| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |\
| |
| |
| |
| |
| |
| |
| | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Add the auto detection for OpenGL in configure for makefile build system on Symbian.
Don't display softkeys/statusbar on FS childs of non-FS windows
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fullscreen child windows of non-fullscreen main application window
displayed softkeys and status bar when they shouldn't. Fixed it so that
fullscreen windows never display these items (except for softkeys when
explicitly specified to do so.)
Task-number: QTTH-1207
Reviewed-by: Sami Merila
|
| |\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
Manually amend latin1 change
QGraphicsItem::ItemStops[ClickFocusProgration|FocusHandling] must not interfer with touch event propagation
suppress warnings about ASCII to QString conversion
take the user defined temp dir into account
Compilation break in MSVC2008 and MSVC2005
Remove useless LFLAGS
Add docs for QNetworkConfigurationManager's time-consuming constructor.
rebuild configure
fix configure.exe option -opengl desktop
use mkdir -p for mingw+sh
Ensure the right font is used for rendering text on Mac
QDirIterator returns hidden directories when it should only return files
Update Polish translations
L10n: Update German translations.
|
| | | |
| | |
| | |
| | |
| | | |
The result of QLatin1String + QLatin1String is a QStringBuilder object,
not a QString. Fall back to QString.arg(). Patch by Ritt Konstantin.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
interfer with touch event propagation
Both flags are only supposed to affect focus handling but not the regular touch
event propagation. In this case, touch begin events were not propagated to any
item beneath the one carrying those flags.
Merge-request: 2588
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | | |
and optimize macroses to use string literal concatenation instead of .arg()
Merge-request: 1038
Reviewed-by: Harald Fernengel <harald.fernengel@nokia.com>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | | |
use user's QT_QWS_TEMP_DIR rather than hardcoded "/tmp" one
Merge-request: 1038
Reviewed-by: Harald Fernengel <harald.fernengel@nokia.com>
|
| | | |
| | |
| | |
| | |
| | | |
Review-by: Iiro Kause
Task-number: QT-4313
|
| | |\ \ |
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
When ATSFontGetName() did not set name to the font name then it would
cause Qt to fall back to a font that was not the right one. Therefore
calling ATSFontFamilyGetName() in this case will help ensure it gets the
right font.
Reviewed-by: Jiang
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
In Qt 4.5.3 and earlier, creating a QDirIterator with QDir::Files |
QDir::Hidden (i.e. search for all files, including hidden files)
returned a list of all hidden and non-hidden files in a directory tree.
In Qt 4.6.0 and later the same filter causes QDirIterator to return a
list of all hidden and non-hidden files and all hidden directories. As
the user is asking only for files, clearly returning hidden directories
too is incorrect behaviour.
Similarly, when asking for QDir::Dirs | QDir::Hidden (i.e. all
directories, including hidden directories), hidden files are also
(wrongly) returned.
Task-number: QTBUG-15421
Reviewed-by: joao
|
| |\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
* '4.7' of scm.dev.nokia.troll.no:qt/qt-qml:
Add missing test file.
Changing width of RTL positioner doesn't relayout
Fix TextInput auto test failure on mac.
PinchArea and Flickable don't work well enough together
Fix auto test failure.
Once Image sourceSize is set there is no way to clear it.
Rotation transform with NaN angle can cause crash
Canceling image download while reading causes crash
Fix width of TextInput micro focus rectangle.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
If the width of the layout changes then the items must be laid out
again.
Change-Id: I2b97bd45d07842fd3da2a0637391473ed6d78aa8
Task-number: QTBUG-18501
Reviewed-by: Michael Brasser
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Allow PinchArea to be more aggressive in grabbing the gesture and
keep the gesture until all touches are released.
Change-Id: Ic80b7c4c478e1ee3b1c3da0772553756d9d5473f
Task-number: QTBUG-17829
Reviewed-by: Bea Lam
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Image sourceSize lacked a RESET method to allow reverting to the
natural size of the image.
Change-Id: I4e9089d8e16991375745db553f891bd377143eab
Task-number: QTBUG-18442
Reviewed-by: Yann Bodson
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
When used in conjunction with Text resulted in massive memory
consunmption.
Change-Id: I5e19cb54bcd57b1ea32d17641f976f8288611a9e
Task-number: QTBUG-18386
Reviewed-by: Charles Yin
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
We were deleting a QObject owned by another thread, which is fine if
there are no pending events, but very bad if there are, e.g.
queued signals due to downloadProgress() updates. Use deleteLater()
which is safe to call in any thread.
Also call QDeclarativePixmap::clear(QObject *obj) where appropriate
to remove connections and ensure we don't handle any unwanted signals
after the download is cancelled.
Change-Id: Ia8fb41a8cd004d9840e7cec35915f1afdb03ac4d
Task-number: QTBUG-18412
Reviewed-by: Aaron Kennedy
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Remove the padding QLineControl::cursorRect() adds for region updates.
QGraphicsView also grew the rectangle by returning the bounding rect
of the transformed rectangle which is fixed by using the same transform
for QRect as is used for QRectF.
Change-Id: I8d8df9dbc6b4250e4e5392871191123a76b304a0
Task-number: QTBUG-18343
Reviewed-by: Martin Jones
|
| |\ \ \ \ \
| | |_|_|/
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Do not set focus unnecessarily at window activation in Symbian
QS60Style: Regression in drawing dialog background
QS60Style: Support menu separator (pt.2)
Support for new softkey in Symbian^3
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Calling QSymbianControl::setFocusSafely(true) on currently focused
window causes the next window in stack to get the focus momentarily,
leading to various softkey problems, such as leaving child window's
softkeys around when child is closed and flickering softkeys when child
is opened. Added a check to QWidget::activateWindow to avoid calling
setFocusSafely for already focused windows.
Task-number: QTBUG-18409
Reviewed-by: Sami Merila
Reviewed-by: axis
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Fix for QTBUG-16858 improved somewhat QComboBox menu drawing, but
unfortunately caused regression in QMenu and dialog background
drawing. These widgets shouldn't reserve space for PM_MenuVMargin.
Task-number: QTBUG-18482
Reviewed-by: Dmitry Trofimov
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Additional fix for this task. Make the separator slightly less wide,
since on some themes the menu background border might be transparent
near the left and right borders.
Task-number: QTBUG-10054
Reviewed-by: Tomi Vihria
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
QSoftKeyManager maps the StandardSoftKey values to Symbian commands.
This way the softkeys are using Symbian defined icons.
Task-number: QT-4620
Reviewed-by: Sami Merila
|
| |\ \ \ \ \
| |_|_|_|/
|/| | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
Fixed not switching to MeeGo graphicssystem.
Fixed rounding of coordinates pre-transformation in CG paintengine.
Don't include qwindowsurface_gl_p.h on QT_OPENGL_ES_1.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
If two windows become maximized at once we might not switch, this is an
attempt at fixing that.
Reviewed-by: Armin Berres
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Rounding the coordinate pre-transformation leads to pretty big
positioning errors when the painter has a large scale, and small source
coordinates are used.
Task-number: QTBUG-18416
Reviewed-by: Eskil Abrahamsen Blomfeldt
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This causes linker errors since qwindowsurface.cpp isn't compiled.
Task-number: QTBUG-14217
Reviewed-by: Eskil Abrahamsen Blomfeldt
|
| |\ \ \ \ \
| | |/ / /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Corrected case mismatch.
Fixed pro file syntax in demos/spectrum
Fix for failing autotest QToolBar/Symbian
Removed useless Exit button in Symbian^3
|
| | |\ \ \ \
| | | | | |
| | | | | |
| | | | | | |
merge-requests/2584
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
In Symbian^3 clearing softkeys with -1 value resulted useless Exit
button. EAknSoftkeyEmpty is used instead to remove the extra button.
There is no flickering with softkeys when using EAknSoftkeyEmtpy with
Symbian^3.
Task-number: QT-4557
Reviewed-by: Sami Merila
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
QS60Style was calculating the minimum size of a toolbutton inside
a toolbar based on number of visible buttons. Of course, when the
non-visible buttons become visible and we are limiting the toolbar
width to screen width it fails (since the orginally visible button is
already claiming to take whole width of toolbar and now-visible button
is taking half the button width).
As a fix, calculate minimum size based on number buttons only,
irregardless of their visibility.
Task-number: QTBUG-17777
Reviewed-by: Miikka Heikkinen
|
| |\ \ \ \ \ \
| | |_|_|/ /
| |/| | | |
| | | | | |
| | | | | |
| | | | | | |
* '4.7' of scm.dev.nokia.troll.no:qt/qt-qml:
Return correct boundaries reasons from QTextBoundaryFinder.
GridView jumps to beginning of list when resized
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
The next character after a boundary is at pos, not pos + 1. Also
consider whether the previous and next character are whitespace in
combination when determing word boudaries otherwise positions between
whitespace characters will return both StartWord and EndWord.
And since there's no need to look ahead one character don't shortcut`
the regular logic for the boundary before the last character.
Change-Id: I2efbf3947066767945f96bf8456ef518d2149191
Task-number: QTBUG-11365
Reviewed-by: Denis Dzyubenko
Reviewed-by: Ritt Konstantin
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Introduced by RTL changes. The list position should only be reset when
the view is completely regenerated, e.g. when the orientation changes.
Change-Id: I748333529c113d95eaebfb8ac9bb9cf543413d3b
Task-number: QTBUG-18441
Reviewed-by: Bea Lam
|
| |\ \ \ \ \ \
| | |_|_|/ /
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-2:
QMeeGoLivePixmapData: Verify dimensions of locked pixmap
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
When a switch from HW rendering to SW rendering is done, all EGL
surfaces backing a live pixmap become invalid because EGL is
terminated. Thanks to a recent patch, all live images are now
automatically invalidated when SW rendering is activated.
Even with this mechanism in place, it is possible for the underlying EGL
surface to be different than what the live pixmap is expecting. This can
happen, for instance, if the X pixmap gets destroyed and a different X
drawable is created using the same XID. Even though this is an unlikely
scenario, it can result in a complete system failure if the surface
dimensions do not match and the live pixmap user ends up writing over
internal graphics driver structures.
This patch adds a safeguard which disallows locking EGL surfaces with
non-matching dimensions.
Fixes Harmattan bug NB#237138.
Merge-request: 2583
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
|
| |\ \ \ \ \ \
| | |_|_|_|/
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1:
Update changes for 4.7.3
directfb: avoid version dependency by setting DSPD_DST blend functions manually
fix -enable-stdcall-fixup usage
Report that an update is needed if no update was done
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Merge-request: 2558
Reviewed-by: Marcel Schuette <marcel.schuette@nokia.com>
|
| | | | | | | |
|
| |\ \ \ \ \ \
| | |_|_|/ /
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
4.7-integration
* '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public:
Show softkeys when opening a dialog that has a fullscreen parent.
Fixed autotests on Windows.
QS60Style: draw QComboBox menu correctly
Remove two compilation warnings for ARM
QTabWidget tabs are resized in wrong way when device rotates
Symbian: QComboBox popup Listbox size is not adjusted
Switched back to using symlinks for all mkspecs on Linux.
QS60Style: Support for Menu separators
Native image handle provider support in QGLPixmapData
Fix an issue with VGImage readback in openvg.
QSysInfo support for next generation Symbian devices.
Call FixNativeOrientation on Symbian for certain fullscreen qml views.
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Now shows the CBA for each window that defines softkeys even if
topmost parent is fullscreen.
Task-number: QTBUG-4953
Reviewed-by: Sami Merila
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Note that this fix only corrects the issues #2 and #3 highlighted
in the report. These are:
a) no longer is a "background" line drawn after last item
b) combobox background is adjusted bigger so that
first and last item
The remaining issue will be fixed as a separate commit (for separate
JIRA task).
Task-number: QTBUG-16858
Reviewed-by: Dmitry Trofimov
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Initialize the rootItem variable to avoid compilation warning.
Reviewed-by: Guoqing Zhang
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
This used to work so that, style would check from actual AVKON
component (navipane) what its size is. Unfortunately, Qt was later
changed so that screen furniture is only created on-demand and thus
asking navipane's size started to fail. Which lead to situation that
tab size would "grow" after navipane was created and tabbar
was re-drawn.
As a fix, correct the pixel metrics with correct values for nHD
and remove size query from native component. The other pixel metrics
for other layouts were correct.
Task-number: QTBUG-16984
Reviewed-by: Tomi Vihria
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Don't stretch the combobox menu across the minimum screen dimension,
but respect the values given by the application.
Task-number: QTBUG-12714
Reviewed-by: Dmitry Trofimov
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Previously QS60Style didn't draw, nor reserve space for menu
separators. Now,one pixel wide line is drawn across menu rectangle.
Line color is picked from active device theme.
Task-number: QTBUG-10054
Reviewed-by: Tomi Vihria
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
QGLPixmapData's fromNativeType() accepts QNativeImageHandleProvider
pointers with type NativeImageHandleProvider from now on, similarly
to OpenVG pixmaps.
Task-number: QTBUG-18209
Reviewed-by: Jani Hautakangas
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
The special case of creating a pixmap from RSgImage via a native
handle provider needs extra attention when hibernating: In this case
the hibernating happens as usual, however doing a readback would
produce unwanted results and must be avoided.
Reviewed-by: Jani Hautakangas
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
A new QSysInfo::S60Version enum value SV_S60_5_3 is introduced.
QSysInfo::symbianVersion still returns SV_SF_3 for the new platform.
The official name of the platform is still pending, so documentation
will be updated later.
Task-number: QT-4593
Reviewed-by: Sami Merila
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
For fullscreen QDeclarativeViews that lock their orientation to the
native orientation of the device an additional performance gain can
be achieved by calling FixNativeOrientation on the underlying RWindow.
Such apps will either work in a single orientation only, or they will
handle rotation themselves, by using QPainter transformations and
opting out from the standard Avkon auto-rotation. In any of these cases
there is no need for the lower level layers of the system to handle
graphics surface rotation, the app is "fixed" to the native orientation
from their point of view, therefore it is safe to call
FixNativeOrientation().
Task-number: QTBUG-17742
Reviewed-by: Jason Barron
|
