Webkit: Avoid null pointer dereference when canceling resource loading

In some cases ResourceLoader::didCancel may run after the load has
already completed which results in a null pointer dereference. If the
load has already completed - do nothing.

This is a partial backport of r86720 in WebKit upstream.

Task-number: QTBUG-30081
Change-Id: I1e41594f6d7a81ec6c3d97367391b722918abae2

Signed-off-by: Simon Hausmann <simon.hausmann@digia.com>

2012-11-28  Zeno Albisser  <zeno@webkit.org>

[Qt][Win] Enable usage of QTWEBKIT_PLUGIN_PATH.
https://bugs.webkit.org/show_bug.cgi?id=103529

This patch enables the same behavior for Windows
as it is currently implemented for all other operating
systems supported by Qt.
Directories listed in the environment variable
QTWEBKIT_PLUGIN_PATH are being searched for additional
plugins.
The only difference is that directories on Windows
need to be separated by a semicolon instead of a colon.

Reviewed by Kenneth Rohde Christiansen.

* plugins/win/PluginDatabaseWin.cpp:
(WebCore):
(WebCore::addQtWebKitPluginPath):
(WebCore::PluginDatabase::defaultPluginDirectories):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@136017 268f45cc-cd09-0410-ab3c-d52691b4dbfc

fix the build on Mac 10.6 with 10.5 sdk

Task-number: QTBUG-23258
Signed-off-by: Simon Hausmann <simon.hausmann@digia.com>

Bump the version

Fix path to OSX Mountain Lion SDK, as pointed out by Mike McQuaid on https://codereview.qt-project.org/#change,31759

Unreviewed Qt/Mac build fix: Use libWebKitSystemInterfaceMountainLion.a
on Mountain Lion.

* WebKit1.pri:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@124002 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:
Source/WebKit/ChangeLog
Source/WebKit/WebKit1.pri

Add a Mountain Lion version of libWebKitSystemInterface.a.

Source/WebKit/mac:

Reviewed by John Sullivan.

* Configurations/DebugRelease.xcconfig: Look for the library under its expected name.

Source/WebKit2:

Reviewed by John Sullivan.

* Configurations/DebugRelease.xcconfig: Look for the library under its expected name.

Tools:

Reviewed by John Sullivan.

* Scripts/copy-webkitlibraries-to-product-directory: Include libWebKitSystemInterfaceMountainLion.a in the list of libraries to copy.

WebKitLibraries:

Update the WebKitSystemInterface header and binaries.

Reviewed by John Sullivan.

* WebKitSystemInterface.h:
* libWebKitSystemInterfaceLion.a:
* libWebKitSystemInterfaceMountainLion.a: Added.
* libWebKitSystemInterfaceSnowLeopard.a:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@122352 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:
Source/WebKit/mac/ChangeLog
Source/WebKit/mac/Configurations/DebugRelease.xcconfig
Source/WebKit2/ChangeLog
Source/WebKit2/Configurations/DebugRelease.xcconfig
Tools/ChangeLog
Tools/Scripts/copy-webkitlibraries-to-product-directory
WebKitLibraries/ChangeLog
WebKitLibraries/WebKitSystemInterface.h
WebKitLibraries/libWebKitSystemInterfaceLion.a
WebKitLibraries/libWebKitSystemInterfaceSnowLeopard.a

Fix build on linux-i386 where gcc would produce warnings

It is the intention to _not_ use -Werror for production builds. Unfortunately
the qmake expression had an error that made the -Werror suppression only work
on x86-64 but not on i386. A surrounding if() as suggested by Ossi fixes this.

In trunk this is already fixed differently by the use of an extra scope.

Task: QTBUG-23987

Bumped the QtWebKit version to 2.2.3

Fix build with GLib 2.31
https://bugs.webkit.org/show_bug.cgi?id=69840

Patch by Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk> on 2011-10-12
Reviewed by Martin Robinson.

.:

* configure.ac: add check for pthread rwlock support.

Source/JavaScriptCore:

* GNUmakefile.list.am: removed ThreadingGtk.cpp.
* wtf/ThreadingPrimitives.h: remove GTK+-specific definitions.
* wtf/gobject/GOwnPtr.cpp: remove GCond and GMutex specializations.
* wtf/gobject/GOwnPtr.h: ditto.
* wtf/gobject/GTypedefs.h: remove GCond and GMutex forward declarations.
* wtf/gtk/ThreadingGtk.cpp: Removed.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@97269 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

ChangeLog
Source/JavaScriptCore/ChangeLog
Source/JavaScriptCore/wtf/gtk/ThreadingGtk.cpp

JSString::resolveRope() should report extra memory cost to the heap.

At the time a JSString is constructed with fibers, it doesn't report
extra memory cost, which is reasonable because it hasn't allocate
new memory. However when the rope is resolved, it should report memory
cost for the new buffer.

Backport patch by Yong Li
Reviewed by Michael Saboff.
http://trac.webkit.org/changeset/109105

Signed-off-by: Simon Hausmann <simon.hausmann@nokia.com>

Implement OSAllocator::commit/decommit in the correct way for QNX

https://bugs.webkit.org/show_bug.cgi?id=77013
We should use mmap(PROT_NONE, MAP_LAZY) instead of posix_madvise() to
implement memory decommitting for QNX.

Backport Patch by Yong Li
Reviewed by Rob Buis.
http://trac.webkit.org/changeset/111234

Signed-off-by: Simon Hausmann <simon.hausmann@nokia.com>

Qt: Doc: Fix typo which marks document to be printed in console font.
https://bugs.webkit.org/show_bug.cgi?id=82893

Patch by Casper van Donderen <casper.vandonderen@nokia.com> on 2012-04-02
Reviewed by Simon Hausmann.

The qtwebkit-bridge.qdoc file contained a typo where a \c {} was
missing its closing curly bracket, this marked the rest of the page as
text to be printed using the code/console font.
* docs/qtwebkit-bridge.qdoc:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@112872 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the PRI macros used in WTF::String formatters to be compatible with Qt and Visual Studio 2005 and newer.
https://bugs.webkit.org/show_bug.cgi?id=76210

Add compile time check for Visual Studio 2005 or newer.

Patch by Aron Rosenberg <arosenberg@logitech.com> on 2012-02-23
Reviewed by Simon Hausmann.

* os-win32/inttypes.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@108641 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Increment version of shared library to match patch release version

Various Qt documentation bug fixes (wk 42)

Task-number: QTBUG-18101
Task-number: QTBUG-8673
Task-number: QTBUG-14194
Task-number: QTBUG-9109
Task-number: QTBUG-9466
Task-number: QTBUG-8323
Task-number: QTBUG-7924
Task-number: QTBUG-20355
Task-number: QTBUG-19367
Task-number: QTBUG-21295
Task-number: QTBUG-14554
Task-number: QTBUG-8331
Task-number: QTBUG-8329
Task-number: QTBUG-8786
Task-number: QTBUG-8787

Change-Id: I4b6403df4a0078fa385abbfab5b6c3a94f175295
Reviewed-by: Casper van Donderen <casper.vandonderen@nokia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@nokia.com>
Signed-off-by: Simon Hausmann <simon.hausmann@nokia.com>

Fixes Webkit compilation on QNX

Adds missing include headers and platform-specific thread suspend/resume code.

Original Qt Gerrit Change-Id: I16ad04956d3d321cf856365c73b7926e30ba4a12

Reviewed-by: Simon Hausmann

Fix build issue on Windows when Qt is configured with -ltcg

Patch by Andy Shaw. See:
https://bugs.webkit.org/show_bug.cgi?id=75003

Upstreaming patch 9a74fb08fbfc1157210ac23a6b168022be8cdfd7

Patch by Pasi Pentikainen <pasi.a.pentikainen@nokia.com>

Symbian-linux crosscompilation case changes
Update the cases to match actual files in Symbian, and add EPOCROOT
prefixes to data files in iby.

Reviewed-by: Juha Kukkonen
Task-number: QTBUG-14361

Upstreaming patch e64f927cc1f34763f6b61ac05a80e9277200fe3c

Patch by Frederik Gladhorn <frederik.gladhorn@nokia.com>

Fix more permissions (no execute bit).

Upstreaming 81ae9027b72dcf3635d2fb9f20e63d60b35d62e8

Patch by hjk <qtc-committer@nokia.com>

Webkit: Work around C++11 string literal source incompatibility.
ISO/IEC 14882 2.2 Phases of translations: "user defined string literals"
are recognized as single preprocessor token in Phase 3, left unchanged by
"...preprocessing executed" in Phase 4. Previously, there were no user
defined liteal, and the intended concatanation was done in Phase 6
"adjacent string literal tokens are concatenated", after the macro expansion.

Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@nokia.com
Task-number: QTBUG-22847

Upstreaming patch b6fe7690e2b7eada35d9240f260bfb5bfb20b585

Patch by artoka <arto.katajasalo@digia.com>

QWebElement example missed information
In the example there was told to use
button.evaluateJavaScript("click()"); which doesn't work.
Instead of "click()" it should be "this.click()".
Changed to "this.click()".

Task-number: QTBUG-17029

Merge-request: 2698
Reviewed-by: Casper van Donderen <casper.vandonderen@nokia.com>

Upstreaming patch 0455cf2f9c4d82d1badcc568a6c6f4401111c62b

Patch by Casper van Donderen <casper.vandonderen@nokia.com>

Update translation sources.
Reviewed-By: Geir Vattekar

Change the QtWebKit Version to 2.2.2

Change the QtWebKit Version to 2.2.1

Workaround MSVC2010 problems when linking QtWebKit

Include MSVC2010 in the list of compilers where incremental
build is disabled (INCREMENTAL:NO).

Patch by Ademar de Souza Reis Jr.
Change suggested by Simo Falt <simo.falt@nokia.com>

dos2unix on a webkit source file (fix support for Visual Studio)

webkit/Source/WebCore/bindings/js/JSExceptionBase.h had CRLF
terminations, which Visual Studio didn't like for some reason.

Patch by Ademar de Souza Reis Jr.
Reported by Simo Falt. Patch is also being submited to upstream
(webkit.org)

Disable -Werror for standalone packages.

Reviewed by Simon Hausmann.

* Source/WebKit.pri:

Fixed build with enabled workers and disabled inspector
https://bugs.webkit.org/show_bug.cgi?id=62461

Patch by: Konstantin Tokarev  <ktokarev@smartlabs.tv>

Reviewed by Joseph Pecoraro.

* inspector/InspectorWorkerAgent.cpp: Added inspector guard

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@88547 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Unreviewed: Fix build on Mac.
https://bugreports.qt.nokia.com/browse/QTBUG-20619

This bug was introduced by the previous commit
for the same bugreport.

* Source/WebKit.pri:

[Qt] Unreviewed: Fix debug_and_release build on Windows.
https://bugreports.qt.nokia.com/browse/QTBUG-20619

This bug was introduced by the previous commit
for the same bugreport.

* Source/WebKit.pri:

Patch by Jesus Sanchez-Palencia <jesus@webkit.org> on 2011-11-03
Reviewed by Kenneth Rohde Christiansen.

[Qt] QtWebKit should have documentation clarifying its mobile features usage
https://bugs.webkit.org/show_bug.cgi?id=41465

Add "QtWebKit Goes Mobile" documentation

* docs/qtwebkit-goes-mobile.qdoc: Added.
* docs/qtwebkit.qdoc: Linked to qtwebkit-goes-mobile documentation
* docs/webkitsnippets/qtwebkit_goes_mobile_snippets.cpp: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@99196 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Unreviewed: Fix universal binary build on Mac.
https://bugreports.qt.nokia.com/browse/QTBUG-20619

When building a universal binary the debug version
exceeds the maximum size of a static library.
Therefore we do not allow this configuration.

* Source/WebKit.pri:

Unreviewed Qt documentation fix.

This signal is Qt 4.8 material.

* Api/qwebpage.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@98951 268f45cc-cd09-0410-ab3c-d52691b4dbfc

bad codegen, pointer diff in JSC::JSCallbackConstructor::JSCallbackConstructor https://bugs.webkit.org/show_bug.cgi?id=60951

Adjust symbols visibility for WebCore.

Patch by Zeno Albisser <zeno.albisser@nokia.com> on 2011-11-01
Reviewed by Simon Hausmann.

* WebCore.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@98947 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Unreviewed: fix the changelog.

[Qt] Unreviewed: Build fix for MSVC2005

Patch by Andy Shaw <andy.shaw@digia.com>

* platform/win/SystemTimeWin.cpp:

[Qt] WebKit build does not respect QMAKE_MAC_SDK variable.
https://bugs.webkit.org/show_bug.cgi?id=70596

Instead of only relying on DARWIN_MAJOR_VERSION we also
check QMAKE_MAC_SDK. In case QMAKE_MAC_SDK is not defined
we are still falling back to DARWIN_MAJOR_VERSION.

Patch by Andy Shaw <andy.shaw@digia.com>

Patch by Zeno Albisser <zeno.albisser@nokia.com> on 2011-10-21
Reviewed by Noam Rosenthal.

* QtWebKit.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@98114 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] QtWebKit build error for Mac 32bit
https://bugs.webkit.org/show_bug.cgi?id=69914

In WebCore.pro and QtWebKit.pro we define NSGEOMETRY_TYPES_SAME_AS_CGGEOMETRY_TYPES
when compiling with ENABLE_VIDEO and with WTF_USE_QTKIT.
But this is meant to be defined in NSGeometry.h under certain preconditions.
Without setting NS_BUILD_32_LIKE_64 none of these preconditions is
met and therefore NSGeometry.h will create several conflicting type definitions.

With this patch we create consistent definitions again.
Due to the order of include files we cannot remove
NSGEOMETRY_TYPES_SAME_AS_CGGEOMETRY_TYPES from WebCore.pro.

Patch by Zeno Albisser <zeno.albisser@nokia.com> on 2011-10-13
Reviewed by Andreas Kling.

* WebCore.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@97387 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Unreviewed: Unskip xmlhttprequest-missing-file-exception.html (Qt-4.8)

[Qt]http/tests/xmlhttprequest/xmlhttprequest-missing-file-exception.html
fails with Qt 4.8
https://bugs.webkit.org/show_bug.cgi?id=60176

This test was skipped because of a regression in Qt-4.8. The problem has
been fixed and the bots have been updated, so there's no need to keep
the test skipped.

Fix out-of-bounds access in Gradient::sortStopsIfNecessary
https://bugs.webkit.org/show_bug.cgi?id=67958

Reviewed by Darin Adler.

Reported by Valgrind in http://crbug.com/77049.

The errant code was added as an optimization in r67804.
This patch reverts that one, as all parties agree that the optimization
doesn't seem worthwhile, and there clearly aren't any tests covering
the special case.

No new tests, as existing tests should cover the remaining call to
|std::stable_sort|.

* platform/graphics/Gradient.cpp:
(WebCore::Gradient::sortStopsIfNecessary):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@95010 268f45cc-cd09-0410-ab3c-d52691b4dbfc

frame-loading-via-document-write.html test cases have extra slash in path
https://bugs.webkit.org/show_bug.cgi?id=67784

Removing extra slash in path: file:////tmp/... => file:///tmp/...

Patch by Jarred Nicholls <jarred@sencha.com> on 2011-09-08
Reviewed by Andreas Kling.

* fast/dom/frame-loading-via-document-write.html:
* http/tests/security/frame-loading-via-document-write.html:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94796 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt][DRT] Normalize file:///tmp/LayoutTests in LayoutTestController::pathToLocalResource()
https://bugs.webkit.org/show_bug.cgi?id=67254

Patch by Jarred Nicholls <jarred@sencha.com> on 2011-09-07
Reviewed by Csaba Osztrogonác.

Translate file:///tmp/LayoutTests/* urls into the repository LayoutTests directory, which is
derived from the running location of DumpRenderTree binary.

* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::pathToLocalResource):
(LayoutTestController::setUserStyleSheetLocation): pass the url through pathToLocalResource

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94676 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt][Mac] Skip more new failing tests to paint the bot green

These tests were introduced by r94857 but fail on the
Qt-Mac buildbots. They were skipped in trunk by r94924.

Unreviewed. Chromium, GTK, Qt rebaselines for r94857.

* platform/chromium-linux/fast/lists/list-marker-before-content-table-expected.png: Added.
* platform/chromium-linux/fast/runin/runin-generated-before-content-expected.png: Added.
* platform/chromium-win-xp/fast/lists: Removed.
* platform/chromium-win/fast/lists/list-marker-before-content-table-expected.png: Added.
* platform/chromium-win/fast/runin/runin-generated-before-content-expected.png: Added.
* platform/gtk/fast/lists/list-marker-before-content-table-expected.txt: Added.
* platform/gtk/fast/runin/runin-generated-before-content-expected.txt: Added.
* platform/mac/fast/lists/list-marker-before-content-table-expected.txt: Added.
* platform/mac/fast/runin/runin-generated-before-content-expected.txt: Added.
* platform/qt/fast/lists/list-marker-before-content-table-expected.txt: Added.
* platform/qt/fast/runin/runin-generated-before-content-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94865 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Unreviewed gardening after r94857. Add platform specific expected files.

* platform/qt/fast/lists/list-marker-before-content-table-expected.png: Added.
* platform/qt/fast/lists/list-marker-before-content-table-expected.txt: Added.
* platform/qt/fast/runin/runin-generated-before-content-expected.png: Added.
* platform/qt/fast/runin/runin-generated-before-content-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94860 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Source/WebCore: Assert being hit in AccessibilityRenderObject::addChildren()
https://bugs.webkit.org/show_bug.cgi?id=61805

Patch by Dominic Mazzoni <dmazzoni@google.com> on 2011-09-09
Reviewed by Chris Fleizach.

Fix nextSibling and previousSibling to handle adjacent continuations
properly, otherwise nodes end up appearing in the accessibility
tree twice (or a debug assertion could be raised).

Test: accessibility/adjacent-continuations-cause-assertion-failure.html

* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::previousSibling):
(WebCore::AccessibilityRenderObject::nextSibling):

LayoutTests: Add a test to catch a case where adjacent continuations were
causing nodes to get added to the accessibility tree twice, leading
to an assertion failure or a crash. The test expectations are
currently Mac-specific, so added the test to the Skipped file for
gtk and win.
https://bugs.webkit.org/show_bug.cgi?id=61805

Patch by Dominic Mazzoni <dmazzoni@google.com> on 2011-09-09
Reviewed by Chris Fleizach.

* accessibility/adjacent-continuations-cause-assertion-failure.html: Added.
* platform/mac/accessibility/adjacent-continuations-cause-assertion-failure-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94864 268f45cc-cd09-0410-ab3c-d52691b4dbfc

:before content rendering issues with list markers and run-ins.
https://bugs.webkit.org/show_bug.cgi?id=67735

Source/WebCore:

1) Remove the isAnonymous checks for run-in detection since the
run-in can belong to a node.
2) When the parent has block children, then the list marker will
be enclosed in an anonymous block. In that case, for going to the
next list marker, we need to traverse one level up. We don't need
this check when searching for generated run-in (loop 2), since we
know parent will have inline children, so the list marker wont be
enclosed in an anonymous block.

Reviewed by Dave Hyatt.

Tests: fast/lists/list-marker-before-content-table.html
       fast/runin/runin-generated-before-content.html

* rendering/RenderObjectChildList.cpp:
(WebCore::RenderObjectChildList::beforePseudoElementRenderer):

LayoutTests:

Reviewed by Dave Hyatt.

* fast/lists/list-marker-before-content-table-expected.png: Added.
* fast/lists/list-marker-before-content-table-expected.txt: Added.
* fast/lists/list-marker-before-content-table.html: Added.
* fast/runin/runin-generated-before-content-expected.png: Added.
* fast/runin/runin-generated-before-content-expected.txt: Added.
* fast/runin/runin-generated-before-content.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94857 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Source/WebCore: Crash due to bad data in SVGDocumentExtensions m_pendingResources
https://bugs.webkit.org/show_bug.cgi?id=67488

Patch by Ken Buchanan <kenrb@chromium.org> on 2011-09-10
Reviewed by Nikolas Zimmermann.

Resolving a crash condition caused by the deletion of
elements while pending resource entries for those elements are still
recorded.

* rendering/svg/RenderSVGResourceContainer.cpp:
(WebCore::RenderSVGResourceContainer::registerResource)
* svg/SVGDocumentExtensions.h:
(WebCore::SVGDocumentExtensions::isElementInPendingResources)
* svg/SVGDocumentExtensions.cpp:
(WebCore::SVGDocumentExtensions::addPendingResource)
(WebCore::SVGDocumentExtensions::isElementInPendingResources)
(WebCore::SVGDocumentExtensions::removeElementFromPendingResources)
* svg/SVGStyledElement.h:
(WebCore::SVGStyledElement::clearHasPendingResourcesIfPossible)
* svg/SVGStyledElement.cpp:
(WebCore::SVGStyledElement::buildPendingResourcesIfNeeded)
(WebCore::SVGStyledElement::clearHasPendingResourcesIfPossible)
* svg/SVGUseElement.cpp:
(WebCore::SVGUseElement::svgAttributeChanged)

LayoutTests: Crash due to bad data in SVGDocumentExtensions m_pendingResources.
https://bugs.webkit.org/show_bug.cgi?id=67488

Patch by Ken Buchanan <kenrb@chromium.org> on 2011-09-10
Reviewed by Nikolas Zimmermann.

Test added: validating that the crash referenced in the bug is not present.

* svg/dom/SVGStyledElement-pendingResource-crash.html: Added.
* svg/dom/SVGStyledElement-pendingResource-crash-expected.txt: Added.
* svg/dom/resources/SVGStyledElement-pendingResource-crash.svg: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94905 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-05-23  Thouraya ANDOLSI  <thouraya.andolsi@st.com>

        Reviewed by Gavin Barraclough.

        [SH4] AssemblerLabel does not name a type
        https://bugs.webkit.org/show_bug.cgi?id=59927

        SH4Assembler.h file shoold be included before AbstractMacroAssembler.h.

        * assembler/MacroAssemblerSH4.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@87076 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    REGRESSION (r66874): Missing RefPtr in ScriptController
        https://bugs.webkit.org/show_bug.cgi?id=67748

        Reviewed by Adam Barth.

        * bindings/ScriptControllerBase.cpp: (WebCore::ScriptController::executeScript):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94820 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Update layout test results for newer Qt 4.8.x
https://bugs.webkit.org/show_bug.cgi?id=64527

Patch by Csaba Osztrogonác <ossy@webkit.org>

Because of http://qt.gitorious.org/qt/qt/commit/429f0507a68f4aa686449308af770ce5cf2c913f
we have to update expected files for 504 layout tests. Additionally it made 1 test fail.

* platform/qt-4.8/Skipped: Skip svg/text/lengthAdjust-text-metrics.html, because it fails.
* platform/qt-4.8/ [...] : Add Qt 4.8 specific expected results for 504 tests.

[Qt][Mac] Skip new failing tests to paint the bot green

These tests were introduced by r94543, but fail on the
Qt-Mac buildbots. They were skipped in trunk by r94570.

Style not propagated to anonymous boxes and anonymous
inline-blocks.
https://bugs.webkit.org/show_bug.cgi?id=67364

Source/WebCore:

Reviewed by James Robinson.

Share propagateStyleToAnonymousChildren with RenderBlock::styleDidChange.

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::styleDidChange):
* rendering/RenderObject.cpp:
(WebCore::RenderObject::propagateStyleToAnonymousChildren):
* rendering/RenderObject.h:

LayoutTests:

Add comments explaining what the test does.

Reviewed by James Robinson.

* fast/ruby/ruby-block-style-not-updated-with-before-after-content.html:
* fast/ruby/ruby-block-style-not-updated.html:
* fast/ruby/ruby-inline-style-not-updated-with-before-after-content.html:
* fast/ruby/ruby-inline-style-not-updated.html:
* fast/table/table-before-child-style-update.html:
* fast/table/table-row-before-child-style-update.html:
* fast/table/table-row-style-not-updated-with-after-content.html:
* fast/table/table-row-style-not-updated-with-before-content.html:
* fast/table/table-style-not-updated.html:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94582 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed. Qt rebaselines for r94543.

* platform/qt/fast/css-generated-content/table-before-after-child-add-expected.txt:
* platform/qt/fast/css-generated-content/table-cell-before-after-child-add-expected.txt:
* platform/qt/fast/css-generated-content/table-row-before-after-child-add-expected.txt:
* platform/qt/fast/css-generated-content/table-table-before-after-child-add-expected.txt:
* platform/qt/fast/ruby/ruby-block-style-not-updated-expected.txt: Added.
* platform/qt/fast/ruby/ruby-block-style-not-updated-with-before-after-content-expected.txt: Added.
* platform/qt/fast/ruby/ruby-inline-style-not-updated-expected.txt: Added.
* platform/qt/fast/ruby/ruby-inline-style-not-updated-with-before-after-content-expected.txt: Added.
* platform/qt/fast/table/table-after-child-in-table-expected.txt:
* platform/qt/fast/table/table-before-child-in-table-expected.txt:
* platform/qt/fast/table/table-cell-after-child-in-table-expected.txt:
* platform/qt/fast/table/table-cell-before-child-in-table-expected.txt:
* platform/qt/fast/table/table-row-after-child-in-table-expected.txt:
* platform/qt/fast/table/table-row-before-child-in-table-expected.txt:
* platform/qt/fast/table/table-row-style-not-updated-expected.txt: Added.
* platform/qt/fast/table/table-row-style-not-updated-with-after-content-expected.txt: Added.
* platform/qt/fast/table/table-row-style-not-updated-with-before-content-expected.txt: Added.
* platform/qt/fast/table/table-style-not-updated-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94547 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-06-28  Pierre Rossi  <pierre.rossi@gmail.com>

        Reviewed by Eric Seidel.

        Warnings in JSC's JIT on 32 bit
        https://bugs.webkit.org/show_bug.cgi?id=63259

        Fairly straightforward, just use ASSERT_JIT_OFFSET_UNUSED when it applies.

        * jit/JITPropertyAccess32_64.cpp:
        (JSC::JIT::emit_op_method_check):
        (JSC::JIT::compileGetByIdHotPath):
        (JSC::JIT::emit_op_put_by_id):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@89985 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/JavaScriptCore/jit/JITPropertyAccess32_64.cpp

[Qt][Symbian] REGRESSION[94105] DumpRenderTree.exe doesn't build on Symbian
https://bugs.webkit.org/show_bug.cgi?id=67644

Patch by Ademar de Souza Reis Jr. <ademar.reis@openbossa.org> on 2011-09-06
Reviewed by Csaba Osztrogonác.

* symbian/eabi/QtWebKitu.def: add missing entry for
FrameLoaderClientQt::dumpProgressFinishedCallback(bool)

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94588 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed. Rebaselines for r94541.

Author: Abhishek Arya <inferno@chromium.org>

* platform/qt/fast/multicol/span/anonymous-split-block-crash-expected.txt:
* platform/qt/fast/multicol/span/clone-anonymous-block-non-inline-child-crash-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94546 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

LayoutTests/platform/qt/fast/multicol/span/anonymous-split-block-crash-expected.txt
LayoutTests/platform/qt/fast/multicol/span/clone-anonymous-block-non-inline-child-crash-expected.txt

    (rebaselined the tests with the results from the build-bots)

Style not propagated to anonymous boxes and anonymous
inline-blocks.
https://bugs.webkit.org/show_bug.cgi?id=67364

Author: Abhishek Arya <inferno@chromium.org>
Reviewed by James Robinson.

Source/WebCore:

Tests: fast/ruby/ruby-block-style-not-updated-with-before-after-content.html
       fast/ruby/ruby-block-style-not-updated.html
       fast/ruby/ruby-inline-style-not-updated-with-before-after-content.html
       fast/ruby/ruby-inline-style-not-updated.html
       fast/table/table-row-style-not-updated-with-after-content.html
       fast/table/table-row-style-not-updated-with-before-content.html
       fast/table/table-row-style-not-updated.html
       fast/table/table-style-not-updated.html

* rendering/RenderObject.cpp:
(WebCore::RenderObject::propagateStyleToAnonymousChildren):
* rendering/RenderObject.h:
(WebCore::RenderObject::isBeforeAfterContent):
* rendering/RenderRuby.cpp:
(WebCore::RenderRubyAsInline::styleDidChange):
(WebCore::RenderRubyAsBlock::styleDidChange):
* rendering/RenderRuby.h:
* rendering/RenderTable.cpp:
(WebCore::RenderTable::styleDidChange):
* rendering/RenderTableRow.cpp:
(WebCore::RenderTableRow::styleDidChange):
(WebCore::RenderTableRow::addChild):
* rendering/RenderTableSection.cpp:
(WebCore::RenderTableSection::styleDidChange):
(WebCore::RenderTableSection::addChild):
* rendering/RenderTableSection.h:

LayoutTests:

None of the different color boxes should overlap. If they overlap, test fails.

* fast/ruby/ruby-block-style-not-updated-expected.png: Added.
* fast/ruby/ruby-block-style-not-updated-expected.txt: Added.
* fast/ruby/ruby-block-style-not-updated-with-before-after-content-expected.png: Added.
* fast/ruby/ruby-block-style-not-updated-with-before-after-content-expected.txt: Added.
* fast/ruby/ruby-block-style-not-updated-with-before-after-content.html: Added.
* fast/ruby/ruby-block-style-not-updated.html: Added.
* fast/ruby/ruby-inline-style-not-updated-expected.png: Added.
* fast/ruby/ruby-inline-style-not-updated-expected.txt: Added.
* fast/ruby/ruby-inline-style-not-updated-with-before-after-content-expected.png: Added.
* fast/ruby/ruby-inline-style-not-updated-with-before-after-content-expected.txt: Added.
* fast/ruby/ruby-inline-style-not-updated-with-before-after-content.html: Added.
* fast/ruby/ruby-inline-style-not-updated.html: Added.
* fast/table/table-row-style-not-updated-expected.png: Added.
* fast/table/table-row-style-not-updated-expected.txt: Added.
* fast/table/table-row-style-not-updated-with-after-content-expected.png: Added.
* fast/table/table-row-style-not-updated-with-after-content-expected.txt: Added.
* fast/table/table-row-style-not-updated-with-after-content.html: Added.
* fast/table/table-row-style-not-updated-with-before-content-expected.png: Added.
* fast/table/table-row-style-not-updated-with-before-content-expected.txt: Added.
* fast/table/table-row-style-not-updated-with-before-content.html: Added.
* fast/table/table-row-style-not-updated.html: Added.
* fast/table/table-style-not-updated-expected.png: Added.
* fast/table/table-style-not-updated-expected.txt: Added.
* fast/table/table-style-not-updated.html: Added.
* platform/mac/fast/css-generated-content/table-before-after-child-add-expected.txt:
* platform/mac/fast/css-generated-content/table-cell-before-after-child-add-expected.txt:
* platform/mac/fast/css-generated-content/table-row-before-after-child-add-expected.txt:
* platform/mac/fast/css-generated-content/table-table-before-after-child-add-expected.txt:
* platform/mac/fast/table/table-after-child-in-table-expected.txt:
* platform/mac/fast/table/table-before-child-in-table-expected.txt:
* platform/mac/fast/table/table-cell-after-child-in-table-expected.txt:
* platform/mac/fast/table/table-cell-before-child-in-table-expected.txt:
* platform/mac/fast/table/table-row-after-child-in-table-expected.txt:
* platform/mac/fast/table/table-row-before-child-in-table-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94543 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/WebCore/rendering/RenderTableSection.cpp

Crash in RenderObjectChildList::destroyLeftOverChildren()
https://bugs.webkit.org/show_bug.cgi?id=64753

Author: Abhishek Arya <inferno@chromium.org>
Reviewed by James Robinson.

Source/WebCore:

If any of the ancestors between column span element and containing
column's block is a continuation, then don't attempt to render the
column span by splitting the block into continuations.

Test: fast/multicol/column-span-parent-continuation-crash.html

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::columnsBlockForSpanningElement):

LayoutTests:

anonymous-split-block-crash rendering was already wrong. The fix prevents
the tree to go bad and hence does not do the column-span rendering. same issue
with clone-anonymous-block-non-inline-child-crash test.

* fast/multicol/column-span-parent-continuation-crash-expected.txt: Added.
* fast/multicol/column-span-parent-continuation-crash.html: Added.
* platform/mac/fast/multicol/span/anonymous-split-block-crash-expected.png:
* platform/mac/fast/multicol/span/anonymous-split-block-crash-expected.txt:
* platform/mac/fast/multicol/span/clone-anonymous-block-non-inline-child-crash-expected.png:
* platform/mac/fast/multicol/span/clone-anonymous-block-non-inline-child-crash-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94541 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

LayoutTests/platform/mac/fast/multicol/span/anonymous-split-block-crash-expected.txt
LayoutTests/platform/mac/fast/multicol/span/clone-anonymous-block-non-inline-child-crash-expected.txt

Removed m_owner accessed in custom scrollbars.
https://bugs.webkit.org/show_bug.cgi?id=64737

Reviewed by David Hyatt.

Problem does not reproduce in DRT, even with Eventhandler tricks
and gc(). So, adding a manual test.

* manual-tests/custom-scrollbar-renderer-removed-crash.html: Added.
* page/FrameView.cpp:
(WebCore::FrameView::clearOwningRendererForCustomScrollbars):
* page/FrameView.h:
* rendering/RenderBox.cpp:
(WebCore::RenderBox::willBeDestroyed): when this renderbox is getting
destroyed, clear the custom scrollbar in this frameview having this renderbox
as its owning renderer.
* rendering/RenderScrollbar.cpp:
(WebCore::RenderScrollbar::getScrollbarPseudoStyle): fix the null check.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94107 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/WebCore/page/FrameView.h
Source/WebCore/rendering/RenderBox.cpp

Also changed Source/WebCore/rendering/RenderWidget.cpp because the code
from RenderBox is duplicated there (in this branch, it was refactored in
trunk).

Crash in Range::processAncestorsAndTheirSiblings.
https://bugs.webkit.org/show_bug.cgi?id=67556

Reviewed by Ryosuke Niwa.

Source/WebCore:

Create a temporary RefPtr Node vector to keep all the ancestor's
childs so that we don't access removed child nodes.

Test: fast/dom/Range/range-delete-contents-event-fire-crash.html

* dom/Range.cpp:
(WebCore::Range::processContents):
(WebCore::Range::processAncestorsAndTheirSiblings):

LayoutTests:

Tests that we do not crash when removing contents of
a range from the document.

* fast/dom/Range/range-delete-contents-event-fire-crash-expected.txt: Added.
* fast/dom/Range/range-delete-contents-event-fire-crash.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94511 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Flaky crash with JavaScript URLs
https://bugs.webkit.org/show_bug.cgi?id=66360

Reviewed by Nate Chapin.

Source/WebCore:

This patch teaches DocumentWriter::begin to make a copy of the URL
before beginning a new document to avoid a crash.

Test: fast/loader/javascript-url-iframe-crash.html

* loader/DocumentWriter.cpp:
(WebCore::DocumentWriter::begin):

LayoutTests:

* fast/loader/javascript-url-iframe-crash-expected.txt: Added.
* fast/loader/javascript-url-iframe-crash.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94112 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add test for testing non-existent eventsource and
fix typo in previous test.
https://bugs.webkit.org/show_bug.cgi?id=49448

Reviewed by Alexey Proskuryakov.

* http/tests/eventsource/eventsource-status-error-iframe-crash-expected.txt: Removed.
* http/tests/eventsource/eventsource-status-error-iframe-crash.html: Removed.
* http/tests/eventsource/existent-eventsource-status-error-iframe-crash-expected.txt: Copied from LayoutTests/http/tests/eventsource/eventsource-status-error-iframe-crash-expected.txt.
* http/tests/eventsource/existent-eventsource-status-error-iframe-crash.html: Copied from LayoutTests/http/tests/eventsource/eventsource-status-error-iframe-crash.html.
* http/tests/eventsource/non-existent-eventsource-status-error-iframe-crash-expected.txt: Added.
* http/tests/eventsource/non-existent-eventsource-status-error-iframe-crash.html: Added.
* http/tests/eventsource/resources/request-existent-eventsource-error.html: Copied from LayoutTests/http/tests/eventsource/resources/request-status-error.htm.
* http/tests/eventsource/resources/request-non-existent-eventsource-error.html: Added.
* http/tests/eventsource/resources/request-status-error.htm: Removed.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94258 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed: dos2unix on a recently added test that is failing

document-destruction-within-unload.html's expected file was added with
CRLF line terminations, causing failures in our 2.2 buildbots.

I guess it doesn't fail in trunk because of the new run-webkit-tests
script.

[Qt] Unskip API test for load signals order
https://bugs.webkit.org/show_bug.cgi?id=67285

Reviewed by Andreas Kling.

* tests/qwebpage/tst_qwebpage.cpp:
(tst_QWebPage::loadSignalsOrder):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94182 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Emit last progress notification before calling dispatchDidFinishLoad
https://bugs.webkit.org/show_bug.cgi?id=28851

Reviewed by Adam Barth.

Source/WebCore:

Original patch by Xan Lopez. Change FrameLoader to dispatch the notifications in
a more natural sequence.

Test: http/tests/loading/progress-finished-callback.html

* loader/FrameLoader.cpp:
(WebCore::FrameLoader::checkLoadCompleteForThisFrame):

Source/WebKit/qt:

Add infrastructure to dump progressFinishedNotification callback in DRT.

* WebCoreSupport/DumpRenderTreeSupportQt.cpp:
(DumpRenderTreeSupportQt::dumpProgressFinishedCallback):
* WebCoreSupport/DumpRenderTreeSupportQt.h:
* WebCoreSupport/FrameLoaderClientQt.cpp:
(WebCore::FrameLoaderClientQt::postProgressFinishedNotification):
* WebCoreSupport/FrameLoaderClientQt.h:

Tools:

Implement dumpProgressFinishedCallback() in the layoutTestController for Qt,
Chromium and GTK ports. The first two were simple changes.

Some refactoring was needed in GTK port to assure that the dumps are emitted
correctly and the processing after the page load finished is in the right place.

To detect the load finished, we use the callback for the "load-status" notifier
signal of the top loading frame. And we use the deprecated "load-finished" signal
for dumping (if enabled) the progress finished expected output.

* DumpRenderTree/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(dumpProgressFinishedCallbackCallback):
(LayoutTestController::staticFunctions):
* DumpRenderTree/LayoutTestController.h:
(LayoutTestController::dumpProgressFinishedCallback):
(LayoutTestController::setDumpProgressFinishedCallback):
* DumpRenderTree/chromium/LayoutTestController.cpp:
(LayoutTestController::LayoutTestController):
(LayoutTestController::dumpProgressFinishedCallback):
(LayoutTestController::reset):
* DumpRenderTree/chromium/LayoutTestController.h:
(LayoutTestController::shouldDumpProgressFinishedCallback):
(LayoutTestController::setShouldDumpProgressFinishedCallback):
* DumpRenderTree/chromium/WebViewHost.cpp:
(WebViewHost::didStopLoading):
* DumpRenderTree/gtk/DumpRenderTree.cpp:
(webViewLoadFinished):
(topLoadingFrameLoadFinished):
(webFrameLoadStatusNotified):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::reset):
(LayoutTestController::dumpProgressFinishedCallback):
* DumpRenderTree/qt/LayoutTestControllerQt.h:

LayoutTests:

Keep track of the order expected for the progressFinishedNotification, which is
before the didFinishLoadForFrame.

* http/tests/loading/progress-finished-callback-expected.txt: Added.
* http/tests/loading/progress-finished-callback.html: Added.
* platform/mac/Skipped:
* platform/win/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94105 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

LayoutTests/platform/mac/Skipped
LayoutTests/platform/win/Skipped
Tools/DumpRenderTree/chromium/LayoutTestController.h

[Qt][Symbian] REGRESSION(93235) Missing .def update
https://bugs.webkit.org/show_bug.cgi?id=67307

Patch by Jade Han <jade.han@nokia.com> on 2011-09-02
Reviewed by Laszlo Gombos.

* symbian/eabi/QtWebKitu.def:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94462 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Add test for correct order of load signals in QWebPage
https://bugs.webkit.org/show_bug.cgi?id=66016

Reviewed by Benjamin Poulain.

Add API test to ensure the order of load signals: loadStarted() needs to be emitted
first, then loadProgress(100), followed by loadFinished().

The test is skipped since this right now is broken, the bug
https://bugs.webkit.org/show_bug.cgi?id=28851 tracks one possible way to fix.

* tests/qwebpage/tst_qwebpage.cpp:
(SpyForLoadSignalsOrder::SpyForLoadSignalsOrder):
(SpyForLoadSignalsOrder::isFinished):
(SpyForLoadSignalsOrder::onLoadProgress):
(tst_QWebPage::loadSignalsOrder_data):
(tst_QWebPage::loadSignalsOrder):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@92963 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Source/WebCore: Unload events can crash us when they blank out a parent frame.
https://bugs.webkit.org/show_bug.cgi?id=64741

Author: Nate Chapin <japhet@chromium.org>
Reviewed by Adam Barth.

Test: fast/loader/document-destruction-within-unload.html

* loader/FrameLoader.cpp:
(WebCore::FrameLoader::stopLoading): Prevent unload events
    from going into infinite recursion.
(WebCore::FrameLoader::setDocumentLoader): Ensure we don't
    let set m_documentLoader to a DocumentLoader with a null Frame*.
(WebCore::FrameLoader::detachChildren): Save off a vector of
    children to detach, rather than doing it inline.

LayoutTests: Test for https://bugs.webkit.org/show_bug.cgi?id=64741.

Reviewed by Adam Barth.

* fast/loader/document-destruction-within-unload-expected.txt: Added.
* fast/loader/document-destruction-within-unload.html: Added.
* fast/loader/resources/document-destruction-within-unload-iframe.html: Added.
* fast/loader/resources/document-destruction-within-unload.svg: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93521 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/WebCore/loader/FrameLoader.cpp

[Qt] Update a couple of layout test results

Part of 881f9ca3b79397e9c2d7d762c698759fc4b9d2ed (r94172).
(from trunk)

[Qt] Unreviewed gardening. Add and update platform specific expected files after r94109.

Based on digging of Gabor Rapcsanyi.

* platform/qt/fast/table/table-after-child-in-table-expected.txt:
* platform/qt/fast/table/table-before-child-in-table-expected.txt:
* platform/qt/fast/table/table-before-child-style-update-expected.png: Added.
* platform/qt/fast/table/table-before-child-style-update-expected.txt: Added.
* platform/qt/fast/table/table-row-after-child-in-block-expected.txt:
* platform/qt/fast/table/table-row-after-child-in-table-expected.txt:
* platform/qt/fast/table/table-row-before-child-in-block-expected.txt:
* platform/qt/fast/table/table-row-before-child-in-table-expected.txt:
* platform/qt/fast/table/table-row-before-child-style-update-expected.png: Added.
* platform/qt/fast/table/table-row-before-child-style-update-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94171 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Source/WebCore: Style not updated for table parts in :before, :after content.
https://bugs.webkit.org/show_bug.cgi?id=66141

Reviewed by Dave Hyatt.

Tests: fast/table/table-before-child-style-update.html
       fast/table/table-row-before-child-style-update.html

* rendering/RenderObjectChildList.cpp:
(WebCore::RenderObjectChildList::updateBeforeAfterContent):

LayoutTests: Style not updated for table parts in :before, :after content.
https://bugs.webkit.org/show_bug.cgi?id=66141

1. 2 new tests show that the blue and green boxes don't overlap
after the fix.
2. update r93794 tests showing that style info is propagated.

Reviewed by Dave Hyatt.

* fast/table/table-before-child-style-update-expected.png: Added.
* fast/table/table-before-child-style-update-expected.txt: Added.
* fast/table/table-before-child-style-update.html: Added.
* fast/table/table-row-before-child-style-update-expected.png: Added.
* fast/table/table-row-before-child-style-update-expected.txt: Added.
* fast/table/table-row-before-child-style-update.html: Added.
* platform/mac/fast/table/table-after-child-in-table-expected.txt:
* platform/mac/fast/table/table-before-child-in-table-expected.txt:
* platform/mac/fast/table/table-row-after-child-in-block-expected.txt:
* platform/mac/fast/table/table-row-after-child-in-table-expected.txt:
* platform/mac/fast/table/table-row-before-child-in-block-expected.txt:
* platform/mac/fast/table/table-row-before-child-in-table-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94109 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Adding a few tests for problems affecting trunk but not QtWebKit-2.2

Extracted from these bugs:
    https://bugs.webkit.org/show_bug.cgi?id=66799
    (r93708, 3a8c5f158f923d9cf8f3a3086de17073f0729889)

    https://bugs.webkit.org/show_bug.cgi?id=66572
    (r93487, 146961c0d96c4d15a8c123c10a37e21ffa059366)

[Qt] Unreviewed gardening. Add Qt specific expected results for new tests introduced in r93794.

Author: Csaba Osztrogonác <ossy@webkit.org>

* platform/qt/Skipped:
* platform/qt/fast/table/table-after-child-in-table-expected.png: Added.
* platform/qt/fast/table/table-after-child-in-table-expected.txt: Added.
* platform/qt/fast/table/table-before-child-in-table-expected.png: Added.
* platform/qt/fast/table/table-before-child-in-table-expected.txt: Added.
* platform/qt/fast/table/table-cell-after-child-in-block-expected.png: Added.
* platform/qt/fast/table/table-cell-after-child-in-block-expected.txt: Added.
* platform/qt/fast/table/table-cell-after-child-in-table-expected.png: Added.
* platform/qt/fast/table/table-cell-after-child-in-table-expected.txt: Added.
* platform/qt/fast/table/table-cell-before-child-in-block-expected.png: Added.
* platform/qt/fast/table/table-cell-before-child-in-block-expected.txt: Added.
* platform/qt/fast/table/table-cell-before-child-in-table-expected.png: Added.
* platform/qt/fast/table/table-cell-before-child-in-table-expected.txt: Added.
* platform/qt/fast/table/table-row-after-child-in-block-expected.png: Added.
* platform/qt/fast/table/table-row-after-child-in-block-expected.txt: Added.
* platform/qt/fast/table/table-row-after-child-in-table-expected.png: Added.
* platform/qt/fast/table/table-row-after-child-in-table-expected.txt: Added.
* platform/qt/fast/table/table-row-before-child-in-block-expected.png: Added.
* platform/qt/fast/table/table-row-before-child-in-block-expected.txt: Added.
* platform/qt/fast/table/table-row-before-child-in-table-expected.png: Added.
* platform/qt/fast/table/table-row-before-child-in-table-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93859 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

LayoutTests/platform/qt/Skipped

Incorrect layout of :before and :after content, with display
table, table-row and table-cell.
https://bugs.webkit.org/show_bug.cgi?id=66699

Reviewed by David Hyatt.

Source/WebCore:

Tests: fast/table/table-after-child-in-table.html
       fast/table/table-before-child-in-table.html
       fast/table/table-cell-after-child-in-block.html
       fast/table/table-cell-after-child-in-table.html
       fast/table/table-cell-before-child-in-block.html
       fast/table/table-cell-before-child-in-table.html
       fast/table/table-row-after-child-in-block.html
       fast/table/table-row-after-child-in-table.html
       fast/table/table-row-before-child-in-block.html
       fast/table/table-row-before-child-in-table.html

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::addChildIgnoringAnonymousColumnBlocks):
Fix the looping condition to detect :after child correctly.
isAnonymousBlock() does not apply to tables, instead
using isAnonymous().
* rendering/RenderTableRow.cpp:
(WebCore::RenderTableRow::addChild): Don't add the new child
in the generatedContainer with :before, :after content.
* rendering/RenderTableSection.cpp:
(WebCore::RenderTableSection::addChild): Don't add the new child
in the generatedContainer with :before, :after content.

LayoutTests:

* fast/table/table-after-child-in-table.html: Added.
* fast/table/table-before-child-in-table.html: Added.
* fast/table/table-cell-after-child-in-block.html: Added.
* fast/table/table-cell-after-child-in-table.html: Added.
* fast/table/table-cell-before-child-in-block.html: Added.
* fast/table/table-cell-before-child-in-table.html: Added.
* fast/table/table-row-after-child-in-block.html: Added.
* fast/table/table-row-after-child-in-table.html: Added.
* fast/table/table-row-before-child-in-block.html: Added.
* fast/table/table-row-before-child-in-table.html: Added.
* platform/mac/fast/table/table-after-child-in-table-expected.png: Added.
* platform/mac/fast/table/table-after-child-in-table-expected.txt: Added.
* platform/mac/fast/table/table-before-child-in-table-expected.png: Added.
* platform/mac/fast/table/table-before-child-in-table-expected.txt: Added.
* platform/mac/fast/table/table-cell-after-child-in-block-expected.png: Added.
* platform/mac/fast/table/table-cell-after-child-in-block-expected.txt: Added.
* platform/mac/fast/table/table-cell-after-child-in-table-expected.png: Added.
* platform/mac/fast/table/table-cell-after-child-in-table-expected.txt: Added.
* platform/mac/fast/table/table-cell-before-child-in-block-expected.png: Added.
* platform/mac/fast/table/table-cell-before-child-in-block-expected.txt: Added.
* platform/mac/fast/table/table-cell-before-child-in-table-expected.png: Added.
* platform/mac/fast/table/table-cell-before-child-in-table-expected.txt: Added.
* platform/mac/fast/table/table-row-after-child-in-block-expected.png: Added.
* platform/mac/fast/table/table-row-after-child-in-block-expected.txt: Added.
* platform/mac/fast/table/table-row-after-child-in-table-expected.png: Added.
* platform/mac/fast/table/table-row-after-child-in-table-expected.txt: Added.
* platform/mac/fast/table/table-row-before-child-in-block-expected.png: Added.
* platform/mac/fast/table/table-row-before-child-in-block-expected.txt: Added.
* platform/mac/fast/table/table-row-before-child-in-table-expected.png: Added.
* platform/mac/fast/table/table-row-before-child-in-table-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93794 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Crash in InlineBox::deleteLine due to accessing removed renderer.
https://bugs.webkit.org/show_bug.cgi?id=66015

Author: Abhishek Arya <inferno@chromium.org>
Reviewed by Simon Fraser.

Source/WebCore:

Test: fast/block/line-layout/inline-box-wrapper-crash.html

* rendering/RenderBox.cpp:
(WebCore::RenderBox::positionLineBox): Make sure that the
previous inline box wrapper is destroyed properly, before
setting a new one.

LayoutTests:

* fast/block/line-layout/inline-box-wrapper-crash-expected.txt: Added.
* fast/block/line-layout/inline-box-wrapper-crash.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94001 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/WebCore/rendering/RenderBox.cpp

Revert "Add build flags to reduce the size of Webcore library"

This reverts commit 16cbdf38f280feb8fcc073b5c461673dffe0feab.

    [Qt] libwebcore.a source is compiled without -fvisibility=hidden -fvisibility-inlines-hidden
    https://bugs.webkit.org/show_bug.cgi?id=65470

CONFIG += bsymbolic_functions broke most plugin layout tests on Linux.
Without this CONFIG flag, the library size is the same as before, so we
lose the main motivation for the patch.

Details are in the aforementioned bug.

[Qt] Do not unconditionally use pkg-config in .pro files
https://bugs.webkit.org/show_bug.cgi?id=67055

Reviewed by Andreas Kling.

Original patch from Rohan McGovern <rohan.mcgovern@nokia.com>

Using the first pkg-config in PATH is prone to errors when cross
compiling inside the Qt repository (using Qt's build-system).

This patch protect calls for pkg-config with
!contains(QT_CONFIG, no-pkg-config). no-pkg-config is added to
QT_CONFIG by Qt's 'configure' when cross-compiling on systems
without pkg-config.

The respective change in Qt's configure has been submited already.

No new tests as this is just a build change.

Source/JavaScriptCore:

* wtf/wtf.pri: protect pkg-config calls

Source/WebCore:

* features.pri: protect pkg-config calls

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@94075 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix build error when compiling with MinGW-w64 by disabling JIT
on Windows 64-bit
https://bugs.webkit.org/show_bug.cgi?id=61235

Patch by Jonathan Liu <net147@gmail.com> on 2011-08-28
Reviewed by Gavin Barraclough.

The fixed mmap executable allocator for JIT on x86_64 requires
sys/mman.h which is not available on Windows.

* wtf/Platform.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93950 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Revert "[Qt] Fix tst_QWebFrame::setHtmlWithResource() API test" [1/2]

This reverts commit 83cad9be6897798567d0e477ab7f928c93fceed1.

The patch was not needed in qtwebkit-2.2, as the failure was
specific to trunk. This is the first commit, which added XFAILs
to the test.

Revert "[Qt] Fix tst_QWebFrame::setHtmlWithResource() API test" [2/2]

This reverts commit dfe73c30b98223a75f91050ae55ccb87afcf2c73.

The patch was not needed in qtwebkit-2.2, as the failure was
specific to trunk. This is the second commit, which "fixes"
the test.

Add build flags to reduce the size of Webcore library

Patch by Holger Freyther <zecke@selfish.org>
Rubberstamped by Andreas Kling.

(qtwebkit-2.2 only by now)

[Qt] libwebcore.a source is compiled without -fvisibility=hidden -fvisibility-inlines-hidden
https://bugs.webkit.org/show_bug.cgi?id=65470

WTF_USE_EXPORT_MACROS is still not enabled on trunk because there are
pending issues with WebKit2, but it appears to be safe to enable this
on QtWebKit, so this change is being added to the qtwebkit-2.2 branch.

The patch also adds CONFIG += hide_symbols and on Unix, +=
bsymbolic_functions, which should help reduce the library binary size.

Testing this is quite easy, so if there's any problem we'll soon
know. :-)

[Qt] Enable password echo on Symbian builds
https://bugs.webkit.org/show_bug.cgi?id=66954

Reviewed by Chang Shu.

No tests, as this is just a build change.

* config.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93817 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed QtWebKit.pro fix for when building inside Qt

Patch by Simo Fält <simo.falt@nokia.com>

The QtWebKit version was being overwritten by a global Qt version when
QtWebKit was built inside Qt. Fixed by moving the version
definition after the inclusion of qbase.pri.

* QtWebKit.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93787 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add missing libWebKitSystemInterfaceLion.a to QtWebKit-2.2 branch

This file is required when building QtWebKit using the QTKit multimedia
backend on Lion, as it provides a bridge to some system interfaces. It
was added on trunk a while ago with several other changes, but all we
need is the file itself.

Added support for momentarily revealing last typed character in password input.
Code change was partially based on Apple's iOS code and Samuel Nevala's work.
https://bugs.webkit.org/show_bug.cgi?id=32509

Patch by Chang Shu <cshu@webkit.org> on 2011-08-23
Reviewed by Alexey Proskuryakov.

Source/WebCore:

* editing/InsertIntoTextNodeCommand.cpp:
(WebCore::InsertIntoTextNodeCommand::doApply):
* rendering/RenderText.cpp:
(WebCore::SecureTextTimer::SecureTextTimer):
(WebCore::SecureTextTimer::restartWithNewText):
(WebCore::SecureTextTimer::invalidate):
(WebCore::SecureTextTimer::lastTypedCharacterOffset):
(WebCore::SecureTextTimer::fired):
(WebCore::RenderText::willBeDestroyed):
(WebCore::RenderText::setTextInternal):
(WebCore::RenderText::secureText):
(WebCore::RenderText::momentarilyRevealLastTypedCharacter):
* rendering/RenderText.h:
(WebCore::RenderText::isSecure):
* testing/Internals.cpp:
(WebCore::Internals::setPasswordEchoEnabled): Fixed some silly coding in Internals.
(WebCore::Internals::setPasswordEchoDurationInSeconds):
(WebCore::Internals::reset):

LayoutTests:

Updated failed expected files.

* editing/input/password-echo-passnode-expected.txt:
* editing/input/password-echo-passnode2-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93656 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/WebCore/editing/InsertIntoTextNodeCommand.cpp
Source/WebCore/testing/Internals.cpp

Source/JavaScriptCore: Rename "makeSecure" to "fill" and remove the support for displaying last character
to avoid layering violatation.
https://bugs.webkit.org/show_bug.cgi?id=59114

Patch by Chang Shu <cshu@webkit.org> on 2011-07-08
Reviewed by Alexey Proskuryakov.

* JavaScriptCore.exp:
* JavaScriptCore.order:
* wtf/text/StringImpl.cpp:
(WTF::StringImpl::fill):
* wtf/text/StringImpl.h:
* wtf/text/WTFString.h:
(WTF::String::fill):

Source/WebCore: Update calling sites after function renamed.
https://bugs.webkit.org/show_bug.cgi?id=59114

Patch by Chang Shu <cshu@webkit.org> on 2011-07-08
Reviewed by Alexey Proskuryakov.

No new tests, just refactoring.

* editing/visible_units.cpp:
(WebCore::previousBoundary):
(WebCore::nextBoundary):
* rendering/RenderText.cpp:
(WebCore::RenderText::setTextInternal):

Source/WebKit/win: Update calling sites after function renamed.
https://bugs.webkit.org/show_bug.cgi?id=59114

Patch by Chang Shu <cshu@webkit.org> on 2011-07-08
Reviewed by Alexey Proskuryakov.

* WebKitGraphics.cpp:
(WebDrawText):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@90643 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/JavaScriptCore/JavaScriptCore.order

Add support of setPasswordEchoEnabled and setPasswordEchoDuration for password echo feature
https://bugs.webkit.org/show_bug.cgi?id=66052

Patch by Chang Shu <cshu@webkit.org> on 2011-08-18
Reviewed by Alexey Proskuryakov.

Source/WebCore:

Added runtime settings in WebCore.
Added support in window.internals for testing.

Tests: editing/input/password-echo-passnode.html
       editing/input/password-echo-passnode2.html
       editing/input/password-echo-passnode3.html
       editing/input/password-echo-textnode.html

* page/Settings.cpp:
(WebCore::Settings::Settings):
* page/Settings.h:
(WebCore::Settings::setPasswordEchoEnabled):
(WebCore::Settings::passwordEchoEnabled):
(WebCore::Settings::setPasswordEchoDurationInSeconds):
(WebCore::Settings::passwordEchoDurationInSeconds):
* testing/Internals.cpp:
(WebCore::Internals::Internals):
(WebCore::Internals::setPasswordEchoEnabled):
(WebCore::Internals::setPasswordEchoDurationInSeconds):
(WebCore::Internals::reset):
* testing/Internals.h:
* testing/Internals.idl:

Source/WebKit/qt:

Enable password echo under the build flag.

* Api/qwebsettings.cpp:
(QWebSettingsPrivate::apply):

LayoutTests:

Added tests.

* editing/input/password-echo-passnode-expected.txt: Added.
* editing/input/password-echo-passnode.html: Added.
* editing/input/password-echo-passnode2-expected.txt: Added.
* editing/input/password-echo-passnode2.html: Added.
* editing/input/password-echo-passnode3-expected.txt: Added.
* editing/input/password-echo-passnode3.html: Added.
* editing/input/password-echo-textnode-expected.txt: Added.
* editing/input/password-echo-textnode.html: Added.
* editing/input/resources: Added.
* editing/input/resources/password-echo.js: Added.
(secureChar):
(secureText):
(log):
(assert):
(run.else):
(run):
(init):
* platform/wk2/Skipped: No support yet.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93291 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

LayoutTests/platform/wk2/Skipped
Source/WebCore/page/Settings.cpp
Source/WebCore/page/Settings.h
Source/WebCore/testing/Internals.cpp
Source/WebCore/testing/Internals.h
Source/WebCore/testing/Internals.idl

2011-08-16  Chang Shu  <cshu@webkit.org>

        Support reset in WebCore::Internals
        https://bugs.webkit.org/show_bug.cgi?id=66307

        Reviewed by Dimitri Glazkov.

        New tests will be added when function reset is implemented.

        Added framework code in WebCoreTestSupport. The real implementation of
        Internals::reset() depends on the need from the settings that require a reset.

        * testing/Internals.cpp:
        (WebCore::Internals::reset):
        * testing/Internals.h:
        * testing/js/WebCoreTestSupport.cpp:
        (WebCoreTestSupport::resetInternalsObject):
        * testing/js/WebCoreTestSupport.h:
        * testing/v8/WebCoreTestSupport.cpp:
        (WebCoreTestSupport::resetInternalsObject):
        * testing/v8/WebCoreTestSupport.h:

2011-08-16  Chang Shu  <cshu@webkit.org>

        Support reset in WebCore::Internals
        https://bugs.webkit.org/show_bug.cgi?id=66307

        Reviewed by Dimitri Glazkov.

        Added framework code in WebKit.

        * public/WebTestingSupport.h:
        * src/WebTestingSupport.cpp:
        (WebKit::WebTestingSupport::resetInternalsObject):

2011-08-16  Chang Shu  <cshu@webkit.org>

        Support reset in WebCore::Internals
        https://bugs.webkit.org/show_bug.cgi?id=66307

        Reviewed by Dimitri Glazkov.

        Added framework code in WebKit.

        * WebCoreSupport/DumpRenderTreeSupportQt.cpp:
        (DumpRenderTreeSupportQt::resetInternalsObject):
        * WebCoreSupport/DumpRenderTreeSupportQt.h:

2011-08-16  Chang Shu  <cshu@webkit.org>

        Support reset in WebCore::Internals
        https://bugs.webkit.org/show_bug.cgi?id=66307

        Reviewed by Dimitri Glazkov.

        Added call to resetInternalsObject in DRT.

        * DumpRenderTree/chromium/TestShell.cpp:
        (TestShell::resetTestController):
        * DumpRenderTree/mac/DumpRenderTree.mm:
        (resetWebViewToConsistentStateBeforeTesting):
        * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
        (WebCore::WebPage::resetSettings):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93235 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/WebCore/testing/Internals.cpp
Source/WebCore/testing/Internals.h
Source/WebKit/chromium/public/WebTestingSupport.h
Source/WebKit/qt/WebCoreSupport/DumpRenderTreeSupportQt.h
Source/WebKit/qt/WebCoreSupport/DumpRenderTreeSupportQt.cpp

2011-06-09  Robert Hogan  <robert@webkit.org>

        Reviewed by Andreas Kling.

        Teach Qt about window.internals
        https://bugs.webkit.org/show_bug.cgi?id=61074

        Unskip fast/harness/internals-object.html

        * platform/qt/Skipped:
2011-06-09  Robert Hogan  <robert@webkit.org>

        Reviewed by Andreas Kling.

        Teach Qt about window.internals
        https://bugs.webkit.org/show_bug.cgi?id=61074

        A weakness of the Qt DRT setup is that things like JSContextRef are abstracted
        away from the QtWebKit API so we need DumpRenderTreeSupportQt to access WebCore internals.
        Since the window.internals object requires JSContextRef we need to implement it in DumpRenderTreeSupportQt
        where we can access it. DumpRenderTreeSupportQt cannot be compiled outside Qt's WebCore and as it
        is our only possible route into the WebCoreTestSupport class neither can the new window.internals plumbing.
        Likewise we can't put the accessor in WebCoreTestSupport because it would then need to know about QWebFrame
        and others. The only alternative seems like a compile time guard which we would have to teach the bots about.

        * CodeGenerators.pri:
        * WebCore.pri:
        * WebCore.pro:
2011-06-09  Robert Hogan  <robert@webkit.org>

        Reviewed by Andreas Kling.

        Teach Qt about window.internals
        https://bugs.webkit.org/show_bug.cgi?id=61074

        * WebCoreSupport/DumpRenderTreeSupportQt.cpp:
        (DumpRenderTreeSupportQt::injectInternalsObject):
        * WebCoreSupport/DumpRenderTreeSupportQt.h:
2011-06-09  Robert Hogan  <robert@webkit.org>

        Reviewed by Andreas Kling.

        Teach Qt about window.internals
        https://bugs.webkit.org/show_bug.cgi?id=61074

        * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
        (WebCore::DumpRenderTree::initJSObjects):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@88446 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

LayoutTests/platform/qt/Skipped
Source/WebKit/qt/WebCoreSupport/DumpRenderTreeSupportQt.h

2011-06-02  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Darin Adler.

        Add build logistics and plumbing for window.internals object.
        https://bugs.webkit.org/show_bug.cgi?id=60313

        * fast/dom/Window/script-tests/window-property-descriptors.js: Added internals object to list of properties to skip.
        * fast/dom/Window/window-properties.html: Ditto.
        * fast/dom/script-tests/prototype-inheritance-2.js: Ditto.
        * fast/dom/script-tests/prototype-inheritance.js: Ditto.
        * fast/harness/internals-object-expected.txt: Added.
        * fast/harness/internals-object.html: Added.
        * platform/gtk/Skipped: Skipped the newly added test until platform supports window.internals.
        * platform/mac-wk2/Skipped: Ditto.
        * platform/qt/Skipped: Ditto.
        * platform/win/Skipped: Ditto.
2011-06-02  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Darin Adler.

        Add build logistics and plumbing for window.internals object.
        https://bugs.webkit.org/show_bug.cgi?id=60313

        Test: fast/harness/internals-object.html

        * Configurations/WebCoreTestSupport.xcconfig: Added.
        * DerivedSources.make: Added support for generating from Internals.idl.
        * WebCore.gyp/WebCore.gyp: Added new webcore_test_support library.
        * WebCore.gypi: Ditto.
        * WebCore.xcodeproj/project.pbxproj: Added WebCoreTestSupport library.
        * testing/Internals.cpp: Added.
        * testing/Internals.h: Added.
        * testing/Internals.idl: Added.
        * testing/js/WebCoreTestSupport.cpp: Added.
        * testing/js/WebCoreTestSupport.h: Added.
        * testing/v8/WebCoreTestSupport.cpp: Added.
        * testing/v8/WebCoreTestSupport.h: Added.
2011-06-02  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Darin Adler.

        Add build logistics and plumbing for window.internals object.
        https://bugs.webkit.org/show_bug.cgi?id=60313

        * WebKit.gyp: Added linking new webkit_test_support library.
        * public/WebTestingSupport.h: Added.
        * src/WebTestingSupport.cpp: Added.
2011-06-02  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Darin Adler.

        Add build logistics and plumbing for window.internals object.
        https://bugs.webkit.org/show_bug.cgi?id=60313

        * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added linking new WebCoreTestSupport library.
        * DumpRenderTree/chromium/TestShell.cpp:
        (TestShell::bindJSObjectsToWindow): Added injection code.
        * DumpRenderTree/mac/FrameLoadDelegate.mm:
        (-[FrameLoadDelegate didClearWindowObjectInStandardWorldForFrame:]): Ditto.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@87948 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

LayoutTests/platform/gtk/Skipped
LayoutTests/platform/mac-wk2/Skipped
LayoutTests/platform/qt/Skipped
LayoutTests/platform/win/Skipped

Add a missing trailing whitespace to a test expectation

svg/dom/stylesheet-candidate-node-crash-main.html is a crash test and it
fails because of a missing trailing whitespace. For some reason it doesn't
fail on trunk (maybe because of new run-webkit-tests or something in our
DRT?).

Skip failing test introduced in r92966

SHA1: 0e9f086adb423a785038bc12e7a646620a890f7a

editing/selection/select-start-remove-root-crash.html fails (timeout) in
our 2.2 branch. It's a crash test and the original test-case doesn't
crash the QtTestBrowser. I guess there's something missing in our DRT.
Pending investigation.

Crash in FocusController::advanceFocusInDocumentOrder
https://bugs.webkit.org/show_bug.cgi?id=66678

Source/WebCore:

RefPtr the focusable node to prevent getting deleted by mutation
event.

Author: Abhishek Arya <inferno@chromium.org>
Reviewed by Dave Hyatt.

Test: fast/frames/focus-controller-crash-change-event.html

* page/FocusController.cpp:
(WebCore::FocusController::advanceFocusInDocumentOrder):

LayoutTests:

Reviewed by Dave Hyatt.

* fast/frames/focus-controller-crash-change-event-expected.txt: Added.
* fast/frames/focus-controller-crash-change-event.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93514 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/WebCore/page/FocusController.cpp

SimplifiedBackwardsTextIterator returns incorrect offset with first-letter rule
https://bugs.webkit.org/show_bug.cgi?id=66086

Reviewed by Darin Adler.

Source/WebCore:

The bug was caused by SimplifiedBackwardsTextIterator's not taking care of first-letter at all.
Fixing the bug by detecting RenderTextFragment in handleTextNode.

Also added m_shouldHandleFirstLetter to SimplifiedBackwardsTextIterator to keep track of whether or not
the next call to handleTextNode needs to process the first-letter part of the text fragment.

Test: editing/text-iterator/backward-textiterator-first-letter-crash.html

* editing/TextIterator.cpp:
(WebCore::firstRenderTextInFirstLetter): Extracted from handleTextNodeFirstLetter.
(WebCore::TextIterator::handleTextNodeFirstLetter): Calls firstRenderTextInFirstLetter.
(WebCore::SimplifiedBackwardsTextIterator::SimplifiedBackwardsTextIterator):
(WebCore::SimplifiedBackwardsTextIterator::handleTextNode):
(WebCore::SimplifiedBackwardsTextIterator::handleFirstLetter): Added.
* editing/TextIterator.h:

LayoutTests:

Added a test to ensure WebKit does not hit assertions in SimplifiedBackwardsTextIterator.
Also fixed a bug in first-letter-word-boundary.html and updated expected offsets for move backward by word
from 0 to 1 because there is unrendered space before "hello".

* editing/text-iterator/backward-textiterator-first-letter-crash-expected.txt: Added.
* editing/text-iterator/backward-textiterator-first-letter-crash.html: Added.
* editing/text-iterator/first-letter-word-boundary-expected.txt:
* editing/text-iterator/first-letter-word-boundary.html:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93347 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Fix build on Lion

https://bugs.webkit.org/show_bug.cgi?id=66770

Reviewed by Andreas Kling.

We were mistakenly picking up mac/WebCoreSystemInterface.h instead of
the Qt one, and building on Lion revealed this when a typedef for
IOSurfaceRef was wrapped in PLATFORM(MAC).

For now we fix this by including WebCoreSystemInterface using
brackets, so that we'll pick up the right file based on the
include paths. This also means exposing a few missing enums
in our own version of the file, so those were added.

Lasty, we need to link against the right system interface library
on Lion.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93600 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove counter nodes from the tree and fix-up children when they are removed from the counter map.
https://bugs.webkit.org/show_bug.cgi?id=65346

Reviewed by Adam Barth.

Covered by existing CSS counter tests.

* rendering/CounterNode.cpp:
(WebCore::CounterNode::~CounterNode):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@92630 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Move towards supporting user controlled prototypes on CanvasPixelArray
https://bugs.webkit.org/show_bug.cgi?id=66429

Author: Oliver Hunt <oliver@apple.com>
Reviewed by Gavin Barraclough.

Start using a per-global object structure for canvas pixel array.

* bindings/js/JSImageDataCustom.cpp:
(WebCore::toJS):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93264 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/WebCore/bindings/js/JSImageDataCustom.cpp

Source/WebCore: Crash in Document::recalcStyleSelector
https://bugs.webkit.org/show_bug.cgi?id=66335

Author: Abhishek Arya <inferno@chromium.org>
Reviewed by Simon Fraser.

When node is getting destroyed and its removedFromDocument
is not called due to entire document structure torn down(using
removeAllChildren), make sure to clear out the stylesheet
candidate node from document's structures in its destructor.

Test: svg/dom/stylesheet-candidate-node-crash-main.html

* dom/ProcessingInstruction.cpp:
(WebCore::ProcessingInstruction::~ProcessingInstruction):
* html/HTMLLinkElement.cpp:
(WebCore::HTMLLinkElement::~HTMLLinkElement):
* html/HTMLStyleElement.cpp:
(WebCore::HTMLStyleElement::~HTMLStyleElement):
* svg/SVGStyleElement.cpp:
(WebCore::SVGStyleElement::~SVGStyleElement):

LayoutTests: Tests that we do not crash when iterating through stylesheet
candidate list hashset.
https://bugs.webkit.org/show_bug.cgi?id=66335

Reviewed by Simon Fraser.

* svg/dom/resources/stylesheet-candidate-node-crash.svg: Added.
* svg/dom/stylesheet-candidate-node-crash-main-expected.txt: Added.
* svg/dom/stylesheet-candidate-node-crash-main.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@93227 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Conflicts:

Source/WebCore/html/HTMLLinkElement.cpp