2009-08-02  Kevin Ollivier  <kevino@theolliviers.com>

        Reviewed by Jan Alonzo.

        Scripts needed for the waf build.
        https://bugs.webkit.org/show_bug.cgi?id=27619

        * wx/build: Added.
        * wx/build/build_utils.py: Added.
        * wx/build/settings.py: Added.
        * wx/build/waf_extensions.py: Added.
        * wx/build/wxpresets.py: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46693 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-02  Yong Li  <yong.li@torchmobile.com>

        Reviewed by George Staikos.

        DateMath depends on strftime and localtime, which need to be imported manually on WinCE
        https://bugs.webkit.org/show_bug.cgi?id=26558

        * wtf/DateMath.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46692 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Laszlo Gombos  <laszlo.1.gombos@nokia.com>

        Reviewed by Tor Arne Vestbø.

        [Qt] Initialize HTML5 Application Cache for QtWebKit to a sensible default
        https://bugs.webkit.org/show_bug.cgi?id=27866

        * Api/qwebpage.cpp:
        (initializeApplicationCachePathIfNecessary):
        (QWebPagePrivate::QWebPagePrivate):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46691 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Joseph Pecoraro  <joepeck02@gmail.com>

        Reviewed by Timothy Hatcher.

        Inspector: let me *edit* css styles in the web inspector.
        https://bugs.webkit.org/show_bug.cgi?id=27124

        * English.lproj/localizedStrings.js:
        * inspector/front-end/ElementsPanel.js:
        (WebInspector.ElementsPanel): added property stylesheet pointing to 1 stylesheet added to the page if needed
        * inspector/front-end/StylesSidebarPane.js:
        (WebInspector.StylesSidebarPane.prototype.update): handle blank sections
        (WebInspector.StylesSidebarPane.prototype.addBlankSection): blank section in the correct position
        (WebInspector.StylesSidebarPane.prototype.appropriateSelectorForNode): helper to get a nice selector for the selectd node
        (WebInspector.StylePropertiesSection):
        (WebInspector.StylePropertiesSection.prototype.expand):
        (WebInspector.StylePropertiesSection.prototype.isPropertyInherited):
        (WebInspector.StylePropertiesSection.prototype.isPropertyOverloaded):
        (WebInspector.StylePropertiesSection.prototype.addNewBlankProperty):
        (WebInspector.StylePropertiesSection.prototype._dblclickEmptySpace): easily create new properties
        (WebInspector.StylePropertiesSection.prototype._dblclickSelector):
        (WebInspector.StylePropertiesSection.prototype.startEditingSelector): allow for editing selectors
        (WebInspector.StylePropertiesSection.prototype.editingSelectorCommitted):
        (WebInspector.StylePropertiesSection.prototype.editingSelectorCancelled):
        (WebInspector.StylePropertiesSection.prototype._doesSelectorAffectSelectedNode): helper to check if a selector applies to the selected node
        (WebInspector.BlankStylePropertiesSection.prototype._dblclick):
        (WebInspector.BlankStylePropertiesSection.prototype.startEditing):
        (WebInspector.BlankStylePropertiesSection.prototype.editingCancelled):
        (WebInspector.BlankStylePropertiesSection.prototype.editingCommitted):
        (WebInspector.BlankStylePropertiesSection.prototype.makeNormal): morph into a StylePropertiesSection
        (WebInspector.StylePropertyTreeElement.prototype):
        (WebInspector.StylePropertyTreeElement.prototype.):
        * inspector/front-end/inspector.css:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46690 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Anton Muhin  <antonm@chromium.org>

        Reviewed by Adam Barth.

        [v8] check if proxy is present before invoking a handler
        https://bugs.webkit.org/show_bug.cgi?id=27883

        Do not invoke handler function if proxy is null pointer (that would lead to access violation
        anyway)

        * bindings/v8/custom/V8CustomEventListener.cpp:
        (WebCore::V8EventListener::callListenerFunction):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46689 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        Make pixel test results more consistent across Mac OS X versions

        * DumpRenderTree/cg/ImageDiffCG.cpp:
        (createDifferenceImage): Avoid color correction when rendering the
            images into graphics contexts, so that the actual color values are
            compared, regardless of how the images are tagged.
        * DumpRenderTree/mac/DumpRenderTree.mm:
        (createWebViewAndOffscreenWindow): Set the window’s color space to the
            main screen’s color space, even though the window is off-screen.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46688 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  Pavel Feldman  <pfeldman@chromium.org>

        Reviewed by Timothy Hatcher.

        Web Inspector: Reimplement Elements Panel so that its
        interaction with DOM is serialized.

        This is a first cut wuth read support for DOM elements
        tree as well as some limited editing. No properties, no
        styles, no search capabilities are implemented yet.
        Set Preferences.useDOMAgent to true in order to try it
        out, otherwise disabled by default.

        https://bugs.webkit.org/show_bug.cgi?id=27771

        * GNUmakefile.am:
        * WebCore.gypi:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:
        * inspector/InspectorBackend.cpp:
        (WebCore::InspectorBackend::getChildNodes):
        (WebCore::InspectorBackend::setAttribute):
        (WebCore::InspectorBackend::removeAttribute):
        (WebCore::InspectorBackend::setTextNodeValue):
        * inspector/InspectorBackend.h:
        * inspector/InspectorBackend.idl:
        * inspector/InspectorController.cpp:
        (WebCore::InspectorController::windowScriptObjectAvailable):
        (WebCore::InspectorController::setFrontendProxyObject):
        (WebCore::InspectorController::close):
        (WebCore::InspectorController::populateScriptObjects):
        (WebCore::InspectorController::didCommitLoad):
        * inspector/InspectorController.h:
        (WebCore::InspectorController::domAgent):
        * inspector/InspectorDOMAgent.cpp: Added.
        * inspector/InspectorDOMAgent.h: Added.
        * inspector/InspectorFrontend.cpp:
        * inspector/InspectorFrontend.h:
        * inspector/front-end/Callback.js: Added.
        * inspector/front-end/DOMAgent.js: Added.
        * inspector/front-end/ElementsPanel.js:
        * inspector/front-end/WebKit.qrc:
        * inspector/front-end/inspector.html:
        * inspector/front-end/inspector.js:
        (WebInspector.loaded):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46687 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Try again to make the Qt build bot happy.

* BuildSlaveSupport/test-result-archive: Create a placeholder file inside the newly-created
directory to prevent 'zip' from failing due to the directory containing no files.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46686 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Try and make the Qt build bot happy.

* BuildSlaveSupport/test-result-archive: Handle the case where run-webkit-tests generates
no output files at all and thus does not create the layout-test-results directory.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46685 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Move the include of DumpRenderTreeSupport to after DumpRenderTree has been built.

* Scripts/run-webkit-tests:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46684 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    wtf/Threading.h: added include of Platform.h

        Reviewed by Mark Rowe.

        * wtf/Threading.h: Added #include "Platform.h" since this header
        uses PLATFORM() and other macros.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46683 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 27923: run-webkit-tests should not invoke "ps" repeatedly on Mac OS X
Fix <https://bugs.webkit.org/show_bug.cgi?id=27923>.

Reviewed by Sam Weinig and David D. Kilzer.

run-webkit-tests invokes "ps" after each test on Mac OS X to determine whether DumpRenderTree is
in the process of crashing.  We can do the same test more efficiently with a simple call to sysctl.
I couldn't find any built-in way of doing this directly from perl, so we do it from a simple C module
instead.  This speeds up run-webkit-tests by around 8%.

* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
* DumpRenderTree/mac/PerlSupport/DumpRenderTreeSupport.c: Added.
(processIsCrashing):
* DumpRenderTree/mac/PerlSupport/DumpRenderTreeSupportTiger.pm: Added.  Tiger does not have the SWIG tool
that is used to generate the Perl binding to the C code, so we fall back to the old implementation on Tiger.
* DumpRenderTree/mac/PerlSupport/Makefile: Added.
* Scripts/run-webkit-tests:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46682 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  Ryosuke Niwa  <rniwa@webkit.org>

        Reviewed by Adele Peterson.

        isContentEditable function in htmlediting.* must be removed
        https://bugs.webkit.org/show_bug.cgi?id=27870

        This patch removes isContentEditable function in htmlediting.cpp and
        replaces all usage by calling Node::isContentEditable or its derivatives.

        * editing/AppendNodeCommand.cpp: ditto
        (WebCore::AppendNodeCommand::AppendNodeCommand): ditto
        * editing/IndentOutdentCommand.cpp: ditto
        (WebCore::IndentOutdentCommand::outdentParagraph): ditto
        * editing/InsertNodeBeforeCommand.cpp: ditto
        (WebCore::InsertNodeBeforeCommand::InsertNodeBeforeCommand): ditto
        * editing/htmlediting.cpp: ditto
        (WebCore::enclosingNodeWithTag): ditto
        (WebCore::enclosingNodeOfType): ditto
        (WebCore::canMergeLists): ditto
        * editing/htmlediting.h: Updated prototype

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46681 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Display the reminder once rather than once per ChangeLog.

Reviewed by David D. Kilzer.

* Scripts/prepare-ChangeLog:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46680 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Roll out r46668 as it was misinformed.  ScopeChain is only used with placement new.

Rubber-stamped by Oliver Hunt.

* runtime/ScopeChain.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46679 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Roll out r46676.  Tiger is why we can't have nice things :-(

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46678 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  Brian Weinstein  <bweinstein@apple.com>

        Rubber-stamped by Mark Rowe.

        Remove a non-existent test from the Windows Skipped List.

        * platform/win/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46677 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 27923: run-webkit-tests should not invoke "ps" repeatedly on Mac OS X
Fix <https://bugs.webkit.org/show_bug.cgi?id=27923>.

Reviewed by Sam Weinig.

run-webkit-tests invokes "ps" after each test on Mac OS X to determine whether DumpRenderTree is
in the process of crashing.  We can do the same test more efficiently with a simple call to sysctl.
I couldn't find any built-in way of doing this directly from perl, so we do it from a simple C module
instead.  This speeds up run-webkit-tests by around 8%.

* DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
* DumpRenderTree/mac/PerlSupport/DumpRenderTreeSupport.c: Added.
(processIsCrashing):
* DumpRenderTree/mac/PerlSupport/Makefile: Added.
* Scripts/run-webkit-tests:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46676 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  Zoltan Horvath  <hzoltan@inf.u-szeged.hu>

        Allow custom memory allocation control for JavaScriptCore's HashMap class
        http://bugs.webkit.org/show_bug.cgi?id=27871

        Inherits HashMap class from FastAllocBase because it has been
        instantiated by 'new' in JavaScriptCore/API/JSClassRef.cpp:148.

        * wtf/RefPtrHashMap.h:
        (WTF::):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46675 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  Brady Eidson  <beidson@apple.com>

        For my new test I landed last night, check in Tiger-specific results, and skip on Windows.

        * platform/mac-tiger/http/tests/misc/favicon-loads-with-images-disabled-expected.txt: Added.
        * platform/win/Skipped: Add to the skipped list because the failure needs exploration.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46674 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  John Abd-El-Malek  <jam@chromium.org>

        Reviewed by Eric Seidel.

        Update test to work with multi-process message ports.

        https://bugs.webkit.org/show_bug.cgi?id=27819

        * fast/events/message-channel-gc.html:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46673 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    resolve-ChangeLogs: add error checking to git ls-files command on close()

        Reviewed by Eric Seidel.

        Item 2 of <https://bugs.webkit.org/show_bug.cgi?id=18599#c0>.

        * Scripts/resolve-ChangeLogs: Added error checking to close()
        after running git ls-files.  Added error checking to all
        system() calls by checking for a non-zero WEXITSTATUS($?).
        Changed "|| die;" expressions to "or die $!;".

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46672 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    <webkit.org/b/18599> resolve-ChangeLogs doesn't work with relative paths

        Reviewed by Eric Seidel.

        * Scripts/resolve-ChangeLogs: Used chdirReturningRelativePath()
        and determineVCSRoot() to change directories to the root of the
        project before running the command and to provide a path for
        removeChangeLogArguments() to make sure any ChangeLog arguments
        on the command line are still found.
        (canonicalRelativePath): Added.  Returns a canonical path (e.g.,
        stripping 'dir/../' from the path) relative to the current
        directory.
        (removeChangeLogArguments): Added argument which contains a
        relative path that must be prepended to any ChangeLog arguments.
        Used canonicalRelativePath() and File::Spec->catfile() to
        construct a normalized, relative path to each file.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46671 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Implement VCSUtils::chdirReturningRelativePath()

        Reviewed by Eric Seidel.

        Step 2 to fix:
        <http://webkit.org/b/18599> resolve-ChangeLogs doesn't work with relative paths

        * Scripts/VCSUtils.pm:
        (VCSUtils::chdirReturningRelativePath): Moved here from
        chdirAndGetDifference() in svn-create-patch.
        * Scripts/svn-create-patch: Switched to use
        chdirReturningRelativePath() instead of chdirAndGetDifference().
        (chdirAndGetDifference): Removed.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46670 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Implement VCSUtils::determineVCSRoot()

        Reviewed by Eric Seidel.

        Step 1 to fix:
        <http://webkit.org/b/18599> resolve-ChangeLogs doesn't work with relative paths

        * Scripts/VCSUtils.pm: Removed reference to webkitdirs module.
        (VCSUtils::EXPORT): Added &determineVCSRoot.  Realphabetized.
        (VCSUtils::determineGitRoot): Added.  Based on code in
        commit-log-editor.
        (VCSUtils::determineVCSRoot): Implemented using
        determineGitRoot() and determineSVNRoot().
        * Scripts/commit-log-editor: Replaced use of
        topLevelSourceDirectory() with determineVCSRoot().  Resorted
        use statements.
        (topLevelSourceDirectory): Removed.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46669 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  Zoltan Horvath  <hzoltan@inf.u-szeged.hu>

        Allow custom memory allocation control for JavaScriptCore's ScopeChain class
        https://bugs.webkit.org/show_bug.cgi?id=27834

        Inherits ScopeChain class from FastAllocBase because it has been
        instantiated by 'new' in JavaScriptCore/runtime/JSFunction.h:109.

        * runtime/ScopeChain.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46668 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  Zoltan Horvath  <hzoltan@inf.u-szeged.hu>

        Reviewed by Darin Adler.

        Allow custom memory allocation control for JavaScriptCore's RegExpConstructorPrivate struct
        https://bugs.webkit.org/show_bug.cgi?id=27833

        Inherits RegExpConstructorPrivate class from FastAllocBase because it has been
        instantiated by 'new' in JavaScriptCore/runtime/RegExpConstructor.cpp:152.

        * runtime/RegExpConstructor.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46667 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  John Abd-El-Malek  <jam@chromium.org>

        Reviewed by Adam Barth.

        Fix a message port handle always getting marked as reachable once it gets entangled.

        https://bugs.webkit.org/show_bug.cgi?id=27824

        No tests since this only arises with the Chromium multi-process message port implementation
        which isn't in this repostiory.

        * bindings/v8/V8GCController.cpp:
        (WebCore::GCEpilogueVisitor::visitDOMWrapper):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46666 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-08-01  Evan Martin  <evan@chromium.org>

        Reviewed by Adam Barth.

        Add a regression test for a crash when a page has smart quotes in a
        mime type.

        * fast/parser/smart-quotes-in-tag-expected.txt: Added.
        * fast/parser/smart-quotes-in-tag.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46665 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Daniel Bates  <dbates@intudata.com>

        Reviewed by Eric Seidel.

        https://bugs.webkit.org/show_bug.cgi?id=27757

        Resolves Perl redefinition warnings about functions tmpfile and tmpnam which are
        defined in both File::Temp and POSIX.

        This stops the warnings by preventing the "import" method of File::Temp from being called,
        which would otherwise add these functions into the primary namespace. Note, "use POSIX"
        will import these functions into the primary namespace anyway. We only need one instance
        of these functions in our namespace (or we get redefinition warnings), so lets use the
        one's in the POSIX package.

        * Scripts/update-webkit-auxiliary-libs: Changed "use File::Temp" to "use File::Temp ()".
        * Scripts/update-webkit-support-libs: ditto

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46664 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Yong Li  <yong.li@torchmobile.com>

        Reviewed by Adam Barth.

        Fix Geolocation permission problem
        https://bugs.webkit.org/show_bug.cgi?id=26993

        * page/Geolocation.cpp:
        (WebCore::Geolocation::requestPermission):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46663 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Laszlo Gombos  <laszlo.1.gombos@nokia.com>

        Reviewed by Eric Seidel.

        Add a top level build option to control datagrid support
        https://bugs.webkit.org/show_bug.cgi?id=27915

        * Scripts/build-webkit:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46662 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Norbert Leser  <norbert.leser@nokia.com>

        Reviewed by Eric Seidel.

        Added project properties as needed by symbian platform

        * WebCore.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46661 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Yong Li  <yong.li@torchmobile.com>

        Reviewed by George Staikos.

        Resurrect the old GetTickCount implementation of currentTime, controlled by WTF_USE_QUERY_PERFORMANCE_COUNTER
        currentSystemTime taken from older WebKit; currentTime written by Yong Li <yong.li@torchmobile.com>; cleanup by Joe Mason <joe.mason@torchmobile.com>
        https://bugs.webkit.org/show_bug.cgi?id=27848

        * wtf/CurrentTime.cpp:
        (WTF::currentSystemTime): get current time with GetCurrentFT
        (WTF::currentTime): track msec elapsed since first currentSystemTime call using GetTickCount
        * wtf/Platform.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46660 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Holger Freyther.

        Fix compiler warning.

        Initialize member variables in the correct order.

        * plugins/PluginView.cpp:
        (WebCore::PluginView::PluginView):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46659 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Adam Treat  <adam.treat@torchmobile.com>

        Reviewed by Holger Hans Peter Freyther.

        The reason these four tests fail is a mystery to me right, but placing
        them in the Skipped file and the Qt port is... for the first time ever,
        finally green!

        https://bugs.webkit.org/show_bug.cgi?id=27914

        * platform/qt/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46658 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Adam Treat  <adam.treat@torchmobile.com>

        Reviewed by Eric Seidel.

        Printing related test needs to go in Skipped for now so we can
        get Qt buildbot green for first time.

        https://bugs.webkit.org/show_bug.cgi?id=27913

        * platform/qt/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46657 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Adam Treat  <adam.treat@torchmobile.com>

         Reviewed by Eric Seidel.

         Sometimes they pass, sometimes they fail.  In attempt to get Qt buildbot
         completely green for first time I'm placing these two in the Skipped file.

        https://bugs.webkit.org/show_bug.cgi?id=27912

         * platform/qt/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46656 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Adam Treat  <adam.treat@torchmobile.com>

        Reviewed by George Staikos and Eric Seidel.

        Place some audio related tests into the Skipped list in
        attempt to get Qt buildbot finally green.

        https://bugs.webkit.org/show_bug.cgi?id=27910

        * platform/qt/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46655 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Jon Honeycutt  <jhoneycutt@apple.com>

        Windows build fix!

        Unreviewed.

        * page/win/PageWin.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46654 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Jon Honeycutt  <jhoneycutt@apple.com>

        Mac build fix.

        Unreviewed.

        * page/Page.cpp:
        Move setCanStartPlugins() from here...
        * page/win/PageWin.cpp:
        (WebCore::Page::setCanStartPlugins):
        ... to here.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46653 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  David Levin  <levin@chromium.org>

        Reviewed by Adam Treat.

        check-webkit-style --git-commit has bugs if you select a commit in the past
        https://bugs.webkit.org/show_bug.cgi?id=27908

        * Scripts/check-webkit-style:
        (main): Made the git-commit command only select the first commit
        if a range is given (and print a warning).
        * Scripts/modules/scm.py:
        (SCM::create_patch_from_local_commit):
        (SCM::create_patch_since_local_commit):
        (SCM::commit_locally_with_message):
        (SCM::discard_local_commits):
        Added an error messages in case these aren't overriden.

        (Git::create_patch_since_local_commit):
        Added this command to allow doing a diff of everything after a given commit.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46652 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Jan Michael Alonzo  <jmalonzo@webkit.org>

        Speculative Qt build fix. Not reviewed.

        Add LayoutTestController::disableImageLoading() in jsobjects.h.

        * DumpRenderTree/qt/jsobjects.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46651 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Jan Michael Alonzo  <jmalonzo@webkit.org>

        Skip http/tests/misc/favicon-loads-with-images-disabled.html as
        this requires resource load callbacks in Gtk DRT. I've raised bug
        #27905 to track this.

        * platform/gtk/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46650 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-07-29  Jon Honeycutt  <jhoneycutt@apple.com>

        <rdar://problem/5698113> Safari shouldn't auto-activate plug-ins in
        background tabs (make Win consistent with Mac)

        https://bugs.webkit.org/show_bug.cgi?id=27855

        Reviewed by Anders Carlsson.

        * loader/FrameLoaderClient.h:
        (WebCore::FrameLoaderClient::dispatchDidFailToStartPlugin):
        Declare a new function to dispatch failures to start plug-ins.

        * page/Page.cpp:
        (WebCore::Page::Page):
        Initialize new member variable.
        (WebCore::Page::addUnstartedPlugin):
        Add the PluginView to the set of unstarted plug-ins.
        (WebCore::Page::removeUnstartedPlugin):
        Remove the PluginView from the set of unstarted plug-ins.
        (WebCore::Page::setCanStartPlugins):
        If we can now start plug-ins, iterate the set of unstarted plug-ins,
        starting them. If a plug-in fails to start, dispatch a failed-to-start-
        plug-in error. Clear the list of unstarted plug-ins.

        * page/Page.h:
        Added m_canStartPlugins to track whether we can currently start
        plug-ins. Added m_unstartedPlugins to keep track of plug-ins in the
        page that are waiting to be started.
        (WebCore::Page::canStartPlugins):

        * plugins/PluginView.cpp:
        (WebCore::PluginView::start):
        m_mimeType is now a WebCore::String, so we call .utf8(). If NPP_New()
        fails, set m_status to reflect this; this used to be set by init(), but
        start() will not be called by init() if we are delaying start().
        If NPP_New() succeeds, set m_status to reflect this, and call
        platformStart().
        (WebCore::PluginView::startOrAddToUnstartedList):
        If we cannot start plug-ins, add this view to the Page's list of
        unstarted plug-ins and return true. Otherwise, call start() and return
        the result.
        (WebCore::PluginView::removeFromUnstartedListIfNecessary):
        If we've been started or we don't have a Page, return early. Remove
        ourselves from the Page's list of unstarted plug-ins.
        (WebCore::PluginView::bindingInstance):
        If we haven't been started yet, return 0. This matches the Mac
        behavior.
        (WebCore::PluginView::setParameters):
        If we find the "plug-ins page" parameter, record it; it may be used if
        we fail to start the plug-in.
        (WebCore::PluginView::PluginView):
        Initialize new members m_haveUpdatedPluginWidget. Initialize m_mimeType
        with mimeType directly; m_mimeType is now a WebCore::String.

        * plugins/PluginView.h:
        Made start() public, so Page can call it when we start unstarted
        plug-ins. Changed m_mimeType from CString to String and added
        m_pluginsPage; these are needed by dispatchDidFailToStartPlugin().
        (WebCore::PluginView::pluginsPage):
        Added for dispatchDidFailToStartPlugin().
        (WebCore::PluginView::mimeType):
        Ditto.
        (WebCore::PluginView::url):
        Ditto.

        * plugins/gtk/PluginViewGtk.cpp:
        (WebCore::PluginView::platformStart):
        Stubbed.

        * plugins/mac/PluginViewMac.cpp:
        (WebCore::PluginView::platformStart):
        Stubbed.

        * plugins/qt/PluginViewQt.cpp:
        (WebCore::PluginView::platformStart):
        Stubbed.

        * plugins/win/PluginViewWin.cpp:
        (WebCore::PluginView::updatePluginWidget):
        Use m_haveUpdatedPluginWidget to ensure that we will call MoveWindow()
        and SetWindowRgn() to position the plug-in and set its clip rect if an
        earlier call to updatePluginWidget() was made before
        setPlatformPluginWidget() was called to set the plug-in's HWND.
        (WebCore::PluginView::~PluginView):
        Call removeFromUnstartedListIfNecessary() to remove the plug-in from
        the unstarted list if the plug-in is destroyed before it is started.
        (WebCore::PluginView::init):
        Call startOrAddtoUnstartedList to start the plug-in or, if plug-ins
        cannot be started, to add it to the list of plug-ins waiting to be
        started. Set m_status to indicate success; this is also set by start(),
        but that may be delayed, and we don't want the loader to dispatch a
        failed-to-start-plug-in error when we return from init().
        (WebCore::PluginView::platformStart):
        Added; code moved from init(). This is now called by start() to perform
        any platform-specific tasks that need to take place after the plug-in
        has started. Added a call to updatePluginWidget() to ensure we position
        the plug-in and set its clip rect after the plug-in has started.
        Removed the setting of m_status; start() does this before calling this
        function (and we assert this at the head of the function).

WebKit/win:

2009-07-29  Jon Honeycutt  <jhoneycutt@apple.com>

        Part of <rdar://problem/5698113> Safari shouldn't auto-activate
        plug-ins in background tabs (make Win consistent with Mac)

        https://bugs.webkit.org/show_bug.cgi?id=27855

        Reviewed by Adam Roben and Anders Carlsson.

        * Interfaces/IWebViewPrivate.idl:
        Add setCanStartPlugins().

        * WebCoreSupport/WebFrameLoaderClient.cpp:
        (WebFrameLoaderClient::dispatchDidFailToStartPlugin):
        Added; code moved from createPlugin(). Modified to obtain the plug-in's
        page URL, the plug-ins "plug-ins page" parameter value, and the
        plug-in's MIME type by calling getter functions on the passed
        PluginView.
        (WebFrameLoaderClient::createPlugin):
        If the plug-in status is not "success," call
        dispatchDidFailToStartPlugin().

        * WebCoreSupport/WebFrameLoaderClient.h:
        Declare an override of dispatchDidFailToStartPlugin().

        * WebView.cpp:
        (WebViewWndProc):
        Add a missing ) to a comment.
        (WebView::setCanStartPlugins):
        Call the Page's setCanStartPlugins() function.

        * WebView.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46649 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fixes the Windows release-PGO build.

        Reviewed by Jon Honeycutt.

        * JavaScriptCore.vcproj/WTF/WTF.vcproj: Suppresses the warning about unreachable code that we get by adding "return 0" to WTF::TCMalloc_PageHeap::runScavengerThread().
        * wtf/FastMalloc.cpp:
        (WTF::TCMalloc_PageHeap::runScavengerThread): Fixes the error about the method not returning a value in the release-PGO build.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46648 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-07-30  Darin Adler  <darin@apple.com>

        Reviewed by David Levin.

        Use checked casts for render tree
        https://bugs.webkit.org/show_bug.cgi?id=23522

        Next step: Add casts for all the RenderTable classes and use them everywhere.

        Also added a few uses in places that were still using static_cast.
        Also made some virtual member functions private.
        Also changed RenderTable::m_tableLayout to use OwnPtr.

        * accessibility/AccessibilityTable.cpp:
        (WebCore::AccessibilityTable::isTableExposableThroughAccessibility):
        (WebCore::AccessibilityTable::addChildren):
        (WebCore::AccessibilityTable::cellForColumnAndRow):
        * accessibility/AccessibilityTableCell.cpp:
        (WebCore::AccessibilityTableCell::parentTable):
        (WebCore::AccessibilityTableCell::rowIndexRange):
        (WebCore::AccessibilityTableCell::columnIndexRange):
        (WebCore::AccessibilityTableCell::titleUIElement):
        * accessibility/AccessibilityTableColumn.cpp:
        (WebCore::AccessibilityTableColumn::headerObject):
        * accessibility/AccessibilityTableRow.cpp:
        (WebCore::AccessibilityTableRow::parentTable):
        * editing/DeleteSelectionCommand.cpp:
        (WebCore::DeleteSelectionCommand::removeNode):
        * editing/TextIterator.cpp:
        (WebCore::shouldEmitTabBeforeNode):
        (WebCore::shouldEmitNewlinesBeforeAndAfterNode):
        * html/HTMLTableCellElement.cpp:
        (WebCore::HTMLTableCellElement::parseMappedAttribute):
        * html/HTMLTableColElement.cpp:
        (WebCore::HTMLTableColElement::parseMappedAttribute):
        * page/Frame.cpp:
        (WebCore::Frame::searchForLabelsAboveCell):
        * page/mac/FrameMac.mm:
        (WebCore::Frame::searchForNSLabelsAboveCell):
        * rendering/AutoTableLayout.cpp:
        (WebCore::AutoTableLayout::recalcColumn):
        (WebCore::AutoTableLayout::fullRecalc):
        (WebCore::shouldScaleColumns):
        * rendering/FixedTableLayout.cpp:
        (WebCore::FixedTableLayout::calcWidthArray):
        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::calcPrefWidths):
        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::calcPercentageHeight):
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::addChild):
        (WebCore::RenderObject::containingBlock):
        * rendering/RenderTableCell.cpp:
        (WebCore::RenderTableCell::collapsedBottomBorder):
        * rendering/RenderTableCol.cpp:
        (WebCore::RenderTableCol::table):
        * rendering/RenderTableRow.cpp:
        (WebCore::RenderTableRow::addChild):
        (WebCore::RenderTableRow::layout):
        (WebCore::RenderTableRow::paint):
        * rendering/RenderTableSection.cpp:
        (WebCore::RenderTableSection::addChild):
        (WebCore::RenderTableSection::layoutRows):
        (WebCore::RenderTableSection::lowestPosition):
        (WebCore::RenderTableSection::rightmostPosition):
        (WebCore::RenderTableSection::leftmostPosition):
        (WebCore::RenderTableSection::paintObject):
        (WebCore::RenderTableSection::recalcCells):
        * rendering/RenderTreeAsText.cpp:
        (WebCore::operator<<):
        (WebCore::writeTextRun):
        Use checked casts.

        * rendering/RenderTable.cpp:
        (WebCore::RenderTable::styleDidChange): Updated to use OwnPtr.
        (WebCore::RenderTable::addChild): Use checked cast.
        (WebCore::RenderTable::layout): Ditto.
        (WebCore::RenderTable::setCellWidths): Ditto.
        (WebCore::RenderTable::paintObject): Ditto.
        (WebCore::RenderTable::splitColumn): Ditto.
        (WebCore::RenderTable::appendColumn): Ditto.
        (WebCore::RenderTable::colElement): Ditto.
        (WebCore::RenderTable::recalcSections): Ditto.
        (WebCore::RenderTable::outerBorderBottom): Ditto.
        (WebCore::RenderTable::outerBorderLeft): Ditto.
        (WebCore::RenderTable::outerBorderRight): Ditto.
        (WebCore::RenderTable::sectionAbove): Ditto.
        (WebCore::RenderTable::sectionBelow): Ditto.

        * rendering/RenderTable.h: Added checked cast. Made virtual
        functions private. Changed m_tableLayout to be a OwnPtr.

        * rendering/RenderTableCell.h: Added checked cast.
        * rendering/RenderTableCol.h: Ditto. Made virtual functions private.
        * rendering/RenderTableRow.h: Ditto.
        * rendering/RenderTableSection.h: Ditto.

WebKit/mac:

2009-07-30  Darin Adler  <darin@apple.com>

        Reviewed by David Levin.

        Use checked casts for render tree
        https://bugs.webkit.org/show_bug.cgi?id=23522

        * Misc/WebNSAttributedStringExtras.mm:
        (fileWrapperForElement):
        * Misc/WebNSPasteboardExtras.mm:
        (imageFromElement):
        (-[NSPasteboard _web_declareAndWriteDragImageForElement:URL:title:archive:source:]):
        * WebView/WebFrame.mm:
        (-[WebFrame _computePageRectsWithPrintWidthScaleFactor:printHeight:]):
        (-[WebFrame _accessibilityTree]):
        * WebView/WebRenderNode.mm:
        (copyRenderNode):
        Use checked casts.

WebKit/win:

2009-07-30  Darin Adler  <darin@apple.com>

        Reviewed by David Levin.

        Use checked casts for render tree
        https://bugs.webkit.org/show_bug.cgi?id=23522

        * DOMHTMLClasses.cpp:
        (DOMHTMLInputElement::isUserEdited):
        (DOMHTMLTextAreaElement::isUserEdited):
        Use checked casts.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46647 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-07-31  Brady Eidson  <beidson@apple.com>

        Reviewed by John Sullivan.

        <rdar://problem/6973106> and https://bugs.webkit.org/show_bug.cgi?id=27896
        Favicons are still loaded when automatic image loading is disabled.

        People who want to avoid loading images generally want to avoid loading all images.

        Test: http/tests/misc/favicon-loads-with-images-disabled.html

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::startIconLoader): After committing the URL mapping, don't actually
          perform the load if images shouldn't be loading.

WebKitTools:

2009-07-31  Brady Eidson  <beidson@apple.com>

        Reviewed by John Sullivan.

        <rdar://problem/6973106> and https://bugs.webkit.org/show_bug.cgi?id=27896
        Favicons are still loaded when automatic image loading is disabled.

        * DumpRenderTree/LayoutTestController.cpp:
        (disableImageLoadingCallback):
        (LayoutTestController::staticFunctions):
        * DumpRenderTree/LayoutTestController.h:

        * DumpRenderTree/mac/DumpRenderTree.mm:
        (resetWebViewToConsistentStateBeforeTesting):
        * DumpRenderTree/mac/LayoutTestControllerMac.mm:
        (LayoutTestController::disableImageLoading):

        * DumpRenderTree/win/DumpRenderTree.cpp:
        (resetWebViewToConsistentStateBeforeTesting):
        * DumpRenderTree/win/LayoutTestControllerWin.cpp:
        (LayoutTestController::disableImageLoading):

        * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
        (LayoutTestController::disableImageLoading): Stubbed for now.
        * DumpRenderTree/qt/jsobjects.cpp:
        (LayoutTestController::disableImageLoading): Ditto.

LayoutTests:

2009-07-31  Brady Eidson  <beidson@apple.com>

        Reviewed by John Sullivan.

        <rdar://problem/6973106> and https://bugs.webkit.org/show_bug.cgi?id=27896
        Favicons are still loaded when automatic image loading is disabled.

        * http/tests/misc/favicon-loads-with-images-disabled-expected.txt: Added.
        * http/tests/misc/favicon-loads-with-images-disabled.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46646 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Anders Carlsson.

        Accelerated animations stutter on pages with lots of animations and 3d transforms
        https://bugs.webkit.org/show_bug.cgi?id=27884

        This patch changes the strategy for synchronizing painting view the view,
        and compositing layer updates. Previously the strategy was to disable screen
        updates between the time we updated the layer tree, and painted the view. That
        left screen updates disabled for too long (hundreds of milliseconds) in some
        cases, causing animation stutter.

        The new strategy is to batch up changes to the CA layer tree, and commit them
        all at once just before painting happens (referred to as a "sync" in the code).
        GraphicsLayerCA now keeps a bitmask of changed properties, and then migrates
        the values stored in GraphicsLayer into the CA layer tree at commit time.

        Compositing layers are then synced in FrameView::paintContents(). However, not
        all style/layout changes will result in painting; to deal with style changes that
        touch only compositing properties, we set up a runloop observer that takes care
        of comitting layer changes when no painting happens.

        * WebCore.base.exp: Export FrameView::syncCompositingStateRecursive()

        * loader/EmptyClients.h: scheduleViewUpdate() renamed to syncCompositingStateRecursive()
        * page/ChromeClient.h: scheduleViewUpdate() renamed to syncCompositingStateRecursive()

        * page/FrameView.h:
        * page/FrameView.cpp:
        (WebCore::FrameView::syncCompositingStateRecursive): syncCompositingState() on the
        view and all subviews. Like layoutIfNeededRecursive(). If layout is pending, does not
        sync and returns false, since we only want to sync when layout is done.

        (WebCore::FrameView::paintContents): syncCompositingState() before painting.

        * page/animation/KeyframeAnimation.cpp:
        (WebCore::KeyframeAnimation::endAnimation):
        Call animationPaused() to notify the graphics layers about animation pausing.

        * platform/graphics/FloatPoint3D.h:
        (WebCore::operator==):
        (WebCore::operator!=):
        Add missing comparison operators.

        * platform/graphics/GraphicsLayer.cpp:
        (WebCore::GraphicsLayer::setOpacity):
        (WebCore::GraphicsLayer::setBackgroundColor): Simple setters no longer care about animation info.

        (WebCore::GraphicsLayer::paintGraphicsLayerContents): Null-check client.

        * platform/graphics/GraphicsLayer.h:
        (WebCore::AnimationValue:):
        (WebCore::TransformAnimationValue:):
        (WebCore::KeyframeValueList:):
        (WebCore::KeyframeValueList::insert):
        Cleaned up versions of FloatValue and TransformValue, used to store information
        about keyframes values.

        (WebCore::GraphicsLayer::contentsRect):
        (WebCore::GraphicsLayer::setContentsRect):
        ContentsRect is now a simple setter.

        (WebCore::GraphicsLayer::addAnimation):
        (WebCore::GraphicsLayer::removeAnimationsForProperty):
        (WebCore::GraphicsLayer::removeAnimationsForKeyframes):
        (WebCore::GraphicsLayer::pauseAnimation):
        Simplified animation api.

        (WebCore::GraphicsLayer::setGeometryOrientation):
        (WebCore::GraphicsLayer::geometryOrientation):
        setGeometryOrientation is now just a normal member variable.

        (WebCore::GraphicsLayer::contentsOrientation): add a getter.
        (WebCore::GraphicsLayer::syncCompositingState): Entry into the layer sync code.

        * platform/graphics/GraphicsLayerClient.h: scheduleViewUpdate() renamed to syncCompositingStateRecursive)

        * platform/graphics/mac/GraphicsLayerCA.h:
        * platform/graphics/mac/GraphicsLayerCA.mm:
        Lots of cleanup and refactoring. Main points:
        - Layer changes are all batched, and only committed to the CA layer on syncCompositingState().
        - Bitmask is used to store which properties have changed. More efficient than before.
        - Simpler animation interface; simple setters are no longer confounded with animations.
        - Refactored code that creates CA animations, and stores which animations are running.

        * platform/graphics/transforms/TransformOperations.h:
        (WebCore::TransformOperations::size):
        (WebCore::TransformOperations::at): Useful accessors for size and indexed access.

        * rendering/RenderLayerBacking.h:
        * rendering/RenderLayerBacking.cpp:
        Renamed 'contentsLayer' to 'foregroundLayer' to avoid confusion with GraphicsLayer's
        contentsLayer.
        Adapt to GraphicsLayer's simpler animation API.
        Pass animation pausing through to the graphics layers.
        contentsBox() is no longer a callback via GraphicsLayerClient.

        * rendering/RenderLayerCompositor.h:
        * rendering/RenderLayerCompositor.cpp:
        (WebCore::RenderLayerCompositor::setCompositingLayersNeedRebuild):
        (WebCore::RenderLayerCompositor::scheduleSync):
        (WebCore::RenderLayerCompositor::rebuildCompositingLayerTree):
        scheduleViewUpdate() is no longer required. Instead, we plumb through "compositingLayerSync"
        notifications, which travel up to WebKit and set up a runloop observer.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46645 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Eric Seidel  <eric@webkit.org>

        Reviewed by David Kilzer.

        bugzilla-tool post-commits trunk..head errors out
        https://bugs.webkit.org/show_bug.cgi?id=27847

        Rename commit_ids_from_range_arguments to commit_ids_from_commitish_arguments
        and make it handle the new arguments format proposed in the bug.

        * Scripts/bugzilla-tool:
        * Scripts/modules/scm.py:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46644 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Zoltan Horvath  <hzoltan@inf.u-szeged.hu>

        Change malloc to fastMalloc and free to fastFree in Yarr's RegexInterpreter.cpp
        https://bugs.webkit.org/show_bug.cgi?id=27874

        Use fastMalloc and fastFree instead of malloc and free in RegexInterpreter.cpp's methods.

        * yarr/RegexInterpreter.cpp:
        (JSC::Yarr::Interpreter::allocDisjunctionContext):
        (JSC::Yarr::Interpreter::freeDisjunctionContext):
        (JSC::Yarr::Interpreter::allocParenthesesDisjunctionContext):
        (JSC::Yarr::Interpreter::freeParenthesesDisjunctionContext):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46643 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Greg Bolsinga  <bolsinga@apple.com>

        Reviewed by Eric Seidel.

        Geolocation clean up when no longer updating
        https://bugs.webkit.org/show_bug.cgi?id=27888

        When Geolocation::disconnectFrame() is called, need to call
        Document::setUsingGeolocation(false) to mirror the true call
        when it is set up.

        When handling an error, and there are no more listeners, call
        GeolocationService::stopUpdating().

        * page/Geolocation.cpp:
        (WebCore::Geolocation::disconnectFrame):
        (WebCore::Geolocation::handleError):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46642 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Eric Seidel  <eric@webkit.org>

        Reviewed by David Levin.

        clean up build-webkit to use less copy/paste code and remove unneeded SVG warning
        https://bugs.webkit.org/show_bug.cgi?id=27857

        * Scripts/build-webkit:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46641 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    <webkit.org/b/27894> bugzilla-tool: Bugzilla.clear_attachment_review_flag() fails with ClientForm-0.2.10 and python-2.5

        Reviewed by Eric Seidel.

        * Scripts/modules/bugzilla.py:
        (Bugzilla.clear_attachment_review_flag): Specify which 'comment'
        form field to use when setting its value.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46640 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Jeremy Orlow  <jorlow@chromium.org>

        Reviewed by Adam Barth.

        Add new LocalStorage layout tests.
        https://bugs.webkit.org/show_bug.cgi?id=27778

        These test behavior that I don't think was very well verified earlier.

        * storage/domstorage/localstorage/complex-keys-expected.txt: Added.
        * storage/domstorage/localstorage/complex-keys.html: Added.
              A bunch of corner cases related to keys.
        * storage/domstorage/localstorage/complex-values-expected.txt: Added.
        * storage/domstorage/localstorage/complex-values.html: Added.
              A bunch of corner cases related to values.
        * storage/domstorage/localstorage/remove-item-expected.txt: Added.
        * storage/domstorage/localstorage/remove-item.html: Added.
              Test basic removeItem functionality.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46639 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Anders Carlsson  <andersca@apple.com>

        Reviewed by Dan Bernstein.

        -Fix <rdar://problem/7104097>

        Make EmbeddedWidget::create return a PassRefPtr so it will have the correct ref count.

        * WebCoreSupport/EmbeddedWidget.cpp:
        (EmbeddedWidget::create):
        * WebCoreSupport/EmbeddedWidget.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46638 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Eric Seidel  <eric@webkit.org>

        Reviewed by David Kilzer.

        bugzilla-tool post-commits obsoletes its own work :(
        https://bugs.webkit.org/show_bug.cgi?id=27849

        Simple fix.  Use .add() instead of .update()

        * Scripts/bugzilla-tool:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46637 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Jeremy Orlow  <jorlow@chromium.org>

        Reviewed by David Levin.

        Guard needs //'s between #endif and ENABLE(DATAGRID)
        https://bugs.webkit.org/show_bug.cgi?id=27862

        Compiling this file causes a build break without this change.

        * bindings/v8/custom/V8DataGridColumnListCustom.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46636 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Mike Fenton  <mike.fenton@torchmobile.com>

        Reviewed by Adam Treat.

        Re-Enable tests from Skipped that are passing correctly.

        https://bugs.webkit.org/show_bug.cgi?id=27890

        * platform/qt/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46635 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Mike Fenton  <mike.fenton@torchmobile.com>

        Reviewed by Adam Treat.

        Add LayoutTest results from BuildBot output for tests that currently lack results under Qt.

        https://bugs.webkit.org/show_bug.cgi?id=27882

        * platform/qt/fast/backgrounds/size/backgroundSize20-expected.txt: Copied from LayoutTests/platform/mac/fast/backgrounds/size/backgroundSize20-expected.txt.
        * platform/qt/fast/backgrounds/size/backgroundSize21-expected.txt: Copied from LayoutTests/platform/mac/fast/backgrounds/size/backgroundSize21-expected.txt.
        * platform/qt/fast/backgrounds/size/backgroundSize22-expected.txt: Copied from LayoutTests/platform/mac/fast/backgrounds/size/backgroundSize22-expected.txt.
        * platform/qt/fast/borders/border-radius-constraints-expected.txt: Added.
        * platform/qt/fast/box-shadow/inset-expected.txt: Added.
        * platform/qt/fast/box-shadow/spread-expected.txt: Copied from LayoutTests/platform/mac/fast/box-shadow/spread-expected.txt.
        * platform/qt/fast/canvas/image-pattern-rotate-expected.txt: Added.
        * platform/qt/fast/css/nested-floating-relative-position-percentages-expected.txt: Added.
        * platform/qt/fast/css/pseudo-cache-stale-expected.txt: Added.
        * platform/qt/fast/css/rem-dynamic-scaling-expected.txt: Added.
        * platform/qt/fast/css/rem-units-on-root-expected.txt: Added.
        * platform/qt/fast/dynamic/window-resize-scrollbars-test-expected.txt: Added.
        * platform/qt/fast/dynamic/window-scrollbars-test-expected.txt: Added.
        * platform/qt/fast/forms/isindex-placeholder-expected.txt: Added.
        * platform/qt/fast/forms/search-vertical-alignment-expected.txt: Added.
        * platform/qt/fast/frames/content-opacity-1-expected.txt: Copied from LayoutTests/platform/mac/fast/frames/content-opacity-1-expected.txt.
        * platform/qt/fast/frames/content-opacity-2-expected.txt: Copied from LayoutTests/platform/mac/fast/frames/content-opacity-2-expected.txt.
        * platform/qt/fast/frames/frame-navigation-expected.txt: Added.
        * platform/qt/fast/frames/frame-scrolling-attribute-expected.txt: Added.
        * platform/qt/fast/frames/iframe-scrolling-attribute-expected.txt: Added.
        * platform/qt/fast/gradients/crash-on-zero-radius-expected.txt: Added.
        * platform/qt/fast/lists/ol-start-parsing-expected.txt: Added.
        * platform/qt/fast/overflow/overflow-with-local-background-attachment-expected.txt: Added.
        * platform/qt/fast/repaint/float-new-in-block-expected.txt: Copied from LayoutTests/platform/mac/fast/repaint/float-new-in-block-expected.txt.
        * platform/qt/fast/text/international/khmer-selection-expected.txt: Added.
        * platform/qt/fast/text/shadow-translucent-fill-expected.txt: Added.
        * platform/qt/fast/text/text-letter-spacing-expected.txt: Added.
        * platform/qt/plugins/netscape-dom-access-expected.txt: Added.
        * platform/qt/svg/batik/filters/feTile-expected.txt: Added.
        * platform/qt/svg/custom/createImageElement-expected.txt: Copied from LayoutTests/platform/mac/svg/custom/createImageElement-expected.txt.
        * platform/qt/svg/custom/createImageElement2-expected.txt: Added.
        * platform/qt/svg/custom/js-update-image-and-display-expected.txt: Copied from LayoutTests/platform/mac/svg/custom/js-update-image-and-display-expected.txt.
        * platform/qt/svg/custom/js-update-image-and-display2-expected.txt: Copied from LayoutTests/platform/mac/svg/custom/js-update-image-and-display2-expected.txt.
        * platform/qt/svg/custom/js-update-image-and-display3-expected.txt: Copied from LayoutTests/platform/mac/svg/custom/js-update-image-and-display3-expected.txt.
        * platform/qt/svg/custom/outer-svg-unknown-feature-expected.txt: Copied from LayoutTests/platform/mac/svg/custom/outer-svg-unknown-feature-expected.txt.
        * platform/qt/svg/filters/feMerge-expected.txt: Added.
        * platform/qt/svg/filters/feOffset-expected.txt: Added.
        * platform/qt/svg/filters/feTile-expected.txt: Added.
        * platform/qt/svg/filters/subRegion-one-effect-expected.txt: Added.
        * platform/qt/svg/filters/subRegion-two-effects-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46634 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Greg Bolsinga  <bolsinga@apple.com>

        Reviewed by George Staikos.

        Consolidate GeoLocation code to send positions and errors
        https://bugs.webkit.org/show_bug.cgi?id=27863

        Create helper methods that will send positions and errors to either
        one shots or watchers.

        * page/Geolocation.cpp:
        (WebCore::Geolocation::sendError):
        (WebCore::Geolocation::sendErrorToOneShots):
        (WebCore::Geolocation::sendErrorToWatchers):
        (WebCore::Geolocation::sendPosition):
        (WebCore::Geolocation::sendPositionToOneShots):
        (WebCore::Geolocation::sendPositionToWatchers):
        * page/Geolocation.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46633 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Jakob Petsovits  <jakob.petsovits@torchmobile.com>

        Reviewed by Adam Treat.

        Build QtWebKit without benchmark tests for Qt < 4.5.
        https://bugs.webkit.org/show_bug.cgi?id=27886

        The QBENCHMARK macro is only available from 4.5 on,
        so builds for Qt < 4.5 shouldn't try to build those.

        * tests/tests.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46632 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>

        Reviewed by Anders Carlsson.

        Improve platform-detection in run-webkit-tests

        This prevents miss-detection of mac-* over Qt/Mac.

        * Scripts/run-webkit-tests:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46631 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Kenneth Rohde Christiansen  <kenneth@webkit.org>

        Reviewed by Adam Treat.

        Fix warning by migrating use of QString.sprintf(...) to
        QString.arg().arg()...

        * WebCoreSupport/EditorClientQt.cpp:
        (dumpRange):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46630 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Adam Treat  <adam.treat@torchmobile.com>

        Reviewed by Tor Arne Vestbo.

        Adding -webkit-animation-play-state back in following r46547.

        * platform/qt/fast/css/getComputedStyle/computed-style-expected.txt:
        * platform/qt/fast/css/getComputedStyle/computed-style-without-renderer-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46629 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Kenneth Rohde Christiansen  <kenneth@webkit.org>

        Reviewed by Tor Arne Vestbø.

        Expose the Offline Web Application Cache in the Qt API.

        * Api/qwebsettings.cpp:
        (QWebSettings::QWebSettings):
        (QWebSettings::setOfflineWebApplicationCachePath):
        (QWebSettings::offlineWebApplicationCachePath):
        (QWebSettings::setOfflineWebApplicationCacheQuota):
        (QWebSettings::offlineWebApplicationCacheQuota):
        * Api/qwebsettings.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46628 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Xan Lopez  <xlopez@igalia.com>

        Roll out previous change as it might be causing some weirdness in
        the bots.

        * platform/image-decoders/bmp/BMPImageDecoder.cpp:
        (WebCore::BMPImageDecoder::processFileHeader):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46627 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-31  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Mark Rowe.

        Fix compiler warning.
        https://bugs.webkit.org/show_bug.cgi?id=27851

        GCC does not like multi-character character constants, so use the
        explicit numerical value of 'BM' in the enum.

        * platform/image-decoders/bmp/BMPImageDecoder.cpp:
        (WebCore::BMPImageDecoder::processFileHeader):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46626 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the test case that was landed in r46624.

* http/tests/misc/meta-refresh-stray-single-quote.html:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46625 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-07-30  Brady Eidson  <beidson@apple.com>

        Reviewed by Mark Rowe, but Dan Bernstein also reviewed and asked thoughtful questions.

        <rdar://problem/7106968> and https://bugs.webkit.org/show_bug.cgi?id=27868
        http://www.ruthhuntcandy.com/ goes into infinite refresh in WebKit, works in others.

        Test: http/tests/misc/meta-refresh-stray-single-quote.html

        * platform/network/HTTPParsers.cpp:
        (WebCore::parseHTTPRefresh): Allow for a stray quote character at the start of the URL string.

LayoutTests:

2009-07-30  Brady Eidson  <beidson@apple.com>

        Reviewed by Mark Rowe, but Dan Bernstein also reviewed and asked thoughtful questions.

        <rdar://problem/7106968> and https://bugs.webkit.org/show_bug.cgi?id=27868
        http://www.ruthhuntcandy.com/ goes into infinite refresh in WebKit, works in others.

        * http/tests/misc/meta-refresh-stray-single-quote-expected.txt: Added.
        * http/tests/misc/meta-refresh-stray-single-quote.html: Added.
        * http/tests/misc/resources/success-notify-done.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46624 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Jan Alonzo.

        Fix compiler warning.

        GCC does not like C++-style comments in preprocessor directives.

        * wtf/Platform.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46623 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Martin Robinson  <mrobinson@appcelerator.com>

        Reviewed by Xan Lopez, Gustavo Noronha and Jan Alonzo.

        [GTK] new-window-policy-decision-requested provides no information about the target frame
        https://bugs.webkit.org/show_bug.cgi?id=27792

        Add frameName property to the WebKitWebNavigation object.

        * WebCoreSupport/FrameLoaderClientGtk.cpp:
        (WebKit::getNavigationAction):
        (WebKit::FrameLoaderClient::dispatchDecidePolicyForNewWindowAction):
        (WebKit::FrameLoaderClient::dispatchDecidePolicyForNavigationAction):
        * webkit/webkitwebnavigationaction.cpp:
        (webkit_web_navigation_action_get_property):
        (webkit_web_navigation_action_set_property):
        (webkit_web_navigation_action_class_init):
        (webkit_web_navigation_action_get_target_frame):
        (webkit_web_navigation_action_set_target_frame):
        * webkit/webkitwebnavigationaction.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46622 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  John McCall  <rjmccall@apple.com>

        Reviewed by Gavin Barraclough.

        Optimize the X86_64 trampolines: avoid the need for filler arguments
        and move the stub-args area closer to the stack pointer.

        * jit/JIT.h: adjust patch offsets because of slight code-size change
        * jit/JITCode.h:
        (JSC::JITCode::execute): don't pass filler args
        * jit/JITStubs.cpp:
        (ctiTrampoline): (X86_64): push args onto stack, use args directly
        (ctiVMThrowTrampoline): (X86_64): adjust %rsp by correct displacement
        (ctiOpThrowNotCaught): (X86_64): adjust %rsp by correct displacement
        * jit/JITStubs.h:
        (JITStackFrame): (X86_64): move args area earlier
        (ctiTrampoline): remove filler args from prototype

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46621 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Gavin Barraclough  <barraclough@apple.com>

        Reviewed by NOBODY (build fix).

        Temporarily revert r46618 since this is b0rking on Linux.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46620 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Teach buildbot to detect test cases that are missing results.

Reviewed by Cameron Zwarich.

* BuildSlaveSupport/build.webkit.org-config/master.cfg:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46619 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-23  Gavin Barraclough  <barraclough@apple.com>

        Reviewed by Oliver Hunt.

        Make get_by_id/put_by_id/method_check/call defer optimization using a data flag rather than a code modification.
        ( https://bugs.webkit.org/show_bug.cgi?id=27635 )

        This improves performance of ENABLE(ASSEMBLER_WX_EXCLUSIVE) builds by 2-2.5%, reducing the overhead to about 2.5%.
        (No performance impact with ASSEMBLER_WX_EXCLUSIVE disabled).

        * bytecode/CodeBlock.cpp:
        (JSC::printStructureStubInfo):
            - Make StructureStubInfo store the type as an integer, rather than an OpcodeID.

        * bytecode/CodeBlock.h:
        (JSC::):
        (JSC::CallLinkInfo::seenOnce):
        (JSC::CallLinkInfo::setSeen):
        (JSC::MethodCallLinkInfo::seenOnce):
        (JSC::MethodCallLinkInfo::setSeen):
            - Change a pointer in CallLinkInfo/MethodCallLinkInfo to use a PtrAndFlags, use a flag to track when an op has been executed once.

        * bytecode/StructureStubInfo.cpp:
        (JSC::StructureStubInfo::deref):
            - Make StructureStubInfo store the type as an integer, rather than an OpcodeID.

        * bytecode/StructureStubInfo.h:
        (JSC::StructureStubInfo::StructureStubInfo):
        (JSC::StructureStubInfo::initGetByIdSelf):
        (JSC::StructureStubInfo::initGetByIdProto):
        (JSC::StructureStubInfo::initGetByIdChain):
        (JSC::StructureStubInfo::initGetByIdSelfList):
        (JSC::StructureStubInfo::initGetByIdProtoList):
        (JSC::StructureStubInfo::initPutByIdTransition):
        (JSC::StructureStubInfo::initPutByIdReplace):
        (JSC::StructureStubInfo::seenOnce):
        (JSC::StructureStubInfo::setSeen):
            - Make StructureStubInfo store the type as an integer, rather than an OpcodeID, add a flag to track when an op has been executed once.

        * bytecompiler/BytecodeGenerator.cpp:
        (JSC::BytecodeGenerator::emitGetById):
        (JSC::BytecodeGenerator::emitPutById):
            - Make StructureStubInfo store the type as an integer, rather than an OpcodeID.

        * jit/JIT.cpp:
        (JSC::JIT::privateCompileCTIMachineTrampolines):
        (JSC::JIT::unlinkCall):
            - Remove the "don't lazy link" stage of calls.

        * jit/JIT.h:
        (JSC::JIT::compileCTIMachineTrampolines):
            - Remove the "don't lazy link" stage of calls.

        * jit/JITCall.cpp:
        (JSC::JIT::compileOpCallSlowCase):
            - Remove the "don't lazy link" stage of calls.

        * jit/JITStubs.cpp:
        (JSC::JITThunks::JITThunks):
        (JSC::JITThunks::tryCachePutByID):
        (JSC::JITThunks::tryCacheGetByID):
        (JSC::JITStubs::DEFINE_STUB_FUNCTION):
        (JSC::JITStubs::getPolymorphicAccessStructureListSlot):
            - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check.

        * jit/JITStubs.h:
        (JSC::JITThunks::ctiStringLengthTrampoline):
        (JSC::JITStubs::):
            - Remove the "don't lazy link" stage of calls, and the "_second" stage of get_by_id/put_by_id/method_check.

        * wtf/PtrAndFlags.h:
        (WTF::PtrAndFlags::PtrAndFlags):
        (WTF::PtrAndFlags::operator!):
        (WTF::PtrAndFlags::operator->):
            - Add ! and -> operators, add constuctor with pointer argument.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46618 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 27828: Title attribute is not respected on option elements
Fix <https://bugs.webkit.org/show_bug.cgi?id=27828> for Mac.

Reviewed by Adele Peterson and Jon Honeycutt.

No tests added as it is not clear how to test a tool tip from DumpRenderTree.

* platform/PopupMenuClient.h: Add a method for retrieving the tool tip of an item.
* platform/mac/PopupMenuMac.mm:
(WebCore::PopupMenu::populate): Set the tool tip of the menu item to that of the represented item.
* rendering/RenderMenuList.cpp:
(WebCore::RenderMenuList::itemToolTip): Expose the title attribute of the element as the tool tip.
* rendering/RenderMenuList.h:
* rendering/RenderTextControlSingleLine.h:
(WebCore::RenderTextControlSingleLine::itemToolTip): Return an empty string, indicating no tool tip, for
the popup menu in RenderTextControlSingleLine.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46617 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Things work better when you create the correct directories.

* Scripts/run-webkit-tests:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46616 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Include some information about the error when we die.

* Scripts/run-webkit-tests:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46615 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Adam Treat  <adam.treat@torchmobile.com>

        Add expected results file for new test.

        * platform/gtk/fast/forms/button-style-color-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46614 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Ensure that the path exists before writing results to it.
Hopefully this will make the Windows build bot happy.

* Scripts/run-webkit-tests:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46613 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Adam Treat  <adam.treat@torchmobile.com>

        Add expected results file for new test.

        * platform/mac/fast/forms/button-style-color-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46612 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Xiaomei Ji  <xji@chromium.org>

        Reviewed by Dan Bernstein.

        Remove ChromeClientChromium::setToolTip().
        https://bugs.webkit.org/show_bug.cgi?id=27861

        This patch is just to remove a temporarily introduced overloaded empty
        virtual function. No test is needed.

        * page/chromium/ChromeClientChromium.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46611 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Geoffrey Garen  <ggaren@apple.com>

        Reviewed by Gavin Barraclough.

        Fixed failing tests seen on Windows buildbot.

        * jit/JITStubs.cpp:
        (JSC::DEFINE_STUB_FUNCTION):
        * jit/JITStubs.h:
        (JSC::): Use "int" instead of "bool" to guarantee a 32-bit result,
        regardless of compiler. gcc on mac uses 32-bit values for bool,
        but gcc on linux and MSVC on Windows use 8-bit values.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46610 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Michael Nordman  <michaeln@google.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=27821

        ApplicationCacheHost refactoring.

        1) Better encapsulate the interfaces between webcore common code
        and the appcache system within a new class ApplicationCacheHost.

        2) Use that interface throughout the loader system, replacing inline appcache logic.

        3) Implement the interface in terms of webcore's appcache system.

        4) Add the new files to various makefiles.

        5) Implement protocolHostAndPortAreEqual() in KURLGoogle.cpp

        No new features, no new tests. The existing layout tests all pass.

        * GNUmakefile.am:
        * WebCore.base.exp:
        * WebCore.gypi:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:
        * html/HTMLHtmlElement.cpp:
        (WebCore::HTMLHtmlElement::insertedIntoDocument):
        * loader/DocumentLoader.cpp:
        (WebCore::DocumentLoader::DocumentLoader):
        (WebCore::DocumentLoader::~DocumentLoader):
        (WebCore::DocumentLoader::mainReceivedError):
        (WebCore::DocumentLoader::detachFromFrame):
        (WebCore::DocumentLoader::setPrimaryLoadComplete):
        * loader/DocumentLoader.h:
        (WebCore::DocumentLoader::applicationCacheHost):
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::canCachePageContainingThisFrame):
        (WebCore::FrameLoader::logCanCacheFrameDecision):
        (WebCore::FrameLoader::loadResourceSynchronously):
        * loader/MainResourceLoader.cpp:
        (WebCore::MainResourceLoader::didReceiveResponse):
        (WebCore::MainResourceLoader::didReceiveData):
        (WebCore::MainResourceLoader::didFinishLoading):
        (WebCore::MainResourceLoader::didFail):
        (WebCore::MainResourceLoader::load):
        * loader/MainResourceLoader.h:
        * loader/ResourceLoader.cpp:
        (WebCore::ResourceLoader::load):
        (WebCore::ResourceLoader::willSendRequest):
        (WebCore::ResourceLoader::didReceiveResponse):
        (WebCore::ResourceLoader::didFail):
        * loader/ResourceLoader.h:
        * loader/appcache/ApplicationCacheGroup.cpp:
        (WebCore::ApplicationCacheGroup::selectCache):
        (WebCore::ApplicationCacheGroup::selectCacheWithoutManifestURL):
        (WebCore::ApplicationCacheGroup::finishedLoadingMainResource):
        (WebCore::ApplicationCacheGroup::failedLoadingMainResource):
        (WebCore::ApplicationCacheGroup::disassociateDocumentLoader):
        (WebCore::ApplicationCacheGroup::update):
        (WebCore::ApplicationCacheGroup::didFinishLoadingManifest):
        (WebCore::ApplicationCacheGroup::manifestNotFound):
        (WebCore::ApplicationCacheGroup::checkIfLoadIsComplete):
        (WebCore::ApplicationCacheGroup::startLoadingEntry):
        (WebCore::ApplicationCacheGroup::associateDocumentLoaderWithCache):
        (WebCore::CallCacheListenerTask::create):
        (WebCore::CallCacheListenerTask::performTask):
        (WebCore::CallCacheListenerTask::CallCacheListenerTask):
        (WebCore::ApplicationCacheGroup::postListenerTask):
        * loader/appcache/ApplicationCacheGroup.h:
        * loader/appcache/ApplicationCacheHost.cpp: Added.
        * loader/appcache/ApplicationCacheHost.h: Added.
        * loader/appcache/ApplicationCacheStorage.cpp:
        (WebCore::ApplicationCacheStorage::transferApplicationCache):
        * loader/appcache/ApplicationCacheStorage.h:
        * loader/appcache/DOMApplicationCache.cpp:
        (WebCore::DOMApplicationCache::DOMApplicationCache):
        (WebCore::DOMApplicationCache::disconnectFrame):
        (WebCore::DOMApplicationCache::applicationCacheHost):
        (WebCore::DOMApplicationCache::status):
        (WebCore::DOMApplicationCache::update):
        (WebCore::DOMApplicationCache::swapCache):
        (WebCore::DOMApplicationCache::addEventListener):
        (WebCore::DOMApplicationCache::removeEventListener):
        (WebCore::DOMApplicationCache::dispatchEvent):
        (WebCore::DOMApplicationCache::callListener):
        (WebCore::DOMApplicationCache::toEventName):
        (WebCore::DOMApplicationCache::toEventType):
        * loader/appcache/DOMApplicationCache.h:
        (WebCore::DOMApplicationCache::):
        (WebCore::DOMApplicationCache::setAttributeEventListener):
        (WebCore::DOMApplicationCache::getAttributeEventListener):
        (WebCore::DOMApplicationCache::clearAttributeEventListener):
        (WebCore::DOMApplicationCache::callEventListener):
        (WebCore::DOMApplicationCache::setOnchecking):
        (WebCore::DOMApplicationCache::onchecking):
        (WebCore::DOMApplicationCache::setOnerror):
        (WebCore::DOMApplicationCache::onerror):
        (WebCore::DOMApplicationCache::setOnnoupdate):
        (WebCore::DOMApplicationCache::onnoupdate):
        (WebCore::DOMApplicationCache::setOndownloading):
        (WebCore::DOMApplicationCache::ondownloading):
        (WebCore::DOMApplicationCache::setOnprogress):
        (WebCore::DOMApplicationCache::onprogress):
        (WebCore::DOMApplicationCache::setOnupdateready):
        (WebCore::DOMApplicationCache::onupdateready):
        (WebCore::DOMApplicationCache::setOncached):
        (WebCore::DOMApplicationCache::oncached):
        (WebCore::DOMApplicationCache::setOnobsolete):
        (WebCore::DOMApplicationCache::onobsolete):
        (WebCore::DOMApplicationCache::~DOMApplicationCache):
        * platform/KURLGoogle.cpp:
        (WebCore::protocolHostAndPortAreEqual):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46609 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Adam Treat  <adam.treat@torchmobile.com>

        Reviewed by Mark Rowe.

        Revert previous commit to test out new functionality of buildbots and
        run-webkit-tests made possible by r46607.

        * platform/gtk/Skipped:
        * platform/mac/Skipped:
        * platform/win/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46608 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Teach run-webkit-tests to always store the actual results in the results directory for new tests.

Reviewed by Adam Treat and Jon Honeycutt.

This allows us to give a working link to the results from the result summary page, even if we're
not generating new results in to the tree.

* Scripts/run-webkit-tests:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46607 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Simon Fraser  <simon.fraser@apple.com>

        Minor change to earlier commit suggested by Darin Adler.
        Use the variable rather than dereferencing 'it' again.

        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::keyframeStylesForAnimation):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46606 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Adam Treat  <adam.treat@torchmobile.com>

        Reviewed by George Staikos.

        Add new test with platform specific results to the Skipped files of ports
        that haven't generated them yet.  This is to make the buildbot green again.

        * platform/gtk/Skipped:
        * platform/mac/Skipped:
        * platform/win/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46605 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 27854 - crash at WebCore::AXObjectCache::notificationPostTimerFired
https://bugs.webkit.org/show_bug.cgi?id=27854

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46604 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dan Bernstein.

        Animation with a timing function property in a keyframe eats CPU
        https://bugs.webkit.org/show_bug.cgi?id=27856
        <rdar://problem/7104476> Animation demo uses lots of CPU

        Don't include animation-timing-function in the list of properties to
        animate in a keyframe animation, because this property is not animated;
        instead, it describes the timing function to apply to this keyframe.

        This prevents the animation code from thinking that there's a property
        that it has to software-animate, and thus firing the animation timer frequently.

        Not testable because there is no visible impact.

        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::keyframeStylesForAnimation): Don't add
        CSSPropertyWebkitAnimationTimingFunction to the list of properties to animate.

        * page/animation/KeyframeAnimation.cpp
        (WebCore::KeyframeAnimation::getKeyframeAnimationInterval): Add a comment

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46603 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add a Qt Linux build bot.

* BuildSlaveSupport/build.webkit.org-config/config.json:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46602 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Geoffrey Garen  <ggaren@apple.com>

        Windows build fix: added missing symbols on Windows.

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46601 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:
2009-07-30  Mike Fenton  <mike.fenton@torchmobile.com>

         Reviewed by Adam Treat.

         Apply colour style to buttons that use Theme settings in Qt.

         https://bugs.webkit.org/show_bug.cgi?id=27814

         * platform/qt/RenderThemeQt.cpp:
         (WebCore::RenderThemeQt::adjustButtonStyle):

LayoutTests:
2009-07-30  Mike Fenton  <mike.fenton@torchmobile.com>

         Reviewed by Adam Treat.

         Add Button Colour style test and results for Qt based upon fix for bug.
         Unskip tables/mozilla/bugs/bug92647-2.html which passes with bug fix.
         Update expected results for qt for test fast/dom/HTMLTableColElement/resize-table-using-col-width-expected
         which also passes with the the bug fix.

         https://bugs.webkit.org/show_bug.cgi?id=27814

         * fast/forms/button-style-color.html: Added.
         * platform/qt/Skipped:
         * platform/qt/fast/dom/HTMLTableColElement/resize-table-using-col-width-expected.txt:
         * platform/qt/fast/forms/button-style-color-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46600 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Geoffrey Garen  <ggaren@apple.com>

        Windows build fix: removed stale symbols on Windows.

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore_debug.def:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46599 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Merged nitro-extreme branch into trunk.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46598 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Jakob Petsovits  <jakob.petsovits@torchmobile.com>

         Reviewed by Adam Roben.

         Add --minimal option to webkit-build.
         https://bugs.webkit.org/show_bug.cgi?id=27852

         This option disables all optional build features unless
         they are explicitly enabled.

         * Scripts/build-webkit:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46597 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Dean McNamee  <deanm@chromium.org>

        Reviewed by Dimitri Glazkov.

        Don't try to hash member function pointers, instead use a precomputed value based on the field.
        https://bugs.webkit.org/show_bug.cgi?id=27843

        * bindings/scripts/CodeGeneratorV8.pm:
        * bindings/v8/V8SVGPODTypeWrapper.h:
        (WebCore::PODTypeWrapperCacheInfo::PODTypeWrapperCacheInfo):
        (WebCore::PODTypeWrapperCacheInfo::operator==):
        (WebCore::PODTypeWrapperCacheInfoHash::hash):
        (WebCore::V8SVGDynamicPODTypeWrapperCache::lookupOrCreateWrapper):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46596 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-30  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Simon Fraser.

        Fix compiler warning.

        * dom/SelectElement.cpp:
        (WebCore::SelectElement::menuListDefaultEventHandler):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46595 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-07-29  Matt Perry  <mpcomplete@chromium.org>

        Reviewed by Adam Barth.

        Add a way to register V8 extensions for Isolated Worlds only.
        https://bugs.webkit.org/show_bug.cgi?id=27785

        * bindings/v8/ScriptController.cpp:
        (WebCore::ScriptController::evaluateInNewWorld):
        (WebCore::ScriptController::evaluateInNewContext):
        * bindings/v8/ScriptController.h:
        * bindings/v8/V8IsolatedWorld.cpp:
        (WebCore::V8IsolatedWorld::evaluate):
        * bindings/v8/V8IsolatedWorld.h:
        * bindings/v8/V8Proxy.cpp:
        (WebCore::V8Proxy::evaluateInNewWorld):
        (WebCore::V8Proxy::evaluateInNewContext):
        (WebCore::V8Proxy::createNewContext):
        (WebCore::V8Proxy::initContextIfNeeded):
        (WebCore::V8Proxy::registerExtensionWithV8):
        (WebCore::V8Proxy::registerExtension):
        * bindings/v8/V8Proxy.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@46594 268f45cc-cd09-0410-ab3c-d52691b4dbfc