2009-01-27  Paul Godavari  <paul@chromium.org>

        Reviewed by Eric Seidel.

        https://bugs.webkit.org/show_bug.cgi?id=23561
        Explicitly send the MIME type from an image dragged from a web
        page to the desktop, to Chromium so that it can properly generate
        a file name.

        * platform/chromium/ChromiumDataObject.cpp:
        (WebCore::ChromiumDataObject::clear):
        (WebCore::ChromiumDataObject::hasData):
        * platform/chromium/ChromiumDataObject.h:
        * platform/chromium/ClipboardChromium.cpp:
        (WebCore::writeImageToDataObject):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40294 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-01-27  Mads Ager  <ager@chromium.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=23300
        Follow Firefox in throwing exceptions when using selection
        functions on HTMLInputElements that cannot have selection.

        This requires custom setSelectionRange, setSelectionStart and
        setSelectionEnd implementations, but it gets rid of the custom
        getOwnPropertySlot.

        * bindings/js/JSHTMLInputElementCustom.cpp:
        (WebCore::JSHTMLInputElement::selectionStart):
        (WebCore::JSHTMLInputElement::selectionEnd):
        (WebCore::JSHTMLInputElement::setSelectionRange):
        * html/HTMLInputElement.idl:

LayoutTests:

2009-01-27  Mads Ager  <ager@chromium.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=23300
        Update test of selection functions on HTMLInputElements to expect
        the selection functions to throw exceptions when used on buttons.
        This is consistent with the Firefox behavior.

        * fast/forms/selection-functions-expected.txt:
        * fast/forms/selection-functions.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40293 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-27  Evan Stade  <estade@chromium.org>

        Reviewed by Nikolas Zimmermann.

        https://bugs.webkit.org/show_bug.cgi?id=23435
        Make spreadMethod a member of Gradient rather than GraphicsContext.
        Implement spreadMethod for Skia.

        * platform/graphics/Gradient.cpp:
        (WebCore::Gradient::Gradient):
        * platform/graphics/Gradient.h:
        (WebCore::Gradient::setSpreadMethod):
        (WebCore::Gradient::spreadMethod):
        * platform/graphics/GraphicsContext.cpp:
        * platform/graphics/GraphicsContext.h:
        * platform/graphics/GraphicsContextPrivate.h:
        * platform/graphics/GraphicsTypes.h:
        (WebCore::):
        * platform/graphics/cairo/GradientCairo.cpp:
        (WebCore::Gradient::platformGradient):
        * platform/graphics/cairo/GraphicsContextCairo.cpp:
        (WebCore::GraphicsContext::fillPath):
        (WebCore::GraphicsContext::strokePath):
        * platform/graphics/qt/GradientQt.cpp:
        (WebCore::Gradient::platformGradient):
        * platform/graphics/qt/GraphicsContextQt.cpp:
        (WebCore::GraphicsContext::fillPath):
        (WebCore::GraphicsContext::strokePath):
        * platform/graphics/skia/GradientSkia.cpp:
        (WebCore::Gradient::platformGradient):
        * svg/SVGLinearGradientElement.cpp:
        (WebCore::SVGLinearGradientElement::buildGradient):
        * svg/SVGRadialGradientElement.cpp:
        (WebCore::SVGRadialGradientElement::buildGradient):
        * svg/graphics/SVGPaintServerGradient.cpp:
        (WebCore::SVGPaintServerGradient::SVGPaintServerGradient):
        (WebCore::SVGPaintServerGradient::setup):
        (WebCore::SVGPaintServerGradient::externalRepresentation):
        * svg/graphics/SVGPaintServerGradient.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40292 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-27  Ariya Hidayat  <ariya.hidayat@trolltech.com>

        Rubber-stamped by Simon Hausmann.

        http://www.qtsoftware.com/developer/task-tracker/index_html?id=238662&method=entry

        [Qt] Map function keys F1 to F24.

        * platform/qt/PlatformKeyboardEventQt.cpp:
        (WebCore::windowsKeyCodeForKeyEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40291 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-27  Ariya Hidayat  <ariya.hidayat@trolltech.com>

        Reviewed by Simon Hausmann.

        http://www.qtsoftware.com/developer/task-tracker/index_html?id=238391&method=entry

        [Qt] If QPainter fails to start on a QPrinter instance, do not
        continue printing.

        * Api/qwebframe.cpp:
        (QWebFrame::print):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40290 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-01-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Dan Bernstein

        Rework FrameLoaderClient to work on a CachedFrame basis instead of CachedPage

        * WebCore.base.exp:
        * history/CachedPage.cpp:
        * history/CachedPage.h:
        (WebCore::CachedPage::cachedMainFrame):

        * loader/EmptyClients.h:
        (WebCore::EmptyFrameLoaderClient::savePlatformDataToCachedFrame):
        (WebCore::EmptyFrameLoaderClient::transitionToCommittedFromCachedFrame):

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::restoreDocumentState):
        (WebCore::FrameLoader::commitProvisionalLoad):
        (WebCore::FrameLoader::transitionToCommitted):
        (WebCore::FrameLoader::cachePageForHistoryItem):
        * loader/FrameLoaderClient.h:

WebKit/gtk:

2009-01-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Dan Bernstein

        Rework FrameLoaderClient to work on a CachedFrame basis instead of CachedPage

        * WebCoreSupport/FrameLoaderClientGtk.cpp:
        (WebKit::FrameLoaderClient::savePlatformDataToCachedFrame):
        (WebKit::FrameLoaderClient::transitionToCommittedFromCachedFrame):
        * WebCoreSupport/FrameLoaderClientGtk.h:

WebKit/mac:

2009-01-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Dan Bernstein

        Rework FrameLoaderClient to work on a CachedFrame basis instead of CachedPage

        * History/WebHistoryItem.mm:

        * WebCoreSupport/WebFrameLoaderClient.h:
        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::savePlatformDataToCachedFrame):
        (WebFrameLoaderClient::transitionToCommittedFromCachedFrame):
        * WebKit.order:

WebKit/qt:

2009-01-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Dan Bernstein

        Rework FrameLoaderClient to work on a CachedFrame basis instead of CachedPage

        * WebCoreSupport/FrameLoaderClientQt.cpp:
        (WebCore::FrameLoaderClientQt::savePlatformDataToCachedFrame):
        (WebCore::FrameLoaderClientQt::transitionToCommittedFromCachedFrame):
        * WebCoreSupport/FrameLoaderClientQt.h:

WebKit/win:

2009-01-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Dan Bernstein

        Rework FrameLoaderClient to work on a CachedFrame basis instead of CachedPage

        * WebCoreSupport/WebFrameLoaderClient.cpp:
        (WebFrameLoaderClient::savePlatformDataToCachedFrame):
        (WebFrameLoaderClient::transitionToCommittedFromCachedFrame):
        * WebCoreSupport/WebFrameLoaderClient.h:

        * WebFrame.cpp:
        * WebFrame.h:

WebKit/wx:

2009-01-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Dan Bernstein

        Rework FrameLoaderClient to work on a CachedFrame basis instead of CachedPage

        * WebKitSupport/FrameLoaderClientWx.cpp:
        (WebCore::FrameLoaderClientWx::savePlatformDataToCachedFrame):
        (WebCore::FrameLoaderClientWx::transitionToCommittedFromCachedFrame):
        * WebKitSupport/FrameLoaderClientWx.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40289 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dave Hyatt

        https://bugs.webkit.org/show_bug.cgi?id=23360

        Add variants of all the methods that compute dirty rects for repainting
        with versions that compute those rects relative to some container
        RenderBox that acts as a painting root.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40288 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by David Hyatt

        Back out r40285, because it was checked in with no bug number, no
        testcase, is rendering change that did not get thorough review,
        and broke the Mac build.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40287 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin Adler

        -Make the recently added back/forward cache logging much better by actually
         walking the entire frame tree and indenting the resulting output.
        -Fix a null-termination bug in HistoryItem tree logging

        * history/HistoryItem.cpp:
        (WebCore::HistoryItem::showTreeWithIndent):

        * loader/FrameLoader.cpp:
        (WebCore::pageCacheLogPrefix):
        (WebCore::pageCacheLog):
        (WebCore::FrameLoader::logCanCachePageDecision):
        (WebCore::FrameLoader::logCanCacheFrameDecision):
        * loader/FrameLoader.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40286 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Allow option of hit testing frame content without clipping to the visible
viewport.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40285 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix misspelling of "lowercase" in font-face-locally-installed.html

        Reviewed by Dan Bernstein.

        * fast/css/font-face-locally-installed.html: Updated.
        * platform/mac/fast/css/font-face-locally-installed-expected.checksum: Updated.
        * platform/mac/fast/css/font-face-locally-installed-expected.png: Updated.
        * platform/mac/fast/css/font-face-locally-installed-expected.txt: Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40284 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Sam did _not_ review this

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40283 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Adele Peterson  <adele@apple.com>

        Build fix.

        * debugger/Debugger.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40282 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Anders Carlsson  <andersca@apple.com>

        Reviewed by Sam Weinig.

        Add the ability for plug-ins to make WebKit operate in "modal mode"

        * Plugins/Hosted/NetscapePluginHostProxy.h:
        * Plugins/Hosted/NetscapePluginHostProxy.mm:
        (WebKit::NetscapePluginHostProxy::NetscapePluginHostProxy):
        (WebKit::NetscapePluginHostProxy::pluginHostDied):
        If the plug-in crashes while we're modal, make sure to leave the modal mode.

        (WebKit::NetscapePluginHostProxy::setModal):
        (WKPCSetModal):
        * Plugins/Hosted/WebKitPluginClient.defs:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40281 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  John Sullivan  <sullivan@apple.com>

        fixed <rdar://problem/6530053> REGRESSION (Leopard): Shift-tab in http authentication window gets
        stuck in the Name field rather than cycling around

        Reviewed by Dan Bernstein

        * Panels/English.lproj/WebAuthenticationPanel.nib/designable.nib:
        * Panels/English.lproj/WebAuthenticationPanel.nib/keyedobjects.nib:
        The two static text fields and the last button all had their "next key view" outlets set to the
        name field, which caused shift-tab from the name field to do the wrong thing. Fixed by making each
        selectable view have exactly one "next key view" set to it.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40280 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Gavin Barraclough  <barraclough@apple.com>

        Reviewed by Darin Adler.

        Fixes for eq null & neq null, on 64-bit JIT.
        https://bugs.webkit.org/show_bug.cgi?id=23559

        This patch degrades 64-bit JIT performance on some benchmarks,
        due to the whole not-being-incorrect thing.

        * jit/JIT.cpp:
        (JSC::JIT::privateCompileMainPass):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40279 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-01-26  Sam Weinig  <sam@webkit.org>

        Reviewed by Dan Bernstein.

        Fix for https://bugs.webkit.org/show_bug.cgi?id=18141
        Acid3 tests 77 and 78 fail on reload due to SVG font race
        <rdar://problem/6504899>

        Match Opera's behavior of loading SVGFonts as soon as <font-face-uri>
        tag is inserted into the document and not at the first use of the font
        as is done for CSS web fonts.

        This fixes a race condition noticeable by some in the Acid3 test that
        happened because the onload event handler for an iframe containing a
        <font-face-uri> with an external reference was not delayed by the pending
        load of the font.

        Test: http/tests/misc/SVGFont-delayed-load.html

        * svg/SVGFontFaceUriElement.cpp:
        (WebCore::SVGFontFaceUriElement::~SVGFontFaceUriElement):
        (WebCore::SVGFontFaceUriElement::parseMappedAttribute): Call loadFont if
        the href attribute changes.
        (WebCore::SVGFontFaceUriElement::insertedIntoDocument): Call loadFont.
        (WebCore::SVGFontFaceUriElement::loadFont): Force a load of the
        font specified in href attribute.
        * svg/SVGFontFaceUriElement.h:

LayoutTests:

2009-01-26  Sam Weinig  <sam@webkit.org>

        Reviewed by Dan Bernstein.

        Test for https://bugs.webkit.org/show_bug.cgi?id=18141
        Acid3 tests 77 and 78 fail on reload due to SVG font race
        <rdar://problem/6504899>

        * http/tests/misc/SVGFont-delayed-load-expected.txt: Added.
        * http/tests/misc/SVGFont-delayed-load.html: Added.
        * http/tests/misc/resources/SVGFont-delayed-loader.svg: Added.
        * svg/custom/acid3-test-77.html:
        * svg/custom/resources/Acid3Font-loader.svg:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40278 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Anders Carlsson  <andersca@apple.com>

        Reviewed by Sam Weinig.

        Add the ability for a plug-in to show or hide the menu bar.

        * Plugins/Hosted/NetscapePluginHostProxy.h:
        * Plugins/Hosted/NetscapePluginHostProxy.mm:
        (WebKit::NetscapePluginHostProxy::NetscapePluginHostProxy):
        (WebKit::NetscapePluginHostProxy::pluginHostDied):
        (WebKit::NetscapePluginHostProxy::setMenuBarVisible):
        (WKPCSetMenuBarVisible):
        * Plugins/Hosted/WebKitPluginClient.defs:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40277 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Pierre-Olivier Latour  <pol@apple.com>

        Updated the reference images for the video tests that rely on the H264 test movie,
        as QuickTime 7.6 apparently slightly changed the results from the 7.5.5 version.

        * platform/mac/media/video-aspect-ratio-expected.checksum:
        * platform/mac/media/video-aspect-ratio-expected.png:
        * platform/mac/media/video-controls-rendering-expected.checksum:
        * platform/mac/media/video-controls-rendering-expected.png:
        * platform/mac/media/video-display-toggle-expected.checksum:
        * platform/mac/media/video-display-toggle-expected.png:
        * platform/mac/media/video-layer-crash-expected.checksum:
        * platform/mac/media/video-layer-crash-expected.png:
        * platform/mac/media/video-transformed-expected.checksum:
        * platform/mac/media/video-transformed-expected.png:
        * platform/mac/media/video-zoom-expected.checksum:
        * platform/mac/media/video-zoom-expected.png:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40276 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Change Web Archive serialization back to the binary property list format.

        <rdar://problem/6127865> Web archives saved with Safari 4 cannot be read by Safari 3

        Reviewed by Brady Eidson.

        * loader/archive/cf/LegacyWebArchive.cpp:
        (WebCore::LegacyWebArchive::rawDataRepresentation):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40275 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Cameron Zwarich  <cwzwarich@uwaterloo.ca>

        Reviewed by Gavin Barraclough.

        Bug 23552: Dashcode evaluator no longer works after making ExecStates actual call frames
        <https://bugs.webkit.org/show_bug.cgi?id=23552>
        <rdar://problem/6398839>

        Dashcode will crash when using the evaluator because it saves a global call
        frame, even after global code has finished executing, and then uses this as
        a launching pad to execute new JS in the evaluator. The fix is to detect
        when Dashcode is attempting to do this and execute code from a global call
        frame instead.

        JavaScriptCore:

        * JavaScriptCore.exp:
        * debugger/Debugger.cpp:
        (JSC::evaluateInGlobalCallFrame): Added so that WebScriptCallFrame can
        evaluate JS starting from a global call frame.
        * debugger/Debugger.h:

        WebKit/mac:

        * ForwardingHeaders/runtime/Protect.h: Added.
        * WebView/WebScriptDebugDelegate.mm:
        (-[WebScriptCallFrame _initWithGlobalObject:debugger:caller:debuggerCallFrame:]):
        Added debugger, a WebScriptDebugger* argument.
        (-[WebScriptCallFrame evaluateWebScript:]): Detect when Dashcode is using
        a stale WebScriptCallFrame to execute new JS and evaluate it starting from
        the global object's global call frame instead.
        * WebView/WebScriptDebugger.h:
        (WebScriptDebugger::globalObject): Added.
        (WebScriptDebugger::globalCallFrame): Added.
        * WebView/WebScriptDebugger.mm:
        (WebScriptDebugger::WebScriptDebugger): Initialize m_globalObject.
        (WebScriptDebugger::initGlobalCallFrame): Created as a clone of callEvent
        so that the global call frame can be saved immediately after being created.
        (WebScriptDebugger::callEvent): Pass 'this' as the debugger argument of
        WebScriptCallFrame's _initWithGlobalObject method.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40274 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Anders Carlsson  <andersca@apple.com>

        Reviewed by Oliver Hunt.

        Make WKPCInvoke a simpleroutine.

        * Plugins/Hosted/NetscapePluginHostProxy.mm:
        (WKPCInvoke):
        * Plugins/Hosted/WebKitPluginClient.defs:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40273 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Anders Carlsson  <andersca@apple.com>

        Fix 64-bit build.

        * WebCore.LP64.exp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40272 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Dmitry Titov  <dimich@chromium.org>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=23511
        Move DOMTimer from bindings/js to page/.
        This was prepared by moving all JSC dependencies from DOMTimer
        to ScheduledAction earlier, and helps unfork the Chromuim version
        of the DOMTimer.

        * GNUmakefile.am:
        * WebCore.pro:
        * WebCore.scons:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:
        * bindings/js/DOMTimer.cpp: Removed.
        * bindings/js/DOMTimer.h: Removed.
        * page/DOMTimer.cpp: Copied from WebCore/bindings/js/DOMTimer.cpp.
        * page/DOMTimer.h: Copied from WebCore/bindings/js/DOMTimer.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40271 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  David Hyatt  <hyatt@apple.com>

        Eliminate rightmostPosition, lowestPosition and leftmostPosition from RenderFlow.  Move the code into
        RenderBlock's methods.

        Reviewed by Eric Seidel

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::lowestPosition):
        (WebCore::RenderBlock::rightmostPosition):
        (WebCore::RenderBlock::leftmostPosition):
        * rendering/RenderFlow.cpp:
        * rendering/RenderFlow.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40270 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Darin Fisher  <darin@chromium.org>

        Reviewed by Eric Seidel.

        https://bugs.webkit.org/show_bug.cgi?id=23553
        Upstream changes from jcampan to fix chromium bugs 6692 and 6694.

        * platform/chromium/PopupMenuChromium.cpp:
        (WebCore::PopupListBox::setLoopSelectionNavigation):
        (WebCore::PopupListBox::PopupListBox):
        (WebCore::PopupContainer::setLoopSelectionNavigation):
        (WebCore::PopupListBox::handleKeyEvent):
        (WebCore::PopupListBox::paintRow):
        (WebCore::PopupListBox::clearSelection):
        (WebCore::PopupListBox::selectNextRow):
        (WebCore::PopupListBox::selectPreviousRow):
        * platform/chromium/PopupMenuChromium.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40269 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove extraneous WebCore XCode project setting, introduced in previous commit.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40268 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Brady Eidson  <beidson@apple.com>

        Reviewed by Sam Weinig

        Remove some historical CachedPage time stamp fluff.

        * history/CachedPage.cpp:
        (WebCore::CachedPage::CachedPage):
        * history/CachedPage.h:
        (WebCore::CachedPage::timeStamp):

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40267 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  David Hyatt  <hyatt@apple.com>

        Move absoluteClippedOverflowRect, linesBoundingBox and borderBoundingBox from RenderFlow to RenderInline.

        Reviewed by Eric Seidel

        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::containingBlockHeightForPositioned):
        * rendering/RenderFlow.cpp:
        * rendering/RenderFlow.h:
        * rendering/RenderInline.cpp:
        (WebCore::RenderInline::linesBoundingBox):
        (WebCore::RenderInline::absoluteClippedOverflowRect):
        * rendering/RenderInline.h:
        (WebCore::RenderInline::borderBoundingBox):
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::updateLayerPosition):
        * rendering/RenderTreeAsText.cpp:
        (WebCore::operator<<):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40266 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Eric Seidel.

        https://bugs.webkit.org/show_bug.cgi?id=22650
        Implement support for "Immutable" attribute to CodeGeneratorJS.pm, thus
        allowing to explicitly mark methods as immutable (i.e. returning a new
        instance rather than a reference to an existing instance).

        * bindings/js/JSSVGMatrixCustom.cpp: Removed custom methods.
        * bindings/scripts/CodeGeneratorJS.pm: Added check for Immutable
            attribute when generating function call.
        * svg/SVGMatrix.idl: Replaced Custom attribute with Immutable on
            methods.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40265 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  David Hyatt  <hyatt@apple.com>

        Eliminate isInlineFlow() now that a RenderBlock can never be an inline flow.  isInlineFlow() is now
        equivalent to isRenderInline(), so all calls to isInlineFlow() have been replaced with isRenderInline().

        Reviewed by Eric Seidel

        * dom/Text.cpp:
        (WebCore::Text::rendererIsNeeded):
        * rendering/InlineFlowBox.cpp:
        (WebCore::InlineFlowBox::placeBoxesHorizontally):
        (WebCore::shouldDrawTextDecoration):
        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::hitTestContents):
        (WebCore::InlineMinMaxIterator::next):
        (WebCore::RenderBlock::calcInlinePrefWidths):
        * rendering/RenderBlock.h:
        (WebCore::RenderBlock::isBlockFlow):
        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::clientWidth):
        (WebCore::RenderBox::clientHeight):
        (WebCore::RenderBox::nodeAtPoint):
        (WebCore::RenderBox::imageChanged):
        (WebCore::RenderBox::offsetForPositionedInContainer):
        (WebCore::RenderBox::containingBlockWidthForPositioned):
        (WebCore::RenderBox::containingBlockHeightForPositioned):
        * rendering/RenderBox.h:
        (WebCore::RenderBox::width):
        (WebCore::RenderBox::height):
        (WebCore::RenderBox::size):
        (WebCore::RenderBox::frameRect):
        * rendering/RenderContainer.cpp:
        (WebCore::RenderContainer::updateBeforeAfterContentForContainer):
        * rendering/RenderFlow.cpp:
        (WebCore::RenderFlow::dirtyLinesFromChangedChild):
        (WebCore::RenderFlow::createInlineBox):
        (WebCore::RenderFlow::paintLines):
        (WebCore::RenderFlow::hitTestLines):
        (WebCore::RenderFlow::absoluteClippedOverflowRect):
        (WebCore::RenderFlow::lowestPosition):
        (WebCore::RenderFlow::rightmostPosition):
        (WebCore::RenderFlow::leftmostPosition):
        * rendering/RenderFlow.h:
        (WebCore::RenderFlow::borderBoundingBox):
        * rendering/RenderInline.h:
        (WebCore::RenderInline::isRenderInline):
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::updateLayerPosition):
        (WebCore::RenderLayer::intersectsDamageRect):
        (WebCore::RenderLayer::boundingBox):
        * rendering/RenderListItem.cpp:
        (WebCore::getParentOfFirstLineBox):
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::getAbsoluteRepaintRectWithOutline):
        (WebCore::RenderObject::styleWillChange):
        (WebCore::RenderObject::firstLineStyle):
        * rendering/RenderObject.h:
        * rendering/RenderSVGInline.cpp:
        (WebCore::RenderSVGInline::createInlineBox):
        * rendering/RenderSVGText.cpp:
        (WebCore::RenderSVGText::createInlineBox):
        * rendering/RenderTableRow.cpp:
        (WebCore::RenderTableRow::nodeAtPoint):
        * rendering/RenderTableSection.cpp:
        (WebCore::RenderTableSection::nodeAtPoint):
        * rendering/RenderText.cpp:
        (WebCore::isInlineFlowOrEmptyText):
        * rendering/RenderTreeAsText.cpp:
        (WebCore::operator<<):
        * rendering/bidi.cpp:
        (WebCore::bidiNext):
        (WebCore::bidiFirst):
        (WebCore::appendRunsForObject):
        (WebCore::RenderBlock::createLineBoxes):
        (WebCore::RenderBlock::computeHorizontalPositionsForLine):
        (WebCore::RenderBlock::layoutInlineChildren):
        (WebCore::inlineFlowRequiresLineBox):
        (WebCore::requiresLineBox):
        (WebCore::RenderBlock::skipTrailingWhitespace):
        (WebCore::RenderBlock::skipLeadingWhitespace):
        (WebCore::RenderBlock::findNextLineBreak):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40264 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Anders Carlsson  <andersca@apple.com>

        Reviewed by Sam Weinig.

        Implement using plug-in objects as constructors, and setting and getting properties from a plug-in object.

        * Plugins/Hosted/NetscapePluginHostProxy.mm:
        (WKPCBooleanAndDataReply):
        * Plugins/Hosted/NetscapePluginInstanceProxy.h:
        (WebKit::NetscapePluginInstanceProxy::Reply::):
        (WebKit::NetscapePluginInstanceProxy::BooleanAndDataReply::BooleanAndDataReply):
        Rename NPObjectInvokeReply to BooleanAndDataReply.

        * Plugins/Hosted/NetscapePluginInstanceProxy.mm:
        (WebKit::NetscapePluginInstanceProxy::addValueToArray):
        Fix a cut and paste error.

        (WebKit::NetscapePluginInstanceProxy::demarshalValueFromArray):
        Handle NPObjects.

        * Plugins/Hosted/ProxyInstance.h:
        * Plugins/Hosted/ProxyInstance.mm:
        (WebKit::ProxyField::valueFromInstance):
        (WebKit::ProxyField::setValueToInstance):
        Call the ProxyInstance method.

        (WebKit::ProxyInstance::~ProxyInstance):
        Release the NPObject.

        (WebKit::ProxyInstance::supportsConstruct):
        Ask the plug-in host if an instance supports construct.

        (WebKit::ProxyInstance::fieldValue):
        (WebKit::ProxyInstance::setFieldValue):
        Call the plug-in host methods.

        * Plugins/Hosted/WebKitPluginHostTypes.h:
        Rename ObjectValueType to JSObjectValueType, and add NPObjectValueType.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40263 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Land new layout tests now that display:compact support has been removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40262 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Land new layout tests now that display:compact support has been removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40261 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Land new layout tests now that display:compact support has been removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40260 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Land new layout tests now that display:compact support has been removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40259 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  David Hyatt  <hyatt@apple.com>

        Drop support for display:compact.

        (1) display:compact is no longer part of CSS2.1.
        (2) Shipping Safari has completely broken compact support with severely broken rendering
        (all of the compact text just piles on top of the text in the inserted block and on top of itself as well).
        (3) This has been broken in WebKit for years (ever since blocks converted to painting lines directly) without
        anyone even noticing.

        Reviewed by Eric Seidel

        * rendering/InlineFlowBox.cpp:
        (WebCore::InlineFlowBox::placeBoxesHorizontally):
        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::handleSpecialChild):
        (WebCore::RenderBlock::handleRunInChild):
        (WebCore::RenderBlock::layoutBlockChildren):
        (WebCore::RenderBlock::paint):
        (WebCore::RenderBlock::paintObject):
        (WebCore::RenderBlock::nodeAtPoint):
        (WebCore::RenderBlock::positionForCoordinates):
        (WebCore::shouldCheckLines):
        (WebCore::getHeightForLineCount):
        (WebCore::RenderBlock::renderName):
        * rendering/RenderBlock.h:
        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::sizesToIntrinsicWidth):
        * rendering/RenderBox.h:
        (WebCore::RenderBox::width):
        (WebCore::RenderBox::height):
        (WebCore::RenderBox::size):
        (WebCore::RenderBox::frameRect):
        * rendering/RenderFlow.cpp:
        (WebCore::RenderFlow::paintLines):
        (WebCore::RenderFlow::hitTestLines):
        (WebCore::RenderFlow::absoluteClippedOverflowRect):
        (WebCore::RenderFlow::linesBoundingBox):
        * rendering/RenderObject.h:
        (WebCore::RenderObject::isInline):
        * rendering/bidi.cpp:
        (WebCore::RenderBlock::computeHorizontalPositionsForLine):
        (WebCore::RenderBlock::layoutInlineChildren):
        * rendering/bidi.h:
        (WebCore::BidiRun::BidiRun):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40258 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the build.

Remove -Wformat=2 from the warning flags as newer versions of GCC emit
warnings about non-literal format strings for uses of our UI_STRING macro.

* Configurations/Base.xcconfig:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40257 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Pierre-Olivier Latour  <pol@apple.com>

        Tweaked again earlier fix, this time just to print a warning and not abort if
        attempting to generate pixel results and Perian is installed.

        https://bugs.webkit.org/show_bug.cgi?id=22615

        * Scripts/run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40256 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Adam Langley  <agl@google.com>

        Reviewed by Nikolas Zimmermann.

        gtk2drawing: fix compiler warnings

        g_object_set_data expects a gpointer as the last argument. Cast the
          TRUE/FALSE values to avoid a warning.
        gtk_tree_view_column_new doesn't return a GtkWidget.

        * platform/gtk/gtk2drawing.c:
        (setup_widget_prototype):
        (moz_gtk_get_combo_box_inner_button):
        (moz_gtk_get_combo_box_button_inner_widgets):
        (ensure_combo_box_widgets):
        (moz_gtk_get_combo_box_entry_inner_widgets):
        (moz_gtk_get_combo_box_entry_arrow):
        (ensure_combo_box_entry_widgets):
        (ensure_toolbar_widget):
        (ensure_menu_bar_item_widget):
        (ensure_menu_popup_widget):
        (ensure_menu_item_widget):
        (ensure_image_menu_item_widget):
        (ensure_menu_separator_widget):
        (ensure_check_menu_item_widget):
        (ensure_tree_header_cell_widget):
        (moz_gtk_entry_paint):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40255 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  David Hyatt  <hyatt@apple.com>

        Change display:run-in handling so that we no longer reuse RenderBlock as though it is an inline flow.
        Instead we create a new RenderInline, move the RenderBlock's children underneath it, and then set the
        inline as the new element's renderer.

        Reviewed by Dan Bernstein

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::handleRunInChild):
        * rendering/RenderInline.cpp:
        (WebCore::RenderInline::renderName):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40254 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Clean up after r40240.

Rubber-stamped by Sam Weinig.

* WebKit.xcodeproj/project.pbxproj: Reinstate WebKitPluginHost.defs.

* Configurations/Base.xcconfig: Don't dead code strip in debug builds for now as it leads to link errors.
* Plugins/Hosted/HostedNetscapePluginStream.mm: Revert change that is no longer needed now that WebKitPluginHost.defs
is back in the build.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40253 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix an assertion failure when Safari loads an error page

        Reviewed by Alexey Proskuryakov.

        * WebFrame.cpp:
        (WebFrame::loadData): Use the two-argument version of KURL and
        MarshallingHelpers::BSTRToKURL, for reasons mentioned in r40248.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40252 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Christian Dywan  <christian@twotoasts.de>

        Rubber stamped by Holger Freyther.

        * GtkLauncher/main.c:
        (main): Initialize threads, which is required for libSoup.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40251 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  David Boddie  <dboddie@trolltech.com>

        Reviewed by Simon Hausmann.

        Doc: qdoc fix for a link to a property and added a basic description of the offline caching and storage features from HTML 5.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40250 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=23480
        Add missing method stubs to BackForwardListChromium.

        * history/BackForwardListChromium.cpp:
        (WebCore::BackForwardList::backItem):
        (WebCore::BackForwardList::forwardItem):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40249 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix an assertion on launch in KURL::KURL

        All URL strings passed in to the WebKit API need to be passed to
        MarshallingHelpers::BSTRToKURL so that they may be properly parsed.

        Reviewed by Alexey Proskuryakov.

        * MarshallingHelpers.cpp:
        (MarshallingHelpers::BSTRToKURL): Changed to use the two-argument
        constructor for KURL so that the strings will be parsed rather than
        assumed to be in the correct encoding/form.

        * WebResource.cpp:
        (WebResource::initWithData):
        * WebURLResponse.cpp:
        (WebURLResponse::initWithURL):
        * WebView.cpp:
        (WebView::userAgentForURL):
        (WebView::copyURL):
        Changed to use MarshallingHelpers::BSTRToKURL instead of trying to do
        the work manually.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40248 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-26  Darin Adler  <darin@apple.com>

        Requested by David Hyatt.

        * platform/mac/fast/table/form-with-table-style-expected.checksum: Updated to reflect
        the progression; we now render things we used to fail to render.
        * platform/mac/fast/table/form-with-table-style-expected.png: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40247 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Darin Adler  <darin@apple.com>

        Try to fix GTK build.

        * webkit/webkitwebview.cpp: Added include of FloatQuad.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40246 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKit:

2009-01-25  Darin Adler  <darin@apple.com>

        Try to fix full build on Mac Leopard.

        * WebKit.xcodeproj/project.pbxproj: Removed WebKitPluginClient.defs from the WebKit target.
        Maybe this is temporary, but I had to do it to get the full build to succeed.

WebKit/mac:

2009-01-25  Darin Adler  <darin@apple.com>

        * Plugins/Hosted/HostedNetscapePluginStream.mm: Added a missing extern "C".

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40245 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Darin Adler  <darin@apple.com>

        Try to fix Windows build.

        * WebView.cpp: Added FloatQuad.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40244 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Darin Adler  <darin@apple.com>

        Try to fix Tiger buildbot.

        * WebCore.xcodeproj/project.pbxproj: Turn off cast-qual for WebDashboardRegion.m.
        * page/mac/WebDashboardRegion.m:
        (typeName): Added.
        (-[WebDashboardRegion description]): Use typeName.
        (-[WebDashboardRegion isEqual:]): Added comment.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40243 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=23535
        Strengthen debug checks in KURL

        Covered (and prompted) by existing tests.

        * platform/KURL.cpp:
        (WebCore::checkEncodedString): Check that the first character of the URL is an allowed
        first scheme character, not just that it isn't a slash.
        (WebCore::KURL::KURL): Assert that parsing didn't affect the string (which must be the case
        because of parsing being idempotent).

        * page/SecurityOrigin.cpp:
        (WebCore::SecurityOrigin::create): Added a check for the URL being valid - if it is not, an
        empty security origin is created.
        (WebCore::SecurityOrigin::createFromString): Parse the string as URL, do not just convert
        it - it is not guaranteed to be properly encoded.

        * dom/Document.cpp: (WebCore::Document::updateBaseURL): As documentURI is an arbitrary
        string, it needs to be parsed into an URL, not just converted into one.

        * html/CanvasRenderingContext2D.cpp: (WebCore::CanvasRenderingContext2D::createPattern):
        Create the security origin with createFromString() instead of create(). I do not know where
        the string comes from and whether it might not be properly encoded, but this change can't
        hurt, and makes the code slightly more clear.

        * xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::accessControlCheck): Create the security
        origin with createFromString() instead of create(). The string comes from an HTTP response
        header, so it may not be a properly encoded URL.

        * loader/DocLoader.cpp: (WebCore::DocLoader::requestResource): Do not try to request using
        invalid URLs.

        * page/DOMWindow.cpp: (WebCore::DOMWindow::postMessage): Updated to call
        SecurityOrigin::createFromString() (no change in behavior).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40242 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Darin Adler  <darin@apple.com>

        Discussed with Mark Rowe; not sure he reviewed it.

        * Configurations/Base.xcconfig: Add all the same warnings as in WebCore except for
        -Wcast-qual and -Wunused-parameter, which both need to be off at least for now.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40241 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Improve the consistency of settings in our .xcconfig files.

Rubber-stamped by Dan Bernstein.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40240 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Land table/forms layout test progression.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40239 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-01-25  David Hyatt  <hyatt@apple.com>

        Fix for https://bugs.webkit.org/show_bug.cgi?id=23524, lots of missing content in table sections.

        The new table code created a bug involving  markAllDescendantsWithFloatsForLayout, namely that it could
        end up marking ancestors of a block as needing layout when that block was still in the process of
        doing a layout.

        The fix is to add a parameter to markAllDescendantsWithFloatsForLayout that says whether or not
        we are "mid-layout."  If this flag is set, then the method will make sure to do only local dirtying
        of objects to avoid accidentally marking a clean ancestor as needing layout again.

        Ultimately the second parameter to setNeedsLayout and setChildNeedsLayout should just be removed,
        with a check of whether or not we are mid-layout being done by those methods instead.

        Reviewed by Oliver Hunt

        Added fast/repaint/dynamic-table-vertical-alignment-change.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::collapseMargins):
        (WebCore::RenderBlock::clearFloatsIfNeeded):
        (WebCore::RenderBlock::layoutBlockChildren):
        (WebCore::RenderBlock::markAllDescendantsWithFloatsForLayout):
        * rendering/RenderBlock.h:
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::removeFromObjectLists):
        * rendering/RenderObject.h:
        * rendering/RenderTableSection.cpp:
        (WebCore::RenderTableSection::layoutRows):

LayoutTests:

2009-01-25  David Hyatt  <hyatt@apple.com>

        Add layout test for https://bugs.webkit.org/show_bug.cgi?id=23524.

        Reviewed by Oliver Hunt

        * fast/repaint/dynamic-table-vertical-alignment-change.html: Added.
        * platform/mac/fast/repaint/dynamic-table-vertical-alignment-change-expected.checksum: Added.
        * platform/mac/fast/repaint/dynamic-table-vertical-alignment-change-expected.png: Added.
        * platform/mac/fast/repaint/dynamic-table-vertical-alignment-change-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40238 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

2009-01-25  Darin Adler  <darin@apple.com>

        Reviewed by Mark Rowe.

        Bug 23352: Turn on more compiler warnings in the Mac build
        https://bugs.webkit.org/show_bug.cgi?id=23352

        Turn on the following warnings:

            -Wcast-qual
            -Wextra-tokens
            -Wformat=2
            -Winit-self
            -Wmissing-noreturn
            -Wpacked
            -Wrendundant-decls

        * Configurations/Base.xcconfig: Added the new warnings. Switched to -Wextra instead of
        -W for clarity since we don't have to support the older versions of gcc that require the
        old -W syntax. Since we now use -Wformat=2, removed -Wformat-security. Also removed
        -Wno-format-y2k since we can have that one on now.

WebCore:

2009-01-25  Darin Adler  <darin@apple.com>

        Reviewed by Mark Rowe.

        Bug 23352: Turn on more compiler warnings in the Mac build
        https://bugs.webkit.org/show_bug.cgi?id=23352

        Turn on the following warnings:

            -Wcast-qual
            -Wextra-tokens
            -Wformat=2
            -Winit-self
            -Wmissing-noreturn
            -Wpacked
            -Wrendundant-decls

        * Configurations/Base.xcconfig: Added the new warnings. Switched to -Wextra instead of
        -W for clarity since we don't have to support the older versions of gcc that require the
        old -W syntax. Since we now use -Wformat=2, removed -Wformat-security. Also removed
        -Wno-format-y2k since we can have that one on now.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40237 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Oliver Hunt.

        - fix <rdar://problem/6525618> REGRESSION (r39566): NewsFire crashes in
          string hashing code when opening a Digg RSS article

        Test: fast/loader/null-request-after-willSendRequest.html

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::loadItem): Null-check the request returned from
        the client's willSendRequest method.

LayoutTests:

        Reviewed by Oliver Hunt.

        - test for <rdar://problem/6525618> REGRESSION (r39566): NewsFire crashes in
          string hashing code when opening a Digg RSS article

        * fast/loader/null-request-after-willSendRequest-expected.txt: Added.
        * fast/loader/null-request-after-willSendRequest.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40236 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  David Hyatt  <hyatt@apple.com>

        Fix for <rdar://problem/6525392> REGRESSION (r40180): Scroll wheel events over text scroll the page
        instead of the ancestor overflow (23525)

        Make sure that we always call scroll() on the nearest enclosing RenderBox.  No test case is possible
        here, since wheel scrolling is currently untestable (the offset varies from machine to machine).

        Reviewed by Dan Bernstein

        * page/EventHandler.cpp:
        (WebCore::scrollAndAcceptEvent):
        (WebCore::EventHandler::scrollOverflow):
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::enclosingBox):
        * rendering/RenderObject.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40235 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Darin Adler  <darin@apple.com>

        Reviewed by Dan Bernstein.

        Bug 23538: REGRESSION (r39969): Garbage text instead of blank content rendered when plug-ins are disabled
        https://bugs.webkit.org/show_bug.cgi?id=23538
        rdar://problem/6523719

        I'm not sure how to make a regression test for this, since it depends on plug-ins being disabled.

        The main problem here is that when plug-ins were disabled, the code ignored the classId attribute and
        the <embed> element entirely. That resulted in a page treating a Flash document as plain HTML rather
        than doing fallback.

        * rendering/RenderPartObject.cpp:
        (WebCore::createClassIdToTypeMap): Added. Broke this out into a separate function so we could get rid of an
        inelegant if statement.
        (WebCore::activeXType): Added. To avoid repeating the MIME type for ActiveX and possibly having a typo.
        (WebCore::havePlugin): Added. Helper function to make sure we don't forget the null check.
        (WebCore::serviceTypeForClassId): Give this function a return value since that's a more natural way to
        return a string than an "out" parameter.  Fixed the logic to only prefer the ActiveX type over the
        type guessed from the classId when there actually is an ActiveX plug-in to use. The old function assumed
        there was one, which I presume right for Chrome on Windows when plug-ins are enabled, but wrong in many
        other cases, and wrong all the time for all clients on Mac. We don't want to assume either way. Use the
        new havePlugin function so we handle the case where pluginData is 0.
        (WebCore::shouldUseEmbedDescendant): Renamed. Simplified the comment. Changed to use serviceTypeForClassId
        instead of the old version that used an out parameter. Always use the <embed> if there isn't a plug-in
        that knows how to handle <object> elements.
        (WebCore::RenderPartObject::updateWidget): Removed null checks of pluginData. The two functions that
        use this data still need to be called; they still do something even if no plug-ins are present. And
        they have now been corrected to handle 0 properly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40234 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Darin Adler  <darin@apple.com>

        * rendering/RenderMenuList.cpp: Fix build by adding back needed include.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40233 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Judit Jasz  <jasy@inf.u-szeged.hu>

        Reviewed by Darin Adler.

        Compilation problem fixing
        http://bugs.webkit.org/show_bug.cgi?id=23497

        * jit/JITCall.cpp:
        (JSC::JIT::compileOpCall): Use JSValuePtr::encode.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40232 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

2009-01-25  Darin Adler  <darin@apple.com>

        Reviewed by Sam Weinig.

        Bug 23352: Turn on more compiler warnings in the Mac build
        https://bugs.webkit.org/show_bug.cgi?id=23352

        Fourth patch: Deal with the last few stray warnings.

        * parser/Parser.cpp: Only declare jscyyparse if it's not already declared.
        This makes both separate compilation and all-in-one compilation work with the
        -Wredundant-decls warning.

WebCore:

2009-01-25  Darin Adler  <darin@apple.com>

        Reviewed by Sam Weinig.

        Bug 23352: Turn on more compiler warnings in the Mac build
        https://bugs.webkit.org/show_bug.cgi?id=23352

        Fourth patch: Deal with the last few stray warnings.

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::firstLineBlock): Handle the const_cast a little better.
        Also make the local variable be a RenderBlock so the cast is closer to the code
        that guarantees it's correct.

        * rendering/SVGRenderTreeAsText.cpp:
        (WebCore::writeStyle): Put the enum into a local variable that's a built-in
        type before using it with TextStream, so we don't get a warning due to -Wsign-promo.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40231 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Darin Adler  <darin@apple.com>

        Reviewed by Sam Weinig.

        Bug 23352: Turn on more compiler warnings in the Mac build
        https://bugs.webkit.org/show_bug.cgi?id=23352

        Third patch: Use the noreturn attribute on functions that don't
        return to prepare for the use of the -Wmissing-noreturn warning.

        * jit/JITCall.cpp:
        (JSC::unreachable): Added NO_RETURN.
        * jsc.cpp:
        (functionQuit): Ditto.
        (printUsageStatement): Ditto.
        * wtf/AlwaysInline.h: Added definition of NO_RETURN.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40230 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-01-25  Darin Adler  <darin@apple.com>

        Reviewed by Sam Weinig.

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

        Step one: RenderText.

        Also change set of RenderFlow* to be a ListHashSet as the comment in the
        header indicated.

        * dom/CharacterData.cpp:
        (WebCore::CharacterData::setData): Use toRenderText.
        (WebCore::CharacterData::appendData): Ditto.
        (WebCore::CharacterData::insertData): Ditto.
        (WebCore::CharacterData::deleteData): Ditto.
        (WebCore::CharacterData::replaceData): Ditto.
        * dom/ContainerNode.cpp:
        (WebCore::ContainerNode::getUpperLeftCorner): Ditto.
        (WebCore::ContainerNode::getLowerRightCorner): Ditto.
        * dom/Position.cpp:
        (WebCore::nextRenderedEditable): Ditto.
        (WebCore::previousRenderedEditable): Ditto.
        (WebCore::Position::renderedOffset): Ditto. Also use function members
        instead of data members of InlineTextBox.
        (WebCore::Position::upstream): Ditto.
        (WebCore::Position::downstream): Ditto.
        (WebCore::Position::hasRenderedNonAnonymousDescendantsWithHeight): Ditto.
        (WebCore::Position::inRenderedText): Ditto.
        (WebCore::Position::isRenderedCharacter): Ditto.
        (WebCore::searchAheadForBetterMatch): Ditto.
        (WebCore::Position::getInlineBoxAndOffset): Ditto.

        * dom/Range.cpp: Added include of VisiblePosition.h now that it's no longer
        included in RenderObject.h.

        * dom/Text.cpp:
        (WebCore::Text::splitText): Use toRenderText.
        (WebCore::Text::recalcStyle): Ditto.
        * editing/CompositeEditCommand.cpp:
        (WebCore::CompositeEditCommand::deleteInsignificantText): Ditto.
        * editing/SelectionController.cpp:
        (WebCore::SelectionController::debugRenderer): Ditto.
        * editing/TextIterator.cpp:
        (WebCore::TextIterator::handleTextNode): Ditto.
        (WebCore::TextIterator::handleTextBox): Ditto.
        (WebCore::TextIterator::emitText): Ditto.
        (WebCore::SimplifiedBackwardsTextIterator::handleTextNode): Ditto.

        * editing/VisiblePosition.cpp: Added include of FloatQuad.h now that it's
        no longer included in RenderObject.h.
        * inspector/InspectorController.cpp: Ditto.
        * page/Frame.cpp: Ditto.
        * rendering/RenderBox.cpp: Ditto.
        * rendering/RenderInline.cpp: Ditto.
        * rendering/RenderPath.cpp: Ditto.
        * rendering/RenderSVGContainer.cpp: Ditto.
        * rendering/RenderSVGImage.cpp: Ditto.
        * rendering/RenderSVGInlineText.cpp: Ditto.
        * rendering/RenderSVGTSpan.cpp: Ditto.
        * rendering/RenderSVGText.cpp: Ditto.
        * rendering/RenderSVGTextPath.cpp: Ditto.
        * rendering/RenderTableCell.cpp: Ditto.
        * rendering/RenderView.cpp: Ditto.

        * editing/visible_units.cpp:
        (WebCore::startPositionForLine): Use function members instead of data
        members of InlineTextBox.
        (WebCore::endPositionForLine): Ditto.
        (WebCore::startOfParagraph): Use toRenderText.
        (WebCore::endOfParagraph): Ditto.

        * page/AccessibilityRenderObject.cpp:
        (WebCore::AccessibilityRenderObject::accessibilityIsIgnored): Use toRenderText.
        * rendering/InlineFlowBox.cpp:
        (WebCore::InlineFlowBox::placeBoxesHorizontally): Ditto.
        * rendering/InlineTextBox.cpp:
        (WebCore::InlineTextBox::deleteLine): Ditto.
        (WebCore::InlineTextBox::extractLine): Ditto.
        (WebCore::InlineTextBox::attachLine): Ditto.
        (WebCore::InlineTextBox::placeEllipsisBox): Ditto.
        (WebCore::InlineTextBox::paintDecoration): Ditto.
        (WebCore::InlineTextBox::paintCompositionUnderline): Ditto.
        (WebCore::InlineTextBox::offsetForPosition): Ditto.
        (WebCore::InlineTextBox::positionForOffset): Ditto.

        * rendering/InlineTextBox.h: Made most members private. Use toRenderText.

        * rendering/RenderArena.cpp:
        (WebCore::RenderArena::allocate): Use static_cast instead of C-style cast.
        (WebCore::RenderArena::free): Ditto.

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::addContinuationWithOutline): Use ListHashSet<RenderFlow*>
        instead of RenderFlowSequencedSet.
        (WebCore::RenderBlock::paintContinuationOutlines): Ditto.
        (WebCore::stripTrailingSpace): Ditto.
        (WebCore::RenderBlock::calcInlinePrefWidths): Ditto.
        (WebCore::RenderBlock::updateFirstLetter): Ditto.

        * rendering/RenderContainer.cpp:
        (WebCore::RenderContainer::addChild): Use toRenderText.

        * rendering/RenderFlow.cpp:
        (WebCore::RenderFlow::dirtyLinesFromChangedChild): Use ListHashSet<RenderFlow*>
        instead of RenderFlowSequencedSet.
        (WebCore::RenderFlow::paintLines): Ditto.

        * rendering/RenderMenuList.cpp: Removed some unneeded includes.

        * rendering/RenderMenuList.h: Added now-needed forward declaration of RenderText.
        Made more members private.

        * rendering/RenderObject.cpp: Updated includes.
        (WebCore::RenderObject::isEditable): Use toRenderText.
        (WebCore::RenderObject::positionForPoint): Made non-inline so we wouldn't have to
        include VisiblePosition.h in the header just for this.

        * rendering/RenderObject.h: Removed unneeded includes and forward declarations.
        Changed special vertical position values to be constants rather than an enum.
        Eliminated RenderFlowSequencedSet, and changed clients to use ListHashSet<RenderFlow*>.

        * rendering/RenderText.cpp:
        (WebCore::RenderText::findNextInlineTextBox): Use function members instead of
        data members of InlineTextBox.
        (WebCore::RenderText::positionForCoordinates): Ditto.
        (WebCore::isInlineFlowOrEmptyText): Use toRenderText.
        (WebCore::RenderText::previousCharacter): Use function members instead of
        data members of InlineTextBox.
        (WebCore::RenderText::position): Ditto.
        (WebCore::RenderText::caretMinOffset): Ditto.
        (WebCore::RenderText::caretMaxOffset): Ditto.
        (WebCore::RenderText::caretMaxRenderedOffset): Ditto.

        * rendering/RenderText.h: Added definition of toRenderText.

        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::textWithHardLineBreaks): Use toRenderText.
        * rendering/RenderTreeAsText.cpp:
        (WebCore::operator<<): Use toRenderText and toRenderBox.
        (WebCore::writeTextRun): Ditto.
        (WebCore::write): Ditto.

        * rendering/bidi.cpp:
        (WebCore::InlineIterator::increment): Use toRenderText.
        (WebCore::InlineIterator::current): Ditto.
        (WebCore::checkMidpoints): Ditto.
        (WebCore::RenderBlock::computeHorizontalPositionsForLine): Ditto.
        (WebCore::RenderBlock::layoutInlineChildren): Ditto.
        (WebCore::RenderBlock::determineStartPosition): Ditto.
        (WebCore::shouldSkipWhitespaceAfterStartObject): Ditto.
        (WebCore::RenderBlock::findNextLineBreak): Ditto.

WebKit/mac:

2009-01-25  Darin Adler  <darin@apple.com>

        Reviewed by Sam Weinig.

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

        Step one: RenderText.

        * WebView/WebRenderNode.mm:
        (copyRenderNode): Use toRenderText.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40229 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Urs Wolfer  <uwolfer@kde.org>

        Reviewed by Simon Hausmann.

        Add QWebPage::SelectAll to WebAction enum to select all content.

        https://bugs.webkit.org/show_bug.cgi?id=22946

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40228 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Thiago Macieira  <thiago.macieira@nokia.com>

        Reviewed by Simon Hausmann.

        Fix gcc compiler warning (deprecated conversion from constant string to char*)

        "" cannot be stored in a char*

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40227 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  David Boddie  <dboddie@trolltech.com>

        Reviewed by Simon Hausmann.

        Doc: Fixed the terminology used.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40226 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Thiago Macieira  <thiago.macieira@nokia.com>

        Reviewed by Simon Hausmann.

        Rename links from trolltech.com to qtsoftware.com

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40225 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-25  Tor Arne Vestbø  <tavestbo@trolltech.com>

        Reviewed by Simon Hausmann.

        Make QWebPage auto-test more robust by cleaning up generated files

        Sometimes the auto-test may segfault, leaving behind files because the
        call to cleanupTestCase() never happens, so we also clean up the files
        at initTestCase() just to make sure we're running in a clean environment.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40224 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-12-20  Urs Wolfer  <uwolfer@kde.org>

        Reviewed by Simon Hausmann.

        Fix: compile qwebhistory.h with pendantic comiler flag.

        https://bugs.webkit.org/show_bug.cgi?id=22947

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40223 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove a slightly misleading comment from ChangeLog - the issue is related to bug 23500, but not the same.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40222 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        <rdar://problem/5954398> REGRESSION: 1.1% PLT regression from 33577 and 33578 (encoding fixes)

WebCore:
        Changed single argument KURL constructors back to always expect an already encoded string,
        eliminating extra conversions.

        This is a rather unstable situation, as it is often unclear whether a given string is safe
        to convert to KURL without resolving. I think that going forward, the solution is to try to
        keep encoded URLs as KURL instances, and not as strings.

        * platform/KURL.h: Updated comments.

        * platform/KURL.cpp:
        (WebCore::KURL::KURL): In debug builds, verify that the passed string is ASCII-only. The
        intention is to verify that it is already parsed and encoded by KURL or equivalent code, but
        since encoding is scheme-dependent, such a verification would be quite complicated.
        Don't encode the string as UTF-8, as it supposed to be ASCII-only.
        Removed a hack that made strings beginning with "/" turn into "file:" URLs. I didn't find
        any reason for it to exist, but I saw several cases where this code path was taken
        inadvertently (see examples in LayoutTests/ChangeLog).
        (WebCore::KURL::setProtocol): Using a user-provided string without validation or encoding
        is clearly wrong here (e.g., the "protocol" can be set to a full URL, effectively replacing
        the old one), and an already encoded string is expected by parse().
        In debug builds, non-ASCII input will make an assertion in parse() fail. Added a FIXME.
        (WebCore::KURL::setHost): Ditto.
        (WebCore::KURL::setPort): Ditto.
        (WebCore::KURL::setHostAndPort): Ditto.
        (WebCore::KURL::setUser): Ditto.
        (WebCore::KURL::setPass): Ditto.
        (WebCore::KURL::setRef): Ditto.
        (WebCore::KURL::setQuery): Ditto.
        (WebCore::KURL::setPath): Ditto.  Note that problems described in bug 23500 mask some of
        the problems in release builds currently, as the incorrectly parsed URL is ignored.
        (WebCore::KURL::parse): Verify that the passed string is already encoded.

        * html/HTMLLinkElement.cpp:
        (WebCore::HTMLLinkElement::parseMappedAttribute):
        (WebCore::HTMLLinkElement::process):
        * html/HTMLLinkElement.h:
        Changed to avoid using invalid URLs (this was causing problems on DNS prefetch tests, see
        LayoutTests/ChangeLog).

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::init): Create an empty KURL without indirection for a small speedup.
        (WebCore::FrameLoader::requestFrame): Resolve and encode javascript URLs properly, now that
        String to KURL conversion requires the string to be already encoded.

        * page/DOMWindow.cpp: (WebCore::DOMWindow::postMessage): Resolve and encode the origin.
        HTML5 seems a little unclear on how this should work (it talks about "either parsing it as
        a URL, or resolving it", and then somehow compares unaltered targetOrigin string to a
        security origin object), so I just made the code as close to what we already had as possible.

LayoutTests:
        * http/tests/misc/dns-prefetch-control-expected.txt:
        * http/tests/misc/dns-prefetch-control.html:
        Google documentation for DNS Prefetch makes use of net-path relative URLs (//server-name),
        explaining that scheme is not necessary. This is of course true, but this test uses data:
        subframes, and data: is a non-hierachical scheme, so resolving such URLs fails, resulting
        in a KURL object that is not valid. WebKit used to ignore this, and tried to create a URL
        from this string again, now with a single argument KURL constructor, which resulted in a
        valid file: URL, which was successfully used! Both issues have been corrected in WebCore,
        so I had to change the test to no longer use relative net-path URLs.

        * http/tests/security/postMessage/invalid-origin-throws-exception-expected.txt:
        * http/tests/security/postMessage/invalid-origin-throws-exception.html:
        URLs that start with "/" are no longer converted to "file:" ones, so the results now
        match Firefox.

        * http/tests/uri/resolve-encoding-relative-expected.txt: Added.
        * http/tests/uri/resolve-encoding-relative.html: Added.
        Added a test to cover some cases of relative URL resolving that were not covered before.
        Expected results are taken from Firefox 3, and WebKit doesn't match in how fragments are
        encoded (we use document encoding, while Firefox uses UTF-8). Since fragments are not
        sent in HTTP requests, this is not too dangerous, but the Firefox behavior looks more
        consistent.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40221 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Oops, fix the Qt build.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40220 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-24  Eric Carlson  <eric.carlson@apple.com>

        <video> controls visibility test needs to dump render tree as text because
        the test does not always end at the same time, so the timeline slider thumb
        is not always at the same position.

        * media/video-controls-visible-audio-only.html:
        * platform/mac/media/video-controls-visible-audio-only-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40219 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add QWebFrame::renderContents() method which allows arbitary rendering
of any content region within the QWebFrame.  It also does not draw any
scrollbars.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40218 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-24  Darin Adler  <darin@apple.com>

        Try to fix expected results for new video element test.

        * platform/mac/media/video-controls-visible-audio-only-expected.png: Landed new result
        generated on my machine.
        * platform/mac/media/video-controls-visible-audio-only-expected.txt: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40217 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-24  Darin Adler  <darin@apple.com>

        Try to fix Qt build.

        * WebCore.pro: Added JSTextCustom.cpp.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40216 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    https://bugs.webkit.org/show_bug.cgi?id=23467

        Change styleAvailable functionality to be a linked list
        rather than scanning every CompositeAnimation for
        Animations that are waiting for styleAvailable. This has
        potential for large performance improvement in cases where
        there are many animations firing at once. But it will also
        help in my next fix, which is to synchronize software and
        accelerated animations.

        I also changed slightly when the beginAnimationUpdate and
        endAnimationUpdate calls are made. But this has no real
        effect on performance or behavior.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40215 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Force inlining of Lexer::matchPunctuator

Reviewed by Maciej Stachowiak.

2.2% win when parsing jQuery, Mootools, Prototype, etc

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40214 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        <rdar://problem/6368059> REGRESSION: URL encoding problems on http://www.cineman.ch

        Test: http/tests/xmlhttprequest/encode-request-url-2.html

        * platform/KURL.cpp:
        (WebCore::appendASCII): Added a helper function.
        (WebCore::KURL::KURL): Explicitly encode the URL to call two-argument parse() with better
        specified behavior.
        (WebCore::KURL::init): Ditto. This avoids trying to round-trip an URL encoded into a byte
        stream, but not yet transformed to ASCII-only using percent escapes. Since different parts
        of the byte stream can use different encodings, round-tripping is not possible.
        (WebCore::KURL::parse): Reverted an earlier change that made the single-argument version of
        this function convert the string to utf-8. I think that on the remanining code paths, it is
        correct to assume that the string is all ASCII, but I'm not yet confident enough to drop a
        FIXME warning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40213 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-24  Jan Michael Alonzo  <jmalonzo@webkit.org>

        Gtk build fix after r40170

        * GNUmakefile.am: Removed editing/RemoveNodeAttributeCommand.cpp
        and editing/RemoveNodeAttributeCommand.h

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40212 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the Qt build after r40205.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40211 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-23  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin Adler

        In CachedPage, break off per-Frame pieces of data into a CachedFrame object.

        * GNUmakefile.am:
        * WebCore.base.exp:
        * WebCore.pro:
        * WebCore.scons:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:

        * history/CachedFrame.cpp: Added.
        (WebCore::cachedFrameCounter):
        (WebCore::CachedFrame::CachedFrame):
        (WebCore::CachedFrame::~CachedFrame):
        (WebCore::CachedFrame::restore):
        (WebCore::CachedFrame::clear):
        (WebCore::CachedFrame::setCachedFramePlatformData):
        (WebCore::CachedFrame::cachedFramePlatformData):
        * history/CachedFrame.h: Added.
        (WebCore::CachedFrame::document):
        (WebCore::CachedFrame::documentLoader):
        (WebCore::CachedFrame::view):
        (WebCore::CachedFrame::mousePressNode):
        (WebCore::CachedFrame::url):
        (WebCore::CachedFrame::domWindow):

        * history/CachedPage.cpp:
        (WebCore::CachedPage::CachedPage):
        (WebCore::CachedPage::restore):
        (WebCore::CachedPage::clear):
        (WebCore::CachedPage::setCachedFramePlatformData):
        (WebCore::CachedPage::cachedFramePlatformData):
        * history/CachedPage.h:
        (WebCore::CachedPage::document):
        (WebCore::CachedPage::documentLoader):
        (WebCore::CachedPage::view):
        (WebCore::CachedPage::mousePressNode):
        (WebCore::CachedPage::url):
        (WebCore::CachedPage::domWindow):

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::cachePageForHistoryItem): Setting the DocumentLoader independently of
          creating a CachedPage was a historical relic.  Take the opportunity to clean that up.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40210 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-23  David Hyatt  <hyatt@apple.com>

        Add ASSERTs to catch anyone trying to obtain the width/height of an inline flow.  Fix up all the places
        that triggered the ASSERT (since all of these places are effectively regressions from my landing that
        devirtualized width()/height()).

        Reviewed by Adele

        * dom/Position.cpp:
        (WebCore::Position::hasRenderedNonAnonymousDescendantsWithHeight):
        * editing/DeleteButtonController.cpp:
        (WebCore::isDeletableElement):
        * html/HTMLAnchorElement.cpp:
        (WebCore::HTMLAnchorElement::isKeyboardFocusable):
        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::clientWidth):
        (WebCore::RenderBox::clientHeight):
        (WebCore::RenderBox::scrollWidth):
        (WebCore::RenderBox::scrollHeight):
        (WebCore::RenderBox::absoluteOutlineBounds):
        * rendering/RenderBox.h:
        (WebCore::RenderBox::width):
        (WebCore::RenderBox::height):
        (WebCore::RenderBox::size):
        (WebCore::RenderBox::frameRect):
        (WebCore::RenderBox::borderBoundingBox):
        * rendering/RenderFlow.h:
        (WebCore::RenderFlow::borderBoundingBox):
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::updateLayerPosition):
        * rendering/RenderScrollbarPart.cpp:
        (WebCore::RenderScrollbarPart::styleWillChange):
        * rendering/RenderScrollbarPart.h:
        * wml/WMLAElement.cpp:
        (WebCore::WMLAElement::isKeyboardFocusable):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40209 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-23  Gavin Barraclough  <barraclough@apple.com>

        Reviewed by Geoff Garen.

        Fix for <rdar://problem/6126212>
        Ensure that callbacks out from the JSC interface are only allowed
        to return in reverse-chronological order to that in which they were
        made.  If we allow earlier callbacks to return first, then this may
        result in setions of the RegisterFile in use by another thread
        being trampled.

        See uber-comment in JSLock.h for details.

        * runtime/JSLock.cpp:
        (JSC::JSLock::DropAllLocks::DropAllLocks):
        (JSC::JSLock::DropAllLocks::~DropAllLocks):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40208 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-23  Darin Adler  <darin@apple.com>

        Reviewed by Anders Carlsson.

        Bug 23352: Turn on more compiler warnings in the Mac build
        https://bugs.webkit.org/show_bug.cgi?id=23352

        Second patch: Slightly less simple cases of various warnings.

        * bridge/jni/jni_runtime.h:
        (JSC::Bindings::JavaString::_commonInit): Use reinterpret_cast here and avoid casting
        away const.

        * dom/XMLTokenizerLibxml2.cpp:
        (WebCore::PendingCallbacks::appendStartElementNSCallback): Use static_cast instead of
        reinterpret_cast; this is casting the result of malloc, and the idiom is to use
        static_cast, which works because we're casting from void*.
        (WebCore::PendingCallbacks::PendingStartElementNSCallback::call): Use const_cast instead
        C-style casts.

        * xml/XSLTUnicodeSort.cpp:
        (inlineXSLTTransformError): Tweak soft linking of xsltTransformError to use a macro
        instead of redefining the function.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40207 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-23  Darin Adler  <darin@apple.com>

        Reviewed by Anders Carlsson.

        Bug 23475: fast/text/find-backwards.html fails with ICU disabled SearchBuffer
        https://bugs.webkit.org/show_bug.cgi?id=23475

        * editing/TextIterator.cpp:
        (WebCore::SearchBuffer::search): Set the character start flag to false to
        ensure we won't ever return the same result twice. There are other ways we
        could accomplish this, mentioned in the comment, but this one was the smallest
        and cleanest I could think of.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40206 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-01-23  Darin Adler  <darin@apple.com>

        Reviewed by Anders Carlsson.

        Bug 23510: REGRESSION (r40171): CDATASection objects not getting proper JavaScript wrappers
        https://bugs.webkit.org/show_bug.cgi?id=23510

        That patch fixed the optimization that's supposed to give a faster path for
        allocating the wrappers for DOM Text objects. But CDATASection is the one
        subclass of Text objects.

        * GNUmakefile.am: Added JSCDATASectionCustom.cpp.
        * WebCore.scons: Ditto.
        * WebCore.vcproj/WebCore.vcproj: Ditto.
        * WebCore.xcodeproj/project.pbxproj: Ditto.
        * WebCoreSources.bkl: Ditto.

        * bindings/js/JSCDATASectionCustom.cpp: Copied from bindings/js/JSTextCustom.cpp.
        (WebCore::toJSNewlyCreated): Changed to do CDATASection, not Text.

        * bindings/scripts/CodeGeneratorJS.pm: Added another toJSNewlyCreated function
        for CDATASection. We don't want to have to check when creating a wrapper for a
        Text object whether it's a CDATASection object or not, so we need to do this overload
        at compile time.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40205 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Eric Seidel.

Add a base class for WML form control elements: WMLFormControlElement (as it's designed in HTML as well).
WMLFormControlElement inherits from WMLElement and FormControlElement (just like the HTMLFormControlElement equivalent).

WMLInputElement / WMLOptionElement / WMLOptGroupElement are supposed to inherit from the new WMLFormControlElement class.
Now WMLInputElement & WMLOptGroupElement only inherit from WMLElement and FormControlElement, but WMLOptionElement inherits
from WMLEventHandlingElement (which in turn inherits from WMLElement) and FormControlElement. In order to introduce a shared
base class, WMLEventHandlingElement is no longer allowed to inherit from WMLElement directly. Instead it should turn into
another base class, just like FormControlElement.

Add a base class for form control elements 'with state': FormControlElementWithState, serving as base class for
HTMLFormControlElementWithState and WMLFormControlElementWithState. -> Make Document operate on FormControlElementWithState
objects instead of HTMLFormControlElementWithState objects, to support save/restore control state for non HTMLFormControlElements.

Rename isGenericFormElement() to isFormControlElement(), and move it to Element.h
Add new virtual function isFormControlElementWithState() to Element.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40204 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 23492: Separating the WebKitSystemInterface Calls.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40203 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 23509: Crash at -[WebCoreAXObject doAXNextSentenceEndTextMarkerForTextMarker:]

Move null checking into makeRange

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40202 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the Gtk build too after removal of RenderBox::toConstRenderBox

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40201 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the Qt build after rename of RenderBox::toConstRenderBox -> toRenderBox

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40200 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the Qt build after r40168.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40199 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-01-23  Brady Eidson  <beidson@apple.com>

        Rubberstamped by Darin Adler

        Rename CachedPagePlatformData to CachedFramePlatformData to more accurately reflect its true role.

        * GNUmakefile.am:
        * WebCore.base.exp:
        * WebCore.order:
        * WebCore.xcodeproj/project.pbxproj:

        * history/CachedFramePlatformData.h: Copied from WebCore/history/CachedPagePlatformData.h.
        (WebCore::CachedFramePlatformData::~CachedFramePlatformData):
        * history/CachedPagePlatformData.h: Removed.

        * history/CachedPage.cpp:
        (WebCore::CachedPage::clear):
        (WebCore::CachedPage::setCachedFramePlatformData):
        (WebCore::CachedPage::cachedFramePlatformData):
        * history/CachedPage.h:

WebKit:

2009-01-23  Brady Eidson  <beidson@apple.com>

        Rubberstamped by Darin Adler

        Rename CachedPagePlatformData to CachedFramePlatformData to more accurately reflect its true role.

        * WebKit.xcodeproj/project.pbxproj:

WebKit/mac:

2009-01-23  Brady Eidson  <beidson@apple.com>

        Rubberstamped by Darin Adler

        Rename CachedPagePlatformData to CachedFramePlatformData to more accurately reflect its true role.

        * WebCoreSupport/WebCachedFramePlatformData.h: Copied from WebKit/mac/WebCoreSupport/WebCachedPagePlatformData.h.
        (WebCachedFramePlatformData::WebCachedFramePlatformData):
        * WebCoreSupport/WebCachedPagePlatformData.h: Removed.

        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::savePlatformDataToCachedPage):
        (WebFrameLoaderClient::transitionToCommittedFromCachedPage):

        * WebKit.order:

WebKit/win:

2009-01-23  Brady Eidson  <beidson@apple.com>

        Rubberstamped by Darin Adler

        Rename CachedPagePlatformData to CachedFramePlatformData to more accurately reflect its true role.

        * WebCachedFramePlatformData.h: Copied from WebKit/win/WebCachedPagePlatformData.h.
        (WebCachedFramePlatformData::WebCachedFramePlatformData):
        * WebCachedPagePlatformData.h: Removed.

        * WebCoreSupport/WebFrameLoaderClient.cpp:
        (WebFrameLoaderClient::savePlatformDataToCachedPage):

        * WebKit.vcproj/WebKit.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40198 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Kevin Ollivier.

Fix drawing in situations where sub-portions of the bitmap are to be drawn to the screen.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40197 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Build fixes for recent changes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40196 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix Windows bustage.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@40195 268f45cc-cd09-0410-ab3c-d52691b4dbfc