what we want.
This patch avoids constructing these panes in QApplication, and manually
creates them later when the first window is shown in a state where they
would be needed.
Reviewed-by: Janne Anttila
Reviewed-by: mread
|
| |
|
|
|
|
|
|
|
|
| |
setGeometry() has the side effect that it internally sets both the
WA_Moved and WA_Resized attributes and these attributes should only be
used when the application has explicitly set a size or position. When
a widget is resized due to being maximized or fullscreened then we
should resize the native window handle instead of calling setGeometry.
Reviewed-by: Janne Anttila
|
| |
|
|
|
|
|
|
| |
The native window is used quite early in this function so as soon as we
determine that this does in fact require a native window, we should
create it immediately.
Reviewed-by: Janne Anttila
|
| |\
| |
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
QMenu: Behavior regression
Mac: Graphics View starving user events and sometimes crashes.
Fixed a crash on Windows 7 systems with invalid PrinterPorts entries.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Problem was that we called processEvents() and excluded user input
events. The idea was to only process pending update
requests, but by doing that there's also a chance that user input events
will starve. Also, there's no gurantee that only update requests
will be processed by processEvents(), so a safer solution is to call
"HIViewRender" on Carbon and "displayIfNeeded" on Cocoa. This will for
sure dispatch pending update requests and nothing else (which is exactly
what we want).
No auto test regressions.
Fixes tst_qgraphicsproxywidget::updateAndDelete failure on Carbon.
Benchmarks indicate an increase in performance.
Task-number: QTBUG-7502
Reviewed-by: richard
|
| |\ \
| |/
|/|
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Symbol visibility fixes for RVCT4 on Symbian
Symbol visibility fixes for RVCT4 on Symbian
Symbol visibility fixes for RVCT4 on Symbian
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
RVCT 4 is far more strict with regards to symbol visiblity that RVCT
2.2, and will hide symbols unless all references have default visibility
in the object files. Update the various places in Qt code where the
symbol visibility was set incorrectly for DLL-based platforms (those that
use __declspec(dllimport) and (dllexport).
Note: QtWebkit and QtScript are fixed in different commits.
Task-number: QTBUG-9903
Reviewed-by: Jason Barron
|
| |/
|
|
|
|
|
|
|
| |
Help event was not propagated to the widget. Also fixes tooltip value
propagation when setting it on the proxy or on the widget.
Autotest included.
Task-number: QTBUG-5349
Reviewed-by: bnilsen
|
| |
|
|
|
|
|
|
|
|
|
| |
Triggered signal is useful for detecting native 'Options' menu
launches in Symbian. QMenu::aboutToShow event is currently also not
generated, but that is part of another bug report. And
QMenu::aboutToShow would not even be generated for 'Options' menu
itself but only for its sub/cascade menus.
Task-number: QTBUG-9669
Reviewed-by: Sami Merila
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As described in QTBUG-8698, Qt for Symbian has been generating
incorrect focus and window activation events. This has happened since
launching menu from QSoftkeyManager with TryDisplayMenuBarL, invokes
eventually QSymbianControl::FocusChanged. But when the FocusChanged is
called menu being launched is not yet set to visible, meaning that
IsDisplayingMenuOrDialog returns false.
Because there is no way in platform to detect that menu is being
launhced, the fix is to add a new flag QS60Data, which can be used to
detect if FocusChanged event is received due to the fact that menu
is being constructed/launched.
Task-number: QTBUG-8698
* Fixes issues 2, 3 and 4
Reviewed-by: Sami Merila
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This change includes a fix for QTBUG-4887 and other exception safety
problems found while testing it.
The QTBUG-4887 fix is to qimage.cpp. QImage doesn't throw exceptions on
failure like a proper class should, instead it tries to fail "nice".
What happens here is that setAlphaChannel would crash on OOM as after
the convertToFormat call, d could be NULL. This new version checks the
result of the conversion before using it.
The other fixes are all cases where exceptions were thrown from
destructors. I added code to the test app to help debug these cases,
and I fixed all the problems I found.
With these changes, tst_exceptionsafety_objects runs and passes on the
Symbian emulator.
Reviewed-by: Shane Kearns
|
| |
|
|
|
|
|
|
|
|
|
| |
During softkey refactoring this piece of code get changes, but the
changed code did not work correctly if window had no softkey actions.
Switched back to old way where softkey source is stored to variable,
and windowType is asked from stored variable. If softkey source is
dialog or popup, "Exit" is not added automatically to RSK.
Task-number: QT-2203
Reviewed-by: Sami Merila
|
| |
|
|
|
| |
Merge-request: 544
Reviewed-by: Benjamin Poulain <benjamin.poulain@nokia.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When running on WINSCW, the nativePaintMode variable was becoming
corrupted. The point at which this corruption occurs is unclear,
but it can be reproduced as follows:
1. Launch qmediaplayer
2. In the Phonon::MMF::DsaVideoOutput constructor, the
nativePaintMode flag is set to QWExtra::ZeroFill
3. Open a video clip
4. During start of playback, QSymbianControl::Draw is called on the
native control corresponding to the DsaVideoOutput object. This
checks the value of nativePaintMode; it does not match any of
the valid NativePaintMode values, so an assertion fails.
This crash does not occur on target, suggesting that the cause may
be an error in the WINSCW compiler. Although the C++ standard allows
bitfields to have boolean, integral or enumeration type, the latter
is not permitted by C99.
Neither increasing the number of bits allocated to nativePaintMode,
nor changing it position in the list of platform-specific bitfields
has any effect.
After making nativePaintMode into a normal field rather than a
bitfield, the crash no longer happens. Note that, since bitfields
must be kept together, nativePaintMode is moved to the end of the
structure.
|
| |
|
|
| |
Reviewed-By: TrustMe
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously softkey implementation assumed that source for softkey
actions is focused widget or activewindow if there is no focused
widget. Since for example pop-up menu does not take focus immediately
in touch enabled devices, the underlying widget softkeys were visible
and usable when menu was open. This lead to problem that underlying
widget could be interacted via softkeys when popup menu was open.
For example as reported in QTBUG-8688, it was possible to close the
underlying filedialog from which the currently active context menu was
launched. It was also possible to open "Options" menu of underlying
widget when context menu was open, leading to situation where several
context menus could be launched via "Actions" item of "Options" menu.
This was reported as an issue: QTBUG-6167
In addition when user started navigating in context menu via keypad,
the menu got focus and softkeys changed to partially correct. Only
partially correct, since context menu "Select" action and "Options"
action had same priority and both associated to LSK. In addition the
context menu action "Cancel" associated to RSK was set to invisible,
meaning that it was also disabled and appeared in sofkeys as dimmed.
All of these issues were fixed by making making the popup and modal
dialogs highest priority softkey source. In case the focused widget
is inside popup or modal dialog it is being used as an initial softkey
source. In addition the softkeys created with
QSoftKeyManager::createAction or QSoftKeyManager::createKeyedAction are
now by default invisible i.e. not visible in context menu and have
special property set to make them still normally enabled in softkeys.
Task-number: QTBUG-6167
Task-number: QTBUG-8688
Task-number: QTBUG-9144
Reviewed-by: Sami Merila
|
| |
|
|
|
|
|
|
|
|
|
| |
availableGeometry returns incorrect area after cba is made invisible,
it seems that this happens because CAknToolbar is not made invisible.
Fixed the problem by reverting back to SetExtentToWholeScreen usage
when the widget is really fullscreen, i.e.
Qt::WindowSoftkeysVisibleHint is not set.
Task-number: QTBUG-9038
Reviewed-by: Miikka Heikkinen
|
| |
|
|
|
|
| |
Added the missing \row tag
Reviewed-by: TrustMe
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
We shouldn't really care about ReparentNotify when waiting for a window to be
"managed" by the window manager.
We only need to wait until the window is Mapped and Exposed for the first time,
then we (most likely) can safely assume that the window manager has finished
managing the window.
Task-number: QTBUG-9097
Reviewed-by: Bradley T. Hughes
(cherry picked from commit fb6cfbe48bc4f2148062d50d4df95616e06f9324)
|
| |\
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Initialize the graphics system before creating the style.
Added support for video surfaces to Phonon MMF backend
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Previously if the style was creating any pixmaps they were going to the
createDefaultPixmapData() function which creates a default platform
pixmap without considering the graphics system. If these pixmaps are
used they may end up going through a suboptimal code path because they
are not the same type of pixmap that is expected by the graphics
system. The fix here is to intialize the graphics system before the
style is created.
Reviewed-by: Trond
|
| |\ \
| |/
|/|
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
QToolTip may be placed incorrectly when using FullScreen mode - Mac OS X
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When in full screen mode, there is no Dock nor Menu so we can use
the whole screen for displaying the tooltip. However when not in
full screen mode we need to save space for the dock, so we use
availableGeometry instead.
Task-number: QTBUG-8402
Reviewed-by: Prasanth
|
| |/
|
|
|
|
|
|
|
|
|
| |
This also fixed the issue where an input panel would not be requested
by the widget, because it thought it was being focused over and over,
instead of focused, and then clicked (focus followed by click is
default input panel popup behavior on Symbian).
AutoTest: Included and passed
RevBy: yoann
RevBy: Jan-Arve
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of using traditional dropdown menus, use popup menu, like
native side does.
Related to https://qtrequirements.europe.nokia.com/browse/QT-2274
Popup is positioned according to S60 popup position rules:
- in portrait at the bottom, with square popup (dimension is
screen width minus border)
- in landscape, centered on Y-axis, on X-axis to the right
border (with LeftToRight UI direction)
Background behind popup is dimmed.
Popup has automatic scrollbar if content does not fit into the popup.
Popup background graphic is S60 dialog graphic.
Task-number: QT-2274
Reviewed-by: Alessandro Portale
|
| |
|
|
|
|
|
| |
Part of QtP delta reduction effort.
Task-number: QT-3055
Reviewed-by: Janne Koskinen
|
| |\
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/qt-s60-public:
HotFix for fluidlauncher default size caused by 6d44dadd.
Removed compiler warning from qdesktopwidget_s60.cpp.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
It seems that in Symbian we cannot restore Qt::WA_Moved and
Qt::WA_Resized attributes in setWindowState method to the same ones
what the attributes were when method was called. Thus we currently
make sure that Qt::WA_Moved and Qt::WA_Resized attributes are not
touched when normal window state is applied.
There is a new task QTBUG-8977 to sort out when those attributes
should be set and when not.
Reviewed-By: Sami Merila
|
| | |
| |
| |
| | |
Reviewed-By: TrustMe
|
| |\ \
| |/
|/|
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
Updated WebKit from /home/shausman/src/webkit/trunk to qtwebkit/qtwebkit-4.6 ( 266a6c4f1938dd9edf4a8125faf91c62495e3ce2 )
Doc: Removed lie that access to QBuffer is unbuffered.
Minisplitter doesn't paint to the bottom/right
Avoid unnecessary memory allocation in the jpeg handler's image detection
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The problem is not related to the Splitter handle but related to an
optimization in the paintengine.
We were comparing if the size of the widget had changed before setting
the new mask, which might not be correct. We could have changed the mask
without changing the widget size.
Task-number: QTCREATORBUG-753
Reviewed-by: Morten
Reviewed-by: Samuel
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
QDialog::resize() also moved the dialog in Symbian. This occured since
adjustPosition was called as an result of resize and that method did
not check if dialog position was explicitly set.
In addition it was found that in Symbian WA_Resized and WA_Moved
attributes were basically set for almost all top-level widgets
by system. This was also fixed and a new auto test was introduced
to verify these attributes in all platforms. Windows platform also
suffers from bug in this area, and a separate task QTBUG-5897 for
it was created
Task-number: QTBUG-5897
Reviewed-by: Sami Merila
|
| |
|
|
|
|
|
|
| |
XOR'ed cursors under X11 or Mac OS X have never been supported by Qt,
and the doc updates done some time ago where not correct.
Task-number: related to QTBUG-1838
Reviewed-by: Gunnar
|
| |\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
Fix build with Mesa 7.8's EGL implementatioon
Reset the byte order in the iconv codec after using it.
fix "using namespace" recursion crash
Fixed key mappings on X11
QDom: prevent infinite loop when cloning a DTD
QPrintPreviewDialog number of pages is partially blocked from view in OSX
Updated WebKit from /home/shausman/src/webkit/trunk to qtwebkit/qtwebkit-4.6 ( f3110d2f94c825477afac054ed448e45d47f5670 )
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Merge request #1742 (6c8c1c5322a26d789165783d7df3e29c672690cb) changed the
behavior of one of the keys on X11, this fixes it back and documents that on
X11 the key binding is a bit weird - XF86XK_MyComputer is mapped to
Qt::Key_Launch0, XF86XK_Launch0 is mapped to Qt::Key_Launch3 and so on.
Task-number: QTBUG-7063
Reviewed-by: Marius Storm-Olsen
|
| |/
|
|
|
|
|
|
|
| |
Parent window accepted pointer events even though the child was modal.
That was because closing popup didn't restore pointer grab for modal
window.
Task-number: QTBUG-6659
Reviewed-by: axis
|
| |\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
tst_qhostinfo: Compile fix
Document some stuff in QHostInfo and QAbstractSocket
tst_qhostinfo: Mixed up descriptions for test case data
qhostinfo benchmark: Use local _data instead of _global
DNS Cache: Also check inside the DNS threads
qhostinfo benchmark: Benchmark the with-cache and without-cache
tst_qftp: Do not use 1.2.3.4 as IP
Dont use gestures if there is no touch screen.
showFullscreen() broken if UnifiedTitleAndToolBarOnMac is enabled
Fix atk warning on startup in Qt apps wit QGtkStyle
Do not use realpath() with uclibc
|
| | |
| |
| |
| |
| |
| |
| |
| |
| | |
If there is no touch input device attached on Windows7, we shouldn't even
bother subscribing to native gesture events.
Task-number: QTBUG-6007
Reviewed-by: Thierry
Reviewed-by: Prasanth
|
| |/
|
|
|
|
|
|
|
|
|
| |
wacom driver fixed a bug in assigning atom names where it messed
up the device names. Their solution apparently renamed the atoms
The type names used now are "STYLUS", "CURSOR", "ERASER", "PAD"
and "TOUCH" for each respective device. This change detects
those as our tablet input devices.
Reviewed-by: Benjamin Poulain <benjamin.poulain@nokia.com>
Task-number: QTBUG-8599
|
| |\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
Removed unused macro.
Fix runtime warning on Gtk versions < 2.16
Fix compile
Follow-up to c1b067e: fix aliasing in QReadWriteLock too.
Bug with toolbar focus on Mac
|
| | |
| |
| |
| | |
Reviewed-by: Stefano Pironato
|
| | |
| |
| |
| | |
Reviewed-by: cduclos
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Before doing anything we need to make sure that we don't leave anything
in a non-consistent state. When hiding a widget we need to make sure that
no mouse_down events are active, because the mouse_up event will never be
received by a hidden widget or one of its descendants. The solution is
simple, before going through with this we check if there are any
mouse_down events in progress, if so we check if it is related to this
widget or not. If so, we just reset the mouse_down and then we continue.
In X11 and Windows we send a mouse_release event, however we don't do that
here because we were already ignoring that from before. I.e. Carbon did
not send the mouse release event, so we will not send the mouse release
event. There are two ways to interpret this:
1. If we don't send the mouse release event, the widget might get into an
inconsistent state, i.e. it might be waiting for a release event that
will never arrive.
2. If we send the mouse release event, then the widget might decide to
trigger an action that is not supposed to trigger because it is not visible.
Task-number: QTBUG-8604
Reviewed-by: denis
|
| |\ \
| |/
|/|
| |
| |
| |
| | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Reduced S60 softkey flickering caused by softkey image support.
|
| | |
| |
| |
| |
| |
| |
| |
| | |
EikSoftkeyImage::SetLabel call seems to clear and redraw the softkeys
every time when called. This commit removes the unnecessary calls to
the method if given CBA is already in text mode.
Reviewed-By: Sami Merila
|
| |\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
4.6-integration
* '4.6' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
Fix build break in qapplication_x11.cpp.
update russian translations and phrasebook
Allow platform specific values for the double click radius.
Fix strict-aliasing breakage with SunCC: the union trick is a GCC extension.
QNAM HTTP: Fix invoking a method when being destructed right now
QAbstractSocket: Use new faster DNS function
Add DNS caching to QHostInfo
Make the icon visible when set on an action in a QSystemTrayIcon on Mac
Carbon : Setting palette brush to a pixmap does not work.
|
| | | |
| | |
| | |
| | | |
Reviewed-by: TrustMe
|
