<rdar://problem/5722735> Merge fix for SQLITE_FULL error
        is given even if the max_page_count is increased (2920)

        * libWebCoreSQLite3.a:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29980 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-02-04  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Darin.

        Blacklist Silverlight versions older than min required, not newer.

        * plugins/win/PluginPackageWin.cpp:
        (WebCore::PluginPackageWin::isPluginBlacklisted):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29979 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add layout test for selector parsing changes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29978 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add layout test for selector parsing changes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29977 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix for bug 16751, misparsing of html*.test in CSS.

        Reviewed by darin

        Added fast/css/simple-selector-chain-parsing.html

        * css/CSSGrammar.y:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29976 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Suggested by Geoff and Maciej.

        * bindings/js/JSCustomSQLTransactionCallback.cpp: Reworded a misleading
        comment to be correct.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29975 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-02-04  Jon Honeycutt  <jhoneycutt@apple.com>

        Reviewed by Steve, Anders.

        <rdar://problem/5211187> QuickTime and Flash plug-ins draw outside of
        content area when inside an iframe or div with overflow when playing a
        movie and scrolling the iframe/div area

        Clip the update region to the zero rect when scrolling. Don't do this
        for Java, because it results in repaint problems.

        * plugins/PluginQuirkSet.h: Added the DontClipToZeroRectWhenScrolling
        quirk
        (WebCore::):
        * plugins/win/PluginViewWin.cpp:
        (WebCore::PluginViewWin::updateWindow): Readded the old behavior of
        clipping to the zero rect when updating the window during a scroll.
        Added plug-in quirk to ignore this behavior for Java. Swapped order of
        the SetWindowRgn() and MoveWindow() calls to prevent Java from painting
        outside of its container during a scroll.
        (WebCore::PluginViewWin::determineQuirks): If this is Java, add the
        DontClipToZeroRectWhenScrolling quirk.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29974 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Mark Rowe.

        <rdar://problem/5722972> Leopard needs to statically link SQLite

        * Configurations/Base.xcconfig: Always have a header search path for WebCoreSQLite3.
        * Configurations/DebugRelease.xcconfig: Ditto.
        * Configurations/WebCore.xcconfig: Always link against WebCoreSQLite3.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29973 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    * platform/win/Skipped: Added one more failure.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29972 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fixed a typo

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29971 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    * platform/win/Skipped: Added some new failures.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29970 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Windows build fix

        * bindings/scripts/CodeGeneratorCOM.pm: Touched to force a rebuild of
        the COM bindings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29969 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Tim Hatcher.

        - fix <rdar://problem/5715481> REGRESSION (r26499): JavaScript document.lastModified is not supported

        Re-add attribute inadvertantly lost by r26499.

        Test: fast/js/lastModified.html: Added.

        * dom/Document.idl:
        Add lastModified.

LayoutTests:

        Reviewed by Tim Hatcher.

        Test for <rdar://problem/5715481> REGRESSION (r26499): JavaScript document.lastModified is not supported

        * fast/js/lastModified-expected.txt: Added.
        * fast/js/lastModified.html: Added.
        * fast/js/resources/lastModified.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29968 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix wrongly auto-merged change log.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29967 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix the search field if the inpsected Document has overridden
        Document.evaluate or Document.querySelectorAll

        Reviewed by Darin.

        * manual-tests/inspector-document-methods-override.html: Added.
        * page/InspectorController.cpp:
        (WebCore::InspectorController::didCommitLoad): Reworded a comment to
        force WebCore.vcproj to build.
        * page/inspector/inspector.js: Call
        Document.prototype.{evaluate,querySelectorAll}.call instead of calling
        the functions directly on the inspected Document. This ensures we are
        calling the version of these functions we intended to.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29966 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix the search field if the inpsected Document has overridden Document.evaluate or Document.querySelectorAll

        Reviewed by Darin.

        * manual-tests/inspector-document-methods-override.html: Added.
        * page/InspectorController.cpp:
        (WebCore::InspectorController::didCommitLoad): Reworded a comment to
        force WebCore.vcproj to build.
        * page/inspector/inspector.js: Call
        Document.prototype.{evaluate,querySelectorAll}.call instead of calling
        the functions directly on the inspected Document. This ensures we are
        calling the version of these functions we intended to.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29965 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Tim Hatcher.

        - possible fix for <rdar://problem/5714030> Crash in Database::deliverAllPendingCallbacks()
          reloading a page quickly

        I don't fully understand the cause of the crash, but I think this might
        be a helpful change.

        * platform/sql/SQLiteTransaction.cpp:
        (WebCore::SQLiteTransaction::commit): If the commit fails, don't leave this
        transaction and database both marked as "still in progress". As far as I can
        tell this does no good, and also seems to do harm.
        (WebCore::SQLiteTransaction::rollback): Ditto.

        * storage/Database.cpp:
        (WebCore::Database::performTransactionStep): Add some assertions to
        detect databases stuck in the "transaction in progress" state.
        * storage/SQLTransaction.cpp:
        (WebCore::SQLTransaction::openTransactionAndPreflight): Ditto.
        (WebCore::SQLTransaction::postflightAndCommit): Ditto.
        (WebCore::SQLTransaction::cleanupAfterTransactionErrorCallback): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29964 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Geoff.

        - fix <rdar://problem/5715692> REGRESSION (r28570): JavaScript window.scrollTo()
          calls no longer accept 'undefined' values

        By default, we should accept non-numeric parameters and non-integral numbers for
        parameters that expect integers, without throwing exceptions.

        While creating the test for this, I ran into a couple minor bugs with the
        functions involved, and this patch fixes those too.

        Test: fast/dom/non-numeric-values-numeric-parameters.html

        * bindings/js/JSHTMLOptionsCollectionCustom.cpp:
        (WebCore::JSHTMLOptionsCollection::add): Added. This function has unusual behavior
        when passed non-integral values for its second parameter, so it needs to be written
        by hand. I think that [Custom] is better here than inventing a new keyword.

        * bindings/scripts/CodeGeneratorJS.pm: Changed default for "long" and "unsigned long"
        to ignore errors rather than failing due to type differences. Also changed a couple
        functions to use hashes.

        * css/CSSStyleSheet.h:
        (WebCore::CSSStyleSheet::removeRule): Remove overload of removeRule without a second
        parameters. While we do allow this from JavaScript, it's not a true optional parameter,
        but rather just a case of "you can omit parameters and they are treated as undefined"
        combined with "undefined turns into 0 when passed to a function that takes an integer".
        * css/CSSStyleSheet.idl: Removed [Optional] on the index parameter for removeRule.
        This is not truly an optional parameter.

        * dom/ProgressEvent.cpp:
        (WebCore::ProgressEvent::initProgressEvent): Updated this function to match other
        DOM event init functions -- important to do nothing if this is called on the an
        already-dispatched event and we need to respect the bubble and cancelable arguments.
        Also removed initProgressEventNS. We don't support namespaced events, and if we add
        support, it should be across all event classes, not just ProgressEvent.
        * dom/ProgressEvent.h: Removed initProgressEventNS.
        * dom/ProgressEvent.idl: Ditto.

        * dom/Range.cpp:
        (WebCore::Range::createContextualFragment): Added a check for 0. This can happen
        if the passed-in start container is a node that's not an HTML element and also
        does not have a parent.

        * html/HTMLOptionsCollection.idl: Added the [Custom] attribute to add, since the
        rules for processing its parameters are unusual.

        * page/DOMSelection.cpp: Removed the version of setPosition that has only
        one parameter. The offset is not really optional.
        * page/DOMSelection.h: Ditto.
        * page/DOMSelection.idl: Removed the [Optional] keyword for the second parameter of
        setPosition. It's not a true optional parameter (see discussion of removeRule above).

        * page/DOMWindow.cpp:
        (WebCore::DOMWindow::adjustWindowRect): Simplified logic for constraining the X
        and Y coordinates, in a way that makes them work even when the window coordinates
        are infinite. Also strengthened the assertion.

WebKitTools:

        Reviewed by Geoff.

        * Scripts/make-js-test-wrappers: Look at the whole LayoutTests tree, not just the fast
        and svg subdirectories. Added some more exceptions for the benefit of the fast/dom
        subdirectory.

LayoutTests:

        Reviewed by Geoff.

        - test for <rdar://problem/5715692> REGRESSION (r28570): JavaScript window.scrollTo() calls
          no longer accept 'undefined' values

        * fast/dom/non-numeric-values-numeric-parameters-expected.txt: Added.
        * fast/dom/non-numeric-values-numeric-parameters.html: Added.
        * fast/dom/resources/TEMPLATE.html: Copied from fast/dom/Element/resources/TEMPLATE.html.
        * fast/dom/resources/non-numeric-values-numeric-parameters.js: Added.

        * editing/selection/extend-expected.txt: Updated.
        * editing/selection/extend.html: Removed test that extend throws if passed only one parameter.
        There's no reason for us to be strict about this -- we generally treat missing parameters the
        same as if "undefined" had been passed with few exceptions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29963 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Windows Apple-only build fix

        * DumpRenderTree/win/DumpRenderTree.vcproj: Added a new include
        directory.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29962 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-02-04  Alp Toker  <alp@atoker.com>

        Rubber-stamped by Mark Rowe.

        Remove all trailing whitespace in the GTK+ port and related
        components.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29961 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Set some PNGs' svn:mime-type property to image/png

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29960 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add an additional test for Canvas.transform with non-invertible matrix.

RS=Eric

* fast/canvas/canvas-transform-non-invertible-expected.checksum: Added.
* fast/canvas/canvas-transform-non-invertible-expected.png: Added.
* fast/canvas/canvas-transform-non-invertible-expected.txt: Added.
* fast/canvas/canvas-transform-non-invertible.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29959 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-02-03  Christian Dywan  <christian@imendio.com>

        Reviewed by Alp Toker.

        http://bugs.webkit.org/show_bug.cgi?id=17046
        [GTK] Context menu fixes and customisation suport

        Provide standard GTK+ context menu items where appropriate.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29958 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Typo fix. Don't use the provisionalDocumentLoader() after the whole page has already been loaded.
http://bugs.webkit.org/show_bug.cgi?id=17157

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29957 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 17169: Support transform on Canvas

Reviewed by Maciej.

Nice and simple patch as the cross-platform code to apply
a transform was already there.

Tests: fast/canvas/canvas-transform-identity.html
       fast/canvas/canvas-transform-infinity.html
       fast/canvas/canvas-transform-multiply.html
       fast/canvas/canvas-transform-nan.html
       fast/canvas/canvas-transform-skewed.html

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29956 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Eric.
Add missing layout test results.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29955 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Oliver.

Fix mistake in SVGImageElement, breaking "Dock" example of carto.net
It was comparing against the wrong attribute names in svgAttributeChanged().

Added testcase: svg/custom/js-update-image.svg

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29954 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by darin.

        Acid3 expects textNode.localName === null
        http://bugs.webkit.org/show_bug.cgi?id=17060

        Test: fast/dom/Node/initial-values.html

        * dom/Comment.cpp: remove localName implementation
        * dom/Comment.h:
        * dom/Node.cpp: return nullAtom instead of emptyAtom
        * dom/Text.cpp: remove localName implementation
        * dom/Text.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29953 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by darin.

        Make createElementNS and createAttributeNS follow the (vague) DOM Core 2 spec
        by throwing exceptions for more types of invalid qualified names.
        http://bugs.webkit.org/show_bug.cgi?id=16833

        Tests: fast/dom/Document/createAttributeNS-namespace-err.html
               fast/dom/Document/createElementNS-namespace-err.html

        * dom/Document.cpp:
        (WebCore::Document::createElement):
        (WebCore::hasNamespaceError):
        (WebCore::Document::createElementNS):
        (WebCore::Document::createAttributeNS):
        * dom/Document.idl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29952 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Eric.

Fixes: http://bugs.webkit.org/show_bug.cgi?id=15394
Dramatically improve dynamic update performance in DOM / SVG DOM.

The notifyAttributeChange() sledgehammer is gone now. It was implemented on quite a lot of
SVG*Element classes and blindly reacted on any property change caused by DOM / SVG DOM
by rebuilding style/renderer etc. without actually checking what changed. SVG used a hack
for years that attributeChanged() called notifyAttributeChange() - which results in poor
scripting performance and/or dynamic creation/modification of elements using SVG DOM.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29951 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Build fix for last wx commit (a couple things were not committed). Also fix some style issues.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29950 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Clean up overlooked coding guideline issues from last commit.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29949 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Widget needs to subclass wxWindow, not wxScrolledWindow for wx port, to fix several scrolling-related issues.
http://bugs.webkit.org/show_bug.cgi?id=17165

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29948 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - updated some more only-child and only-of-type tests to expect success
          (failures seen on buildbot)

        The tests were now succeeding, but the results and placement reflected failure.
        I'm not sure why we have such a tiny set of tests in the css3 directory.
        The expected results are updated as well as moved.

        * css3/css3-modsel-36.html: Copied from css3/expected_failures/css3-modsel-36.html.
        * css3/css3-modsel-37.html: Copied from css3/expected_failures/css3-modsel-37.html.

        * platform/mac/css3/css3-modsel-36-expected.checksum: Copied from platform/mac/css3/expected_failures/css3-modsel-36-expected.checksum.
        * platform/mac/css3/css3-modsel-36-expected.png: Copied from platform/mac/css3/expected_failures/css3-modsel-36-expected.png.
        * platform/mac/css3/css3-modsel-36-expected.txt: Copied from platform/mac/css3/expected_failures/css3-modsel-36-expected.txt.
        * platform/mac/css3/css3-modsel-37-expected.checksum: Copied from platform/mac/css3/expected_failures/css3-modsel-37-expected.checksum.
        * platform/mac/css3/css3-modsel-37-expected.png: Copied from platform/mac/css3/expected_failures/css3-modsel-37-expected.png.
        * platform/mac/css3/css3-modsel-37-expected.txt: Copied from platform/mac/css3/expected_failures/css3-modsel-37-expected.txt.
        * platform/qt/css3/css3-modsel-36-expected.txt: Copied from platform/qt/css3/expected_failures/css3-modsel-36-expected.txt.
        * platform/qt/css3/css3-modsel-37-expected.txt: Copied from platform/qt/css3/expected_failures/css3-modsel-37-expected.txt.

        * css3/expected_failures: Removed.
        * platform/mac/css3/expected_failures: Removed.
        * platform/qt/css3/expected_failures: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29947 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver.

        Update the selector benchmark/test now that WebKit supports :only-child and :last-child;
        Unfortunately, the javascript libraries don't appear to implement :*-of-type correctly, so that one remains disabled for now.

        * perf/slickspeed/header.html: Typo fix, and updated the note to clarify that WebKit is no longer the only implementation causing us to leave some selectors disabled.
        * perf/slickspeed/selectors.list: Added the new selectors

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29946 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add the test files. Oops.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29945 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Fix for bug 5468, support CSS3 :only-child and :only-of-type selectors.

        Reviewed by olliej

        Added fast/css/only-child-pseudo-class.html, fast/css/only-of-type-pseudo-class.html

        * css/CSSSelector.cpp:
        (WebCore::CSSSelector::extractPseudoType):
        * css/CSSSelector.h:
        (WebCore::CSSSelector::):
        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::checkOneSelector):

LayoutTests:

        Add test for bug 5468, supporting :only-child and :only-of-type in CSS3.

        Reviewed by olliej

        * platform/mac/fast/css/only-child-pseudo-class-expected.checksum: Added.
        * platform/mac/fast/css/only-child-pseudo-class-expected.png: Added.
        * platform/mac/fast/css/only-child-pseudo-class-expected.txt: Added.
        * platform/mac/fast/css/only-of-type-pseudo-class-expected.checksum: Added.
        * platform/mac/fast/css/only-of-type-pseudo-class-expected.png: Added.
        * platform/mac/fast/css/only-of-type-pseudo-class-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29944 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

        Reviewed by Geoff Garen.

        PLT speedup related to <rdar://problem/5659272> REGRESSION: PLT .4%
        slower due to r28884 (global variable symbol table optimization)

        Geoff's theory is that the slowdown was due to copying hash tables when
        putting things into the back/forward cache. If that's true, then this
        should fix the problem.

        (According to Geoff's measurements, in a PLT that exaggerates the
        importance of symbol table saving during cached page creation, this
        patch is a ~3X speedup in cached page creation, and a 9% speedup overall.)

        * JavaScriptCore.exp: Updated.

        * kjs/JSVariableObject.cpp:
        (KJS::JSVariableObject::saveLocalStorage): Updated for changes to SavedProperty,
        which has been revised to avoid initializing each SavedProperty twice when building
        the array. Store the property names too, so we don't have to store the symbol table
        separately. Do this by iterating the symbol table instead of the local storage vector.
        (KJS::JSVariableObject::restoreLocalStorage): Ditto. Restore the symbol table as
        well as the local storage vector.

        * kjs/JSVariableObject.h: Removed save/restoreSymbolTable and do that work inside
        save/restoreLocalStorage instead. Made restoreLocalStorage a non-const member function
        that takes a const reference to a SavedProperties object.

        * kjs/LocalStorage.h: Changed attributes to be unsigned instead of int to match
        other declarations of attributes elsewhere.

        * kjs/property_map.cpp:
        (KJS::SavedProperties::SavedProperties): Updated for data member name change.
        (KJS::PropertyMap::save): Updated for data member name change and to use the new
        inline init function instead of setting the fields directly. This allows us to
        skip initializing the SavedProperty objects when first allocating the array, and
        just do it when we're actually setting up the individual elements.
        (KJS::PropertyMap::restore): Updated for SavedProperty changes.

        * kjs/property_map.h: Changed SavedProperty from a struct to a class. Set it up so
        it does not get initialized at construction time to avoid initializing twice when
        creating an array of SavedProperty. Removed the m_ prefixes from the members of
        the SavedProperties struct. Generally we use m_ for class members and not struct.

WebCore:

        Reviewed by Geoff Garen.

        PLT speedup related to <rdar://problem/5659272> REGRESSION: PLT .4%
        slower due to r28884 (global variable symbol table optimization)

        * history/CachedPage.cpp:
        (WebCore::CachedPage::CachedPage): Removed saveSymbolTable call.
        (WebCore::CachedPage::restore): Removed restoreSymbolTable call.
        (WebCore::CachedPage::clear): Removed clear of m_windowSymbolTable.

        * history/CachedPage.h: Removed m_windowSymbolTable, since save/restoreLocalStorage
        now takes care of the symbol table. Also removed many unnecessary includes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29943 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by eseidel.

        Add an include for for <objidl.h> which isn't included
        by default with WIN32_LEAN_AND_MEAN.
        * platform/graphics/FontCache.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29942 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by darin.  Landed by eseidel.

        Add #define guards for WIN32_LEAN_AND_MEAN and _CRT_RAND_S.

        * kjs/config.h:
        * wtf/FastMalloc.cpp:
        * wtf/TCSpinLock.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29941 268f45cc-cd09-0410-ab3c-d52691b4dbfc

wxFont is reference counted and meant to be used as as a value object; when using wxFont*, ref-counting was not happening properly and the font object would be destroyed while still in use. Also, fix memory leak in wx's FrameData::clear() method.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29940 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - updated some more last-child and last-of-type tests to expect success
          (failures seen on buildbot)

        The tests were now succeeding, but the results and placement reflected failure.
        I'm not sure why we have such a tiny set of tests in the css3 directory.
        The expected results are updated as well as moved.

        * css3/css3-modsel-33.html: Copied from css3/expected_failures/css3-modsel-33.html.
        * css3/css3-modsel-35.html: Copied from css3/expected_failures/css3-modsel-35.html.
        * css3/expected_failures/css3-modsel-33.html: Removed.
        * css3/expected_failures/css3-modsel-35.html: Removed.

        * platform/mac/css3/css3-modsel-33-expected.checksum: Copied from platform/mac/css3/expected_failures/css3-modsel-33-expected.checksum.
        * platform/mac/css3/css3-modsel-33-expected.png: Copied from platform/mac/css3/expected_failures/css3-modsel-33-expected.png.
        * platform/mac/css3/css3-modsel-33-expected.txt: Copied from platform/mac/css3/expected_failures/css3-modsel-33-expected.txt.
        * platform/mac/css3/css3-modsel-35-expected.checksum: Copied from platform/mac/css3/expected_failures/css3-modsel-35-expected.checksum.
        * platform/mac/css3/css3-modsel-35-expected.png: Copied from platform/mac/css3/expected_failures/css3-modsel-35-expected.png.
        * platform/mac/css3/css3-modsel-35-expected.txt: Copied from platform/mac/css3/expected_failures/css3-modsel-35-expected.txt.
        * platform/mac/css3/expected_failures/css3-modsel-33-expected.checksum: Removed.
        * platform/mac/css3/expected_failures/css3-modsel-33-expected.png: Removed.
        * platform/mac/css3/expected_failures/css3-modsel-33-expected.txt: Removed.
        * platform/mac/css3/expected_failures/css3-modsel-35-expected.checksum: Removed.
        * platform/mac/css3/expected_failures/css3-modsel-35-expected.png: Removed.
        * platform/mac/css3/expected_failures/css3-modsel-35-expected.txt: Removed.

        * platform/qt/css3/css3-modsel-33-expected.txt: Copied from platform/qt/css3/expected_failures/css3-modsel-33-expected.txt.
        * platform/qt/css3/css3-modsel-35-expected.txt: Copied from platform/qt/css3/expected_failures/css3-modsel-35-expected.txt.
        * platform/qt/css3/expected_failures/css3-modsel-33-expected.txt: Removed.
        * platform/qt/css3/expected_failures/css3-modsel-35-expected.txt: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29939 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add WordPress plugin to give more control over blog comment settings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29938 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Don't use callback-based timers on Windows, they eat crashes and put the application in an undefined state (and since the crash was eaten, it's impossible to determine the original cause). This is a version of r28500 for the wx port.

Reviewed by Adam Roben.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29937 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Qt build fix.  Fix typo.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29936 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Blind Qt build fix. Add Template.h include.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29935 268f45cc-cd09-0410-ab3c-d52691b4dbfc

wx build fix - add missing include file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29934 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Fix for bug 4812. Support last-child and last-of-type CSS3 selectors.  Brings Acid3 score up to 68/100.

        Reviewed by olliej

        Added fast/css/last-child-pseudo-class.html, fast/css/last-of-type-pseudo-class.html

        * css/CSSGrammar.y:
        * css/CSSSelector.cpp:
        (WebCore::CSSSelector::extractPseudoType):
        * css/CSSSelector.h:
        (WebCore::CSSSelector::):
        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::checkOneSelector):

LayoutTests:

        Fix for bug 4812.  Support last-child and last-of-type (and make sure they are properly dynamic).

        Reviewed by olliej

        * fast/css/last-child-pseudo-class.html: Added.
        * fast/css/last-of-type-pseudo-class.html: Added.
        * platform/mac/fast/css/last-child-pseudo-class-expected.checksum: Added.
        * platform/mac/fast/css/last-child-pseudo-class-expected.png: Added.
        * platform/mac/fast/css/last-child-pseudo-class-expected.txt: Added.
        * platform/mac/fast/css/last-of-type-pseudo-class-expected.checksum: Added.
        * platform/mac/fast/css/last-of-type-pseudo-class-expected.png: Added.
        * platform/mac/fast/css/last-of-type-pseudo-class-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29933 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Make :first-child and :first-of-type properly dynamic when the DOM changes.  Brings the Acid3 score up
        to 66/100.

        Reviewed by olliej

        Added fast/css/first-child-pseudo-class.html, fast/css/first-of-type-pseudo-class.html, fast/css/empty-body-test.html

        * css/CSSGrammar.y:
        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::checkOneSelector):
        * dom/Element.cpp:
        (WebCore::Element::recalcStyle):
        (WebCore::Element::childrenChanged):
        * rendering/RenderStyle.cpp:
        (WebCore::RenderStyle::RenderStyle):
        * rendering/RenderStyle.h:
        (WebCore::RenderStyle::childrenAffectedByFirstChildRules):
        (WebCore::RenderStyle::setChildrenAffectedByFirstChildRules):
        (WebCore::RenderStyle::childrenAffectedByLastChildRules):
        (WebCore::RenderStyle::setChildrenAffectedByLastChildRules):
        (WebCore::RenderStyle::childrenAffectedByPositionalRules):
        (WebCore::RenderStyle::setChildrenAffectedByPositionalRules):
        (WebCore::RenderStyle::firstChildState):
        (WebCore::RenderStyle::setFirstChildState):
        (WebCore::RenderStyle::lastChildState):
        (WebCore::RenderStyle::setLastChildState):

LayoutTests:

        Added tests for dynamic :first-child support (Acid3).

        Reviewed by olliej

        * fast/css/empty-body-test.html: Added.
        * fast/css/first-child-pseudo-class.html: Added.
        * fast/css/first-of-type-pseudo-class.html: Added.
        * platform/mac/fast/css/empty-body-test-expected.checksum: Added.
        * platform/mac/fast/css/empty-body-test-expected.png: Added.
        * platform/mac/fast/css/empty-body-test-expected.txt: Added.
        * platform/mac/fast/css/first-child-pseudo-class-expected.checksum: Added.
        * platform/mac/fast/css/first-child-pseudo-class-expected.png: Added.
        * platform/mac/fast/css/first-child-pseudo-class-expected.txt: Added.
        * platform/mac/fast/css/first-of-type-pseudo-class-expected.checksum: Added.
        * platform/mac/fast/css/first-of-type-pseudo-class-expected.png: Added.
        * platform/mac/fast/css/first-of-type-pseudo-class-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29932 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - added Tiger-only results for a test

        * platform/mac-leopard/fast/text/cg-fallback-bolding-expected.txt: Copied from LayoutTests/platform/mac/fast/text/cg-fallback-bolding-expected.txt.
        * platform/mac/fast/text/cg-fallback-bolding-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29931 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Oliver Hunt.

        - fix <rdar://problem/5720637> Missing characters in right-to-left complex text where different fonts are used in the same run

        * platform/graphics/win/UniscribeController.cpp:
        (WebCore::UniscribeController::advance): Fixed an off-by-1 error in the
        start offset of sub-runs of RTL runs. Changed to update
        m_currentCharacter to the first character of the sub-run before calling
        itemizeShapeAndPlace. In RTL runs, m_currentCharacter decreases as
        sub-runs are processed from left to right.
        (WebCore::UniscribeController::itemizeShapeAndPlace): Removed the
        updating of m_currentCharacter because advance() does it now.

LayoutTests:

        Reviewed by Oliver Hunt.

        - changed a test to use the same fonts on Windows as on Mac OS X

        * fast/text/atsui-spacing-features.html: Changed to use Lucida Grande
        for Hebrew.
        * platform/win/Skipped: Re-categorized the above test. It is still
        failing on Windows, but due to a different bug.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29930 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin's rubberstamp

        Move the SecurityOriginHash out into a separate header as it will soon be used in different places

        * WebCore.xcodeproj/project.pbxproj:
        * platform/SecurityOriginHash.h: Added.
        (WebCore::SecurityOriginHash::hash):
        (WebCore::SecurityOriginHash::equal):
        (WebCore::SecurityOriginTraits::deletedValue):
        (WebCore::SecurityOriginTraits::emptyValue):
        * storage/DatabaseTracker.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29929 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Dave Hyatt.

        - address the remaining tests listed under <rdar://problem/5075790> Layout tests that depend on certain fonts don't match Mac results

        * fast/text/cg-fallback-bolding.html: Changed to use Ahem instead of
        Geneva.
        * fast/text/international/bidi-fallback-font-weight.html: Made Mac-only.
        * platform/mac/fast/text/cg-fallback-bolding-expected.checksum: Updated.
        * platform/mac/fast/text/cg-fallback-bolding-expected.png: Updated
        * platform/mac/fast/text/cg-fallback-bolding-expected.txt: Updated.
        * platform/mac/fast/text/international/bidi-fallback-font-weight.html: Copied from LayoutTests/fast/text/international/bidi-fallback-font-weight.html.
        * platform/win/Skipped: Removed passing tests and categorized remaining
        failures.
        * platform/win/svg/W3C-SVG-1.1: Added.
        * platform/win/svg/W3C-SVG-1.1/resources: Added.
        * platform/win/svg/W3C-SVG-1.1/resources/Mac-compatible-font-fallback.css: Added.
        * platform/win/svg/W3C-SVG-1.1/resources/run-webkit-tests-epilogue.html: Added.
        * platform/win/svg/W3C-SVG-1.1/resources/run-webkit-tests-prologue.html: Added.
        * svg/text/text-tselect-02-f.svg: Added @font-face rule to use Lucida
        Grande for Hebrew.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29928 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Tim.

        - fixed mistake in the JSLock fix that caused an assertion every time

        * bindings/js/JSCustomSQLTransactionCallback.cpp:
        (WebCore::JSCustomSQLTransactionCallback::JSCustomSQLTransactionCallback): Added JSLock.
        (WebCore::unprotectOnMainThread): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29927 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Antti and Darin.

        Change Text::createWithLengthLimit to take a UChar pointer instead of a string. This
        lets us avoid making a copy of the buffer in TextDocument.cpp.

        * dom/Text.cpp:
        (WebCore::Text::createWithLengthLimit):
        * dom/Text.h:
        * html/HTMLParser.cpp:
        (WebCore::HTMLParser::parseToken):
        * loader/TextDocument.cpp:
        (WebCore::TextTokenizer::write):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29926 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        * platform/network/mac/ResourceHandleMac.mm:
        (WebCore::ResourceHandle::loadsBlocked): Fix grammar and word typo.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29925 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam.

        - fix <rdar://problem/4527931> showModalDialog calls from onload functions fail (Aspect)

        No automated regression test because showModalDialog doesn't work in DumpRenderTree.

        * platform/network/mac/ResourceHandleMac.mm:
        (WebCore::CallbackGuard::CallbackGuard): Added.
        (WebCore::CallbackGuard::~CallbackGuard): Added.
        (WebCore::ResourceHandle::supportsBufferedData): Simplified to take advantage of how
        static initialization works in C++.
        (WebCore::ResourceHandle::loadsBlocked): Always return false on Leopard. The problem in
        NSURLConnection that created the need to block loads is fixed in Leoaprd. This is the
        bug fix.
        (-[WebCoreResourceHandleAsDelegate connection:willSendRequest:redirectResponse:]):
        Use CallbackGuard instead of directly incrementing the count; allows us to omit the code
        entirely on Leopard.
        (-[WebCoreResourceHandleAsDelegate connection:didReceiveAuthenticationChallenge:]):
        Ditto.
        (-[WebCoreResourceHandleAsDelegate connection:didCancelAuthenticationChallenge:]):
        Ditto.
        (-[WebCoreResourceHandleAsDelegate connection:didReceiveResponse:]):
        Ditto.
        (-[WebCoreResourceHandleAsDelegate connection:didReceiveData:lengthReceived:]):
        Ditto.
        (-[WebCoreResourceHandleAsDelegate connection:willStopBufferingData:]):
        Ditto.
        (-[WebCoreResourceHandleAsDelegate connectionDidFinishLoading:]):
        Ditto.
        (-[WebCoreResourceHandleAsDelegate connection:didFailWithError:]):
        Ditto.
        (-[WebCoreResourceHandleAsDelegate connection:willCacheResponse:]):
        Ditto. Fixes a problem where this could leave inNSURLConnectionCallback set
        permanently in one of the code paths; this would break showModalDialog from that
        point on. This problem doesn't affect Safari.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29924 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    * platform/network/mac/ResourceHandleMac.mm:
        (WebCore::ResourceHandle::loadsBlocked): Roll out accidentally checked in file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29923 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Geoff and Brady.

        - fix <rdar://problem/5680469> Each database worker thread takes a JSLock,
          which slows down all JavaScript execution until the thread terminates

        * bindings/js/JSCustomSQLTransactionCallback.cpp:
        (WebCore::JSCustomSQLTransactionCallback::JSCustomSQLTransactionCallback):
        Explicitly gcProtect, since we aren't using ProtectedPtr any more.
        (WebCore::unprotectOnMainThread): Added.
        (WebCore::JSCustomSQLTransactionCallback::~JSCustomSQLTransactionCallback):
        Use unprotectOnMainThread rather than doing a gcUnprotect here.
        * bindings/js/JSCustomSQLTransactionCallback.h: Changed m_callback from
        a ProtectedPtr to a plain old JSObject*; we gcProtect by hand now.

        * loader/icon/IconDatabase.cpp:
        (WebCore::IconDatabase::notifyPendingLoadDecisionsOnMainThread):
        Renamed. Changed to takea a void* with the icon database pointer.
        (WebCore::IconDatabase::notifyPendingLoadDecisions): Renamed.
        (WebCore::IconDatabase::performURLImport): Updated name, and pass the
        icon database pointer in.
        * loader/icon/IconDatabase.h: Renamed and added void* parameter.

        * platform/Threading.h: Changed callOnMainThread to take a function with a
        single parameter rather than a function with no parameters. Added a typedef.

        * platform/gtk/ThreadingGtk.cpp:
        (WebCore::callFunctionOnMainThread): Changed to use a structure with both
        a function pointer and a context pointer, and delete it after calling.
        (WebCore::callOnMainThread): Changed to create the structure with both
        the function and context pointer.

        * platform/mac/Threading.mm:
        (-[WebCoreFunctionWrapper initWithFunction:context:]): Added context.
        (-[WebCoreFunctionWrapper invoke]): Updated to do both function and context.
        Renamed from _call.
        (WebCore::callOnMainThread): Updated to take both function and context.

        * platform/qt/ThreadingQt.cpp:
        (WebCore::PerformFunctionEvent::PerformFunctionEvent): Added context.
        (WebCore::PerformFunctionEvent::invoke): Added. Calls function with context pointer.
        (WebCore::MainThreadInvoker::event): Changed to call invoke.
        (WebCore::callOnMainThread): Updated to take both function and context.

        * platform/win/ThreadingWin.cpp: Added FunctionWithContext.
        (WebCore::callFunctionsOnMainThread): Changed to pass call function with context.
        (WebCore::callOnMainThread): Updated to queue both function and context.

        * platform/wx/ThreadingWx.cpp:
        (WebCore::callOnMainThread): Updated to take both function and context.

        * storage/Database.cpp:
        (WebCore::Database::scheduleTransactionCallback): Pass 0 for context.
        (WebCore::Database::deliverAllPendingCallbacks): Changed to take ignored context pointer.
        * storage/Database.h: Added void* parameter to deliverAllPendingCallbacks.

        * storage/DatabaseTracker.cpp:
        (WebCore::DatabaseTracker::scheduleForNotification): Pass 0 for context.
        (WebCore::DatabaseTracker::notifyDatabasesChanged): Changed to take ignored context pointer.
        * storage/DatabaseTracker.h: Added void* parameter to notifyDatabasesChanged.

WebKit/win:

        - fix <rdar://problem/5680469> Each database worker thread takes a JSLock,
          which slows down all JavaScript execution until the thread terminates

        * WebIconDatabase.cpp:
        (WebIconDatabase::scheduleNotificationDelivery): Updated for new callOnMainThread.
        (WebIconDatabase::deliverNotifications): Ditto.
        * WebIconDatabase.h: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29922 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-02-01  Jon Honeycutt  <jhoneycutt@apple.com>

        Rubber-stamped by Ollie.

        * plugins/win/PluginPackageWin.cpp:
        (WebCore::PluginPackageWin::load): Fix two function pointers

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29921 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Alexey Proskuryakov.

        - change fast/xsl/xslt-text to dump as text

        * fast/xsl/resources/xslt-text.xml: Copied from LayoutTests/fast/xsl/xslt-text.xml.
        * fast/xsl/resources/xslt-text.xsl: Copied from LayoutTests/fast/xsl/xslt-text.xsl.
        * fast/xsl/xslt-processor.html:
        * fast/xsl/xslt-text-expected.txt: Added.
        * fast/xsl/xslt-text.html: Added.
        * fast/xsl/xslt-text.xml: Removed.
        * fast/xsl/xslt-text.xsl: Removed.
        * platform/mac/fast/xsl/xslt-text-expected.checksum: Removed.
        * platform/mac/fast/xsl/xslt-text-expected.png: Removed.
        * platform/mac/fast/xsl/xslt-text-expected.txt: Removed.
        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29920 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by John Sullivan, Oliver Hunt.

        Fixed <rdar://problem/5688039> REGRESSION (r29428): Weather widget
        fails to load due to reliance on "var location" quirk

        Added a Dashboard quirk.

        * bindings/js/kjs_window.cpp:
        (KJS::Window::put): In DB backwards compatibility mode, make assignment
        to window.location in top-level windows shadow the DOM location API
        instead of performing a navigation. This makes a reasonable amount of
        sense, because widgets can't navigate their top-level frames, anyway.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29919 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Fix for bug 11387, CSS3 :empty selector is not dynamic.

        Add support for dynamically updating elements whose style is affected by :empty when the child count changes.

        Reviewed by olliej, mitzpettel

        fast/css/empty-pseudo-class.html

        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::locateSharedStyle):
        (WebCore::CSSStyleSelector::checkOneSelector):
        * dom/Element.cpp:
        (WebCore::Element::childrenChanged):
        * dom/Element.h:
        * rendering/RenderStyle.h:
        (WebCore::RenderStyle::affectedByEmpty):
        (WebCore::RenderStyle::emptyState):
        (WebCore::RenderStyle::setEmptyState):

LayoutTests:

        Add a new layout test for dynamic :empty selectors.

        Reviewed by olliej

        * fast/css/empty-pseudo-class.html: Added.
        * platform/mac/fast/css/empty-pseudo-class-expected.checksum: Added.
        * platform/mac/fast/css/empty-pseudo-class-expected.png: Added.
        * platform/mac/fast/css/empty-pseudo-class-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29918 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix typo.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29917 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    <rdar://problem/5717523> Don't set DEP opt-in flag (data execution prevention) since it is incompaible with the video plugin used on CNN.com

        Rubber-stamped by John Honeycutt.

        * win/tools/vsprops/common.vsprops:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29916 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        - switch line breaking from using Carbon Unicode Utilities to using ICU
          on Leopard
          <http://bugs.webkit.org/show_bug.cgi?id=4628>

        fast/text/international/thai-line-breaks.html results do not change.

        * rendering/break_lines.cpp:
        (WebCore::nextBreakablePosition): Changed to use Carbon only on Tiger.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29915 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-02-01  Rodney Dawes  <dobey@wayofthemonkey.com>

        Reviewed by Darin.

        Update npfunctions.h to export the proper function signatures for UNIX and add
        the appropriate #if around the Windows versions

        * plugins/npfunctions.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29914 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver.

        Add getSVGDocument to frame and iframe.

        Test: svg/custom/frame-getSVGDocument.html

        * html/HTMLEmbedElement.cpp:
        * html/HTMLEmbedElement.h:
        * html/HTMLFrameElement.idl:
        * html/HTMLFrameOwnerElement.cpp:
        (WebCore::HTMLFrameOwnerElement::getSVGDocument):
        * html/HTMLFrameOwnerElement.h:
        * html/HTMLIFrameElement.idl:
        * html/HTMLObjectElement.cpp:
        * html/HTMLObjectElement.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29913 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29912 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by darin.

        Beat CSSParser with the RefPtr stick.  Hopefully squashing any leaks in the process.
        (Deploy RefPtr and PassRefPtr throughout CSSParser)
        http://bugs.webkit.org/show_bug.cgi?id=17108

        No functional changes, thus no tests.

        * css/CSSParser.cpp:
        (WebCore::CSSParser::parseColor):
        (WebCore::CSSParser::parseValue):
        (WebCore::CSSParser::addBackgroundValue):
        (WebCore::CSSParser::parseBackgroundShorthand):
        (WebCore::CSSParser::addTransitionValue):
        (WebCore::CSSParser::parseTransitionShorthand):
        (WebCore::CSSParser::parseBackgroundPosition):
        (WebCore::CSSParser::parseBackgroundProperty):
        (WebCore::CSSParser::parseTransitionProperty):
        (WebCore::CSSParser::parseFontFaceSrc):
        (WebCore::ShadowParseContext::ShadowParseContext):
        (WebCore::CSSParser::parseShadow):
        (WebCore::BorderImageParseContext::BorderImageParseContext):
        (WebCore::BorderImageParseContext::commitBorderImage):
        * css/CSSParser.h:
        * css/SVGCSSParser.cpp:
        (WebCore::CSSParser::parseSVGValue):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29910 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-01-31  Alp Toker  <alp@atoker.com>

        Reviewed by Mark Rowe.

        Fix the pkg-config file to follow GTK+ package naming and versioning
        conventions.

        Remove unneeded dependency listings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29909 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add platform/Locker.h to vcproj

RS=Brady

* WebCore.vcproj/WebCore.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29908 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Build fix

        * WebCore.xcodeproj/project.pbxproj: Header -> WebKit

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29907 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Maciej

        Abstracted the concept of a "Locker" as upcoming work will rely on it.

        * platform/Locker.h: Added.  Template class to "lock()" and "unlock()" some arbitrary object
        (WebCore::Locker::Locker): Lock the object
        (WebCore::Locker::~Locker): Unlock it
        * platform/Threading.h: `class MutexLocker` is now `typedef Locker<Mutex>`
        * WebCore.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29906 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        <rdar://problem/5716912> REGRESSION (r29816): Can't navigate back to
        Gmail Inbox from Gmail Compose page
        Rolling out the change that caused the regression.

        * WebCore.base.exp:
        * bindings/js/JSHTMLDocumentCustom.cpp:
        (WebCore::JSHTMLDocument::open):
        * dom/DOMImplementation.cpp:
        (WebCore::DOMImplementation::createHTMLDocument):
        * dom/Document.cpp:
        (WebCore::Document::recalcStyle):
        (WebCore::Document::open):
        (WebCore::Document::write):
        (WebCore::Document::clear):
        * dom/Document.h:
        * history/HistoryItem.cpp:
        (WebCore::HistoryItem::HistoryItem):
        * history/HistoryItem.h:
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::didExplicitOpen):
        (WebCore::FrameLoader::load):
        (WebCore::FrameLoader::reloadAllowingStaleData):
        (WebCore::FrameLoader::reload):
        (WebCore::FrameLoader::tokenizerProcessedData):
        (WebCore::FrameLoader::post):
        (WebCore::FrameLoader::continueLoadAfterNewWindowPolicy):
        (WebCore::FrameLoader::shouldTreatURLAsSameAsCurrent):
        (WebCore::FrameLoader::loadItem):
        * loader/FrameLoader.h:
        * platform/text/CharacterNames.h:
        * xml/DOMParser.cpp:
        (WebCore::DOMParser::parseFromString):
        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::getResponseXML):
        * xml/XSLTProcessor.cpp:
        (WebCore::XSLTProcessor::createDocumentFromSource):

LayoutTests:

        <rdar://problem/5716912> REGRESSION (r29816): Can't navigate back to
        Gmail Inbox from Gmail Compose page
        Rolling out the change that caused the regression.

        * http/tests/navigation/document-open-adds-history-item-expected.txt: Removed.
        * http/tests/navigation/document-open-adds-history-item.html: Removed.
        * http/tests/navigation/document-open-delayed-adds-history-item-expected.txt: Removed.
        * http/tests/navigation/document-open-delayed-adds-history-item.html: Removed.
        * http/tests/navigation/document-open-new-window-adds-history-item-expected.txt: Removed.
        * http/tests/navigation/document-open-new-window-adds-history-item.html: Removed.
        * http/tests/navigation/document-open-replace-no-history-item-expected.txt: Removed.
        * http/tests/navigation/document-open-replace-no-history-item.html: Removed.
        * http/tests/navigation/resources/document-open-page-2.html: Removed.
        * http/tests/navigation/resources/document-open.js: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29905 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - re-categorize three skipped tests

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29904 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix <rdar://5713172> Focus lost from Gmail message body after toggling
        tabs

        This regressed in r29583. Prior to r29583, the code path that resulted
        in calling focusController->setFocusedFrame(0) was never getting hit.
        r29583 changed the code to adhere to the comment above the code and
        actually call setFocusedFrame(0) when losing focus to another part of
        the window. It turns out this is wrong behavior (and not what
        WebKit/mac does). The only time we should clear the focused frame is
        when we tab out of the WebView, and that's already taken care of by
        FocusController::advanceFocus. The code to clear the focused frame in
        WebView was written before FocusController::advanceFocus, so it was
        correct at one point but no longer is.

        We now never call setFocusedFrame(0) when the WebView receives a
        WM_KILLFOCUS message. Instead we just call setFocused(false) on the
        focused frame so that blur events will be sent. This matches what
        WebKit/mac does.

        Reviewed by Darin and Adele.

        * WebView.cpp:
        (WebViewWndProc): Also reworded a couple of comments to make them
        clearer.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29903 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Tim Hatcher.

        Bug 14960: Cannot copy text in Web Inspector Network panel
        http://bugs.webkit.org/show_bug.cgi?id=14960

        * page/inspector/inspector.css:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29902 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam Roben.

        Bug 14514: Add full URL view on mouseOver in the network page
        http://bugs.webkit.org/show_bug.cgi?id=14514
        <rdar://problem/5712841>

        * page/inspector/inspector.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29901 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Geoffrey Garen.

        - change the font used in fast/text/monospace-width-cache.html again to
          a font that contains the U+2500 character in Tiger too

        * fast/text/monospace-width-cache.html:
        * platform/mac/fast/text/monospace-width-cache-expected.checksum:
        * platform/mac/fast/text/monospace-width-cache-expected.png:
        * platform/mac/fast/text/monospace-width-cache-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29900 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix r29898 ChangeLog entries: rubber-stamped, not reviewed

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29899 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-01-31  Alp Toker  <alp@atoker.com>

        Reviewed by Adam Roben.

        http://bugs.webkit.org/show_bug.cgi?id=17006
        [GTK] Header path should be webkit/webkit.h

        Move the GTK+ API sources as needed and update the build systems.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29898 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Geoff Garen.

        No need to get the prototype just to compare compare property names.

        * bindings/js/JSHistoryCustom.cpp:
        (WebCore::JSHistory::customGetOwnPropertySlot):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29897 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Fix non-mac builds.

        * GNUmakefile.am:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCoreSources.bkl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29896 268f45cc-cd09-0410-ab3c-d52691b4dbfc

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29895 268f45cc-cd09-0410-ab3c-d52691b4dbfc

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29894 268f45cc-cd09-0410-ab3c-d52691b4dbfc

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29893 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix for bug 17101.

        List items need to ignore the line box shrinking quirk, since IE and Firefox both do.

        Reviewed by Eric

        fast/lists/list-item-line-height.html

        * rendering/InlineFlowBox.h:
        (WebCore::InlineFlowBox::InlineFlowBox):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29892 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam Roben.

        - change several tests that use Mac-only fonts to use fonts that are
          available on both Windows and Mac OS X

        * fast/text/monospace-width-cache.html:
        * platform/mac/fast/text/monospace-width-cache-expected.checksum:
        * platform/mac/fast/text/monospace-width-cache-expected.png:
        * platform/mac/fast/text/monospace-width-cache-expected.txt:
        * platform/mac/tables/mozilla/bugs/bug10633-expected.checksum:
        * platform/mac/tables/mozilla/bugs/bug10633-expected.png:
        * platform/mac/tables/mozilla/bugs/bug10633-expected.txt:
        * platform/mac/tables/mozilla/bugs/bug113424-expected.checksum:
        * platform/mac/tables/mozilla/bugs/bug113424-expected.png:
        * platform/mac/tables/mozilla/bugs/bug113424-expected.txt:
        * platform/mac/tables/mozilla/bugs/bug23151-expected.checksum:
        * platform/mac/tables/mozilla/bugs/bug23151-expected.png:
        * platform/mac/tables/mozilla/bugs/bug23151-expected.txt:
        * platform/mac/tables/mozilla/core/bloomberg-expected.checksum:
        * platform/mac/tables/mozilla/core/bloomberg-expected.png:
        * platform/mac/tables/mozilla/core/bloomberg-expected.txt:
        * platform/mac/tables/mozilla_expected_failures/bugs/bug11331-expected.checksum:
        * platform/mac/tables/mozilla_expected_failures/bugs/bug11331-expected.png:
        * platform/mac/tables/mozilla_expected_failures/bugs/bug11331-expected.txt:
        * platform/win/Skipped:
        * tables/mozilla/bugs/bug10633.html:
        * tables/mozilla/bugs/bug113424.html:
        * tables/mozilla/bugs/bug23151.html:
        * tables/mozilla/core/bloomberg.html:
        * tables/mozilla_expected_failures/bugs/bug11331.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29891 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin Adler.

        Fix for <rdar://problem/5708993> Mutability of the History object

        - Don't allow cross-domain get access to any of the history objects properties
          except the back(), forward() and go() methods.
        - Don't allow cross-domain put access to any of the history objects properties.
        - Don't allow cross-domain enumeration of the History or Location objects.

        Tests: http/tests/security/cross-frame-access-history-get-override.html
               http/tests/security/cross-frame-access-history-get.html
               http/tests/security/cross-frame-access-history-put.html

        * WebCore.xcodeproj/project.pbxproj:
        * bindings/js/JSDOMWindowCustom.cpp: Remove unnessary KJS::'s
        (WebCore::JSDOMWindow::customGetOwnPropertySlot):
        (WebCore::JSDOMWindow::customPut):
        (WebCore::JSDOMWindow::getPropertyNames): Moved implementation from KJS::Window now that the declaration is autogenerated
        using the new CustomGetPropertyNames.
        (WebCore::JSDOMWindow::postMessage):

        * bindings/js/JSHistoryCustom.cpp: Added.
        (WebCore::allowsAccessFromFrame):
        (WebCore::JSHistory::customGetOwnPropertySlot): Only allow getting the declared functions back(), forward() and go() from cross-domain.
        Deny all other gets.
        (WebCore::JSHistory::customPut): Don't allow putting cross-domain.
        (WebCore::JSHistory::getPropertyNames): Don't allow enumeration cross-domain.

        * bindings/js/JSLocation.cpp:
        (WebCore::allowsAccessFromFrame):
        (WebCore::JSLocation::getPropertyNames): Don't allow enumeration cross-domain.
        * bindings/js/JSLocation.h:

        * bindings/js/kjs_window.cpp:
        * bindings/js/kjs_window.h:

        * bindings/scripts/CodeGeneratorJS.pm:
        Add support for new CustomGetPropertNames extended attribute and changed the logic of CustomPutFunction
        to create an overrided put() function even if no read-write properties exist.

        * page/DOMWindow.idl: Added CustomGetPropertNames
        * page/History.idl: Added CustomGetPropertNames

LayoutTests:

        Reviewed by Darin Adler.

        Tests for <rdar://problem/5708993> Mutability of the History object

        * http/tests/security/cross-frame-access-enumeration-expected.txt:
        * http/tests/security/cross-frame-access-enumeration.html:
        * http/tests/security/cross-frame-access-history-expected.txt: Removed.
        * http/tests/security/cross-frame-access-history-get-expected.txt: Renamed from LayoutTests/http/tests/security/cross-frame-access-history-expected.txt.
        * http/tests/security/cross-frame-access-history-get-override-expected.txt: Added.
        * http/tests/security/cross-frame-access-history-get-override.html: Added.
        * http/tests/security/cross-frame-access-history-get.html: Renamed from LayoutTests/http/tests/security/cross-frame-access-history.html.
        * http/tests/security/cross-frame-access-history-put-expected.txt: Added.
        * http/tests/security/cross-frame-access-history-put.html: Added.
        * http/tests/security/cross-frame-access-history.html: Removed.
        * http/tests/security/resources/cross-frame-access.js:
        * http/tests/security/resources/cross-frame-iframe-for-enumeration-test.html:
        * http/tests/security/resources/cross-frame-iframe-for-history-get-override-test.html: Added.
        * http/tests/security/resources/cross-frame-iframe-for-history-get-test.html: Added.
        * http/tests/security/resources/cross-frame-iframe-for-history-put-test.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29890 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin Adler.

        <rdar://problem/5708115> REGRESSION: Words selected with a double click and copied won't paste into Mail

        * page/mac/WebCoreFrameBridge.h: Re-exposed smartInsertForString:, it's used by a WebKit method used by Mail.

WebKit/mac:

        Reviewed by Darin Adler.

        <rdar://problem/5708115> REGRESSION: Words selected with a double click and copied won't paste into Mail

        * WebView/WebHTMLView.mm:
        (-[WebHTMLView _smartInsertForString:replacingRange:beforeString:afterString:]): Brought
        this back, it's used by Mail.
        (-[WebHTMLView _canSmartReplaceWithPasteboard:]): This WebInternal method is
        also used by Mail.  Moved to WebPrivate.
        * WebView/WebHTMLViewPrivate.h: Expose two methods that Mail uses here, so that we don't
        accidently remove them in the future.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29889 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Add line box drawing code to InspectorController::drawNodeHighlight

        This makes drawNodeHighlight a complete replacement for the painting
        code in WebKit/mac/WebNodeHighlightView.mm, and also brings line box
        rects to Windows for the first time.

        Reviewed by Darin.

        * dom/Node.h: Changed isSVGElement to always exist, but to only be
        virtual when ENABLE(SVG) is true. This way you can always call
        node->isSVGElement() without checking ENABLE(SVG).
        * page/InspectorController.cpp:
        (WebCore::InspectorController::drawNodeHighlight): Ported line box
        rect code from the Mac implementation in WebNodeHighlightView.mm.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29888 268f45cc-cd09-0410-ab3c-d52691b4dbfc

 Let WebCore take care of the highlight drawing entirely

 WebCore:

         Put more knowledge about the node highlight in WebCore

         InspectorController now calculates the overlay rect and node rect when
         drawing the node highlight instead of having them be passed in.
         InspectorController now holds onto the highlighted node so that it can
         determine these rects.

         Once all platforms are calling down to drawNodeHighlight instead of
         drawing the highlight themselves, we can change
         InspectorClient::highlight(Node*) to something like
         InspectorClient::updateAndShowHighlight().

         This also fixes Bug 14264: Node highlight makes it impossible to
         scroll the page
         <http://bugs.webkit.org/show_bug.cgi?id=14264>
         <rdar://5712788>

         Reviewed by Darin.

         * page/InspectorController.cpp:
         (WebCore::InspectorController::highlight): Store the node for use in
         drawNodeHighlight.
         (WebCore::InspectorController::drawNodeHighlight): Changed to be a
         const instance method. Now calculates the overlay rect and node rect
         instead of having them passed in.
         * page/InspectorController.h:

 WebKit/win:

         Let WebCore take care of the highlight drawing entirely

         Reviewed by Darin.

         * WebInspectorClient.cpp:
         (WebInspectorClient::highlight): We now just show our highlight
         window and let WebCore figure out what/where to paint. Once all ports
         follow suit the Node* parameter to this method should be removed
         entirely, and the name should probably change to
         updateAndShowHighlight or something similar.
         * WebNodeHighlight.cpp: Renamed m_webView to m_inspectedWebViewWindow.
         (WebNodeHighlight::WebNodeHighlight): Now takes a WebView* parameter.
         (WebNodeHighlight::show): Renamed from highlight(). Now gets the
         WebView's HWND. Updated for member rename/removal.
         (WebNodeHighlight::updateWindow): Updated for member rename and for
         InspectorController changes.
         * WebNodeHighlight.h: Added m_inspectedWebView member, removed m_rect
         member, renamed highlight(Node*) -> show().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29887 268f45cc-cd09-0410-ab3c-d52691b4dbfc

 Move node highlight drawing code to WebCore

 WebCore:

         Add node highlight drawing code to InspectorController

         The code came from WebKit/win/WebNodeHighlight.cpp. It's not quite as
         complete as the Mac implementation (in particular, it doesn't handle
         line-box rects), but it's a start.

         Reviewed by Darin.

         * page/InspectorController.cpp:
         (WebCore::InspectorController::drawNodeHighlight): Added.
         * page/InspectorController.h:

 WebKit/win:

         Move node highlight drawing code to WebCore

         Reviewed by Darin.

         * WebNodeHighlight.cpp:
         (WebNodeHighlight::updateWindow): Call into WebCore to do the node
         highlight drawing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29886 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Dave Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=17107
          <rdar://problem/5716722> REGRESSION (r29834): Article text on redhat.com magazine site appears to be painting twice

        Test: fast/block/float/intruding-painted-twice.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::layoutBlock): Pass 'false' for the new
        makeChildPaintOtherFloats parameter to addOverhangingFloats() because at
        this point we are only taking away floats from the child.
        (WebCore::RenderBlock::layoutBlockChildren): Pass 'true' for the new
        makeChildPaintOtherFloats parameter to addOverhangingFloats() iff the
        child was not laid out again. Only in that case, it may have overhanging
        floats that it does not paint because they used to be overhanging from
        the parent, but now they are not.
        (WebCore::RenderBlock::addOverhangingFloats): Refined the conditions for
        making the child paint the float: require that the float be a descendant
        of the child (the other case is when it intrudes into the child from
        another sibling) and that it does not have a layer (in which case it
        paints itself). In addition, do the check only if the caller passed
        'true' for the makeChildPaintOtherFloats parameter.
        * rendering/RenderBlock.h:

LayoutTests:

        Reviewed by Dave Hyatt.

        - pixel test for http://bugs.webkit.org/show_bug.cgi?id=17107
          <rdar://problem/5716722> REGRESSION (r29834): Article text on redhat.com magazine site appears to be painting twice

        * fast/block/float/intruding-painted-twice.html: Added.
        * platform/mac-leopard/fast/block/float/intruding-painted-twice-expected.checksum: Added.
        * platform/mac-leopard/fast/block/float/intruding-painted-twice-expected.png: Added.
        * platform/mac/fast/block/float/intruding-painted-twice-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29885 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Sam Weinig.

        - change the interpretation of unicode-range values in "from-to" form
          to include the "to" character.

        Test: fast/css/font-face-unicode-range.html

        * css/CSSParser.cpp:
        (WebCore::CSSParser::parseFontFaceUnicodeRange):
        * css/CSSSegmentedFontFace.cpp:
        (WebCore::CSSSegmentedFontFace::overlayRange):
        * platform/graphics/GlyphPageTreeNode.cpp:
        (WebCore::GlyphPageTreeNode::initializePage):
        * platform/graphics/SegmentedFontData.cpp:
        (WebCore::SegmentedFontData::fontDataForCharacter):
        (WebCore::SegmentedFontData::containsCharacters):

LayoutTests:

        Reviewed by Sam Weinig.

        - add a test for the unicode-range @font-face property
        - update @font-face rules in existing tests after changing unicode-range
          values in "from-to" form to include the "to" character

        * fast/css/font-face-unicode-range.html: Added.
        * fast/css/rtl-ordering.html:
        * fast/text/international/resources/Mac-compatible-font-fallback.css:
        * platform/mac-leopard/fast/css/font-face-unicode-range-expected.checksum: Added.
        * platform/mac-leopard/fast/css/font-face-unicode-range-expected.png: Added.
        * platform/mac/fast/css/font-face-unicode-range-expected.txt: Added.
        * platform/win/css2.1/resources/Mac-compatible-font-fallback.css:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29884 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        - prune references to custom fonts' SimpleFontData from the glyph page
          tree when they are destroyed

        * css/CSSFontFaceSource.cpp:
        (WebCore::CSSFontFaceSource::pruneTable):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29883 268f45cc-cd09-0410-ab3c-d52691b4dbfc

This test started failed with my last checkin.  The new results are *correct* but I don't know why my changes would have fixed the test, so I want to investigate a bit before I check in new results.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29882 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Tim Hatcher.

        New fix for <rdar://problem/5688428> Reproducible assertion failure
        in SQLTransaction::performNextStep() (16876)

        Fixes a deadlock that was happening for all Database use. All
        storage layout tests pass.

        * storage/Database.cpp:
        (WebCore::CurrentThreadSetter::CurrentThreadSetter): New helper class
        to set the current thread and zero it on destruction.
        (WebCore::CurrentThreadSetter::~CurrentThreadSetter): Set threadIdentifierStorage to 0.
        (WebCore::Database::Database): Set m_transactionStepThread to 0.
        (WebCore::Database::performTransactionStep): Set m_transactionStepThread to currentThread().
        (WebCore::Database::scheduleTransactionCallback): Assert m_transactionStepThread
        is currentThread().
        * storage/Database.h: Add m_transactionStepThread for debug builds
        to track which thread performTransactionStep() was called on.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29881 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-01-30  Stephanie  <slewis@apple.com>

        Reviewed by Geoff.

        Add an option to run-webkit-tests to merge leaks results and print the number of unique leaks found under a certain call depth.  This should make it easier for the buildbots to track new leaks.

        * Scripts/run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@29880 268f45cc-cd09-0410-ab3c-d52691b4dbfc