GitLab

git-svn-id: svn://svn.chromium.org/blink/trunk@18831 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by hyatt.

        RichDraw demo does not live-update in Safari
        (SVG elements should ignore CSS position)
        http://bugs.webkit.org/show_bug.cgi?id=10364

        * css/cssstyleselector.cpp:
        (WebCore::CSSStyleSelector::adjustRenderStyle): ignore position for everything but <svg>
        * ksvg2/svg/SVGStyledElement.cpp:
        (WebCore::SVGStyledElement::rebuildRenderer): move path generation into layout()
        * rendering/RenderContainer.cpp:
        (WebCore::RenderContainer::layout): ASSERT kids marked as having completed layout
        * rendering/RenderPath.cpp:
        (WebCore::RenderPath::layout): move path generation into layout()
        * rendering/RenderSVGContainer.cpp:
        (WebCore::RenderSVGContainer::layout): ASSERT kids marked as having completed layout


git-svn-id: svn://svn.chromium.org/blink/trunk@18830 bbb929c8-8fbe-4397-9dbb-9b2b20218538

git-svn-id: svn://svn.chromium.org/blink/trunk@18829 bbb929c8-8fbe-4397-9dbb-9b2b20218538

git-svn-id: svn://svn.chromium.org/blink/trunk@18828 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by bdash

        fast/multicol/column-rules.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::paintColumns):
        * rendering/RenderStyle.h:
        (WebCore::RenderStyle::columnRuleIsTransparent):



git-svn-id: svn://svn.chromium.org/blink/trunk@18827 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by andersca

        Added fast/multicol/float-multicol.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::layoutBlock):
        (WebCore::RenderBlock::paintColumns):
        (WebCore::RenderBlock::paintObject):
        * rendering/RenderBlock.h:



git-svn-id: svn://svn.chromium.org/blink/trunk@18826 bbb929c8-8fbe-4397-9dbb-9b2b20218538

git-svn-id: svn://svn.chromium.org/blink/trunk@18825 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Maciej.

        - test for fast/css/line-height.html
          REGRESSION (r18816-r18818): Assertion failure in
          WebCore::Length::Length[in-charge](int, WebCore::LengthType, bool)

        * fast/css/line-height-expected.checksum: Added.
        * fast/css/line-height-expected.png: Added.
        * fast/css/line-height-expected.txt: Added.
        * fast/css/line-height.html: Added.

WebCore:

        Reviewed by Maciej.

        - fix http://bugs.webkit.org/show_bug.cgi?id=12239
          REGRESSION (r18816-r18818): Assertion failure in 
          WebCore::Length::Length[in-charge](int, WebCore::LengthType, bool)

        Test: fast/css/line-height.html

        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::createInnerTextStyle): Use a float to create a percent Length.



git-svn-id: svn://svn.chromium.org/blink/trunk@18824 bbb929c8-8fbe-4397-9dbb-9b2b20218538

git-svn-id: svn://svn.chromium.org/blink/trunk@18823 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Darin.

        * wtf/HashTraits.h:
        Add hash traits for unsigned long and unsigned long long.

WebCore:

        Reviewed by Darin.

        Use unsigned long for identifiers instead of Objective-C objects.
        
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::loadResourceSynchronously):
        (WebCore::FrameLoader::startLoadingMainResource):
        (WebCore::FrameLoader::startLoading):
        (WebCore::FrameLoader::assignIdentifierToInitialRequest):
        (WebCore::FrameLoader::opened):
        (WebCore::FrameLoader::sendRemainingDelegateMessages):
        (WebCore::FrameLoader::requestFromDelegate):
        (WebCore::FrameLoader::loadedResourceFromMemoryCache):
        * loader/FrameLoader.h:
        * loader/FrameLoaderClient.h:
        * loader/ResourceLoader.cpp:
        (WebCore::ResourceLoader::ResourceLoader):
        (WebCore::ResourceLoader::willSendRequest):
        * loader/ResourceLoader.h:
        (WebCore::ResourceLoader::setIdentifier):
        (WebCore::ResourceLoader::identifier):
        * page/Page.cpp:
        (WebCore::Page::Page):
        (WebCore::Page::createUniqueIdentifier):
        * page/Page.h:
        * platform/graphics/svg/SVGImageEmptyClients.h:
        (WebCore::SVGEmptyFrameLoaderClient::assignIdentifierToInitialRequest):
        (WebCore::SVGEmptyFrameLoaderClient::dispatchWillSendRequest):
        (WebCore::SVGEmptyFrameLoaderClient::dispatchDidReceiveAuthenticationChallenge):
        (WebCore::SVGEmptyFrameLoaderClient::dispatchDidCancelAuthenticationChallenge):
        (WebCore::SVGEmptyFrameLoaderClient::dispatchDidReceiveResponse):
        (WebCore::SVGEmptyFrameLoaderClient::dispatchDidReceiveContentLength):
        (WebCore::SVGEmptyFrameLoaderClient::dispatchDidFinishLoading):
        (WebCore::SVGEmptyFrameLoaderClient::dispatchDidFailLoading):
        (WebCore::SVGEmptyFrameLoaderClient::incrementProgress):
        (WebCore::SVGEmptyFrameLoaderClient::completeProgress):

WebKit:

        Reviewed by Darin.

        Add a HashMap between unsigned longs and Objective-C objects and use it for
        the resource load delegate.
        
        * WebCoreSupport/WebFrameLoaderClient.h:
        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::assignIdentifierToInitialRequest):
        (WebFrameLoaderClient::dispatchIdentifierForInitialRequest):
        (WebFrameLoaderClient::dispatchWillSendRequest):
        (WebFrameLoaderClient::dispatchDidReceiveAuthenticationChallenge):
        (WebFrameLoaderClient::dispatchDidCancelAuthenticationChallenge):
        (WebFrameLoaderClient::dispatchDidReceiveResponse):
        (WebFrameLoaderClient::dispatchDidReceiveContentLength):
        (WebFrameLoaderClient::dispatchDidFinishLoading):
        (WebFrameLoaderClient::dispatchDidFailLoading):
        (WebFrameLoaderClient::incrementProgress):
        (WebFrameLoaderClient::completeProgress):
        * WebView/WebView.mm:
        (-[WebViewPrivate init]):
        (-[WebViewPrivate dealloc]):
        (-[WebView _addObject:forIdentifier:]):
        (-[WebView _objectForIdentifier:]):
        (-[WebView _removeObjectForIdentifier:]):
        * WebView/WebViewInternal.h:



git-svn-id: svn://svn.chromium.org/blink/trunk@18822 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Darin.

        - fix http://bugs.webkit.org/show_bug.cgi?id=12235
          REGRESSION (r18815): Marquees do not scroll

        * css/cssstyleselector.cpp:
        (WebCore::CSSStyleSelector::applyProperty): Apply the '-webkit-marquee' value
        of 'overflow-x' and 'overflow-y'.


git-svn-id: svn://svn.chromium.org/blink/trunk@18821 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by hyatt

        <rdar://problem/4875425>
        EBay, GMail, and GoogleDocs RTEs font size popups don't work

        Added to test changes:
        * editing/style/fontsize-1-expected.checksum: Added.
        * editing/style/fontsize-1-expected.png: Added.
        * editing/style/fontsize-1-expected.txt: Added.
        * editing/style/fontsize-1.html: Added.

        Removed use of "px" and "pt" values passed to
        execCommand("FontSize", ...) (other browsers don't
        support them): 
        * editing/input/attributed-substring-from-range-lines.html-disabled:
        * editing/style/block-style-004-expected.checksum:
        * editing/style/block-style-004-expected.png:
        * editing/style/block-style-004-expected.txt:
        * editing/style/block-style-004.html:
        * editing/style/block-style-005-expected.checksum:
        * editing/style/block-style-005-expected.png:
        * editing/style/block-style-005-expected.txt:
        * editing/style/block-style-005.html:
        * editing/style/block-style-006-expected.checksum:
        * editing/style/block-style-006-expected.png:
        * editing/style/block-style-006-expected.txt:
        * editing/style/block-style-006.html:

WebCore:

        Reviewed by hyatt
        
        <rdar://problem/4875425>
        EBay, GMail, and GoogleDocs RTEs font size popups don't work

        Changed execCommand("FontSize", ...) to match the spec 
        and the behavior of other browsers.  It expects a number, 
        n, which, the spec says can be 1 through 7.  Other browsers 
        perform the font size change by creating a font element 
        wrapper with size="n".  We convert n into one of the absolute 
        size CSS keywords (x-small, medium, etc.).

        * editing/JSEditor.cpp: Call cssValueFromFontSizeNumber.  If
        successful, perform the style change.
        * html/HTMLFontElement.cpp:
        (WebCore::HTMLFontElement::cssValueFromFontSizeNumber):  Added.
        Moved this code from parsedMappedAttribute so that it could be
        called from the Editor.
        (WebCore::HTMLFontElement::parseMappedAttribute): Moved code into
        cssValueFromFontSizeNumber.
        * html/HTMLFontElement.h:



git-svn-id: svn://svn.chromium.org/blink/trunk@18820 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        buttons, list boxes, and dropdown lists.  Make sure list box respects
        scrollLeft/Top/Width/Height as well as clientWidth/Height.

        Fix bugs in RenderStyle::diff.  CursorData was not being compared properly.
        vertical-align had a non-inline display check that was wrong (preventing
        dynamic changes to keyword-based vertical-align from working).

        Reviewed by mitz

        * page/EventHandler.cpp:
        (WebCore::EventHandler::selectClosestWordFromMouseEvent):
        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::layout):
        (WebCore::RenderBlock::layoutBlock):
        (WebCore::RenderBlock::determineHorizontalPosition):
        (WebCore::RenderBlock::layoutBlockChildren):
        (WebCore::RenderBlock::paint):
        (WebCore::RenderBlock::fillHorizontalSelectionGap):
        (WebCore::RenderBlock::nodeAtPoint):
        (WebCore::RenderBlock::layoutColumns):
        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::computeAbsoluteRepaintRect):
        * rendering/RenderBox.h:
        * rendering/RenderButton.cpp:
        (WebCore::RenderButton::setStyle):
        (WebCore::RenderButton::controlClipRect):
        * rendering/RenderButton.h:
        (WebCore::RenderButton::hasControlClip):
        * rendering/RenderFlexibleBox.cpp:
        (WebCore::RenderFlexibleBox::layoutBlock):
        (WebCore::RenderFlexibleBox::layoutHorizontalBox):
        (WebCore::RenderFlexibleBox::layoutVerticalBox):
        * rendering/RenderFlow.cpp:
        (WebCore::RenderFlow::addFocusRingRects):
        * rendering/RenderListBox.cpp:
        (WebCore::RenderListBox::itemBoundingBoxRect):
        (WebCore::RenderListBox::paintObject):
        (WebCore::RenderListBox::itemHeight):
        (WebCore::RenderListBox::verticalScrollbarWidth):
        (WebCore::RenderListBox::scrollWidth):
        (WebCore::RenderListBox::scrollHeight):
        (WebCore::RenderListBox::scrollLeft):
        (WebCore::RenderListBox::setScrollLeft):
        (WebCore::RenderListBox::scrollTop):
        (WebCore::RenderListBox::setScrollTop):
        (WebCore::RenderListBox::controlClipRect):
        * rendering/RenderListBox.h:
        (WebCore::RenderListBox::hasControlClip):
        * rendering/RenderMenuList.cpp:
        (WebCore::RenderMenuList::setStyle):
        (WebCore::RenderMenuList::controlClipRect):
        * rendering/RenderMenuList.h:
        (WebCore::RenderMenuList::hasControlClip):
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::verticalScrollbarWidth):
        (WebCore::RenderObject::horizontalScrollbarHeight):
        (WebCore::RenderObject::clientWidth):
        (WebCore::RenderObject::clientHeight):
        (WebCore::RenderObject::setStyle):
        * rendering/RenderObject.h:
        (WebCore::RenderObject::hasControlClip):
        (WebCore::RenderObject::controlClipRect):
        (WebCore::RenderObject::contentWidth):
        (WebCore::RenderObject::contentHeight):
        * rendering/RenderStyle.cpp:
        (WebCore::StyleVisualData::StyleVisualData):
        (WebCore::cursorDataEqvuialent):
        (WebCore::StyleInheritedData::operator==):
        (WebCore::CursorList::operator==):
        (WebCore::RenderStyle::diff):
        * rendering/RenderStyle.h:
        (WebCore::LengthBox::operator!=):
        (WebCore::StyleVisualData::operator==):
        (WebCore::CursorData::operator==):
        (WebCore::CursorData::operator!=):
        (WebCore::CursorList::operator!=):
        (WebCore::RenderStyle::setTableLayout):
        * rendering/bidi.cpp:
        (WebCore::RenderBlock::layoutInlineChildren):



git-svn-id: svn://svn.chromium.org/blink/trunk@18819 bbb929c8-8fbe-4397-9dbb-9b2b20218538

git-svn-id: svn://svn.chromium.org/blink/trunk@18818 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        text fields.

        Reviewed by darin

        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::createInnerTextStyle):



git-svn-id: svn://svn.chromium.org/blink/trunk@18816 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Hyatt.  Patch by Mitz.

        - test updated for http://bugs.webkit.org/show_bug.cgi?id=12215
          Treat 'overflow' and '-webkit-border-radius' as shorthands when parsing and
          when removing.

        * fast/css/remove-shorthand-expected.txt:
        * fast/css/remove-shorthand.html:

WebCore:

        Reviewed by Hyatt.  Patch by Mitz.

        - http://bugs.webkit.org/show_bug.cgi?id=12215
          Treat 'overflow' and '-webkit-border-radius' as shorthands when parsing and
          when removing.

        Test: fast/css/remove-shorthand.html (updated)

        * css/CSSMutableStyleDeclaration.cpp:
        (WebCore::initShorthandMap):
        * css/cssparser.cpp:
        (WebCore::CSSParser::parseValue):



git-svn-id: svn://svn.chromium.org/blink/trunk@18815 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Do not mix up JSSVGPolygonElement and JSSVGPolylineElement.



git-svn-id: svn://svn.chromium.org/blink/trunk@18814 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by andersca.

        Fixed: <rdar://problem/4701326>21.2 ms launch time regression spent in +[WebPluginDatabase installedPlugins]

        * dom/DOMImplementation.cpp:
        (WebCore::DOMImplementation::createDocument): Added a guard clause for text/html which used to be in WebKit before a lot of the 
        loader plumbing was moved to WebCore.  Pinging the plug-in database to ask it if it supports a given mime type has the side 
        effect of reading from disk on first call, therefore such a check should be a last resort.  


git-svn-id: svn://svn.chromium.org/blink/trunk@18813 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        
        Rolling back in r18786 with leaks fixed, and these renames slightly reworked:

        Because they can return 0:
        rootObjectForImp => findRootObject (overloaded for JSObject* and Interpreter*)
        rootObjectForInterpreter => findRootObject (ditto)
        findReferenceSet => findProtectCountSet



git-svn-id: svn://svn.chromium.org/blink/trunk@18811 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Darin.  Patch by Mitz.

        - fix http://bugs.webkit.org/show_bug.cgi?id=12223
          REGRESSION: Leaks under CSSStyleSelector::applyProperty

        No test possible (no change to functionality).

        * rendering/RenderObject.h: Removed unused forward declaration.
        * rendering/RenderStyle.cpp:
        (WebCore::StyleRareNonInheritedData::~StyleRareNonInheritedData):
        Delete shadow data.



git-svn-id: svn://svn.chromium.org/blink/trunk@18810 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        
        * WebCore.xcodeproj/project.pbxproj:
        Remove SubresourceLoaderMac.mm.



git-svn-id: svn://svn.chromium.org/blink/trunk@18806 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Make some more of the loader code platform independent.

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::clear):
        (WebCore::FrameLoader::checkLoadCompleteForThisFrame):
        (WebCore::FrameLoader::urlSelected):
        (WebCore::FrameLoader::createEmptyDocument):
        (WebCore::FrameLoader::setTitle):
        (WebCore::FrameLoader::originalRequestURL):
        (WebCore::FrameLoader::referrer):
        * loader/SubresourceLoader.cpp:
        (WebCore::SubresourceLoader::SubresourceLoader):
        (WebCore::SubresourceLoader::load):
        (WebCore::SubresourceLoader::create):
        (WebCore::SubresourceLoader::willSendRequest):
        (WebCore::SubresourceLoader::didReceiveResponse):
        (WebCore::SubresourceLoader::didReceiveData):
        (WebCore::SubresourceLoader::didFinishLoading):
        (WebCore::SubresourceLoader::didFail):
        (WebCore::SubresourceLoader::didCancel):
        (WebCore::SubresourceLoader::stopLoading):
        * loader/SubresourceLoader.h:
        * loader/mac/FrameLoaderMac.mm:
        * loader/mac/SubresourceLoaderMac.mm: Removed.
        * loader/qt/FrameLoaderQt.cpp:



git-svn-id: svn://svn.chromium.org/blink/trunk@18805 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Get DumpRenderTree to work again for the Qt build.

        Make run-webkit-tests a little less verbose when testing
        Qt, and add an option to run DumpRenderTree inside valgrind
        (useful for debugging)



git-svn-id: svn://svn.chromium.org/blink/trunk@18804 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Valgrind reported uninitialized variable.

        * platform/network/qt/ResourceHandleManagerQt.cpp:
        (WebCore::FileLoader::request):



git-svn-id: svn://svn.chromium.org/blink/trunk@18803 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        - add missing null check in new fixed table layout code



git-svn-id: svn://svn.chromium.org/blink/trunk@18802 bbb929c8-8fbe-4397-9dbb-9b2b20218538

make clean before make as they fix stale dependencies (for buildbot mainly)


git-svn-id: svn://svn.chromium.org/blink/trunk@18801 bbb929c8-8fbe-4397-9dbb-9b2b20218538

git-svn-id: svn://svn.chromium.org/blink/trunk@18800 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Fix for bug 9569.  Make sure the selection offset for selection roots excludes border and padding.

        Reviewed by Eric

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::leftSelectionOffset):
        (WebCore::RenderBlock::rightSelectionOffset):



git-svn-id: svn://svn.chromium.org/blink/trunk@18799 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Fix for bug 8705, readonly text fields should not get focus rings.  Add a rule to that effect to
        html4.css and then also add readonly as a special case for style sharing so that we can retain the
        performance boost from sharing styles on inputs.

        Reviewed by bdash

        * css/cssstyleselector.cpp:
        (WebCore::CSSStyleSelector::canShareStyleWithElement):
        (WebCore::CSSStyleSelector::checkOneSelector):
        * css/html4.css:



git-svn-id: svn://svn.chromium.org/blink/trunk@18798 bbb929c8-8fbe-4397-9dbb-9b2b20218538

git-svn-id: svn://svn.chromium.org/blink/trunk@18797 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Fix for bug 11595.  Autoscrolling of layers didn't really work at all unless the layer was coincidentally
        in the root layer's coordinate space.  The event point was an absolute position within the document, but
        the layer's rect was only local coords.  This patch makes sure to use absolute bounds for the layer
        when comparing it with the point.

        This is covered by the scrollRevealButton test case already in the tree... the results change to be more
        correct.

        Reviewed by ggaren

        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::scrollRectToVisible):



git-svn-id: svn://svn.chromium.org/blink/trunk@18796 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        
        Rolling out r18786 because it caused leaks.



git-svn-id: svn://svn.chromium.org/blink/trunk@18795 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        -Move more HTTP Auth handling towards ResourceHandle, away from ResourceHandleClient
        -Make more of ResourceLoader cross-platform

        * loader/ResourceLoader.cpp:
        (WebCore::ResourceLoader::ResourceLoader):
        (WebCore::ResourceLoader::releaseResources):
        (WebCore::ResourceLoader::didCancel):
        (WebCore::ResourceLoader::didReceiveAuthenticationChallenge):
        (WebCore::ResourceLoader::didCancelAuthenticationChallenge):
        (WebCore::ResourceLoader::receivedCancellation):
        * loader/ResourceLoader.h:
        (WebCore::ResourceLoader::willStopBufferingData):
        (WebCore::ResourceLoader::didCancelAuthenticationChallenge):
        (WebCore::ResourceLoader::willCacheResponse):
        * loader/mac/ResourceLoaderMac.mm:
        * platform/network/ResourceHandle.h:
        * platform/network/ResourceHandleClient.h:
        (WebCore::ResourceHandleClient::willStopBufferingData):
        * platform/network/ResourceHandleInternal.h:
        (WebCore::ResourceHandleInternal::ResourceHandleInternal):
        * platform/network/mac/ResourceHandleMac.mm:
        (WebCore::ResourceHandle::didReceiveAuthenticationChallenge):
        (WebCore::ResourceHandle::didCancelAuthenticationChallenge):
        (WebCore::ResourceHandle::receivedCredential):
        (WebCore::ResourceHandle::receivedRequestToContinueWithoutCredential):
        (WebCore::ResourceHandle::receivedCancellation):
        (WebCore::ResourceHandle::clearAuthentication):
        (-[WebCoreResourceHandleAsDelegate connection:didReceiveAuthenticationChallenge:]):
        (-[WebCoreResourceHandleAsDelegate connection:didCancelAuthenticationChallenge:]):
        (-[WebCoreResourceHandleAsDelegate useCredential:forAuthenticationChallenge:]):
        (-[WebCoreResourceHandleAsDelegate continueWithoutCredentialForAuthenticationChallenge:]):
        (-[WebCoreResourceHandleAsDelegate cancelAuthenticationChallenge:]):
        


git-svn-id: svn://svn.chromium.org/blink/trunk@18794 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        A hit test is returning a null renderer.  Make sure to null-check it in the autoscroll code.
        This really shouldn't happen, but I'm just going to bullet proof for now given that it's happening
        on the front page of yahoo.com!

        Reviewed by mark

        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::autoscroll):



git-svn-id: svn://svn.chromium.org/blink/trunk@18793 bbb929c8-8fbe-4397-9dbb-9b2b20218538

       Force column-width to parse in strict mode so that there is no confusion when it is used in the columns
       shorthand (it can be mistaken for column-count otherwise).

       Reviewed by Mark
       
        * fast/multicol/columns-shorthand-parsing-expected.checksum: Added.
        * fast/multicol/columns-shorthand-parsing-expected.png: Added.
        * fast/multicol/columns-shorthand-parsing-expected.txt: Added.
        * fast/multicol/columns-shorthand-parsing.html: Added.

WebCore:

        Force column-width to parse in strict mode so that there is no confusion when it is used in the columns
        shorthand (it can be mistaken for column-count otherwise).

        Reviewed by Mark

        fast/multicol/columns-shorthand-parsing.html

        * css/cssparser.cpp:
        (WebCore::CSSParser::parseValue):



git-svn-id: svn://svn.chromium.org/blink/trunk@18792 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=12124
          RenderTable::section{Above,Below}() can return the wrong answer for the {footer,header}

        * rendering/RenderTable.cpp:
        (WebCore::resetSectionPointerIfNotBefore): Added this function to help with
        checking if a section being inserted is the going to be the first of its kind.
        (WebCore::RenderTable::addChild): Changed to maintain the pointers to the
        header, the footer, the caption and the first table body properly when objects
        are inserted in the middle. Also changed to make the first caption win over
        any other captions. This matches Firefox and what already used to happen when a
        section recalc took place.
        (WebCore::RenderTable::layout): Changed to use sectionBelow() for iterating
        over the sections from top to bottom.
        (WebCore::RenderTable::calcBorderLeft): Changed to really find the section
        containing the first row.
        (WebCore::RenderTable::calcBorderRight): Ditto.
        (WebCore::RenderTable::sectionAbove): Fixed a bug when the last section was
        skipped if the input was the footer.
        (WebCore::RenderTable::sectionBelow): Ditto for "first" and "header"
        respectively.

2007-01-11  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Hyatt.

        - updated results for http://bugs.webkit.org/show_bug.cgi?id=12124
          RenderTable::section{Above,Below}() can return the wrong answer for the {footer,header}

        Collapsed borders changed in the "marvin" tests, as the footer cells now
        collapse correctly with the last row of the body. Visual differences not
        captured by the pixel results since they occurred below 600px.

        The "other" test has changed since in a table with multiple captions,
        now the first one wins. No visual difference in this case since both captions
        are the same.

        * tables/mozilla/marvin/backgr_layers-opacity-expected.txt:
        * tables/mozilla/marvin/backgr_position-table-expected.txt:
        * tables/mozilla/marvin/backgr_simple-table-cell-expected.txt:
        * tables/mozilla/marvin/backgr_simple-table-column-expected.txt:
        * tables/mozilla/marvin/backgr_simple-table-column-group-expected.txt:
        * tables/mozilla/marvin/backgr_simple-table-expected.txt:
        * tables/mozilla/marvin/backgr_simple-table-row-expected.txt:
        * tables/mozilla/marvin/backgr_simple-table-row-group-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_border-table-cell-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_border-table-column-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_border-table-column-group-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_border-table-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_border-table-quirks-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_border-table-row-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_border-table-row-group-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_fixed-bg-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_layers-hide-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_layers-show-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_position-table-cell-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_position-table-column-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_position-table-column-group-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_position-table-row-expected.txt:
        * tables/mozilla_expected_failures/marvin/backgr_position-table-row-group-expected.txt:
        * tables/mozilla_expected_failures/other/test4-expected.txt:


git-svn-id: svn://svn.chromium.org/blink/trunk@18791 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=12115
          RTL text is jumbled on display

        Fixed fixed table layout. Implemented an algorithm that matches the behavior
        of Firefox, up to rounding errors.

        Fixes three existing table tests.

        * rendering/FixedTableLayout.cpp:
        Undid accidental search-and-replace of "width" to "m_width" in the comments.
        Removed an outdated comment, since WebKit no longer honors table-layout:fixed
        for variable-width tables.
        (WebCore::FixedTableLayout::calcWidthArray): Fixed the logic that finds the
        first row. Changed to divide the width specified for a cell that spans multiple
        columns between the columns rather than assign its entire width to each column.
        (WebCore::FixedTableLayout::calcMinMaxWidth):
        (WebCore::FixedTableLayout::layout): Implemented the behavior observed in
        Firefox.
        * rendering/FixedTableLayout.h:

2007-01-11  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Hyatt.

        - updated results for http://bugs.webkit.org/show_bug.cgi?id=12115
          RTL text is jumbled on display
          (fixing fixed table layout)

        * tables/mozilla/bugs/bug157890-expected.checksum:
        * tables/mozilla/bugs/bug157890-expected.png:
        * tables/mozilla/bugs/bug157890-expected.txt:
        * tables/mozilla/bugs/bug44505-expected.checksum:
        * tables/mozilla/bugs/bug44505-expected.png:
        * tables/mozilla/bugs/bug44505-expected.txt:
        * tables/mozilla_expected_failures/bugs/bug7243-expected.checksum:
        * tables/mozilla_expected_failures/bugs/bug7243-expected.png:
        * tables/mozilla_expected_failures/bugs/bug7243-expected.txt:


git-svn-id: svn://svn.chromium.org/blink/trunk@18790 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Hyatt.

        - http://bugs.webkit.org/show_bug.cgi?id=10249
          Temporarily disable tests that are causing kernel panics

        Changed the ImageDiff tool not to use CoreImage filters.

        * DumpRenderTree/ImageDiff.m:
        (main):
        (createImageFromStdin):
        (compareImages):
        (getDifferenceBitmap):


git-svn-id: svn://svn.chromium.org/blink/trunk@18789 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=8438
          iExploder(#460): Assertion failure in RenderObject::drawBorder()

        Test: fast/block/basic/fieldset-stretch-to-legend.html

        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::calcWidth): Stretch to the minimum intrinsic
        width if the object requires it.
        (WebCore::RenderBox::calcHorizontalMargins): Some style cleanup.
        (WebCore::RenderBox::calcAbsoluteHorizontal): Stretch to the minimum
        intrinsic width if the object requires it.
        * rendering/RenderBox.h:
        (WebCore::RenderBox::stretchesToMinIntrinsicWidth): Added. Returns false.
        * rendering/RenderFieldset.cpp:
        (WebCore::RenderFieldset::calcMinMaxWidth): Overloaded RenderBlock's
        implementation to ensure that the legend's will fit in the fieldset's
        minimum intrinsic width.
        (WebCore::RenderFieldset::findLegend): Made const.
        (WebCore::RenderFieldset::paintBoxDecorations):
        (WebCore::RenderFieldset::paintBorderMinusLegend): Fixed the behavior when
        the legend does overflow the fieldset. This is the part of the fix that
        prevents the assertion failure. If the legend intersects with a corner of
        the fieldset's border, that corner will not be painted, even if it is not
        completely occluded. This edge case can be improved with clipping in the
        future. Added FIXME to implement border-radius.
        * rendering/RenderFieldset.h:
        (WebCore::RenderFieldset::stretchesToMinIntrinsicWidth): Overloaded to
        return true.

2007-01-11  Mitz Pettel  <mitz@webkit.org>

        Reviewed by Hyatt.

        - test for http://bugs.webkit.org/show_bug.cgi?id=8438
          iExploder(#460): Assertion failure in RenderObject::drawBorder()

        * fast/block/basic/fieldset-stretch-to-legend-expected.checksum: Added.
        * fast/block/basic/fieldset-stretch-to-legend-expected.png: Added.
        * fast/block/basic/fieldset-stretch-to-legend-expected.txt: Added.
        * fast/block/basic/fieldset-stretch-to-legend.html: Added.


git-svn-id: svn://svn.chromium.org/blink/trunk@18788 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Even more cleanup in preparation for fixing <rdar://problem/4608404> 
        WebScriptObject's _executionContext lack of ownership policy causes 
        crashes (e.g., in Dashcode)
        
        Layout tests pass.
        
        Renames:
            ReferencesSet | ProtectCounts => ProtectCountSet (because it's a typename for a set of GC protect counts)
            ReferencesByRootMap => RootObjectMap (because RootObjectToProtectCountSetMap would have been confusing)
            pv => protectedValues
            rootObjectForImp => getRootObject (overloaded for JSObject* and Interpreter*)
            rootObjectForInterpreter => getRootObject (ditto)
            findReferenceSet => getProtectCountSet
            imp => jsObject
        
        (KJS::Bindings::getRootObjectMap): Changed to take advantage of built-in
        facility for initializing static variables.

        (KJS::Bindings::getProtectCountSet):
        (KJS::Bindings::destroyProtectCountSet): Added. Helps encapsulate the fact
        that getting a ProtectCountSet entails adding a RootObject to a hash table,
        and destroying one entails the reverse.

        (KJS::Bindings::getRootObject): Removed spurious NULL check.
        
        (KJS::Bindings::findReferenceSet): Renamed. Changed to use getRootObject()
        instead of iterating on its own.

        (KJS::Bindings::addNativeReference): Changed to use an early return instead
        of indenting the whole function.
        (KJS::Bindings::removeNativeReference): Ditto.



git-svn-id: svn://svn.chromium.org/blink/trunk@18786 bbb929c8-8fbe-4397-9dbb-9b2b20218538