1. 06 Feb, 2009 26 commits
    • eric@webkit.org's avatar
      2009-02-06 David Levin <levin@chromium.org> · 8e38883c
      eric@webkit.org authored
              Reviewed by Eric Seidel.
      
              <https://bugs.webkit.org/show_bug.cgi?id=23741>
      
              Bug 23798: KURLGoogle needs support for protocolInHTTPFamily
              <https://bugs.webkit.org/show_bug.cgi?id=23798>
      
              Add protocolInHTTPFamily() to the KURLGoogle implementation.
              This is another chromium build fix to mirror r40553.
      
              While in this file, I did some style fixes and consistency fixes:
              + fixed many_occurences_of_underscoring_naming
              + usages of NULL
              + comparisons to 0
              + changed the copyright remove an extraneous , and capitalize the (C)
                which was suggested for other files that I've submitted.
      
              No observable change in behavior, so no test.
      
              * platform/KURL.h:
              (WebCore::KURL::protocolInHTTPFamily):
              * platform/KURLGoogle.cpp:
              (WebCore::KURLCharsetConverter::ConvertFromUTF16):
              (WebCore::lowerCaseEqualsASCII):
              (WebCore::KURLGooglePrivate::KURLGooglePrivate):
              (WebCore::KURLGooglePrivate::setUtf8):
              (WebCore::KURLGooglePrivate::setAscii):
              (WebCore::KURLGooglePrivate::init):
              (WebCore::KURLGooglePrivate::initProtocolInHTTPFamily):
              (WebCore::KURLGooglePrivate::copyTo):
              (WebCore::KURLGooglePrivate::replaceComponents):
              (WebCore::KURL::KURL):
              (WebCore::KURL::createCFURL):
              (WebCore::KURL::isEmpty):
              (WebCore::KURL::protocolInHTTPFamily):
              (WebCore::KURL::query):
              (WebCore::decodeURLEscapeSequences):
              (WebCore::KURL::protocolIs):
              (WebCore::encodeWithURLEscapeSequences):
              (WebCore::KURL::invalidate):
              (WebCore::equalIgnoringRef):
              * platform/KURLGooglePrivate.h:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40735 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      8e38883c
    • ddkilzer@apple.com's avatar
      Bug 23741: StyleRareNonInheritedData::operator==() should not compare... · 1e3a190f
      ddkilzer@apple.com authored
              Bug 23741: StyleRareNonInheritedData::operator==() should not compare ContentData objects by pointer
      
              <https://bugs.webkit.org/show_bug.cgi?id=23741>
      
              Reviewed by Darin Adler.
      
              * rendering/style/RenderStyle.cpp:
              (WebCore::RenderStyle::contentDataEquivalent): Moved method to
              WebCore::StyleRareNonInheritedData class.
              * rendering/style/RenderStyle.h:
              (WebCore::InheritedFlags::contentDataEquivalent): Replaced declaration
              with inline method that calls contentDataEquivalent() on
              WebCore::StyleRareNonInheritedData instead.
              * rendering/style/StyleRareNonInheritedData.cpp:
              (WebCore::StyleRareNonInheritedData::operator==): This is the bug fix!
              Replaced pointer comparison of ContentData objects with call to
              StyleRareNonInheritedData::contentDataEquivalent().
              (WebCore::StyleRareNonInheritedData::contentDataEquivalent): Added.
              * rendering/style/StyleRareNonInheritedData.h:
              (WebCore::StyleRareNonInheritedData::contentDataEquivalent): Added
              declaration.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40734 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      1e3a190f
    • andersca@apple.com's avatar
      2009-02-06 Anders Carlsson <andersca@apple.com> · f1fd1c90
      andersca@apple.com authored
              Reviewed by Sam Weinig.
      
              <rdar://problem/6562220> 
              CrashTracer: [USER] 21 crashes in Safari at com.apple.WebKit • WebKit::NetscapePluginHostProxy::port
              
              Make the handling of crashes in the plug-in host more robust.
              
              * Plugins/Hosted/NetscapePluginHostProxy.h:
              Add m_portSet.
              
              * Plugins/Hosted/NetscapePluginHostProxy.mm:
              (WebKit::NetscapePluginHostProxy::NetscapePluginHostProxy):
              Initialize m_portSet.
              
              (WebKit::NetscapePluginHostProxy::~NetscapePluginHostProxy):
              Free m_portSet.
              
              (WebKit::NetscapePluginHostProxy::processRequests):
              Listen for messages on the port set. If we get a message to the port death notification port,
              then call pluginHostDied. Otherwise, process the message.
              
              * Plugins/Hosted/NetscapePluginInstanceProxy.h:
              * Plugins/Hosted/NetscapePluginInstanceProxy.mm:
              ...
      f1fd1c90
    • simon.fraser@apple.com's avatar
      2009-02-06 Simon Fraser <simon.fraser@apple.com> · 9583c054
      simon.fraser@apple.com authored
              Reviewed by Eric Seidel
      
              Factor code that checks whether the object is rooted in
              repaint() and repaintRectangle() into a common isRooted()
              method.
      
              * rendering/RenderObject.cpp:
              (WebCore::RenderObject::repaint):
              (WebCore::RenderObject::repaintRectangle):
              (WebCore::RenderObject::isRooted):
              * rendering/RenderObject.h:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40732 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      9583c054
    • simon.fraser@apple.com's avatar
      2009-02-06 Simon Fraser <simon.fraser@apple.com> · f0b50ae3
      simon.fraser@apple.com authored
              Reviewed by Eric Seidel
      
              Expose an isRootLayer() method on RenderLayer that works whether the
              layer is rooted or not, and use that in a few places in the
              accelerated compositing code.
      
              * rendering/RenderLayer.h:
              (WebCore::RenderLayer::isRootLayer):
              * rendering/RenderLayerBacking.cpp:
              (WebCore::RenderLayerBacking::paintingGoesToWindow):
              * rendering/RenderLayerCompositor.cpp:
              (WebCore::RenderLayerCompositor::rebuildCompositingLayerTree):
              (WebCore::RenderLayerCompositor::requiresCompositingLayer):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40731 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      f0b50ae3
    • simon.fraser@apple.com's avatar
      2009-02-06 Simon Fraser <simon.fraser@apple.com> · e99c986f
      simon.fraser@apple.com authored
              Reviewed by Eric Seidel
      
              Make RenderStyle::transitionForProperty() const.
      
              * rendering/style/RenderStyle.cpp:
              (WebCore::RenderStyle::transitionForProperty):
              * rendering/style/RenderStyle.h:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40730 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      e99c986f
    • mitz@apple.com's avatar
      Reviewed by Darin Adler. · 2955e088
      mitz@apple.com authored
              - fix <rdar://problem/6552338> FontCache::releaseFontData() called on a small caps font that is not in the cache
      
              Most platforms do not retrieve the small caps font from the FontCache,
              but cross-platform code in ~SimpleFontData() assumes the the small caps
              font data did come from the cache and therefore calls releaseFontData()
              on it.
      
              * platform/graphics/SimpleFontData.cpp:
              (WebCore::SimpleFontData::~SimpleFontData): Reordered to call
              platformDestroy() before releasing the small caps font and pruning the
              glyph page tree. This allows platformDestroy() to reset
              m_smallCapsFontData to 0 if it was not obtained from the cache.
              * platform/graphics/chromium/SimpleFontDataLinux.cpp:
              (WebCore::SimpleFontData::platformDestroy): Set m_smallCapsFontData to
              0 after deleting it.
              * platform/graphics/gtk/SimpleFontDataGtk.cpp:
              (WebCore::SimpleFontData::platformDestroy): Ditto.
              * platform/graphics/gtk/SimpleFontDataPango.cpp:
              (WebCore::SimpleFontData::platformDestroy): Ditto.
              * platform/graphics/win/SimpleFontDataWin.cpp:
              (WebCore::SimpleFontData::platformCommonDestroy): Ditto.
              * platform/graphics/wx/SimpleFontDataWx.cpp:
              (WebCore::SimpleFontData::platformDestroy): Ditto.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40729 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      2955e088
    • simon.fraser@apple.com's avatar
      2009-02-06 Simon Fraser <simon.fraser@apple.com> · f82e9ad8
      simon.fraser@apple.com authored
              Fix build when ACCELERATED_COMPOSITING is turned on.
      
              Remove font smoothing changes that will be committed
              separately.
      
              * platform/graphics/mac/WebLayer.mm:
              (WebCore::if):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40728 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      f82e9ad8
    • mitz@apple.com's avatar
      - try to fix the Tiger build · 76cda1c6
      mitz@apple.com authored
              * Misc/WebNSArrayExtras.h:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40727 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      76cda1c6
    • mitz@apple.com's avatar
      Reviewed by Anders Carlsson. · 4dc6aad8
      mitz@apple.com authored
              - fix an assertion failure in Vector::at() beneath
                WebHistoryItem::dictionaryRepresentation.
      
              * WebHistoryItem.cpp:
              (WebHistoryItem::dictionaryRepresentation): Give the numbers vector initial
              size. Also reduced the inline capacity of the vector used for weekly visit
              counts to 5, which is the expected maximum size.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40726 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      4dc6aad8
    • andersca@apple.com's avatar
      2009-02-06 Anders Carlsson <andersca@apple.com> · 605a9ccf
      andersca@apple.com authored
              Build fix.
              
              * bridge/npapi.h:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40725 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      605a9ccf
    • darin@chromium.org's avatar
      2009-02-06 Darin Fisher <darin@chromium.org> · a687d8ed
      darin@chromium.org authored
              Reviewed by Eric Seidel.
      
              https://bugs.webkit.org/show_bug.cgi?id=23790
              Custom -> JSCCustom in SVGPointList.idl
      
              * svg/SVGPointList.idl:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40724 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      a687d8ed
    • andersca@apple.com's avatar
      2009-02-06 Anders Carlsson <andersca@apple.com> · 36914638
      andersca@apple.com authored
              Reviewed by Dan Bernstein.
      
              <rdar://problem/6563650> 
              Add Netscape plug-in API to tell the browser not to load streams (some plug-ins handle network loading manually)
                      
              * Plugins/WebNetscapePluginView.mm:
              (-[WebNetscapePluginView loadStream]):
              (-[WebNetscapePluginView pluginView:receivedData:]):
              (-[WebNetscapePluginView _shouldCancelSrcStream]):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40723 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      36914638
    • dglazkov@chromium.org's avatar
      2009-02-06 Anantanarayanan Iyengar <ananta@chromium.org> · 83f56c1f
      dglazkov@chromium.org authored
              Reviewed by Eric Seidel.
      
              https://bugs.webkit.org/show_bug.cgi?id=23309
              Fix whitespace indent errors. The rest of the original patch has already
              landed.
      
              * platform/graphics/cairo/ImageSourceCairo.cpp:
              (WebCore::ImageSource::clear):
              * platform/graphics/cg/ImageSourceCG.cpp:
              (WebCore::ImageSource::clear):
              * platform/graphics/qt/ImageSourceQt.cpp:
              (WebCore::ImageSource::clear):
              * platform/graphics/wx/ImageSourceWx.cpp:
              (WebCore::ImageSource::clear):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40722 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      83f56c1f
    • krit@webkit.org's avatar
      2009-02-06 Dirk Schulze <krit@webkit.org> · 13bb7deb
      krit@webkit.org authored
              Reviewed by Simon Hausmann.
      
              Fix bug in clearRect(). Use fillRect() instead of eraseRect() to get
              the context transparent.
      
              [QT] clearRect fill's a given rect with white
              https://bugs.webkit.org/show_bug.cgi?id=23728
      
              * platform/graphics/qt/GraphicsContextQt.cpp:
              (WebCore::GraphicsContext::clearRect):
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40721 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      13bb7deb
    • bfulgham@webkit.org's avatar
      Build fix. Re-enable missing file in Release_Cairo target. · d6f5f8e6
      bfulgham@webkit.org authored
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40720 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d6f5f8e6
    • simon.fraser@apple.com's avatar
      2009-02-06 Simon Fraser <simon.fraser@apple.com> · 8f9c77ba
      simon.fraser@apple.com authored
              Reviewed by Dave Hyatt
      
              https://bugs.webkit.org/show_bug.cgi?id=23781
      
              When the document is attached and detached, and when it becomes active/inactive
              (e.g. coming out or going into the page cache) we need to call
              RenderView::didMoveOnscreen()/willMoveOffscreen() to hook and unhook the composited
              layers.
      
              Also, if Document::recalcStyle() does not do a layout, it needs to
              update the compositing layers.
      
              * dom/Document.cpp:
              (WebCore::Document::updateRendering):
              (WebCore::Document::updateLayout):
              (WebCore::Document::attach):
              (WebCore::Document::detach):
              (WebCore::Document::documentWillBecomeInactive):
              (WebCore::Document::documentDidBecomeActive):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40719 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      8f9c77ba
    • hausmann@webkit.org's avatar
      2009-02-06 Kavindra Palaraja <kavindra.palaraja@nokia.com> · 3ab6f4c7
      hausmann@webkit.org authored
              Reviewed by Simon Hausmann.
      
              Doc - made 2 functions internal as they are not part of the API anymore (and this fixes 2 qdoc warnings)
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40718 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      3ab6f4c7
    • hausmann@webkit.org's avatar
      WebCore: · 5256c46a
      hausmann@webkit.org authored
          2009-02-06  Simon Hausmann  <simon.hausmann@nokia.com>
      
                  Reviewed by Tor Arne Vestbø.
      
                  Added support for different ownership models for wrapped QObjects in
                  the JavaScript environment.
      
      WebKit/qt:
      
          2009-02-06  Simon Hausmann  <simon.hausmann@nokia.com>
      
                  Reviewed by Tor Arne Vestbø.
      
                  Added an overload of QWebFrame::addToJavaScriptWindowObject that takes a QScriptEngine::ValueOwnership parameter.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40717 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      5256c46a
    • ap@webkit.org's avatar
      2009-02-06 Xan Lopez <xan@gnome.org> · 8085721d
      ap@webkit.org authored
              Reviewed by Alexey Proskuryakov.
      
              https://bugs.webkit.org/show_bug.cgi?id=23769
      
              Do not use empty ResourceError errors.
      
              Rather create bogus but non-null errors, since some codepaths
              expect these. For example, see DocumentLoader::mainReceivedError.
      
              * WebCoreSupport/FrameLoaderClientGtk.cpp:
              (WebKit::FrameLoaderClient::cancelledError):
              (WebKit::FrameLoaderClient::blockedError):
              (WebKit::FrameLoaderClient::cannotShowURLError):
              (WebKit::FrameLoaderClient::interruptForPolicyChangeError):
              (WebKit::FrameLoaderClient::cannotShowMIMETypeError):
              (WebKit::FrameLoaderClient::fileDoesNotExistError):
              (WebKit::FrameLoaderClient::pluginWillHandleLoadError):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40716 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      8085721d
    • ap@webkit.org's avatar
      2009-02-06 Xan Lopez <xan@gnome.org> · 3125c2a4
      ap@webkit.org authored
              Reviewed by Alexey Proskuryakov.
      
              https://bugs.webkit.org/show_bug.cgi?id=23761
      
              Use two-arg KURL ctor.
      
              We are using the one-arg ctor currently, but:
      
              - It assumes the strings are already encoded, which is not
              necesarily the case for us.
      
              - The single-argument KURL ctors expect their input to already be
              the output of a previous KURL::parse call, so for the general
              case (ie, random user input) we need to use the two-arg ctor
              anyway.
      
              * webkit/webkitwebframe.cpp:
              * webkit/webkitwebview.cpp:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40715 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      3125c2a4
    • zecke@webkit.org's avatar
      2009-02-06 Aaron Boodman <aa@chromium.org> · d7cbb5ce
      zecke@webkit.org authored
              Reviewed by Holger Freyther.
      
              * WebCoreSupport/FrameLoaderClientQt.cpp:
              (WebCore::FrameLoaderClientQt::documentElementAvailable):)
              Fix compile error in Qt build introduced by the below change (r40694).
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40714 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d7cbb5ce
    • mjs@apple.com's avatar
      2009-02-06 Maciej Stachowiak <mjs@apple.com> · 4c0f5a52
      mjs@apple.com authored
              Rubber stamped by Dan Bernstein.
      
              - fix obvious problem in previous commit (|| used instead of &&)
      
              * WebHistoryItem.cpp:
              (WebHistoryItem::initFromDictionaryRepresentation):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40713 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      4c0f5a52
    • mjs@apple.com's avatar
      WebCore: · 13d04dd5
      mjs@apple.com authored
      2009-02-05  Maciej Stachowiak  <mjs@apple.com> and Brady Eidson  <beidson@apple.com>
      
              Reviewed by Dan Bernstein and Geoff Garen.
      
              - WebCore code to track per-day and per-week visit counts in history
      
              The basic idea is as follows: we track daily visits for up to 13
              days, and weekly visits for up to 5 weeks past the end of the
              current daily visits. As soon as the number of daily counts goes
              over 13, we squish them down into weeks, and then prune the excess weeks.
      
              * history/HistoryItem.cpp:
              (WebCore::HistoryItem::HistoryItem): Remove bogus whitespace.
              (WebCore::timeToDay): Helper function: convert time to days from the epoch.
              (WebCore::HistoryItem::padDailyCountsForNewVisit): Insert 0 counts if
              a site is visited after a span of days.
              (WebCore::HistoryItem::collapseDailyVisitsToWeekly): Core day/week upgrade
              algorithm as described above.
              (WebCore::HistoryItem::recordVisitAtTime): Use above helpers to record all
              info about a visit to this history item.
              (WebCore::HistoryItem::setLastVisitedTime): Updated.
              (WebCore::HistoryItem::visited): Updated.
              (WebCore::HistoryItem::recordInitialVisit): New function to
              handle recording data about the very first visit, replacing the
              prior practice of explicitly setting visit count to 1.
              (WebCore::HistoryItem::adoptVisitCounts): Take ownership of
              daily and weekly visit count vectors.
              (WebCore::HistoryItem::mergeAutoCompleteHints): Added a comment
              noting that this function doesn't work properly now (though
              this shouldn't matter much in practice.)
              * history/HistoryItem.h:
              (WebCore::HistoryItem::dailyVisitCounts): Accessor for daily counts.
              (WebCore::HistoryItem::weeklyVisitCounts): Ditto for weekly counts.
              * WebCore.base.exp: Add new symbols and sort.
      
      WebKit/win:
      
      2009-02-05  Maciej Stachowiak  <mjs@apple.com> and Brady Eidson  <beidson@apple.com>
      
              Reviewed by Dan Bernstein and Geoff Garen..
      
              - WebKit code to track per-day and per-week visit counts in history
      
              For now this data is only exposed via SPI for performance reasons.
      
              * Interfaces/IWebHistoryItemPrivate.idl: Added new interface.
              * WebHistory.cpp:
              (WebHistory::visitedURL): Use new recordInitialVisit method.
              * WebHistoryItem.cpp:
              (WebHistoryItem::initFromDictionaryRepresentation): Add parsing support
              for new data.
              (WebHistoryItem::dictionaryRepresentation): Add saving support for
              new data.
              (WebHistoryItem::getDailyVisitCounts): SPI accessor.
              (WebHistoryItem::getWeeklyVisitCounts): SPI accessor.
              (WebHistoryItem::recordInitialVisit): Tell WebCore to record an initial visit.
              * WebHistoryItem.h:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40712 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      13d04dd5
    • mjs@apple.com's avatar
      2009-02-05 Maciej Stachowiak <mjs@apple.com> and Brady Eidson <beidson@apple.com> · 5a4dc3cf
      mjs@apple.com authored
              Reviewed by Dan Bernstein and Geoff Garen.
      
              - WebKit code to track per-day and per-week visit counts in history
      
              For now this data is only exposed via SPI for performance reasons.
              
              * History/WebHistoryItem.mm:
              (-[WebHistoryItem initFromDictionaryRepresentation:]): Add parsing support
              for new data.
              (-[WebHistoryItem _recordInitialVisit]): Tell WebCore to record an initial visit.
              (-[WebHistoryItem dictionaryRepresentation]): Add saving support for new data.
              (-[WebHistoryItem _getDailyVisitCounts:]): SPI accessor.
              (-[WebHistoryItem _getWeeklyVisitCounts:]): SPI accessor.
              * History/WebHistoryItemInternal.h: Declare new methods.
              * History/WebHistoryItemPrivate.h: Ditto.
      
              * History/WebHistory.mm:
              (-[WebHistoryPrivate visitedURL:withTitle:]): For the initial visit, use
              the new _recordInitialVisit method instead of setting visit count to 1.
      
              * Misc/WebNSArrayExtras.h:
              * Misc/WebNSArrayExtras.m:
              (-[NSArray _webkit_numberAtIndex:]): Helper to retrieve an NSNumber or nil from an NSArray
              (-[NSArray _webkit_stringAtIndex:]): Helper to retrieve an NSString of nil from an NSArray
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40711 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      5a4dc3cf
    • ap@webkit.org's avatar
      Reviewed by Darin Adler. · 73ecea19
      ap@webkit.org authored
              <rdar://problem/6438271> Improve backspace handling of accented characters
      
              Test: platform/mac/editing/deleting/backward-delete.html
      
              This change makes WebKit behave like AppKit on the Mac. Other platforms are not affected,
              because the current behavior matches their standards (tested on Windows, assumed elsewhere).
      
              * dom/Position.cpp:
              (WebCore::Position::previous):
              (WebCore::Position::next):
              (WebCore::Position::uncheckedPreviousOffsetForBackwardDeletion):
              * dom/Position.h:
              Added a new option for previous(), used to move to a next position for backward deletion.
              Renamed EUsingComposedCharacters to PositionMoveType.
      
              * editing/TypingCommand.cpp: (WebCore::TypingCommand::deleteKeyPressed):
              Use previous(BackwardDeletion) to find a proper range to delete. Also, simplified a check
              surrounding this code a little.
      
              * editing/htmlediting.cpp:
              (WebCore::nextVisuallyDistinctCandidate):
              (WebCore::previousVisuallyDistinctCandidate):
              Adjusted for renamed enum values.
      
              * rendering/RenderObject.cpp:
              * rendering/RenderObject.h:
              * rendering/RenderText.h:
              Added previousOffsetForBackwardDeletion().
      
              * rendering/RenderText.cpp: (WebCore::RenderText::previousOffsetForBackwardDeletion):
              On PLATFORM(MAC), use an algorithm that matches the one AppKit has for backward deletion.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@40710 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      73ecea19
  2. 05 Feb, 2009 14 commits