| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
Rather than try to resolve functions or impose configure time
limitations we simply remove the dependancy entirely and rely on the
fallbac for all platforms.
Task-number: 250731
Reviewed-by: thiago
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
origin/qtwebkit-4.5 ( 4ee8af9348b3f57d3c0f3575ae0a58336cf07a92 )
Changes in WebKit since the last update:
++ b/LayoutTests/ChangeLog
2009-05-20 Holger Hans Peter Freyther <zecke@selfish.org>
Reviewed by Anders Carlsson.
https://bugs.webkit.org/show_bug.cgi?id=24510
Add a test case that Netscape Plugin can properly
determine if a given object has a specific method
and property.
The test will ask the plugin to call hasproperty and
hasmethod and will compare the result with what is
expected. This approach is taken from
netscape-get-property-return-value.html
* plugins/netscape-invoke-browserfuncs-expected.txt: Added.
* plugins/netscape-invoke-browserfuncs.html: Added.
++ b/WebCore/ChangeLog
2009-05-19 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.org>
Reviewed by Simon Hausmann.
Do not call the parent implementation (Widget::) in show() and hide()
of the PluginViewQt, as it always changes the visible state of the
platformWidget (equal to the platformPluginWidget in the Qt port),
thus ignoring the isParentVisible() test.
* plugins/qt/PluginViewQt.cpp:
(WebCore::PluginView::show):
(WebCore::PluginView::hide):
2009-04-22 Tamas Szirbucz <szirbucz.tamas@stud.u-szeged.hu>
Reviewed by Ariya Hidayat.
https://bugs.webkit.org/show_bug.cgi?id=25023
Delete reply in QNetworkReplyHandler::abort() to avoid leak.
* platform/network/qt/QNetworkReplyHandler.cpp:
(WebCore::QNetworkReplyHandler::abort):
2009-05-20 Holger Hans Peter Freyther <zecke@selfish.org>
Reviewed by Anders Carlsson.
https://bugs.webkit.org/show_bug.cgi?id=24510
Fix a bug where the browserfuncs were not properly assigned,
make hasproperty use _NP_HasProperty and hasmethod _NP_HasMethod.
Test: plugins/netscape-invoke-browserfuncs.html
* plugins/gtk/PluginPackageGtk.cpp:
(WebCore::PluginPackage::load): Fix assignment
* plugins/qt/PluginPackageQt.cpp:
(WebCore::PluginPackage::load): Fix assignment
++ b/WebKit/qt/ChangeLog
2009-05-19 Kenneth Rohde Christiansen <kenneth.christiansen@openbossa.org>
Reviewed by Simon Hausmann.
Fix a plugin bug in the WebKit code, similar to the one in WebCore.
The problem is when a non visible QtPluginWidget would show it self
in a sibling frame. The problem was due to our clipping. In Qt,
if setMask is set with an empty QRegion, no clipping will
be performed, so in that case we hide the PluginContainer
* WebCoreSupport/FrameLoaderClientQt.cpp:
(WebCore::):
++ b/WebKitTools/ChangeLog
2009-05-14 Holger Hans Peter Freyther <zecke@selfish.org>
Reviewed by Anders Carlsson.
https://bugs.webkit.org/show_bug.cgi?id=24510 where
Add testHasProperty and testHasMethod to the existing functions
of the PluginObject to be able to test the browser hasproperty
and hasmethod implementation. Invoke them from pluginInvoke.
Change the defines to an enum to avoid manually updating
NUM_METHOD_IDENTIFIERS and assigning numbers.
* DumpRenderTree/TestNetscapePlugIn.subproj/PluginObject.cpp:
(testHasProperty): test hasproperty
(testHasMethod): test hasmethod
(pluginInvoke): invoke the two
|
|
|
|
|
|
|
|
|
| |
This case it was possible to fix by using a union of the types when we
actually declare the variable.
Besides, this avoids a bunch of #ifdef for IPv6 functionality.
Reviewed-By: Oswald Buddenhagen
|
|
|
|
|
|
|
|
|
|
| |
We currently don't support subpixel hinting when drawing text to a
QImage on Mac. The alpha map that is returned by the font engine
is a plain 8 bit gray mask, which means we have to switch off subpixel
hinting when we draw the glyph for caching.
Task-number: 249178
Reviewed-by: Samuel
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 30f7edc0aab629499b74263391ae529ad31b2ff8.
There is no way to restore float-equal warning using the pragma trick.
This means (as it was mentioned in the said commit log) anyone that
includes qtransform.h will be forced to deal with float-equal.
Reviewed-by: Samuel Rødal
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In 4.5 the toolbutton icon rendering was changed somewhat and the
bottom line of text for tool buttons icons with TextUnderIcon set is
incorrectly clipped on Windows. The style reserves only 5 pixels but
tries to use 6 pixels for text and icon spacing, hence we adjust the
text rect one pixel up. This should be safe considering the fact that
we have a margin on both sides of the icon already and avoids actually
moving the icon positioning.
Task-number: 252554
Reviewed-by: trond
|
|
|
|
|
|
| |
Explicitly cast to integer.
Reviewed-by: Shane McLaughlin <Shane.McLaughlin@trolltech.com>
|
|
|
|
|
|
| |
Found by valgrind, value isn't set but is used, fixes this.
Reviewed-by: Justin McPherson
|
|
|
|
|
|
| |
As found by valgrind. Also add error reporting that was missing.
Reviewed-by: Justin McPherson
|
|
|
|
|
|
|
| |
I tested it with 32 bits compilation and there is no warning any more.
Task-number: 247325
Reviewed-by: ogoffart
|
|
|
|
| |
The typo was in commit 30f7edc0aab629499b74263391ae529ad31b2ff8.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
I made this change to make things look a bit better on the mac. However,
a toolbutton with text only looks strange and as a bonus the arrow
covers some of the text, which no one wants. It seems that similar
consructions in Cocoa don't show the arrow for text button, so we
shouldn't either.
Task-number: 253339
Reviewed-by: Jens Bache-Wiig
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Regression from 4.4 introduced while fixing task 167106
Autotest: tst_QComboBox::task253944_itemDelegateIsReset()
Task-number: 253944
Reviewed-by: jbache
|
| |
| |
| |
| | |
Reviewed-By: TrustMe
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We changed this primarily for the mac as active appearance on
widgets in itemviews should depend on the window activation state and
not on the focus widget. It was explicitly added back for windows only
but has been reported as a bug on X11 as well so we might as well keep
it mac-only for now.
Reviewed-by: mortens
|
| |
| |
| |
| |
| |
| | |
QMessageBox is like QDialog. It doesn't have any extra magic with window
modality on creation, so don't advertise that it does. The note about
using message boxes as sheets is also updated.
|
|/
|
|
| |
Task-number: 252565
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The invariant that QCocoaWindow's lifetime is contained in a QWidget is
simply not true.
A top-level QWidget gets associated with a QCocoaWindow (which is
reference counted). However, it can be the case that we've destroyed our
QWidget, the link is removed, the window is hidden, but the window still
gets an event. In that case we would crash with an eventual null pointer
access. However, we don't really need to do anything in this case, so
just call super and return.
Task-number: 253402
Reviewed-by: Morten Sørvig
|
| | |
|
| |
| |
| |
| |
| |
| | |
Reviewed-by: Olivier Goffart
Reviewed-by: Peter Hartmann
Reviewed-by: João Abecasis
|
| |
| |
| |
| |
| |
| | |
As suggested in gitorius merge request 421, solution supplied by Harald.
Reviewed-by: Harald Fernengel
|
| |
| |
| |
| |
| |
| | |
Allows for cleaner multi-thread working for mysql clients.
Task-number: 253407
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
origin/qtwebkit-4.5 ( 40b523e9eaaba38c182e5a9c319f0069ebf98330 )
Changes in WebKit since the last update:
++ b/WebCore/ChangeLog
2009-05-11 Yael Aharon <yael.aharon@nokia.com>
Reviewed by Holger Freyther.
Change Qt port to match the mac and windows ports, and unregister plugins when plugins are stopped.
Not doing that can cause assersion failure.
https://bugs.webkit.org/show_bug.cgi?id=25702
* plugins/qt/PluginViewQt.cpp:
(WebCore::PluginView::stop):
|
| |
| |
| |
| | |
This was accidentally committed.
|
| |
| |
| |
| |
| |
| | |
origin/qtwebkit-4.5 ( 7b8d6ab6f2b73862d11c2a41ab0223e55585d88f )
Changes in WebKit since the last update:
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
currentFileInfo is only used for returning from the public functions
since the file info used in the algorithm is one step ahead.
nextFileInfo is the one actually used in the algorithm. The bug was
introduced in a compile fix for Windows and broke the stopLinkLoop test
for QDirIterator.
Reviewed-by: Olivier
|
| |
| |
| |
| | |
Reviewed-by: David Boddie
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
origin/qtwebkit-4.5 ( 7b8d6ab6f2b73862d11c2a41ab0223e55585d88f )
Changes in WebKit since the last update:
++ b/WebKit/qt/ChangeLog
2009-03-27 Erik L. Bunce <elbunce@xendom.com>
Reviewed by Simon Hausmann.
https://bugs.webkit.org/show_bug.cgi?id=24746
Improved selection tests.
* tests/qwebpage/tst_qwebpage.cpp:
(tst_QWebPage::textSelection):
|
|/
|
|
|
|
|
|
|
| |
The .mm file is not read by qdoc for packages other than for the MAC.
This problem is overcome by moving the qdoc comments from the .mm file
to a .qdoc file in doc/src, because all these files are read by qdoc
for each of the packages.
#Task-number: 252566
|
|
|
|
|
|
|
|
| |
Adding more details on QUrl::addQueryItem()
Task-number: 234125
Rev-by: Thiago Macieira
|
|
|
|
|
|
|
|
|
|
|
|
| |
Due to optimizations, there are few cases where comparing with a constant
is needed, e.g. in operator*=. G++ will give us a warning for this.
Since we know what we are doing, surpress the warning.
The only drawback for this workaround is if somebody includes
qtransform.h in his code and disable the float-equal warning since
the warning will be activated again.
Reviewed-by: Samuel Rødal
|
|
|
|
| |
Reviewed-by: Marius Storm-Olsen
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
origin/qtwebkit-4.5 ( 1f83e4058bffd5a3fe7e44cf45add01953a772d4 )
Changes in WebKit since the last update:
++ b/LayoutTests/ChangeLog
2009-05-12 Ariya Hidayat <ariya.hidayat@nokia.com>
Reviewed by Dimitri Glazkov.
Added Qt-specific expected result for toDataURL test, since Qt
does not support saving pixmaps to GIF.
* platform/qt/fast/canvas/toDataURL-supportedTypes-expected.txt: Added.
2009-05-12 Ariya Hidayat <ariya.hidayat@nokia.com>
Rubber-stamped by Simon Hausmann.
Added Qt-specific expected result for Canvas getImageData's test.
https://bugs.webkit.org/show_bug.cgi?id=22150
Since Qt is internally using premultiplied ARGB32 for doing
alpha-blending painting, reading the color back will not
necessarily give the same exact color.
* platform/qt/Skipped: Excluded canvas-getImageData.
* platform/qt/fast/canvas/canvas-getImageData-expected.txt: Added.
2009-04-29 Ariya Hidayat <ariya.hidayat@nokia.com>
Reviewed by Simon Fraser.
Updated expected results after Qt's GraphicsContext fixes.
* platform/qt/fast/canvas/set-colors-expected.txt:
2009-04-24 Ariya Hidayat <ariya.hidayat@nokia.com>
Reviewed by Simon Hausmann.
Added Qt-specific expected result for color conversion.
This is to compensate the lack of color profile in Qt to do
color conversion, hence CMYK 0,0,0,1 always give pure black (#000)
instead of very dark black.
* platform/qt/fast/canvas/set-colors-expected.txt: Added.
++ b/WebCore/ChangeLog
2009-05-18 Ariya Hidayat <ariya.hidayat@nokia.com>
Reviewed by Simon Hausmann.
Done together with Balazs Kelemen <kelemen.balazs@stud.u-szeged.hu>.
https://bugs.webkit.org/show_bug.cgi?id=24551
[Qt] Reuse FontPlatformData for the same FontDescription.
This effectively prevents growing heap usage for loading every web page.
* platform/graphics/qt/FontCacheQt.cpp:
(WebCore::qHash): Necessary for FontPlatformDataCache.
(WebCore::FontCache::getCachedFontPlatformData): Reuse the instance if
it exists, otherwise create a new one and insert it in the cache.
2009-05-18 Balazs Kelemen <kelemen.balazs@stud.u-szeged.hu>
Reviewed by Ariya Hidayat.
https://bugs.webkit.org/show_bug.cgi?id=24551
[Qt] Fix the leak in FontFallbackList::fontDataAt() function.
When creating a new instance of SimpleFontData, put it in the font list
so that it can deleted later on.
* platform/graphics/qt/FontFallbackListQt.cpp:
(WebCore::FontFallbackList::invalidate):
(WebCore::FontFallbackList::releaseFontData):
(WebCore::FontFallbackList::fontDataAt):
2009-05-15 Ariya Hidayat <ariya.hidayat@nokia.com>
Reviewed by Holger Freyther.
[Qt] In the image decoder, remove the raw image data represented as QImage
once the image is converted to QPixmap and inserted in the pixmap cache.
This effectively reduces the heap usage when running on graphics system
other than raster (i.e the case where QImage != QPixmap).
* platform/graphics/qt/ImageDecoderQt.cpp:
(WebCore::ImageDecoderQt::imageAtIndex): Nullified the image on purpose.
* platform/graphics/qt/ImageDecoderQt.h: Made m_imageList mutable.
2009-05-15 Ariya Hidayat <ariya.hidayat@nokia.com>
Reviewed by Holger Freyther.
[Qt] Refactor alpha channel detection the image decoder.
Sets the boolean flag as soon as the image is being read.
* platform/graphics/qt/ImageDecoderQt.cpp:
(WebCore::ImageDecoderQt::ImageDecoderQt): Initialized m_hasAlphaChannel.
(WebCore::ImageDecoderQt::setData): Set the flag when appropriate.
(WebCore::ImageDecoderQt::supportsAlpha): Simplified.
(WebCore::ImageDecoderQt::reset): Resetted the flag.
* platform/graphics/qt/ImageDecoderQt.h: Added m_hasAlphaChannel.
2009-05-13 Ariya Hidayat <ariya.hidayat@nokia.com>
Reviewed by Sam Weinig.
[Qt] Fix "lighther" composition mode.
QPainter::CompositionMode_Plus is the right match.
* platform/graphics/qt/GraphicsContextQt.cpp:
(WebCore::toQtCompositionMode):
2009-04-29 Ariya Hidayat <ariya.hidayat@nokia.com>
Reviewed by Simon Fraser.
[Qt] Initialize GraphicsContext's and ImageBuffer's QPainter to match
the default values of canvas attributes.
* platform/graphics/qt/ImageBufferQt.cpp:
(WebCore::ImageBufferData::ImageBufferData):
2009-04-27 Ariya Hidayat <ariya.hidayat@nokia.com>
Reviewed by Tor Arne Vestbø.
https://bugs.webkit.org/show_bug.cgi?id=18475
[Qt] Widget painting should follow the layout direction (LTR, RTL)
of the element style, not the application layout direction.
* platform/qt/RenderThemeQt.cpp:
(WebCore::RenderThemeQt::applyTheme):
|
|
|
|
|
| |
Task-number: 253835
Reviewed-by: David Boddie
|
|
|
|
|
| |
Signed-off-by: Andre Haupt <andre@bitwigglers.org>
Reviewed-By: Thiago Macieira
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's wrong to assume that static_cast<> is allowed everywhere where
dynamic_cast<> is allowed. For example, if class C derives from both A
and B, then you can dynamic_cast<B *>(ptr_to_A), but you can't
static_cast.
So introduce a helper for dynamic casts that doesn't do
static_cast.
Reviewed-by: Olivier Goffart
|
|
|
|
|
|
| |
Usually, "the the" is not proper English
Reviewed-By: Thiago Macieira
|
|
|
|
|
|
|
|
|
|
| |
When expiring cache files use a QMultiMap, when using a QMap not all
files are put into the map because often many files (downloaded or
updated at the same time) will have the same creation QDateTime and
so only one will go into the QMap who's key is QDateTime.
Reviewed-By: Thiago Macieira
Reviewed-By: Peter Hartmann
|
|
|
|
|
|
|
|
|
|
| |
My great metal hack simply needs to hack more and not do the "extra"
assign since I'm doing this through a back door in set attribute. We
probably should have had the brushed metal go via an actual QStyle
subclass instead of through the attribute.
Task-number: 253448
Reviewed-by: ogoffart
|
|
|
|
|
|
|
|
|
|
|
|
| |
if one specify more than one parameter in font-family,
e.g., font-family: Verdana, Arial
Qt should fallback on the second font if the first cannot be found.
QFont::setFamily handle the case when the family name contains a comas,
so we do not need to handle that specially in the css parser code.
Task-number: 252311
Reviewed-by: Thomas Zander
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
while using QOrderedMutexLocker::relock we might unlock our mutex
protecting the 'senders' list for a short moment. Another thread may then
modify or remove element on the list.
Therefore, we need to recheck the consistency of the list once we did
that.
Also, we cannot call removeSender because that will remove every
connections, making impossible for another object destroyed in the same
time to clean up the senders list, so call derefSender instead.
Reviewed-by: Brad
|
|
|
|
|
|
|
|
|
| |
According to the documentation, QCalendarWidget::setDateTextFormat() should
reset the format if the date is not valid.
New tests included for the formating of this widget.
Task-number: 252943
Reviewed-by: Olivier
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When compiling the QWizard on a Windows system where
QT_NO_STYLE_WINDOWSVISTA is not set and running the app
on an old system that has no Vista support, the app would
crash in the paint event of the Vista Back button. In this
scenario, the Vista Back button is not needed (i.e. the
regular Back button next to the Next button is used), so
the fix is simply to avoid instanciating it altogether.
Reviewed-by: janarve
Task-number: 252662
|
|
|
|
|
|
|
|
|
|
| |
CSS properties with more than 1 value was ignored. E.g. the
'stroke-dasharray' attribute is specified by a comma separates list of
numbers. This was previously ignored because the CSS parser split it
into a value array.
Task-number: 253614
Reviewed-by: Kim
|
|
|
|
|
| |
Task-number: 253619
Reviewed-by: Paul
|
|
|
|
|
|
|
|
| |
When the end point of a line in a path is clipped, only closed paths
should have the closing line.
Task-number: 253663
Reviewed-by: Samuel
|
|
|
|
|
|
|
|
| |
For some reason it's only sometimes hit and cannot be deterministically
reproduced.
Task-number: 253722
Reviewed-by: TrustMe
|