1. 21 Jun, 2008 5 commits
  2. 20 Jun, 2008 25 commits
    • timothy@apple.com's avatar
      Fixed an ASSERT(m_actualSelfTime <= m_actualTotalTime) when starting · e99936a5
      timothy@apple.com authored
              and stopping a profile from the Develop menu. Also prevents
              inserting an incorrect parent node as the new head after profiling
              is stopped from the Develop menu.
      
              Reviewed by Dan Bernstein.
      
              * profiler/Profile.cpp:
              (KJS::Profile::stopProfiling): If the current node is already the head
              then there is no more need to record future nodes in didExecute.
              (KJS::Profile::didExecute): Move the code of setupCurrentNodeAsStopped
              into here since this was the only caller. When setting the total time
              keep any current total time while adding the self time of the head.
              (KJS::Profile::setupCurrentNodeAsStopped): Removed.
              * profiler/Profile.h: Removed setupCurrentNodeAsStopped.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34710 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      e99936a5
    • weinig@apple.com's avatar
      2008-06-20 Sam Weinig <sam@webkit.org> · 408f706f
      weinig@apple.com authored
              Fix Windows build.
      
              * bindings/scripts/CodeGeneratorCOM.pm:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34709 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      408f706f
    • jmalonzo@webkit.org's avatar
      2008-06-20 Jan Michael Alonzo <jmalonzo@webkit.org> · bf50b999
      jmalonzo@webkit.org authored
              Gtk build fix: Add files missing in the previous fix (r34705)
      
              * GNUmakefile.am:
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34708 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      bf50b999
    • mitz@apple.com's avatar
      Reviewed by Sam Weinig. · 6f46ccb5
      mitz@apple.com authored
              - fix leak of mask images
      
              * rendering/RenderObject.cpp:
              (WebCore::RenderObject::arenaDelete): Added a call to removeClient() on
              the mask box image.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34707 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      6f46ccb5
    • pewtermoose@webkit.org's avatar
      2008-06-20 Brent Fulgham <bfulgham@gmail.com> · 46655364
      pewtermoose@webkit.org authored
              Reviewed by Darin.
      
              Extend the build-webkit (and set-webkit-configuration) script to
              support Cairo-based webkit builds.  (see http://bugs.webkit.org/show_bug.cgi?17952)
      
              * Scripts/build-webkit:  Add --cairo-win32 to the help message
              * Scripts/webkitdirs.pm:  Extend the 'determinePassedConfiguration
                subroutine to recognize the --cairo-win32 flag.  When present,
                the build configuration is changed from Debug/Release to
                Debug_Cairo/Release_Cairo.  This flag is only active when the
                isCygwin() test is true.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34706 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      46655364
    • jmalonzo@webkit.org's avatar
      2008-06-20 Jan Michael Alonzo <jmalonzo@webkit.org> · b55a99ec
      jmalonzo@webkit.org authored
              Gtk build fix for r34702, r34700 and r34693
              Qt build fix for r34700 and r34693
      
              * GNUmakefile.am:
              * WebCore.pro:
              * platform/gtk/RenderThemeGtk.cpp:
              * platform/gtk/RenderThemeGtk.h:
              * platform/qt/RenderThemeQt.cpp:
              * platform/qt/RenderThemeQt.h:
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34705 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      b55a99ec
    • kevino@webkit.org's avatar
      !USE(MULTIPLE_THREADS) on Darwin build fix · 94e1b3c6
      kevino@webkit.org authored
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34704 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      94e1b3c6
    • kmccullough@apple.com's avatar
      2008-06-20 Kevin McCullough <kmccullough@apple.com> · bdc1acdd
      kmccullough@apple.com authored
              -Leopard Build Fix.
      
              * profiler/Profile.cpp:
              (KJS::Profile::removeProfileStart):
              (KJS::Profile::removeProfileEnd):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34703 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      bdc1acdd
    • weinig@apple.com's avatar
      2008-06-20 Sam Weinig <sam@webkit.org> · 50bca617
      weinig@apple.com authored
              Reviewed by Adele Peterson.
      
              Add 'files' property to the HTMLInputElement, which returns a FileList object 
              (when type=file, null otherwise) containing a list of all the files selected.
              It currently always contains only 0 or 1 files as multifile input is not supported 
              yet.  The list contains File objects which contains the name and size of the file. 
              The inspiration for these interfaces is from:
                  - http://developer.mozilla.org/en/docs/nsIDOMFileList
                  - http://developer.mozilla.org/en/docs/nsIDOMFile
      
              Also fixes <rdar://problem/6022802>
      
              * DerivedSources.make:
              * GNUmakefile.am:
              * WebCore.pro:
              * WebCore.vcproj/WebCore.vcproj:
              * WebCore.xcodeproj/project.pbxproj:
              * WebCoreSources.bkl:
              * bindings/objc/DOMInternal.h:
              * bindings/scripts/CodeGenerator.pm:
              * bindings/scripts/IDLStructure.pm:
              * html/File.cpp: Added.
              (WebCore::File::File):
              (WebCore::File::fileSize):
              * html/File.h: Added.
              (WebCore::File::create):
              (WebCore::File::fileName):
              (WebCore::File::path):
              * html/File.idl: Added.
              * html/FileList.cpp: Added.
              (WebCore::FileList::FileList):
              (WebCore::FileList::item):
              * html/FileList.h: Added.
              (WebCore::FileList::create):
              (WebCore::FileList::length):
              (WebCore::FileList::isEmpty):
              (WebCore::FileList::clear):
              (WebCore::FileList::append):
              * html/FileList.idl: Added.
              * html/HTMLInputElement.cpp:
              (WebCore::HTMLInputElement::setInputType):
              (WebCore::HTMLInputElement::appendFormData):
              (WebCore::HTMLInputElement::value):
              (WebCore::HTMLInputElement::setValue):
              (WebCore::HTMLInputElement::setValueFromRenderer):
              (WebCore::HTMLInputElement::files):
              * html/HTMLInputElement.h:
              * html/HTMLInputElement.idl:
              * page/DOMWindow.idl:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34702 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      50bca617
    • hyatt@apple.com's avatar
      WebCore: · 36b321e5
      hyatt@apple.com authored
      2008-06-20  David Hyatt  <hyatt@apple.com>
      
              Make sure CSS variables work inside the inline style attribute.
      
              Reviewed by Beth
      
              Added fast/css/variables/inline-style-test.html
      
              * css/CSSMutableStyleDeclaration.cpp:
              (WebCore::CSSMutableStyleDeclaration::removeProperty):
              (WebCore::CSSMutableStyleDeclaration::addParsedProperties):
      
      LayoutTests:
      
      2008-06-20  David Hyatt  <hyatt@apple.com>
      
              Make sure CSS variables work inside the inline style attribute.
      
              Reviewed by Beth
      
              * fast/css/variables/inline-style-test.html: Added.
              * platform/mac/fast/css/variables/inline-style-test-expected.checksum: Added.
              * platform/mac/fast/css/variables/inline-style-test-expected.png: Added.
              * platform/mac/fast/css/variables/inline-style-test-expected.txt: Added.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34701 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      36b321e5
    • hyatt@apple.com's avatar
      WebCore: · 04dcb131
      hyatt@apple.com authored
      2008-06-20  David Hyatt  <hyatt@apple.com>
      
              Add support for the CSSVariablesRule and CSSVariablesDeclaration DOM APIs.  These allow querying of
              variables names and values, as well as iteration, setting and removal.
      
              Reviewed by Sam
      
              Added multiple new tests to fast/css/variables/
      
              * DerivedSources.make:
              * WebCore.xcodeproj/project.pbxproj:
              * bindings/js/JSCSSRuleCustom.cpp:
              (WebCore::toJS):
              * bindings/objc/DOMInternal.h:
              * bindings/scripts/CodeGeneratorJS.pm:
              * css/CSSPrimitiveValue.cpp:
              (WebCore::CSSPrimitiveValue::parserValue):
              * css/CSSVariablesDeclaration.cpp:
              (WebCore::CSSVariablesDeclaration::removeVariable):
              (WebCore::CSSVariablesDeclaration::setVariable):
              (WebCore::CSSVariablesDeclaration::setCssText):
              (WebCore::CSSVariablesDeclaration::setChanged):
              * css/CSSVariablesDeclaration.h:
              * css/CSSVariablesDeclaration.idl: Added.
              * css/CSSVariablesRule.h:
              * css/CSSVariablesRule.idl: Added.
      
      LayoutTests:
      
      2008-06-20  David Hyatt  <hyatt@apple.com>
      
              Add tests of the DOM APIs for addition, removal and iteration.
      
              Reviewed by Sam
      
              * fast/css/variables/remove-variable-test.html: Added.
              * fast/css/variables/set-variable-test.html: Added.
              * fast/css/variables/variable-iteration-test.html: Added.
              * platform/mac/fast/css/variables/remove-variable-test-expected.checksum: Added.
              * platform/mac/fast/css/variables/remove-variable-test-expected.png: Added.
              * platform/mac/fast/css/variables/remove-variable-test-expected.txt: Added.
              * platform/mac/fast/css/variables/set-variable-test-expected.checksum: Added.
              * platform/mac/fast/css/variables/set-variable-test-expected.png: Added.
              * platform/mac/fast/css/variables/set-variable-test-expected.txt: Added.
              * platform/mac/fast/css/variables/variable-iteration-test-expected.checksum: Added.
              * platform/mac/fast/css/variables/variable-iteration-test-expected.png: Added.
              * platform/mac/fast/css/variables/variable-iteration-test-expected.txt: Added.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34700 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      04dcb131
    • kmccullough@apple.com's avatar
      2008-06-20 Kevin McCullough <kmccullough@apple.com> · de7f5782
      kmccullough@apple.com authored
              Just giving credit.
      
              * ChangeLog:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34699 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      de7f5782
    • kmccullough@apple.com's avatar
      2008-06-20 Kevin McCullough <kmccullough@apple.com> · 84454c38
      kmccullough@apple.com authored
              Reviewed by Tim.
      
              <rdar://problem/6024846> JSProfiler: ASSERT hit in Profiler.
              - Because InspectorController can call startProfiling() and
              stopProfiling() we cannot assert that console.profile() and
              console.profileEnd() will be in the profile tree.
      
              * profiler/Profile.cpp:
              (KJS::Profile::removeProfileStart):
              (KJS::Profile::removeProfileEnd):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34698 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      84454c38
    • kmccullough@apple.com's avatar
      2008-06-20 Kevin McCullough <kmccullough@apple.com> · 223b56c4
      kmccullough@apple.com authored
              Rubber stamped by Adele. 
      
              * manual-tests/inspector/profiler-test-stop-profiling-after-setTimeout.html: Added.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34697 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      223b56c4
    • kmccullough@apple.com's avatar
      JavaScriptCore: · c9005aea
      kmccullough@apple.com authored
      2008-06-20  Kevin McCullough  <kmccullough@apple.com>
      
              Reviewed by Tim.
      
              <rdar://problem/5958770> JSProfiler: Time incorrectly given to (idle)
              if profiling is started and finished within the same function. (19230)
              - Now we profile one more stack frame up from the last frame to allocate
              the time spent in it, if it exists.
      
              * JavaScriptCore.exp:
              * VM/Machine.cpp: We need to let the profiler know when the JS program 
              has finished since that is what will actually stop the profiler instead
              of just calling stopProfiling().
              (KJS::Machine::execute):
              * profiler/Profile.cpp:
              (KJS::Profile::create): Moved from Profile.h since it was getting pretty
              long.
              (KJS::Profile::Profile): We now have a client, which is a listener who
              we will return this profile to, once it has actually finished.
              (KJS::Profile::stopProfiling): Instead of fully stopping the profiler
              here, we set the flag and keep it profiling in the background.
              (KJS::Profile::didFinishAllExecution): This is where the profiler
              actually finishes and creates the (idle) node if one should be made.
              (KJS::Profile::removeProfileStart): Don't use m_currentNode since it is
              needed by the profiler as it runs silently in the background.
              (KJS::Profile::removeProfileEnd): Ditto.
              (KJS::Profile::willExecute): Don't profile new functions if we have
              stopped profiling.
              (KJS::Profile::didExecute): Only record one more return as all the
              remaining time will be attributed to that function.
              (KJS::Profile::setupCurrentNodeAsStopped): Sets the current node's time.
              * profiler/Profile.h: Added functions and variables for the above
              changes.
              (KJS::Profile::client):
              * profiler/ProfileNode.h:
              (KJS::CallIdentifier::toString): Debug method.
              * profiler/Profiler.cpp: Added support for the ProfilerClient.
              (KJS::Profiler::startProfiling):
              (KJS::Profiler::stopProfiling): No longer return sthe profile.
              (KJS::Profiler::didFinishAllExecution): Now returns the profile to the
              client instead of stopProfiling.
              * profiler/Profiler.h:
              (KJS::ProfilerClient::~ProfilerClient): Clients will implement this
              interface.
      
      WebCore:
      
      2008-06-20  Kevin McCullough  <kmccullough@apple.com>
      
              Reviewed by Tim.
      
              <rdar://problem/5958770> JSProfiler: Time incorrectly given to (idle)
              if profiling is started and finished within the same function. (19230)
              - Now we profile one more stack frame up from the last frame to allocate
              the time spent in it, if it exists.
      
              * page/Console.cpp:
              * manual-tests/inspector/profiler-test-start-and-stop-profiling-in-the-same-function.html: Added.
              (WebCore::Console::profile): When stating the profiler give a client for
              the callback of when the profile actually finishes.
              (WebCore::Console::profileEnd): No longer needs to handle the return of
              the profile object since it will be retruned in the client's callback.
              (WebCore::Console::finishedProfiling): Implemenet the ProfileClient
              callback method.
              * page/Console.h: Inherit from the ProfileClient.
              * page/InspectorController.cpp:
              (WebCore::InspectorController::startUserInitiatedProfiling): Use the
              client callback.
              (WebCore::InspectorController::stopUserInitiatedProfiling): Does not
              need to handle the profile being returned as it is now handled by the
              client callback.
              (WebCore::InspectorController::finishedProfiling): Implement the
              ProfileClient callback method.
              * page/InspectorController.h: Inherit from the ProfileClient.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34696 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c9005aea
    • timothy@apple.com's avatar
      Makes the JavaScript syntax highlighter process lines in chunks · ed74a7a7
      timothy@apple.com authored
              so the user interface isn't blocked for large script files.
      
              https://bugs.webkit.org/show_bug.cgi?id=19677
      
              Reviewed by Adam Roben.
      
              * page/inspector/SourceFrame.js:
              (WebInspector.SourceFrame.prototype._addMessageToSource): Use the cells
              property on the row instead of getElementsByTagName.
              (WebInspector.SourceFrame.prototype.syntaxHighlightJavascript): Use the
              cells property on the row instead of getElementsByTagName. Added
              a nested processChunk function that highlights 10 lines at a time
              This processChunk function is called at an interval of 25ms. The
              code is still highlighted quickly, and the user can't tell it wasn't
              highlighted all at once.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34695 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      ed74a7a7
    • timothy@apple.com's avatar
      Fixes a bug where the source view in the Resources panel had a · 9dce71be
      timothy@apple.com authored
              couple of extra pixels at the bottom when fully scrolled.
      
              Reviewed by Adam Roben.
      
              * page/inspector/SourceFrame.js:
              (WebInspector.SourceFrame.prototype.set autoSizesToFitContentHeight):
              Call removeStyleClass instead of addStyleClass for "webkit-height-sized-to-fit"
              when sizing to fit is being disabled.
              * page/inspector/inspector.css:
              (.resource-view.headers-visible .source-view-frame): Added the
              vertical-align: top property to prevent line alignment from adding extra
              pixels on the bottom.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34694 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      9dce71be
    • timothy@apple.com's avatar
      Fixes: Bug 19679: iframes with a height of 32,768px or greater do not layout correctly · c59ad194
      timothy@apple.com authored
      WebCore:
      
      2008-06-19  Timothy Hatcher  <timothy@apple.com>
      
              Changed all lineHeight, baselinePosition and verticalPositionHint
              calls to return int instead of short. The short was overflowing
              when a value greater than 32,767 was encountered.
      
              Fixes: iframes with a height of 32,768px or greater do not layout correctly
              https://bugs.webkit.org/show_bug.cgi?id=19679
      
              Reviewed by Dave Hyatt.
      
              Test: fast/css/line-height-overflow.html
      
              * rendering/RenderBR.cpp:
              (WebCore::RenderBR::baselinePosition):
              * rendering/RenderBR.h:
              * rendering/RenderBlock.cpp:
              (WebCore::RenderBlock::lineHeight):
              * rendering/RenderBlock.h:
              * rendering/RenderFlow.cpp:
              * rendering/RenderFlow.h:
              * rendering/RenderListBox.cpp:
              * rendering/RenderListBox.h:
              * rendering/RenderListMarker.cpp:
              (WebCore::RenderListMarker::lineHeight):
              * rendering/RenderListMarker.h:
              * rendering/RenderObject.cpp:
              (WebCore::RenderObject::verticalPositionHint):
              * rendering/RenderObject.h:
              (WebCore::):
              * rendering/RenderPath.cpp:
              (WebCore::RenderPath::lineHeight):
              * rendering/RenderPath.h:
              * rendering/RenderReplaced.cpp:
              (WebCore::RenderReplaced::lineHeight):
              * rendering/RenderReplaced.h:
              * rendering/RenderSVGContainer.cpp:
              (WebCore::RenderSVGContainer::lineHeight):
              * rendering/RenderSVGContainer.h:
              * rendering/RenderSVGHiddenContainer.cpp:
              (WebCore::RenderSVGHiddenContainer::lineHeight):
              * rendering/RenderSVGHiddenContainer.h:
              * rendering/RenderSVGRoot.cpp:
              (WebCore::RenderSVGRoot::lineHeight):
              * rendering/RenderSVGRoot.h:
              * rendering/RenderSlider.cpp:
              * rendering/RenderSlider.h:
              * rendering/RenderTableCell.cpp:
              * rendering/RenderTableCell.h:
              * rendering/RenderTableCol.h:
              * rendering/RenderTableRow.h:
              * rendering/RenderTableSection.h:
              * rendering/RenderText.cpp:
              * rendering/RenderText.h:
              * rendering/RenderTextControl.cpp:
              * rendering/RenderTextControl.h:
              * rendering/RenderTheme.cpp:
              * rendering/RenderTheme.h:
              * rendering/RenderThemeMac.h:
              * rendering/RenderThemeMac.mm:
              * rendering/RenderThemeSafari.cpp:
              * rendering/RenderThemeSafari.h:
      
      LayoutTests:
      
      2008-06-19  Timothy Hatcher  <timothy@apple.com>
      
              Test for: iframes with a height of 32,768px or greater do not layout correctly
              https://bugs.webkit.org/show_bug.cgi?id=19679
      
              Reviewed by Dave Hyatt.
      
              * fast/css/line-height-overflow.html: Added.
              * platform/mac/fast/css/line-height-overflow-expected.txt: Added.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34693 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c59ad194
    • mitz@apple.com's avatar
      WebCore: · 9f7cc0e3
      mitz@apple.com authored
              Reviewed by Darin Adler.
      
              - fix https://bugs.webkit.org/show_bug.cgi?id=19519
                <rdar://problem/6007345> DOM modification causes stack exhaustion (BUTTON OBJECT COLGROUP)
      
              Test: fast/table/prepend-in-anonymous-table.html
      
              * rendering/RenderBlock.cpp:
              (WebCore::RenderBlock::addChildToFlow): Added code to handle the case
              of inserting before a child that has been wrapped by an anonymous table,
              in which case if the new child is the type that needs to be in a table,
              it is inserted into the table, and otherwise it is inserted before the
              table.
              * rendering/RenderTable.cpp:
              (WebCore::RenderTable::addChild): Added table-column-group alongside
              table-caption as content that can exist inside an anonymous table
              without being wrapped in a table section.
      
      LayoutTests:
      
              Reviewed by Darin Adler.
      
              - test for https://bugs.webkit.org/show_bug.cgi?id=19519
                <rdar://problem/6007345> DOM modification causes stack exhaustion (BUTTON OBJECT COLGROUP)
      
              * fast/table/prepend-in-anonymous-table.html: Added.
              * platform/mac/fast/table/prepend-in-anonymous-table-expected.checksum: Added.
              * platform/mac/fast/table/prepend-in-anonymous-table-expected.png: Added.
              * platform/mac/fast/table/prepend-in-anonymous-table-expected.txt: Added.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34692 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      9f7cc0e3
    • hyatt@apple.com's avatar
      WebCore: · 4be20eed
      hyatt@apple.com authored
      2008-06-20  David Hyatt  <hyatt@apple.com>
      
              Allow CSS variables to support arbitrary expressions as values instead of just a single term.
      
              Reviewed by Sam
      
              Removed fast/css/variables/single-term-test.html and replaced with multiple-term-test.html
      
              * css/CSSGrammar.y:
              * css/CSSParser.cpp:
              (WebCore::CSSParser::addVariable):
              * css/CSSParser.h:
              * css/CSSStyleSelector.cpp:
              (WebCore::CSSStyleSelector::addMatchedDeclaration):
              * css/CSSVariablesDeclaration.cpp:
              (WebCore::CSSVariablesDeclaration::CSSVariablesDeclaration):
              (WebCore::CSSVariablesDeclaration::getVariableValue):
              (WebCore::CSSVariablesDeclaration::addParsedVariable):
              (WebCore::CSSVariablesDeclaration::getParsedVariable):
              * css/CSSVariablesDeclaration.h:
              (WebCore::CSSVariablesDeclaration::create):
      
      LayoutTests:
      
      2008-06-20  David Hyatt  <hyatt@apple.com>
      
              Allow CSS variables to support arbitrary expressions as values instead of just a single term.
      
              Reviewed by Sam
      
              * fast/css/variables/single-term-test.html: Removed.
              * fast/css/variables/multiple-term-test.html: Added.
              * platform/mac/fast/css/variables/multiple-term-test-expected.checksum: Added.
              * platform/mac/fast/css/variables/multiple-term-test-expected.png: Added.
              * platform/mac/fast/css/variables/multiple-term-test-expected.txt: Added.
              * platform/mac/fast/css/variables/single-term-test-expected.checksum: Removed.
              * platform/mac/fast/css/variables/single-term-test-expected.png: Removed.
              * platform/mac/fast/css/variables/single-term-test-expected.txt: Removed.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34691 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      4be20eed
    • darin@apple.com's avatar
      2008-06-20 Darin Adler <darin@apple.com> · 625e7371
      darin@apple.com authored
              - updated results for some recent changes
      
              * fast/dom/Window/window-properties-expected.txt: Updated results to reflect the
              new rule type, VARIABLES_RULE (Hyatt's CSS variables patch).
      
              * fast/forms/plaintext-mode-1-expected.txt: Updated test and results to expect
              ForeColor to be disabled when the region is plain-text-only. Justin's execCommand
              patch for https://bugs.webkit.org/show_bug.cgi?id=16049.
              * fast/forms/plaintext-mode-1.html: Ditto.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34690 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      625e7371
    • hausmann@webkit.org's avatar
      2008-06-19 Ariya Hidayat <ariya.hidayat@trolltech.com> · cfd638f9
      hausmann@webkit.org authored
              Reviewed by Simon.
      
              Surpress compiler warning (int vs unsigned comparison).
      
              * wtf/unicode/qt4/UnicodeQt4.h:
              (WTF::Unicode::toLower):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34689 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      cfd638f9
    • hausmann@webkit.org's avatar
      2008-06-20 Marc Ordinas i Llopis <marc.ordinasillopis@collabora.co.uk> · 21c9c00c
      hausmann@webkit.org authored
              Reviewed by Simon.
      
              https://bugs.webkit.org/show_bug.cgi?id=19082
              [Qt] Full-page plugins not activated
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34688 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      21c9c00c
    • hausmann@webkit.org's avatar
      2008-06-20 Marco Barisione <marco.barisione@collabora.co.uk> · 83614e73
      hausmann@webkit.org authored
              Reviewed by Simon.
      
              https://bugs.webkit.org/show_bug.cgi?id=19082
              [Qt] Full-page plugins not activated
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34687 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      83614e73
    • hausmann@webkit.org's avatar
      2008-06-19 Ariya Hidayat <ariya.hidayat@trolltech.com> · ffe66d6a
      hausmann@webkit.org authored
              Reviewed by Timothy Hatcher.
      
              Introduce compiler define for MinGW, to have COMPILER(MINGW).
      
              * wtf/Platform.h:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34686 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      ffe66d6a
  3. 19 Jun, 2008 10 commits
    • mitz@apple.com's avatar
      Reviewed by Sam Weinig. · 7c7fe626
      mitz@apple.com authored
              - two more updated results for the last patch
      
              * platform/mac/fast/dynamic/insert-before-table-part-in-continuation-expected.txt:
              * platform/mac/fast/invalid/020-expected.txt:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34685 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      7c7fe626
    • ap@webkit.org's avatar
      Reviewed by Geoff. · 0825ac85
      ap@webkit.org authored
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34684 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      0825ac85
    • mitz@apple.com's avatar
      WebCore: · 85ce3531
      mitz@apple.com authored
              Reviewed by John Sullivan.
      
              - fix a bug where anonymous tables were inserted in the wrong place
      
              Tests: tables/mozilla/bugs/bug2479-2.html
                     tables/mozilla/bugs/bug278266.html
                     tables/mozilla/bugs/bug8411.xml
      
              * rendering/RenderContainer.cpp:
              (WebCore::RenderContainer::addChild): Changed to create an anonymous
              table and insert it before beforeChild unless the insertion point is
              right after an existing anonymous table, in which case the existing
              table is used.
      
      LayoutTests:
      
              Reviewed by John Sullivan.
      
              - updated tests and results after fixing a bug where anonymous tables were inserted in the wrong place
      
              * fast/forms/form-hides-table.html: Wrapped every case in a block in
              order to prevent cross-talk.
              * platform/mac/fast/forms/form-hides-table-expected.checksum:
              * platform/mac/fast/forms/form-hides-table-expected.png:
              * platform/mac/fast/forms/form-hides-table-expected.txt:
              * platform/mac/tables/mozilla/bugs/bug2479-2-expected.checksum: Added.
              * platform/mac/tables/mozilla/bugs/bug2479-2-expected.png: Added.
              * platform/mac/tables/mozilla/bugs/bug2479-2-expected.txt: Added.
              * platform/mac/tables/mozilla/bugs/bug2479-4-expected.txt:
              * platform/mac/tables/mozilla/bugs/bug278266-expected.checksum: Added.
              * platform/mac/tables/mozilla/bugs/bug278266-expected.png: Added.
              * platform/mac/tables/mozilla/bugs/bug278266-expected.txt: Added.
              * platform/mac/tables/mozilla/bugs/bug8411-expected.checksum: Added.
              * platform/mac/tables/mozilla/bugs/bug8411-expected.png: Added.
              * platform/mac/tables/mozilla/bugs/bug8411-expected.txt: Added.
              * platform/mac/tables/mozilla_expected_failures/bugs/bug2479-2-expected.checksum: Removed.
              * platform/mac/tables/mozilla_expected_failures/bugs/bug2479-2-expected.png: Removed.
              * platform/mac/tables/mozilla_expected_failures/bugs/bug2479-2-expected.txt: Removed.
              * platform/mac/tables/mozilla_expected_failures/bugs/bug278266-expected.checksum: Removed.
              * platform/mac/tables/mozilla_expected_failures/bugs/bug278266-expected.png: Removed.
              * platform/mac/tables/mozilla_expected_failures/bugs/bug278266-expected.txt: Removed.
              * platform/mac/tables/mozilla_expected_failures/bugs/bug8411-expected.checksum: Removed.
              * platform/mac/tables/mozilla_expected_failures/bugs/bug8411-expected.png: Removed.
              * platform/mac/tables/mozilla_expected_failures/bugs/bug8411-expected.txt: Removed.
              * platform/mac/tables/mozilla_expected_failures/other/test4-expected.txt:
              * tables/mozilla/bugs/bug2479-2.html: Copied from LayoutTests/tables/mozilla_expected_failures/bugs/bug2479-2.html.
              * tables/mozilla/bugs/bug278266.html: Copied from LayoutTests/tables/mozilla_expected_failures/bugs/bug278266.html.
              * tables/mozilla/bugs/bug8411.xml: Copied from LayoutTests/tables/mozilla_expected_failures/bugs/bug8411.xml.
              * tables/mozilla_expected_failures/bugs/bug2479-2.html: Removed.
              * tables/mozilla_expected_failures/bugs/bug278266.html: Removed.
              * tables/mozilla_expected_failures/bugs/bug8411.xml: Removed.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34683 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      85ce3531
    • justin.garcia@apple.com's avatar
      WebCore: · c04d36ef
      justin.garcia@apple.com authored
      2008-06-19  Justin Garcia  <justin.garcia@apple.com>
      
              Reviewed by Oliver.
      
              <https://bugs.webkit.org/show_bug.cgi?id=16049>
              execCommand('backColor') fails on collapsed selections
      
              * editing/EditorCommand.cpp:
              (WebCore::CommandEntry::): Enable BackColor for caret selections, and
              disable ForeColor in plaintext-only regions.
      
      LayoutTests:
      
      2008-06-19  Justin Garcia  <justin.garcia@apple.com>
      
              Reviewed by Oliver.
              
              <https://bugs.webkit.org/show_bug.cgi?id=16049>
              execCommand('backColor') fails on collapsed selections
      
              * editing/execCommand/16049-expected.txt: Added.
              * editing/execCommand/16049.html: Added.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34682 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c04d36ef
    • justin.garcia@apple.com's avatar
      WebCore: · d1eac9a5
      justin.garcia@apple.com authored
      2008-06-19  Justin Garcia  <justin.garcia@apple.com>
      
              Reviewed by John.
      
              <https://bugs.webkit.org/show_bug.cgi?id=19653>
              Typing style lost when creating list from, indenting or outdenting an empty paragraph
              
              There were two problems.  First, moveParagraphs didn't preserve the style of empty paragraphs.
              Second, indent, outdent and list creation returned false from preservesTypingStyle.
              The second problem couldn't be fixed by just adding preservesTypingStyle() { return true; }
              to those commands, though, because of other bugs.
              
              Cleaned up the implementation of typing style enough to remove FIXMEs for:
              
              <rdar://problem/3769899> Implementation of typing style needs improvement 
              
              which has actually been closed for some time now anyway.
              
              * dom/Document.cpp: Removed an unused header.
              * editing/CompositeEditCommand.cpp:
              (WebCore::CompositeEditCommand::moveParagraphs): Preserve the style of an empty paragraph, too.
              * editing/DeleteSelectionCommand.cpp:
              (WebCore::DeleteSelectionCommand::saveTypingStyleState): Removed FIXME.
              (WebCore::DeleteSelectionCommand::calculateTypingStyleAfterDelete): 
              Removed FIXME.  
              If there is a typing style to apply after a delete, apply it to any line break that acts as a paragraph
              placeholder, not only one that was inserted by deletion.  This fixes a bug where deleting
              <div><b>Bold</b><br></div>, changing the selection and then coming back and typing wouldn't 
              produce bold text.  
              Don't set the EditCommand's typingStyle, it has been removed (more on that later).
              (WebCore::DeleteSelectionCommand::doApply): We don't need to pass calculateTypingStyleAfterDelete
              the inserted placeholder because it will find it.
              (WebCore::DeleteSelectionCommand::preservesTypingStyle): Normally deletion doesn't preserve the 
              typing style that was present before it.  For example, type a character, Bold, then delete the 
              character and start typing.  The Bold typing style shouldn't stick around.  We got this right before 
              purely by chance. Deletion should preserve a typing style that *it* sets, however.
              * editing/DeleteSelectionCommand.h: No longer need to pass calculateTypingStyleAfterDelete
              the inserted placeholder.
              * editing/EditCommand.cpp:
              (WebCore::EditCommand::apply): 
              Clearing or not clearing a removed anchor should not be determined by whether or not a command 
              preservesTypingStyle().  For example, the deletion that removed an anchor (and stored it), may 
              not preserve the typing style, but that doesn't mean that it should then go and clear the removed 
              anchor.  All high level commands, and all commands that a TypingCommand spawns, except for
              text insertions, which should restore a removed anchor, should clear it.
              There is no longer a typing style on EditCommand, removed code that cleared it.  Code that clears the 
              *actual* typing style is now in Editor::appliedEditing, just like before.
              There is no longer a typing style on EditCommand, removed code to set one.
              (WebCore::EditCommand::styleAtPosition): Removed FIXME.
              * editing/EditCommand.h: 
              Removed code assosiated with m_typingStyle.
              Made preservesTypingStyle() public, so that we can call it from Editor::appliedEditing().
              * editing/Editor.cpp:
              (WebCore::Editor::appliedEditing):
              Removed code to preserve the removedAnchor during the call to setSelection, because we no longer
              request to clear the typing style with that call.  Also removed the FIXME about this.
              Before, a command would set the Frame's typing style by setting its own typingStyle, and then 
              letting code here, in appliedEditing pick that up and set it on the Frame.  Now, the command 
              itself sets the Frame's typing style.  Deletion is the only command that does this right now.
              * editing/IndentOutdentCommand.h:
              (WebCore::IndentOutdentCommand::preservesTypingStyle): Added, returns true.
              * editing/InsertLineBreakCommand.cpp:
              (WebCore::InsertLineBreakCommand::doApply): Removed FIXME.
              * editing/InsertListCommand.h:
              (WebCore::InsertListCommand::preservesTypingStyle): Added.
              * editing/InsertParagraphSeparatorCommand.cpp:
              (WebCore::InsertParagraphSeparatorCommand::applyStyleAfterInsertion): Removed FIXME.
              * editing/InsertTextCommand.cpp:
              (WebCore::InsertTextCommand::input): Removed FIXME.
              * editing/ReplaceSelectionCommand.cpp:
              (WebCore::ReplaceSelectionCommand::doApply): Removed FIXME and also added one about clearing
              the typing style here.  It seems like it's something that could wait until after the command
              has been performed, since there is no code between this point and the end of the operation
              that queries or uses the typing style.
      
      LayoutTests:
      
      2008-06-19  Justin Garcia  <justin.garcia@apple.com>
      
              Reviewed by John.
              
              https://bugs.webkit.org/show_bug.cgi?id=19653
              Typing style lost when creating list from, indenting or outdenting an empty paragraph
      
              These demonstrate fixes:
              * editing/execCommand/19653-1-expected.txt: Added.
              * editing/execCommand/19653-1.html: Added.
              * editing/execCommand/19653-2-expected.txt: Added.
              * editing/execCommand/19653-2.html: Added.
              * editing/execCommand/19653-3-expected.txt: Added.
              * editing/execCommand/19653-3.html: Added.
              * editing/execCommand/19653-4-expected.txt: Added.
              * editing/execCommand/19653-4.html: Added.
              Changed to an equivalent but more bloated DOM because of a pre-existing bug in ApplyStyleCommand:
              * platform/mac/editing/deleting/delete-br-011-expected.txt:
              Fixed a bug where the typing style wasn't applied to the placeholder in an empty paragraph,
              so typing would create text with the right style, but if you were to change the selection
              and then come back and start typing, the style would be wrong.  The size of the caret was
              also wrong as a result:
              * platform/mac/editing/deleting/delete-br-012-expected.txt:
              * platform/mac/editing/deleting/delete-br-012-expected.png:
              * platform/mac/editing/deleting/delete-br-012-expected.checksum:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34681 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d1eac9a5
    • mitz@apple.com's avatar
      2008-06-19 Dan Bernstein <mitz@apple.com> · e67e2ec0
      mitz@apple.com authored
              Reviewed by Alexey Proskuryakov.
      
              - fix crash in CSSStyleDeclaration::copyPropertiesInSet()
      
              Covered by many editing tests
      
              * css/CSSStyleDeclaration.cpp:
              (WebCore::CSSStyleDeclaration::copyPropertiesInSet): Re-ordered to avoid
              null pointer deref.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34680 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      e67e2ec0
    • hyatt@apple.com's avatar
      2008-06-19 David Hyatt <hyatt@apple.com> · 148d6fa2
      hyatt@apple.com authored
              Fix out of bounds bug in CSSVariablesDeclaration's item() method.
      
              Reviewed by Sam, Darin
      
              * css/CSSVariablesDeclaration.cpp:
              (WebCore::CSSVariablesDeclaration::item):
              * css/CSSVariablesDeclaration.h:
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34679 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      148d6fa2
    • sullivan@apple.com's avatar
      2008-06-19 John Sullivan <sullivan@apple.com> · c8050d9d
      sullivan@apple.com authored
              Reviewed by Darin
              
              Tweak to previous checkin
      
              * page/FrameView.cpp:
              (WebCore::FrameView::performPostLayoutTasks):
              clear m_firstLayoutCallbackPending before performing callback, to avoid recursion
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34678 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c8050d9d
    • alp@webkit.org's avatar
      2008-06-19 Alp Toker <alp@nuanti.com> · 94d2ca01
      alp@webkit.org authored
              GTK+/autotools build fix. JSGlobalObject.cpp in now in
              AllInOneFile.cpp and shouldn't be built separately.
      
              * GNUmakefile.am:
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34677 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      94d2ca01
    • sullivan@apple.com's avatar
      2008-06-19 John Sullivan <sullivan@apple.com> · 2139b1ba
      sullivan@apple.com authored
              Reviewed by Darin
              
              - fixed <rdar://problem/6021353> Assertion failure (!root->needsLayout()) after certain steps
              
              The didFirstLayout callback was called in a place where trouble would ensue if the client
              did any work that would cause the layout to be dirtied. Fixed by delaying the callback
              until performPostLayoutTasks.
      
              * page/FrameView.cpp:
              new m_firstLayoutCallbackPending instance member variable in FrameViewPrivate
              (WebCore::FrameViewPrivate::reset):
              set m_firstLayoutCallbackPending to false
              (WebCore::FrameView::layout):
              set m_firstLayoutCallbackPending instead of a local variable; don't do didFirstLayout callback here
              (WebCore::FrameView::performPostLayoutTasks):
              if m_firstLayoutCallbackPending is set, do didFirstLayout callback here, then clear m_firstLayoutCallbackPending
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@34676 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      2139b1ba