2008-04-01  Holger Hans Peter Freyther  <holger.freyther@trolltech.com>

        Reviewed by Simon.

        * Change the order of the methods to match with the FrameLoaderClient.h to ease
        removing methods in the future.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31516 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-04-01  Holger Hans Peter Freyther  <holger.freyther@trolltech.com>

        Reviewed by Simon.

        * Remove virtuals in the FrameLoaderClientQt that don't exist in the base class.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31515 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - fix previous attempted fixes

        * platform/qt/TemporaryLinkStubs.cpp:
        (PluginDatabase::getPluginPathsInDirectories):
        * platform/wx/TemporaryLinkStubs.cpp:
        (PluginDatabase::getPluginPathsInDirectories):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31513 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - fix Qt link stubs

        * platform/qt/TemporaryLinkStubs.cpp:
        (PluginDatabase::getPluginPathsInDirectories):
        (PluginDatabase::isPreferredPluginDirectory):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31512 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - try to fix the Wx build

        * platform/wx/TemporaryLinkStubs.cpp:
        (PluginDatabase::getPluginPathsInDirectories):
        (PluginDatabase::isPreferredPluginDirectory):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31511 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - Gtk and Qt build fixes

        * platform/graphics/gtk/FontCustomPlatformData.cpp:
        (WebCore::FontCustomPlatformData::fontPlatformData):
        * platform/graphics/gtk/FontCustomPlatformData.h:
        * platform/graphics/qt/FontCustomPlatformData.cpp:
        (WebCore::FontCustomPlatformData::fontPlatformData):
        * platform/graphics/qt/FontCustomPlatformData.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31510 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - Mac build fix

        * platform/graphics/mac/FontCustomPlatformData.cpp:
        (WebCore::FontCustomPlatformData::fontPlatformData):
        * platform/graphics/mac/FontCustomPlatformData.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31509 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-31  Dan Bernstein  <mitz@apple.com>

        Reviewed by Darin Adler.

        - fix <rdar://problem/5792638> @font-face with downloadable fonts doesn't work with GDI text

        * WebCore.vcproj/WebCore.vcproj: Added GetEOTHeader.{h,cpp}

        * css/CSSFontFaceSource.cpp:
        (WebCore::CSSFontFaceSource::getFontData): Changed to pass the rendering
        mode to CachedFont::platformDataFromCustomData().

        * loader/CachedFont.cpp:
        (WebCore::CachedFont::platformDataFromCustomData): Added a rendering
        mode parameter which is passed down to
        FontCustomPlatformData::fontPlatformData().

        * loader/CachedFont.h:

        * platform/graphics/win/FontCacheWin.cpp:
        (WebCore::FontCache::fontExists): Changed the number 5 to the named
        constant CLEARTYPE_QUALITY.

        * platform/graphics/win/FontCustomPlatformData.cpp:
        (WebCore::FontCustomPlatformData::~FontCustomPlatformData): Added code
        to delete the embedded font reference.
        (WebCore::FontCustomPlatformData::fontPlatformData): Added code to
        create an HFONT from the embedded font reference and pass it along with
        the CGFontRef to the FontPlatformData constructor. Added a rendering
        mode parameter which is passed down to the FontPlatformData constructor.
        (WebCore::EOTStream::EOTStream): Added this helper class used to feed
        data to TTLoadEmbeddedFont().
        (WebCore::EOTStream::read): Added.
        (WebCore::readEmbedProc): Added this TTLoadEmbeddedFont() callback.
        (WebCore::createUniqueFontName): Added. Creates a unique GUID-derived
        font name.
        (WebCore::createFontCustomPlatformData): Added code to make the font
        accessible to GDI by creating an Embedded OpenType stream and
        activating it in the process under a globally-unique name.

        * platform/graphics/win/FontCustomPlatformData.h:
        Added a font reference data member which holds the embedded font.
        (WebCore::FontCustomPlatformData::FontCustomPlatformData): Added a
        rendering mode parameter.

        * platform/graphics/win/FontPlatformData.h:
        (WebCore::FontPlatformData::FontPlatformData): Made this class always
        own the CGFontRef and the HFONT by turning these data members into a
        RetainPtr and a RefPtr around a RefCountedHFONT.
        Added an HFONT parameter to the CGFontRef-based constructor.
        Made platformDataInit() private.
        (WebCore::FontPlatformData::hfont):
        (WebCore::FontPlatformData::cgFont):
        (WebCore::FontPlatformData::hash):
        (WebCore::FontPlatformData::RefCountedHFONT::RefCountedHFONT): Added.
        (WebCore::FontPlatformData::RefCountedHFONT::~RefCountedHFONT):
        (WebCore::FontPlatformData::RefCountedHFONT::hfont):
        (WebCore::FontPlatformData::RefCountedHFONT::hash):

        * platform/graphics/win/FontPlatformDataCGWin.cpp:
        (WebCore::FontPlatformData::platformDataInit): Changed to take
        ownership of the CGFontRef created here.
        (WebCore::FontPlatformData::FontPlatformData): Added an HFONT parameter
        and a useGDI flag.

        * platform/graphics/win/FontPlatformDataWin.cpp:
        (WebCore::FontPlatformData::FontPlatformData): Changed to take ownership
        of the HFONT.

        * platform/graphics/win/GetEOTHeader.cpp: Added.
        (WebCore::BigEndianUShort::operator unsigned short):
        (WebCore::BigEndianULong::operator unsigned):
        (WebCore::appendBigEndianStringToEOTHeader):
        (WebCore::getEOTHeader): Added. Creates an Embedded OpenType (.eot)
        header for the given font data, making it suitable as input for
        TTLoadEmbeddedFont.

        * platform/graphics/win/GetEOTHeader.h: Added.

        * platform/graphics/win/SimpleFontDataCGWin.cpp:
        (WebCore::SimpleFontData::platformDestroy): Removed code to delete parts
        of the FontPlatformData, since they are now always owned and managed
        by the FontPlatformData.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31507 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-31  Darin Adler  <darin@apple.com>

        Reviewed by Anders.

        - fix crash seen on buildbot (regression from a recent speedup)

        * html/HTMLDocument.cpp:
        (WebCore::HTMLDocument::createElement): Check validity before calling isLower, since
        that takes care of the null string case.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31498 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by John Sullivan.

        - fix <rdar://problem/5741075> Google Maps: cursor and placemarks don't match up

        * platform/win/CursorWin.cpp:
        (WebCore::Cursor::Cursor): Set the 1-bit mask for full-alpha cursors
        to fully opaque. The mask is ignored and the alpha values from the
        32-bit image are used except in the case of a fully-transparent image,
        in which the fully-opaque 1-bit mask is required to get transparency.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31491 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Make more WebArchive checks cross platform

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31489 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Turn on WebArchive loading on Windows

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31488 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-31  Mark Rowe  <mrowe@apple.com>

        Add some images that were missed in the WordPress 2.5 upgrade.

        * blog/:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31487 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-31  Mark Rowe  <mrowe@apple.com>

        Upgrade to WordPress 2.5

        * blog/:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31486 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-31  Dan Bernstein  <mitz@apple.com>

        Reviewed by Dave Hyatt.

        - fix http://bugs.webkit.org/show_bug.cgi?id=18076
          Box disappears after CSS Transition completes

        * page/AnimationController.cpp: Changed the BLEND_MAYBE_INVALID_COLOR
        macro to not blend two invalid colors.
        * rendering/RenderStyle.cpp:
        (WebCore::StyleBackgroundData::StyleBackgroundData): Added missing
        copying of the m_color data member.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31485 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-31  Julien Chaffraix  <jchaffraix@webkit.org>

        Reviewed by Darin.

        Bug 18177: m_encoding is used twice in XMLHttpRequest
        http://bugs.webkit.org/show_bug.cgi?id=18177

        - Change XMLHttpRequest variable m_encoding to m_responseEncoding.

        - Change send local variable m_encoding to encoding.

        No test case (no functionnal change).

        * platform/network/ResourceResponseBase.h: Removed unused variable.
        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::send): Changed local m_encoding to encoding.
        (WebCore::XMLHttpRequest::didReceiveResponse): m_encoding -> m_responseEncoding.
        (WebCore::XMLHttpRequest::didReceiveData): Ditto.
        * xml/XMLHttpRequest.h: Change m_encoding to m_responseEncoding to disambiguate
        the name as it is only used for response.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31483 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-31  Jasper Bryant-Greene  <jasper@unix.geek.nz>

        Reviewed by Darin.

        Resolves http://bugs.webkit.org/show_bug.cgi?id=18010
        "WebKitNetworkRequestPrivate is defined in two places unnecessarily"

        Removed unnecessary definition of WebKitNetworkRequestPrivate in
        webkitprivate.h

        * webkit/webkitprivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31480 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Pass the right ExecStates to toJS when creating JS wrappers for objects from the inspected page

        We now use an ExecState from the inspected page to create the JS
        wrappers for objects from the inspected page, rather than use an
        ExecState from the Inspector.

        Reviewed by Tim Hatcher.

        * page/InspectorController.cpp:
        (WebCore::getResourceDocumentNode):
        (WebCore::InspectorController::focusNode):
        (WebCore::InspectorController::addDatabaseScriptResource):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31479 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Stop relying on functions added to prototypes in the Inspector being callable on objects from the inspected window

        Reviewed by Tim Hatcher.

        * page/inspector/Console.js:
        (WebInspector.ConsolePanel._formatnode): Use nodeTitleInfo.call(node)
        instead of node.titleInfo().
        * page/inspector/StylesSidebarPane.js: Changed uses of
        getShorthandValue, getShorthandPriority, getLonghandProperties, and
        getUniqueStyleProperties to call them as functions instead of as
        methods on the CSSStyleDeclaration objects.
        * page/inspector/utilities.js: Removed Node.prototype.titleInfo, and
        changed our CSSStyleDeclaration methods to just be standalone
        functions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31472 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix an exception while adding a message to the console

        Reviewed by Mark Rowe.

        * page/inspector/ConsolePanel.js: Don't call addMessageToSource if the
        panel has no such method. This can happen if we get a message for,
        e.g., a Database.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31470 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-31  Julien Chaffraix  <julien.chaffraix@gmail.com>

        Reviewed by Darin.

        Bug 17665: determineSourceDir() dies if $sourceDir has a trailing backslash

        Remove trailing '/' in $sourceDir in determineSourceDir().

        Fix suggested by Dmitriy Kazachkov.

        * Scripts/webkitdirs.pm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31469 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2008-03-31  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin and Mitz's rubber stamp

        Remove dataForArchivedSelection(WebCore::Frame*) from the EditorClient - only usage is now directly in WebCore

        * page/EditorClient.h:
        * svg/graphics/SVGImageEmptyClients.h:

        * platform/mac/PasteboardMac.mm:
        (WebCore::Pasteboard::writeSelection): Make the archive and get the data directly

WebKit/mac:

2008-03-31  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin and Mitz's rubber stamp

        Remove dataForArchivedSelection(WebCore::Frame*) from the EditorClient - only usage is now directly in WebCore

        * WebCoreSupport/WebEditorClient.mm:
        * WebCoreSupport/WebEditorClient.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31468 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-31  Brady Eidson  <beidson@apple.com>

        Reviewed by Mitz Pettel

        Fix for http://bugs.webkit.org/show_bug.cgi?id=18183

        Despite the bug reporter's instructions, I have still not been able to actually reproduce this crash and
        therefore don't know how to make a layout test for it.

        * loader/archive/ArchiveResource.cpp:
        (WebCore::ArchiveResource::response): Fix potential null dereference

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31467 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix Bug 18208: Acid3 test 65 takes >33ms due to plugin refreshing on Windows

        <http://bugs.webkit.org/show_bug.cgi?id=18208>

        We now keep track of all the plugin paths found each time refresh() is
        called. We'll only instantiate PluginPackages if there are new paths
        or paths with changed timestamps since the last time refresh() was
        called.

        Reviewed by Darin Adler and Anders Carlsson.

        * plugins/PluginDatabase.cpp:
        (WebCore::PluginDatabase::refresh):
          - Only instantiate PluginPackages if there is a new path or a path
            with a changed timestamp since we last ran refresh().
          - Cache the set of plugin paths found and their timestamps for the
            next call to refresh().
          - Only re-register MIME types if our set of plugins changed.
        * plugins/PluginDatabase.h: Added a new member to cache plugin paths
        and their timestamps.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31466 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Separate filesystem crawling from PluginPackage instantiation

        Part of Bug 18208: Acid3 test 65 takes >33ms due to plugin refreshing
        on Windows
        <http://bugs.webkit.org/show_bug.cgi?id=18208>

        refresh() is now the only place where PluginPackages are instantiated.
        refresh() now asks for a set of plugins that no longer exist on disk,
        and the set of all plugin files in our plugin directories. Using these
        two sets we can update our instantiated plugins without copying any
        HashSets. The code in refresh() and in the platform-specific methods
        to crawl the filesystem is now quite a bit simpler.

        PluginDatabase now stores both a PluginSet and a HashMap that maps
        plugin paths to PluginPackages. This allows us to quickly determine
        whether we already have a PluginPackage instantiated for a particular
        path. The new add/remove methods handle the modification of these two
        collections.

        A nice side effect of all this is that refresh() no longer copies any
        HashSets.

        Reviewed by Mitz Pettel.

        * plugins/PluginDatabase.cpp:
        (WebCore::PluginDatabase::refresh): Unload any plugins that have been
        deleted from disk, and add any plugins that either weren't installed
        last time, or have changed since last time.
        (WebCore::PluginDatabase::getDeletedPlugins): Added.
        (WebCore::PluginDatabase::add): Added. Returns whether or not the
        PluginPackage was actually added to the database (duplicates won't be
        added).
        (WebCore::PluginDatabase::remove): Added.
        * plugins/PluginDatabase.h:
        * plugins/PluginPackage.h:
        (WebCore::PluginPackage::path): Added.
        (WebCore::PluginPackage::lastModified): Added.
        * plugins/gtk/PluginDatabaseGtk.cpp:
        (WebCore::PluginDatabase::getPluginPathsInDirectories): Renamed from
        getPluginsInDirectories. Now fills a HashSet of Strings instead of
        instantiated PluginPackages.
        * plugins/win/PluginDatabaseWin.cpp:
        (WebCore::addPluginPathsFromRegistry): Ditto.
        (WebCore::PluginDatabase::getPluginPathsInDirectories): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31465 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Change getPluginsInDirectories to use an out parameter

        This avoids copying a HashSet.

        Reviewed by Mitz Pettel.

        * plugins/PluginDatabase.cpp:
        (WebCore::PluginDatabase::refresh):
        * plugins/PluginDatabase.h:
        * plugins/gtk/PluginDatabaseGtk.cpp:
        (WebCore::PluginDatabase::getPluginsInDirectories):
        * plugins/win/PluginDatabaseWin.cpp:
        (WebCore::PluginDatabase::getPluginsInDirectories):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31464 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Make some PluginPackage methods return const String&

        Reviewed by Mitz Pettel.

        * plugins/PluginPackage.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31463 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rename PluginDatabase's "PluginPaths" to "PluginDirectories"

WebCore:

        Rename PluginDatabase's "PluginPaths" to "PluginDirectories"

        The code expects these paths to all be directories, so it seems good
        to refer to them as such.

        This patch just changes "Paths" to "Directories" and "Path" to
        "Directory".

        Reviewed by Mitz Pettel.

        * platform/qt/TemporaryLinkStubs.cpp:
        * platform/wx/TemporaryLinkStubs.cpp:
        * plugins/PluginDatabase.cpp:
        (WebCore::PluginDatabase::installedPlugins):
        (WebCore::PluginDatabase::refresh):
        * plugins/PluginDatabase.h:
        (WebCore::PluginDatabase::setPluginDirectories):
        * plugins/gtk/PluginDatabaseGtk.cpp:
        (WebCore::PluginDatabase::getPluginsInDirectories):
        * plugins/win/PluginDatabaseWin.cpp:
        (WebCore::PluginDatabase::getPluginsInDirectories):
        (WebCore::safariPluginsDirectory):
        (WebCore::PluginDatabase::defaultPluginDirectories):
        (WebCore::PluginDatabase::isPreferredPluginDirectory):
        * plugins/win/PluginPackageWin.cpp:
        (WebCore::PluginPackage::compare):

WebKit/win:

        Rename IWebViewPrivate::addAdditionalPluginPath to
        addAdditionalPluginDirectory

        Reviewed by Mitz Pettel.

        * Interfaces/IWebViewPrivate.idl:
        * Interfaces/WebKit.idl: Touched to make sure the Interfaces project
        rebuilds.
        * WebView.cpp:
        (WebView::addAdditionalPluginDirectory):
        * WebView.h:

WebKitTools:

        Update for rename of an IWebViewPrivate method

        Reviewed by Mitz Pettel.

        * DumpRenderTree/win/DumpRenderTree.cpp:
        (createWebViewAndOffscreenWindow):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31462 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Remove PlatformFileTime

        This typedef is time_t on all platforms, so we can just get rid of the
        typedef.

        Reviewed by Mitz Pettel.

        * platform/FileSystem.h:
        * plugins/PluginPackage.cpp:
        * plugins/PluginPackage.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31461 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Change PlatformFileTime on Windows to be time_t

        This matches other platforms and our existing functions in
        FileSystem.h. It will also let us get rid of PlatformFileTime
        altogether.

        Reviewed by Mitz Pettel.

        * platform/FileSystem.h: Change PlatformFileTime to time_t on Windows.
        * plugins/win/PluginDatabaseWin.cpp:
        (WebCore::addPluginsFromRegistry): Use getFileModificationTime to get
        the modification time to pass to PluginPackage::create.
        (WebCore::PluginDatabase::getPluginsInPaths): Ditto.
        * plugins/win/PluginPackageWin.cpp:
        (WebCore::PluginPackage::hash): Updated to use m_lastModified as a
        time_t.
        (WebCore::PluginPackage::equal): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31460 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Make PluginPackage work like our other RefCounted classes

        PluginPackage now starts with a ref count of 1, and its createPackage
        method now returns a PassRefPtr instead of a raw pointer.

        Reviewed by Darin Adler.

        * plugins/PluginPackage.cpp:
        (WebCore::PluginPackage::PluginPackage): Removed explicit call to the
        RefCounted constructor so that we'll start with a ref count of 1.
        (WebCore::PluginPackage::createPackage): Changed to return a
        PassRefPtr.
        * plugins/PluginPackage.h:
        * plugins/gtk/PluginDatabaseGtk.cpp:
        (WebCore::PluginDatabase::getPluginsInPaths): Changed to store the
        newly-created PluginPackage in a RefPtr.
        * plugins/win/PluginDatabaseWin.cpp:
        (WebCore::addPluginsFromRegistry): Ditto.
        (WebCore::PluginDatabase::getPluginsInPaths): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31459 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix Bug 18214: WebKit will sometimes load duplicate plugins

        <http://bugs.webkit.org/show_bug.cgi?id=18214>

        We now compare 3 things to determine if two PluginPackages are equal:
          1) Name
          2) Description
          3) Supported MIME types

        This matches Gecko's equality logic for plugins (implemented in
        nsPluginTag::Equals).

        Reviewed by Darin Adler.

        * plugins/win/PluginPackageWin.cpp:
        (WebCore::PluginPackage::hash): Use the above-mentioned 3 criteria to
        calculate the hash.
        (WebCore::PluginPackage::equal): Use the above-mentioned 3 criteria to
        determine equality.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31458 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the Qt build.

ResourceHandleInternal.h references ResourceHandle::fireFailure, so include
ResourceHandle.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31457 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by darin.

        Make matching of regexps using ^ much faster
        http://bugs.webkit.org/show_bug.cgi?id=18086

        * pcre/pcre_compile.cpp:
        (compileBranch):
        (branchNeedsLineStart):
        * pcre/pcre_exec.cpp:
        (match):
        (jsRegExpExecute):
        * pcre/pcre_internal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31454 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin.

        http://bugs.webkit.org/show_bug.cgi?id=15595
        marker property doesn't show up correctly as "shorthand" in inspector

        Make sure parsing the marker property sets the individual
        marker properties as implicit with marker as shorthand.
        Also let CSSStyleDecleration::getPropertyValue return
        the correct value for the marker property.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31453 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver.

        Make AMatthews life easier -- fix the SVG_FILTERS build

        Build fix only, no tests.

        * svg/SVGComponentTransferFunctionElement.cpp:
        (WebCore::SVGComponentTransferFunctionElement::SVGComponentTransferFunctionElement):
        * svg/SVGFEColorMatrixElement.cpp:
        (WebCore::SVGFEColorMatrixElement::SVGFEColorMatrixElement):
        * svg/SVGFEDiffuseLightingElement.cpp:
        * svg/SVGFEGaussianBlurElement.cpp:
        * svg/SVGFEImageElement.cpp:
        (WebCore::SVGFEImageElement::SVGFEImageElement):
        * svg/SVGFESpecularLightingElement.cpp:
        * svg/SVGFETurbulenceElement.cpp:
        * svg/SVGFilterElement.cpp:
        * svg/SVGFilterPrimitiveStandardAttributes.cpp:
        * svg/graphics/filters/SVGFEBlend.cpp:
        * svg/graphics/filters/SVGFEComponentTransfer.cpp:
        * svg/graphics/filters/SVGFEComposite.cpp:
        * svg/graphics/filters/SVGFEDiffuseLighting.cpp:
        * svg/graphics/filters/SVGFEGaussianBlur.cpp:
        * svg/graphics/filters/SVGFEImage.cpp:
        * svg/graphics/filters/SVGFEMerge.cpp:
        * svg/graphics/filters/SVGFEMorphology.cpp:
        * svg/graphics/filters/SVGFEOffset.cpp:
        * svg/graphics/filters/SVGFESpecularLighting.cpp:
        * svg/graphics/filters/SVGFETurbulence.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31451 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Sam Weinig.

        - fix http://bugs.webkit.org/show_bug.cgi?id=18115
          REGRESSION (r31250): Incomplete repaint of GoogleReader sidebar while reading unread posts

        Test: fast/repaint/lines-with-layout-delta.html

        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::repaintRectangle): Made this method take layout
        delta into account.

LayoutTests:

        Reviewed by Sam Weinig.

        - repaint test for http://bugs.webkit.org/show_bug.cgi?id=18115
          REGRESSION (r31250): Incomplete repaint of GoogleReader sidebar while reading unread posts

        * fast/repaint/lines-with-layout-delta.html: Added.
        * platform/mac/fast/repaint/lines-with-layout-delta-expected.checksum: Added.
        * platform/mac/fast/repaint/lines-with-layout-delta-expected.png: Added.
        * platform/mac/fast/repaint/lines-with-layout-delta-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31446 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Maciej Stachowiak.

        - fix assertion failure in CSSParser.cpp's equalIgnoringCase() when
          parsing certain transforms.

        Test: fast/css/transform-function-lowercase-assert.html

        * css/CSSParser.cpp:
        (WebCore::TransformOperationInfo::TransformOperationInfo): Changed to
        always pass a lowercase string as the second argument of
        equalIgnoringCase().

LayoutTests:

        Reviewed by Maciej Stachowiak.

        - test for an assertion failure in CSSParser.cpp's
          equalIgnoringCase() when parsing certain transforms.

        * fast/css/transform-function-lowercase-assert-expected.txt: Added.
        * fast/css/transform-function-lowercase-assert.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31445 268f45cc-cd09-0410-ab3c-d52691b4dbfc

ARM build fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31444 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-30  Hiroyuki Ikezoe  <poincare@ikezoe.net>

        Reviewed by Alp Toker.

        http://bugs.webkit.org/show_bug.cgi?id=18036
        Rendering invalid japanese characters with pango backend.

        * platform/graphics/gtk/FontPlatformDataPango.cpp:
        (FontPlatformData::FontPlatformData): Set PangoFontDescription for
        PangoContext explicitly to use proper font.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31442 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver.

        http://bugs.webkit.org/show_bug.cgi?id=17633
        SVG: modifying x & y attribute of text element from JS fails

        Make SVG text relayout when x, y attribute is changed from js.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31441 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam Roben.

        Fix wrong id.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31440 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Oliver.

        - yet still more bindings speedup

        Make AtomicString straight from UString in handcoded bindings

        * bindings/js/JSElementCustom.cpp:
        (WebCore::JSElement::setAttribute):
        (WebCore::JSElement::setAttributeNS):
        * dom/Document.cpp:
        (WebCore::Document::createElement):
        * dom/Document.h:
        * html/HTMLDocument.cpp:
        (WebCore::HTMLDocument::createElement):
        * html/HTMLDocument.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31439 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-29  Sam Weinig  <sam@webkit.org>

        Reviewed by Darin Adler.

        Fix for <rdar://problem/5828873>

        Since NSURL is allowing invalid urls to be loaded, we need to
        check the URL validity at the ResourceHandle level and fire off
        a cannotShowURL error.

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::cannotShowURLError):
        * loader/FrameLoader.h:
        * loader/ResourceLoader.cpp:
        (WebCore::ResourceLoader::cannotShowURLError):
        (WebCore::ResourceLoader::wasBlocked):
        (WebCore::ResourceLoader::cannotShowURL):
        * loader/ResourceLoader.h:
        * platform/KURL.h:
        (WebCore::KURL::isValid):
        * platform/network/ResourceHandle.cpp:
        (WebCore::ResourceHandle::create):
        (WebCore::ResourceHandle::scheduleFailure):
        (WebCore::ResourceHandle::fireFailure):
        (WebCore::portAllowed):
        * platform/network/ResourceHandle.h:
        (WebCore::ResourceHandle::):
        * platform/network/ResourceHandleClient.h:
        (WebCore::ResourceHandleClient::cannotShowURL):
        * platform/network/ResourceHandleInternal.h:
        (WebCore::ResourceHandleInternal::ResourceHandleInternal):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31438 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-29  Darin Adler  <darin@apple.com>

        Reviewed by Mitz.

        - more bindings speedup

        When I changed the temporary variables to be UString that means that the
        valueToString functions also need to return UString to avoid converting
        from UString to String twice.

        * bindings/js/kjs_binding.cpp:
        (WebCore::valueToStringWithNullCheck): Changed return value to UString.
        (WebCore::valueToStringWithUndefinedOrNullCheck): Ditto.
        * bindings/js/kjs_binding.h: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31437 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-29  Darin Adler  <darin@apple.com>

        Reviewed by Sam.

        * DerivedSources.make: Removed temporary remove-stray-JSRGBColor build rule.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31436 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-29  Darin Adler  <darin@apple.com>

        Reviewed by Sam.

        - tweak handling of the class attribute to speed up code that
          sets it but never needs to parse it

        6% speedup of Acid3 test 26

        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::matchRules): Change to use non-virtual
        classNames function since the hasClass bit can only be set for a
        StyledElement.
        (WebCore::CSSStyleSelector::checkOneSelector): Ditto. Also
        streamlined the code a bit for the tag matching as well.

        * dom/ClassNames.cpp:
        (WebCore::hasNonASCIIOrUpper): Added. This function does what other
        callers seem to want isLower() to do. We should merge this with
        isLower() in a subsequent cleanup pass.
        (WebCore::ClassNamesData::createVector): Renamed from parseClassAttribute.
        Turns the string into a vector.
        (WebCore::ClassNamesData::containsAll): Added. Used by getElementsByClassName.

        * dom/ClassNames.h: Added a separate ClassNameData class so we could hold
        the string and case folding flag as well as the vector. Changed ClassNames
        to have a set function rather than a parseClassAttribute function. Removed
        the "static" from isClassWhitespace. There's no reason to ask for internal linkage.

        * dom/ClassNodeList.cpp:
        (WebCore::ClassNodeList::ClassNodeList): Use constructor instead of the
        parseClassAttribute function.
        (WebCore::ClassNodeList::nodeMatches): Get rid of unnneeded isElementNode
        check, since hasClass will only be true for StyledElement nodes. Use the new
        containsAll function in ClassNames instead of having a loop here.

        * dom/ClassNodeList.h: Removed unneeded forward declaration.

        * dom/Document.cpp: Removed unneeded include.

        * dom/Element.cpp:
        (WebCore::Element::cloneNode): Changed code to copy attributes to use a
        function call instead of the assignment operator. This paves the way to
        making the function more efficient, using a virtual function.
        (WebCore::Element::setAttributeMap): Updated for PassRefPtr and for name change
        (element -> m_element).
        (WebCore::Element::createAttributeMap): Changed to use create function instead of
        calling the constructor directly.
        * dom/Element.h: Removed unneeded virtual getClassNames function.
        Changed the argument to setAttributeMap to be a PassRefPtr.

        * dom/NamedAttrMap.cpp:
        (WebCore::NamedAttrMap::NamedAttrMap): Updated for name change (element -> m_element).
        (WebCore::NamedAttrMap::getNamedItem): Ditto.
        (WebCore::NamedAttrMap::removeNamedItem): Ditto.
        (WebCore::NamedAttrMap::setNamedItem): Ditto.
        (WebCore::NamedAttrMap::item): Ditto.
        (WebCore::NamedAttrMap::detachFromElement): Ditto.
        (WebCore::NamedAttrMap::setAttributes): Renamed this from the assignment operator.
        (WebCore::NamedAttrMap::addAttribute): Ditto.
        (WebCore::NamedAttrMap::removeAttribute): Ditto.
        (WebCore::NamedAttrMap::isReadOnlyNode): Ditto.
        * dom/NamedAttrMap.h: Removed the copy constructor and assignment operator and added
        a new function, setAttributes, possibly to be made virtual in the future. Made
        isReadOnlyNode non-virtual. Renamed element -> m_element. Made constructor protected
        and added a create function.

        * dom/NamedMappedAttrMap.cpp:
        (WebCore::NamedMappedAttrMap::NamedMappedAttrMap): Changed constructor parameter to
        take a StyledElement.
        (WebCore::NamedMappedAttrMap::setClass): Changed to use the new set function.

        * dom/NamedMappedAttrMap.h: Made constructor private and added a create function.
        Replaced theparseClassAttribute function with clearClass and setClass functions.
        Replaced the getClassNames function with a classNames function returning a reference
        instead of a pointer.

        * dom/NamedNodeMap.h: Changed to start refcount at 1 instead of 0. Removed unneeded
        virtual function isReadOnlyNode.

        * dom/StyledElement.cpp:
        (WebCore::StyledElement::parseMappedAttribute): Rewrote class attribute handling to
        use new function names and took out uneeded special case for null attribute.
        (WebCore::StyledElement::createAttributeMap): Changed to use create function instead
        of a direct call to new.

        * dom/StyledElement.h: Replaced the virtual getClassNames function with a non-virtual
        inline classNames function.

        * html/HTMLElement.cpp:
        (WebCore::HTMLElement::cloneNode): Changed code to copy attributes to use a
        function call instead of the assignment operator.

        * html/HTMLTokenizer.cpp:
        (WebCore::Token::addAttribute): Use create function instead of a a direct call
        to new.

        * html/HTMLViewSourceDocument.cpp:
        (WebCore::HTMLViewSourceDocument::createContainingTable): Use RefPtr and the new
        create function.
        (WebCore::HTMLViewSourceDocument::addViewSourceToken): Ditto.
        (WebCore::HTMLViewSourceDocument::addSpanWithClassName): Ditto.
        (WebCore::HTMLViewSourceDocument::addLine): Ditto.
        (WebCore::HTMLViewSourceDocument::addText): Ditto.
        (WebCore::HTMLViewSourceDocument::addLink): Ditto.
        * html/HTMLViewSourceDocument.h: Ditto.

        * svg/SVGUseElement.cpp:
        (WebCore::SVGUseElement::expandSymbolElementsInShadowTree): Changed code to copy
        attributes to use a function call instead of the assignment operator.
        (WebCore::SVGUseElement::transferUseAttributesToReplacedElement): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31435 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2008-03-29  Sam Weinig  <sam@webkit.org>

        Reviewed by Brady Eidson.

        Fix crash when canceling a resource load while port blocked failure
        timer is going.

        Test: fast/loader/cancel-load-during-port-block-timer.html

        * platform/network/ResourceHandle.cpp:
        (WebCore::ResourceHandle::scheduleBlockedFailure):
        (WebCore::ResourceHandle::fireBlockedFailure):
        * platform/network/ResourceHandleInternal.h:
        (WebCore::ResourceHandleInternal::ResourceHandleInternal):
        Make the Timer a member variable of ResourceHandleInternal so
        that it can be canceled if the ResourceHandle is destroyed.

LayoutTests:

2008-03-29  Sam Weinig  <sam@webkit.org>

        Reviewed by Brady Eidson.

        Test for crash when canceling a resource load while port blocked failure
        timer is going

        * fast/loader/cancel-load-during-port-block-timer-expected.txt: Added.
        * fast/loader/cancel-load-during-port-block-timer.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31434 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver Hunt.

        <rdar://problem/5829556> REGRESSION: Leak in KJS::initializeThreading()

        * kjs/InitializeThreading.cpp: (KJS::initializeThreading): There is no guarantee that
        initializeThreading() is called only once; check that the mutex hasn't been already allocated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31433 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 17924: Crash in KJS::ConstDeclNode::evaluate with |with| and |const|
<http://bugs.webkit.org/show_bug.cgi?id=17924>
<rdar://problem/5806933>

Reviewed by Geoff.

It turns out this is trivially avoidable if we just match firefox's
semantics and ensure that an assignment in a const declaration always
writes to the variable object.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31431 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Stephanie  <slewis@apple.com>

        Fix 64bit build

        * WebCore.LP64.exp:
        * loader/archive/cf/LegacyWebArchiveMac.mm:
        (WebCore::createResourceResponseFromMacArchivedData):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31429 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Steve Falkenburg  <sfalken@apple.com>

        Enabled LTCG on files within "page".

        Rubber stamped by Oliver.

        * WebCore.vcproj/WebCore.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31427 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Dave Hyatt.

        - fix a percentage background position regression from r31389

        Test: fast/backgrounds/background-position-1.html

        * rendering/Length.h:
        (WebCore::Length::calcValue): Divide by 100.
        (WebCore::Length::calcMinValue): Ditto.

LayoutTests:

        Reviewed by Dave Hyatt.

        - pixel test for a percentage background position regression from r31389

        * fast/backgrounds/background-position-1.html: Added.
        * platform/mac/fast/backgrounds/background-position-1-expected.checksum: Added.
        * platform/mac/fast/backgrounds/background-position-1-expected.png: Added.
        * platform/mac/fast/backgrounds/background-position-1-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31426 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Oops. Back out unintentional change to RenderBlock.cpp

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31425 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  David Hyatt  <hyatt@apple.com>

        Back out the Ahem font antialiasing hack, since it is now no longer required for LCD antialiased text
        to match the reference rendering on Acid 3.

        * platform/graphics/SimpleFontData.h:
        * platform/graphics/mac/FontMac.mm:
        (WebCore::Font::drawGlyphs):
        * platform/graphics/mac/SimpleFontDataMac.mm:
        (WebCore::SimpleFontData::platformInit):
        * platform/graphics/win/FontCGWin.cpp:
        (WebCore::Font::drawGlyphs):
        * platform/graphics/win/SimpleFontDataCGWin.cpp:
        (WebCore::SimpleFontData::platformInit):
        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::positionNewFloats):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31424 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKit:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Rubberstamped by Darin Adler

        Remove WebArchiver.h/mm

        * WebKit.xcodeproj/project.pbxproj:

WebKit/mac:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Rubberstamped by Darin Adler

        Remove WebArchiver.h/mm

        * WebView/WebArchiver.h: Removed.
        * WebView/WebArchiver.mm: Removed.

        * DOM/WebDOMOperations.mm:
        * WebCoreSupport/WebDragClient.mm:
        * WebCoreSupport/WebEditorClient.mm:
        * WebView/WebDataSource.mm:
        * WebView/WebHTMLView.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31423 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin Adler

        Now that WebCore can create archives from a frame selection directly, we don't need it in WebArchiver anymore

        * WebCoreSupport/WebEditorClient.mm:
        (WebEditorClient::dataForArchivedSelection):

        * WebView/WebArchiver.h: Nuke archiveSelectionInFrame, as there are no remaining users
        * WebView/WebArchiver.mm: Ditto

        * WebView/WebHTMLView.mm:
        (-[WebHTMLView _writeSelectionWithPasteboardTypes:toPasteboard:cachedAttributedString:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31421 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31420 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin Adler

        More Kit->Core WebArchive changes.

        Create an archive from the current selection in a frame

        * WebCore.base.exp:

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

WebKit/mac:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin Adler

        More Kit->Core WebArchive changes.

        Create an archive from the current selection in a frame

        * WebView/WebArchiver.mm: Remove one more *undeclared* method, the last method will drop off
          easily in a followup

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31419 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Kevin McCullough  <kmccullough@apple.com>

        - Somehow managed to duplicate code :(

        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::didFinishLoading):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31418 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Kevin McCullough  <kmccullough@apple.com>

        - Missed some changes in previous checkin.

        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::send): Get identifier from synchronous loader.
        (WebCore::XMLHttpRequest::didFinishLoading):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31415 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Adam.

        -<rdar://problem/5712804> XMLHttpRequests do not show response contents, preview images (14313)
        -<rdar://problem/5712931> XMLHttpRequests should be shown separately and grouped (14315)
        -<rdar://problem/5732836> XMLHttpRequest: Inspector should show network activity/XHR in Console (17233)
        - Send the XMLHttpRequest data to the Inspector Controller.

        * loader/FrameLoader.cpp: Return the identifier for this resource since we need it in the XHR case.
        (WebCore::FrameLoader::loadResourceSynchronously):
        * loader/FrameLoader.h: Ditto.
        * page/InspectorController.cpp:
        (WebCore::XMLHttpRequestResource::XMLHttpRequestResource): We only need the data since the type will
        be XHR and the encoding is not used (we decoded earlier).
        (WebCore::XMLHttpRequestResource::~XMLHttpRequestResource): Locking is necessary when setting and clearing
        the KJS::UString::Rep.
        (WebCore::InspectorResource::setXMLHttpRequestProperties):
        (WebCore::InspectorResource::sourceString): Return the decoded source.
        (WebCore::addSourceToFrame): No longer do the decoding here, it has been encapsulated in the InspectorResource.
        (WebCore::InspectorController::resourceRetrievedByXMLHttpRequest):
        * page/InspectorController.h: Pass a UString instead of copying to a WebCore::String.  In case we
        never ask for the string's contents this will save us an extra copy.
        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::send): In the synchronous load case we normally do not keep the identifier
        but we need it to find the InspectorResource in order to tell it that it was loaded via XHR.
        (WebCore::XMLHttpRequest::didFinishLoading): Tell the InspectorController that its resource was loaded
        via XHR.
        * xml/XMLHttpRequest.h: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31414 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Stephanie Lewis  <slewis@apple.com>

        Update Windows Skipped list.  Bugs filed.

        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31413 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Sam Weinig

        WebArchive saga continues - Can now make archives from ranges in WebCore

        Again, previous layout test coverage exercised this code

        * WebCore.base.exp:

        * editing/markup.cpp:
        (WebCore::createFullMarkup): Added a "markup from range" variant that copies the previous WebKit implementation
        * editing/markup.h:

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

WebKit/mac:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Sam Weinig

        WebArchive saga continues - Can now make archives from ranges in WebCore

        * DOM/WebDOMOperations.mm:
        (-[DOMRange webArchive]):
        (-[DOMRange markupString]):

        * WebView/WebArchiver.h: Remove newly obsolete [WebArchiver archiveRange:]
        * WebView/WebArchiver.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31412 268f45cc-cd09-0410-ab3c-d52691b4dbfc

 2008-03-28  Stephanie Lewis  <slewis@apple.com>

         windows build fix.

         * platform/win/MainThreadWin.cpp:
         (WebCore::initializeThreadingAndMainThread):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31410 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Sam Weinig

        More Kit->Core webarchive code movement

        * WebCore.base.exp:
        * editing/markup.cpp:
        (WebCore::createFullMarkup): Replacement for [DOMNode markupString], creating full markup at this node including
          the document type string, which the WebKit implementation did indirectly
        * editing/markup.h:

WebKit/mac:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Sam Weinig

        More Kit->Core webarchive code movement

        * DOM/WebDOMOperations.mm:
        (-[DOMNode markupString]): Call createFullMarkup() instead

        * WebView/WebFrame.mm: Remove obsolete _markupStringFromNode
        * WebView/WebFrameInternal.h: Ditto

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31409 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Darin Adler  <darin@apple.com>

        Reviewed by Maciej.

        - cut down on copying and refcount churn a little by using references a bit more for
          KURL and String

        * bindings/js/JSXMLHttpRequest.cpp:
        (WebCore::jsXMLHttpRequestPrototypeFunctionOpen): Use const KURL& instead of KURL.
        * css/StyleSheet.h:
        (WebCore::StyleSheet::href): Return const String& instead of String.
        (WebCore::StyleSheet::title): Ditto.
        * dom/Document.cpp:
        (WebCore::Document::completeURL): Use const KURL* for a local variable instead of
        KURL to avoid copying the KURL.
        * dom/Element.cpp:
        (WebCore::Element::baseURI): Use const KURL& instead of KURL.
        * html/HTMLAnchorElement.cpp:
        (WebCore::HTMLAnchorElement::hostname): Ditto.
        * html/HTMLFrameElementBase.cpp:
        (WebCore::HTMLFrameElementBase::isURLAllowed): Ditto.
        * html/HTMLKeygenElement.cpp: Removed unneeded include of KURL.
        * page/ContextMenuController.cpp:
        (WebCore::ContextMenuController::contextMenuItemSelected): Get rid of temporary
        KURL.
        * platform/SecurityOrigin.cpp:
        (WebCore::SecurityOrigin::createForFrame): Use const KURL& instead of KURL.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31405 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Sam Weinig.

        Fix a dtoa thread safety issue.

        WebCore can call kjs_strtod without holding JS lock, but we didn't have thread safety
        compiled in for dtoa.

        This is a 0.5% regression on SunSpider, which Sam Weinig has volunteered to cover with
        his recent improvement.

        * kjs/dtoa.cpp:
        (Bigint::Balloc):
        (Bigint::Bfree):
        Changed to use fastMalloc/fastDelete - they are much faster than the dtoa custom version was
        in the presence of locking (but somewhat slower in single-threaded case).
        (Bigint::pow5mult): Got rid of the dreaded double-checked locking anti-pattern (had to
        restructure the code to avoid significant performance implications).
        (Bigint::lshift): Rewrote to avoid an allocation, if possible.

        (Bigint::rv_alloc):
        (Bigint::kjs_freedtoa):
        (Bigint::kjs_dtoa):
        Check for USE(MULTIPLE_THREADS), not dtoa legacy MULTIPLE_THREADS.

        * kjs/InitializeThreading.cpp: Added.
        (KJS::initializeThreading):
        * kjs/InitializeThreading.h: Added.
        Initialize threading at KJS level, if enabled.

        * kjs/dtoa.h: Expose dtoa mutex for KJS::initializeThreading.

        * kjs/testkjs.cpp: (kjsmain): Call initializeThreading.

        * JavaScriptCore.exp: Export KJS::initializeThreading.

        * GNUmakefile.am:
        * JavaScriptCore.exp:
        * JavaScriptCore.pri:
        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.vcproj:
        * JavaScriptCoreSources.bkl:
        * JavaScriptCore.xcodeproj/project.pbxproj:
        Added InitializeThreading.{h,cpp}.

        * wtf/Threading.h: Removed a using directive for WTF::initializeThreading - it is only
        to be called from KJS::initializeThreading, and having it in the global namespace is useless.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31404 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Brent Fulgham  <bfulgham@gmail.com>

        Reviewed by Maciej, landed by Brady

        Correct build regression (http://bugs.webkit.org/show_bug.cgi?id=18148)

        * platform/graphics/win/GraphicsContextCairoWin.cpp:
        (WebCore::GraphicsContext::getWindowsContext):
        (WebCore::GraphicsContext::releaseWindowsContext):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31403 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Sam Weinig

        Fold [WebArchiver archiveFrame:] into WebDataSource - the last remaining caller

        * WebView/WebArchiver.h:
        * WebView/WebArchiver.mm:
        * WebView/WebDataSource.mm:
        (-[WebDataSource webArchive]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31402 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Sam Weinig

        Remove unused [WebArchiver archiveNode:], made obsolete in r31400

        * WebView/WebArchiver.h:
        * WebView/WebArchiver.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31401 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin

        "Yet another transitional step" to empty out WebKit-based code for archiving.

        With this patch, the key operation of "Creating a WebArchive rooted at a single Node" takes place
        entirely within WebCore, and opens the door to saving WebArchives on Windows.

        This is another "zero behavior change" patch, and current editing/ and webarchive/ layout tests
        covered the relevant code

        * WebCore.base.exp:
        * loader/archive/cf/LegacyWebArchive.cpp:
        (WebCore::LegacyWebArchive::create):
        * loader/archive/cf/LegacyWebArchive.h:

WebKit/mac:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin

        "Yet another transitional step" to empty out WebKit-based code for archiving.

        With this patch, the key operation of "Creating a WebArchive rooted at a single Node" takes place
        entirely within WebCore, and opens the door to saving WebArchives on Windows.

        * DOM/WebDOMOperations.mm:
        * WebView/WebArchiver.mm:
        (+[WebArchiver _archiveWithMarkupString:fromFrame:nodes:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31400 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Jasper Bryant-Greene  <jasper@unix.geek.nz>

        Reviewed by Alp Toker.

        http://bugs.webkit.org/show_bug.cgi?id=18061
        PlatformScrollBarGtk attempts to allocate negative width and/or height for widgets

        Missing clampNegativeToZero() call meant that we tried to allocate widgets with negative width and/or height.

        * platform/gtk/PlatformScrollBarGtk.cpp:
        (PlatformScrollbar::geometryChanged):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31399 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin

        Export Unicode/UTF8.h and convertUTF16ToUTF8() for more flexible conversion in WebCore

        * JavaScriptCore.exp:
        * JavaScriptCore.xcodeproj/project.pbxproj:

WebCore:

2008-03-28  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin

        Using convertUTF16ToUTF8() from WTF, add a helper function that gives you a UTF8 SharedBuffer
        created from a String.

        * ForwardingHeaders/wtf/unicode/UTF8.h: Added.
        * platform/text/PlatformString.h:
        * platform/text/String.cpp:
        (WebCore::utf8Buffer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31398 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Darin Adler  <darin@apple.com>

        Reviewed by Adam.

        - eliminate a tiny bit of unnecessary refcount churn

        * bindings/js/JSEventTargetBase.cpp:
        (WebCore::eventNameForPropertyToken): Return const AtomicString&.
        * bindings/js/JSEventTargetBase.h:
        (WebCore::JSEventTargetBase::getValueProperty): Update for above change.
        (WebCore::JSEventTargetBase::putValueProperty): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31397 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Darin Adler  <darin@apple.com>

        Reviewed by Sam.

        * fast/encoding/invalid-xml-expected.txt: Update to reflect a true success.
        * fast/encoding/resources/invalid-xml.js: Eliminate race condition by not ending
        the test until the js-test-post.js script has loaded.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31396 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Rubber-stamped by Brady.

        Enable server-side includes via AddHandler, hopefully making Apache 1.3 happy.

        * http/tests/misc/.htaccess:
        * http/tests/misc/resources/acid3/.htaccess:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31395 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  Darin Adler  <darin@apple.com>

        * building/build.html: Add another mention of using the Cygwin shell.
        * building/tools.html: Fix a typo on the name Cygwin.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31394 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Not reviewed.

        Enable mod_include for acid3.html - the previous check-in only had it in resources
        subdirectory.

        * http/tests/misc/.htaccess: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31393 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the Qt build.

* rendering/Length.h: Include wtf/MathExtras.h for round().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31392 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by mjs.

        Add Acid3 test as an http test
        http://bugs.webkit.org/show_bug.cgi?id=18176

        * http/tests/misc/acid3-expected.checksum: Added.
        * http/tests/misc/acid3-expected.png: Added.
        * http/tests/misc/acid3-expected.txt: Added.
        * http/tests/misc/acid3.html: Added.
        * http/tests/misc/resources/acid3/.htaccess: Added.
        * http/tests/misc/resources/acid3/empty.css: Added.
        * http/tests/misc/resources/acid3/empty.html: Added.
        * http/tests/misc/resources/acid3/empty.png: Added.
        * http/tests/misc/resources/acid3/empty.txt: Added.
        * http/tests/misc/resources/acid3/empty.xml: Added.
        * http/tests/misc/resources/acid3/font.svg: Added.
        * http/tests/misc/resources/acid3/font.ttf: Added.
        * http/tests/misc/resources/acid3/instructions.inc: Added.
        * http/tests/misc/resources/acid3/reference.html: Added.
        * http/tests/misc/resources/acid3/reference.png: Added.
        * http/tests/misc/resources/acid3/support-a.png.404: Added.
        * http/tests/misc/resources/acid3/support-b.png: Added.
        * http/tests/misc/resources/acid3/support-c.png: Added.
        * http/tests/misc/resources/acid3/svg.xml: Added.
        * http/tests/misc/resources/acid3/xhtml.1: Added.
        * http/tests/misc/resources/acid3/xhtml.2: Added.
        * http/tests/misc/resources/acid3/xhtml.3: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31391 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add layout tests for background position rounding.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31390 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-28  David Hyatt  <hyatt@apple.com>

        Fix a bug where background-position truncates instead of rounding when it evaluates to fractional
        pixel values.  This matches other browsers.

        Reviewed by maciej

        Added fast/backgrounds/background-position-rounding.html

        * rendering/Length.h:
        (WebCore::Length::calcValue):
        (WebCore::Length::calcMinValue):
        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::calculateBackgroundImageGeometry):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31389 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-27  Darin Adler  <darin@apple.com>

        Reviewed by Mark Rowe.

        <rdar://problem/5826236> Regular expressions with large nested repetition counts can have their
        compiled length calculated incorrectly.

        * pcre/pcre_compile.cpp:
        (multiplyWithOverflowCheck):
        (calculateCompiledPatternLength): Check for overflow when dealing with nested repetition counts
        and bail with an error rather than returning incorrect results.

2008-03-27  Mark Rowe  <mrowe@apple.com>

        Reviewed by Adam Roben.

        Tests for <rdar://problem/5826236> Regular expressions with large nested repetition counts can have their
        compiled length calculated incorrectly.

        * fast/js/regexp-overflow-expected.txt:
        * fast/js/resources/regexp-overflow.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31388 268f45cc-cd09-0410-ab3c-d52691b4dbfc

wx build fix. Return a default value for operator ==  when !USE(WXGC) (not implemented in that case).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31387 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-27  Maciej Stachowiak  <mjs@apple.com>

        - fix build

        * svg/SVGGlyphMap.h:
        (WebCore::GlyphMapNode::create):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31386 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-27  Brady Eidson  <beidson@apple.com>

        Forgot a small requested change before committing

        * loader/DocumentLoader.cpp:
        (WebCore::DocumentLoader::mainResource):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31385 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2008-03-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Adam Roben

        Move [WebDataSource mainResource] and [WebDataSource subresources] down into WebCore
        as the push to core-ify WebArchives continues.

        This patch also introduces a behavior change.  WebCore allows ArchiveResources with null or empty data.
        WebKit has had the inexplicable distinction of allowing empty Data in a WebResource, but not null.
        Since WebResource is API, I decided to leave it be to avoid a behavior change.  But internally created resources
        (as in "while archiving a page") are accepting of null or empty data.

        This actually fixes a bug where not all subframes are archived, and resulted in a layout test change.

        * WebCore.base.exp:
        * loader/DocumentLoader.cpp:
        (WebCore::DocumentLoader::archiveResourceForURL): Make this const
        (WebCore::DocumentLoader::mainResource):
        (WebCore::DocumentLoader::subresource): Make this const and relocate
        (WebCore::DocumentLoader::getSubresources):
        * loader/DocumentLoader.h:

WebKit/mac:

2008-03-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Adam Roben

        Move [WebDataSource mainResource] and [WebDataSource subresources] down into WebCore
        as the push to core-ify WebArchives continues.

        This patch also introduces a behavior change.  WebCore allows ArchiveResources with null or empty data.
        WebKit has had the inexplicable distinction of allowing empty Data in a WebResource, but not null.
        Since WebResource is API, I decided to leave it be to avoid a behavior change.  But internally created resources
        (as in "while archiving a page") are accepting of null or empty data.

        This actually fixes a bug where not all subframes are archived, and resulted in a layout test change.

        * WebView/WebDataSource.mm:
        (-[WebDataSource mainResource]): Call DocumentLoader implementation
        (-[WebDataSource subresources]): Ditto
        * WebView/WebFrame.mm: Remove [WebFrame _getAllResourceDatas:andResponses:] as its only caller is obsolete
        * WebView/WebFrameInternal.h:

LayoutTests:

2008-03-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Adam Roben

        Updated results, as we now more consistently archive empty frames/empty resources

        * webarchive/archive-empty-frame-source-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31384 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-27  Maciej Stachowiak  <mjs@apple.com>

        Reviewed by Darin.

        - fixed many buildbot leaks in glyph map code

        * svg/SVGGlyphMap.h:
        (WebCore::GlyphMapNode::create): Use explicit create pattern, to avoid overreffing and therefore leaking
        these objects.
        (WebCore::SVGGlyphMap::add): Call create instead of using new.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31383 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-27  Maciej Stachowiak  <mjs@apple.com>

        Rubber stamped by Oliver.

        - re-fix Acid3, some code was incorrectly moved when fixing the Qt build

        * rendering/SVGRootInlineBox.cpp:
        (WebCore::SVGRootInlineBox::buildLayoutInformationForTextBox):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31381 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-27  Brady Eidson  <beidson@apple.com>

        Reviewed by Adam

        Change the "init from WebCore resource" version of WebResource to take PassRefPtr
        (more efficient)

        * WebView/WebResource.mm:
        (-[WebResource _initWithCoreResource:]):
        * WebView/WebResourceInternal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31377 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix compilation against Qt 4.4 and one missing SVG_FONTS #ifdef.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31376 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the Qt build and the build without ENABLE(SVG_FONTS)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31374 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Removed whitespace that crept in from r31357.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31371 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-26  Timothy Hatcher  <timothy@apple.com>

        Rename some Inspector files to facilitate the up-coming UI refresh changes.

        http://bugs.webkit.org/show_bug.cgi?id=17773

        Rubber-stamped by Adam Roben.

        * WebCore.vcproj/WebCore.vcproj:
        * page/inspector/Console.js: Copied from page/inspector/ConsolePanel.js.
        * page/inspector/ConsolePanel.js: Removed.
        * page/inspector/DocumentPanel.js: Removed.
        * page/inspector/ElementsPanel.js: Copied from page/inspector/DocumentPanel.js.
        * page/inspector/FontPanel.js: Removed.
        * page/inspector/FontView.js: Copied from page/inspector/FontPanel.js.
        * page/inspector/ImagePanel.js: Removed.
        * page/inspector/ImageView.js: Copied from page/inspector/ImagePanel.js.
        * page/inspector/NetworkPanel.js: Removed.
        * page/inspector/ResourcePanel.js: Removed.
        * page/inspector/ResourceView.js: Copied from page/inspector/ResourcePanel.js.
        * page/inspector/ResourcesPanel.js: Copied from page/inspector/NetworkPanel.js.
        * page/inspector/SourcePanel.js: Removed.
        * page/inspector/SourceView.js: Copied from page/inspector/SourcePanel.js.
        * page/inspector/WebKit.qrc:
        * page/inspector/inspector.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31370 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-26  Steve Falkenburg  <sfalken@apple.com>

        Enable LTCG on css, dom files in WebCore

        Reviewed by Stephanie.

        * WebCore.vcproj/WebCore.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31369 268f45cc-cd09-0410-ab3c-d52691b4dbfc

 2008-03-26  Stephanie Lewis  <slewis@apple.com>

         Fix Windows Build

         * WebCore.vcproj/WebCore.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31368 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Speculative GTK build fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31366 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-03-26  Brady Eidson  <beidson@apple.com>

        Build fix - accidentally checked in this change which was work in progress

        * DOM/WebDOMOperations.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31365 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Mac build fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@31363 268f45cc-cd09-0410-ab3c-d52691b4dbfc