Skip to content

GitLab

Switch branch/tag
  1. 10 Sep, 2012 20 commits
  3. 09 Sep, 2012 20 commits
    • ossy@webkit.org's avatar
      [Qt] Unreviewed gardening, skip new failing tests to paint the bots green. · 34066f05
      ossy@webkit.org authored 
      * platform/qt/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128017 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      34066f05
    • keishi@webkit.org's avatar
      REGRESSION(r127558): Fix a crash in DateTimeChooserImpl::didClosePopup · 75a3be7e
      keishi@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=96222

Reviewed by Kent Tamura.

DateTimeChooserClient will delete DateTimeChooserImpl inside didEndChooser()
so we shouldn't be changing the members after that.

Testing covered by fast/forms/date/calendar-picker-appearance-pre-100.html.

* src/DateTimeChooserImpl.cpp:
(WebKit::DateTimeChooserImpl::didClosePopup):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128016 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      75a3be7e
    • commit-queue@webkit.org's avatar
      Source/JavaScriptCore: Fixed a few llint C++ interpreter bugs. · 88e66d6d
      commit-queue@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=96127.

Patch by Mark Lam <mark.lam@apple.com> on 2012-09-09
Reviewed by Geoffrey Garen.

* llint/LLIntCLoop.h:
    CLoop::execute()'s bootstrapOpcodeId does not need a default
    value. There is no case when this function is called without
    that parameter being specified.
* llint/LowLevelInterpreter.asm:
    Moved the dispatchAfterCall() call to where it is needed.
    For the C_LOOP back-end, it generates unreachable code.
* llint/LowLevelInterpreter.cpp:
    #include <wtf/Assertions.h> because LLIntAssembly.h needs it.
(JSC):
    Fixed bug in SIGN_BIT32() macro.
    Placate a MSVC warning for t0, and t1 being uninitialized.
(JSC::CLoop::execute):
    The bootstrapOpcodeId arg should always be specified.
    MSVC doesn't like UNUSED_PARAM() for labels. Switch to using
        the new UNUSED_LABEL() macro.
* offlineasm/cloop.rb:
* offlineasm/generate_offset_extractor.rb:
    Resolved a compiler warning found via MSVC.

Source/WTF: Fixed ASSERT() and ASSERT_AT() macros so that they can be used in
comma expressions. Also, added UNUSED_LABEL().
https://bugs.webkit.org/show_bug.cgi?id=96127.

Patch by Mark Lam <mark.lam@apple.com> on 2012-09-09
Reviewed by Geoffrey Garen.

* wtf/Assertions.h:
* wtf/UnusedParam.h: Added UNUSED_LABEL(). Removed an obsolete comment.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128015 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      88e66d6d
    • paroga@webkit.org's avatar
      Add StringBuilder::appendNumber() and use it · 4d255977
      paroga@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=96030

Reviewed by Eric Seidel.

Also fix a bunch of append() vs. appendLiteral() issues in the surrounding code.

Source/JavaScriptCore:

* API/JSContextRef.cpp:
(JSContextCreateBacktrace):
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
* interpreter/Interpreter.h:
(JSC::StackFrame::toString):

Source/WebCore:

* Modules/websockets/WebSocketHandshake.cpp:
(WebCore::hostName):
* bindings/v8/ScriptController.cpp:
(WebCore::resourceString):
* css/FontFeatureValue.cpp:
(WebCore::FontFeatureValue::customCssText):
* dom/Position.cpp:
(WebCore::Position::formatForDebugger):
* dom/Range.cpp:
(WebCore::Range::formatForDebugger):
* html/FormController.cpp:
(WebCore::FormKeyGenerator::formKey):
* html/HTMLAnchorElement.cpp:
(WebCore::appendServerMapMousePosition):
* html/canvas/CanvasRenderingContext2D.cpp:
(WebCore::CanvasRenderingContext2D::font):
* inspector/IdentifiersFactory.cpp:
(WebCore::IdentifiersFactory::addProcessIdPrefixTo):
* page/FrameTree.cpp:
(WebCore::FrameTree::uniqueChildName):
* page/SecurityOrigin.cpp:
(WebCore::SecurityOrigin::toRawString):
* platform/Decimal.cpp:
(WebCore::Decimal::toString):
* platform/network/ProxyServer.cpp:
(WebCore::appendProxyServerString):
* platform/text/TextStream.cpp:
(WebCore::TextStream::operator<<):
* platform/text/TextStream.h:
* rendering/RenderTreeAsText.cpp:
(WebCore::nodePosition):
* xml/XMLErrors.cpp:
(WebCore::XMLErrors::appendErrorMessage):

Source/WTF:

* wtf/DateMath.cpp:
(WTF::makeRFC2822DateString):
* wtf/text/StringBuilder.cpp:
(WTF::StringBuilder::appendNumber):
* wtf/text/StringBuilder.h:
(StringBuilder):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128014 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      4d255977
    • commit-queue@webkit.org's avatar
      [EFL] Wrong rendering results for square-button-appearance.html · 9c1ffd6a
      commit-queue@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=85493

Patch by KyungTae Kim <ktf.kim@samsung.com> on 2012-09-09
Reviewed by Gyuyoung Kim.

Source/WebCore:

A square-button is an element that has '-webkit-appearance: square-button' style.

If it is a block element,
the RenderBlock::recomputeLogicalWidth recalculate the width from the padding values.

But the recalculated width became wrong because the RenderThemeEfl::adjustSizeConstraints
called from RenderThemeEfl::adjustButtonStyle adds padding values for a button.

So, the adjustSizeConstraints must not be called for a square-button(SquareButtonPart).

Test : fast/css/square-button-appearance.html

* platform/efl/RenderThemeEfl.cpp:
(WebCore::RenderThemeEfl::adjustButtonStyle): Call adjustSizeConstrains only for PushButtonPart and ButtonPart

LayoutTests:

Remove fast/css/square-button-appearance.html from TestExpectations because it became PASS.

* platform/efl/TestExpectations:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128013 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      9c1ffd6a
    • bashi@chromium.org's avatar
      [Chromium] Unreviewed test expectations update after r128002 · 7d823327
      bashi@chromium.org authored 
      * platform/chromium-mac/inspector/console/command-line-api-getEventListeners-expected.txt: Copied from LayoutTests/platform/chromium/inspector/console/command-line-api-getEventListeners-expected.txt.
* platform/chromium-win/inspector/console/command-line-api-getEventListeners-expected.txt: Renamed from LayoutTests/platform/chromium/inspector/console/command-line-api-getEventListeners-expected.txt.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128012 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      7d823327
    • commit-queue@webkit.org's avatar
      [EFL][WK2] Minibrowser locks up when right clicking editable filed · 615ee1e8
      commit-queue@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=96074

Patch by Jinwoo Song <jinwoo7.song@samsung.com> on 2012-09-09
Reviewed by Gyuyoung Kim.

WebProcess crashes when the context menu type is SubmenuType and non-CROSS_PLATFORM_CONTEXT_MENUS is used.
It is because platformSubMenu() in the ContextMenuItemEfl.cpp returns null. This patch implements the setSubMenu()
and platformSubMenu() to set and return the subMenuItems.

* platform/efl/ContextMenuItemEfl.cpp:
(WebCore::ContextMenuItem::setSubMenu):
(WebCore::ContextMenuItem::platformSubMenu):
(WebCore):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128011 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      615ee1e8
    • commit-queue@webkit.org's avatar
      Pass full target idl file path to CodeGenerator as a constructor argument. · c90dee06
      commit-queue@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=96190

Patch by Kangil Han <kangil.han@samsung.com> on 2012-09-09
Reviewed by Kentaro Hara.

Passing idl file path to CodeGenerator would be useful because CodeGenerator* family could manipulate it when creating derived header and cpp files.

* bindings/scripts/CodeGenerator.pm:
(new):
(ProcessDocument):
* bindings/scripts/generate-bindings.pl:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128010 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c90dee06
    • mitz@apple.com's avatar
      Try to fix the build after r128006. · 657ea27f
      mitz@apple.com authored 
      * dom/Element.cpp:
(WebCore::Element::boundsInRootViewSpace): Updated for rename.
(WebCore::Element::getBoundingClientRect): Ditto.
* svg/SVGElement.cpp:
(WebCore::SVGElement::getBoundingBox): Renamed boundingBox() to this, to avoid clashing
with the virtual function on Node. The 'get' prefix also indicates that the result is
returned in an out parameter.
* svg/SVGElement.h:
(SVGElement):



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128009 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      657ea27f
    • paroga@webkit.org's avatar
      Replace WTF::numberToString() with String::numberToStringECMAScript() · 188ed48b
      paroga@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=96130

Reviewed by Benjamin Poulain.

The String class provides a nice wrapper for this function
in the meantime, so just use it to avoid code duplication.

* html/parser/HTMLParserIdioms.cpp:
(WebCore::serializeForNumberType):
* platform/Decimal.cpp:
(WebCore::Decimal::fromDouble):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128008 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      188ed48b
    • bashi@chromium.org's avatar
      [Chromium] Unreviewed gardening after r127907 · 968810bd
      bashi@chromium.org authored 
      Fix CCRenderPassTest::copyShouldBeIdenticalExceptIdAndQuads

* tests/CCRenderPassTest.cpp:
(CCRenderPassSize): Changed type of m_id

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128007 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      968810bd
    • eae@chromium.org's avatar
      Rename Node::getRect/getPixelSnappedRect and remove ContainerNode::getRect · a6f24b06
      eae@chromium.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=81413

Reviewed by David Hyatt.

Source/WebCore:

Rename Node::getRect and getPixelSnappedRect to boundingBox and
pixelSnappedBoundingBox respectively.

Remove overridden getRect method from ContainerNode as it calculates the
same rect as the Node implementation but in a more complicated and
error-prone way. Also remove the getUpperLeftCorner/getLowerRightCorner
support functions as they are no longer needed.

No new tests, no change in functionality.

* accessibility/AccessibilitySlider.cpp:
(WebCore::AccessibilitySliderThumb::elementRect):
Update to call boundingBox.

* dom/ContainerNode.cpp:
* dom/ContainerNode.h:
Remove overridden getRect method and the getUpperLeftCorner/
getLowerRightCorner support methods .

* dom/Element.cpp:
(WebCore::Element::scrollIntoView):
(WebCore::Element::scrollIntoViewIfNeeded):
(WebCore::Element::updateFocusAppearance):
Update to call boundingBox.

* dom/Node.h:
* dom/Node.cpp:
(WebCore::Node::boundingBox):
Rename getRect to boundingBox.

(WebCore::Node::pixelSnappedBoundingBox):
Rename getPixelSnappedRect to pixelSnappedBoundingBox.

* html/ColorInputType.cpp:
(WebCore::ColorInputType::elementRectRelativeToRootView):
* html/ValidationMessage.cpp:
(WebCore::ValidationMessage::buildBubbleTree):
* html/shadow/CalendarPickerElement.cpp:
(WebCore::CalendarPickerElement::openPopup):
* html/shadow/TextControlInnerElements.cpp:
(WebCore::InputFieldSpeechButtonElement::startSpeechInput):
* page/FrameView.cpp:
(WebCore::FrameView::scrollElementToRect):
(WebCore::FrameView::scrollToAnchor):
* page/SpatialNavigation.cpp:
(WebCore::nodeRectInAbsoluteCoordinates):
Update to call boundingBox/pixelSnappedBoundingBox.

Source/WebKit/chromium:

Update WebViewImpl to call boundingBox/pixelSnappedBoundingBox.

* src/WebViewImpl.cpp:
(WebKit::WebViewImpl::computeBlockBounds):
(WebKit::WebViewImpl::applyAutofillSuggestions):
(WebKit::WebViewImpl::refreshAutofillPopup):

Source/WebKit/efl:

Update ewk_frame to call pixelSnappedBoundingBox.

* ewk/ewk_frame.cpp:
(ewk_frame_focused_element_geometry_get):

Source/WebKit/gtk:

Update ChromeClientGtk and webkitwebview to call pixelSnappedBoundingBox.

* WebCoreSupport/ChromeClientGtk.cpp:
(WebKit::ChromeClient::mouseDidMoveOverElement):
* webkit/webkitwebview.cpp:
(getLocationForKeyboardGeneratedContextMenu):
(webkit_web_view_query_tooltip):

Source/WebKit/qt:

Update qwebelement and qwebpage to call pixelSnappedBoundingBox.

* Api/qwebelement.cpp:
(QWebElement::geometry):
(QWebElement::render):
* Api/qwebpage.cpp:
(QWebPagePrivate::TouchAdjuster::findCandidatePointForTouch):

Source/WebKit2:

Update WebPage to call pixelSnappedBoundingBox.

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::editorState):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128006 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      a6f24b06
    • commit-queue@webkit.org's avatar
      [chromium] Do not delete texture backing structures on the main thread · 5e7928f9
      commit-queue@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=96018

Patch by Christopher Cameron <ccameron@chromium.org> on 2012-09-09
Reviewed by James Robinson.

Do not delete CCPrioritizedTexture::Backing structures on the main
thread.  Instead, unlink them from their owning CCPrioritizedTexture
in the main thread, and have the impl thread then delete all unlinked
textures.

This is towards having the main thread not access the m_backings set,
which will allow the impl thread to traverse that set when deleting
resources in response to GPU memory management events.

Tested by existing eviction tests (CCLayerTreeHostTest's
TestEvictTextures, LostContextAfterEvictTextures)

Fix a place where single thread proxy wasn't setting the main thread
blocked state correctly.

* platform/graphics/chromium/cc/CCLayerTreeHost.cpp:
(WebCore::CCLayerTreeHost::unlinkAllContentTextures):
(WebCore):
(WebCore::CCLayerTreeHost::deleteUnlinkedTextures):
* platform/graphics/chromium/cc/CCLayerTreeHost.h:
(CCLayerTreeHost):
* platform/graphics/chromium/cc/CCPrioritizedTextureManager.cpp:
(WebCore::CCPrioritizedTextureManager::clearAllMemory):
(WebCore::CCPrioritizedTextureManager::unlinkAllBackings):
(WebCore):
(WebCore::CCPrioritizedTextureManager::deleteAllUnlinkedBackings):
* platform/graphics/chromium/cc/CCPrioritizedTextureManager.h:
(CCPrioritizedTextureManager):
* platform/graphics/chromium/cc/CCSingleThreadProxy.cpp:
(WebCore::CCSingleThreadProxy::commitAndComposite):
* platform/graphics/chromium/cc/CCThreadProxy.cpp:
(WebCore::CCThreadProxy::beginFrame):
(WebCore::CCThreadProxy::beginFrameCompleteOnImplThread):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128005 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      5e7928f9
    • jamesr@google.com's avatar
      Rebaseline media/video-zoom-controls on chromium-mac for subpixel change. · 5a4264df
      jamesr@google.com authored 
      * platform/chromium-mac/media/video-zoom-controls-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128004 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      5a4264df
    • weinig@apple.com's avatar
      Switch to entering the sandbox directly from main(), rather than waiting for... · 7d9c69d8
      weinig@apple.com authored 
      Switch to entering the sandbox directly from main(), rather than waiting for the initialization message
https://bugs.webkit.org/show_bug.cgi?id=96194

Reviewed by Dan Bernstein.

Source/WebKit2: 

* UIProcess/Launcher/mac/ProcessLauncherMac.mm:
(WebKit::ProcessLauncher::launchProcess):
Pass a client identifier to WebProcess as a command line argument
so that it can be used to create private temporary and cache directories.

* WebProcess/WebProcess.h:
(WebProcess):
* WebProcess/mac/WebProcessMac.mm:
(WebKit::WebProcess::initializeSandbox):
(WebKit::WebProcess::platformInitializeWebProcess):
Expose the initializeSandbox() function and stop calling it from platformInitializeWebProcess()
since it is now going to be called from WebProcessMain.  Also, move changing the current working
directory to sandbox initialization function and replace use of creation parameters with just
the client identifier.

* WebProcess/com.apple.WebProcess.sb.in:
Add some new exceptions that come from entering the sandbox earlier in the AppKit
initialization process. Don't allow access to com.apple.coreservices.appleevents.

* WebProcess/mac/WebProcessMainMac.mm:
(WebKit::WebProcessMainXPC):
(WebKit::WebProcessMain):
Enter the sandbox explicitly, not waiting until the initialization message.  Also,
since we now have an extra parameter for the client identifier, make the WebProcess
launches the UIProcess code path work by extracting a client identifier from the 
client executable path.

WebKitLibraries: 

Update WKSI with SPI for getting the bundle URL from an executable URL.

* WebKitSystemInterface.h:
* libWebKitSystemInterfaceLion.a:
* libWebKitSystemInterfaceMountainLion.a:
* libWebKitSystemInterfaceSnowLeopard.a:
Adds WKCopyBundleURLForExecutableURL().


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128003 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      7d9c69d8
    • kling@webkit.org's avatar
      EventListenerMap: Use Vector instead of HashMap as backend. · 80eb5f60
      kling@webkit.org authored 
      <http://webkit.org/b/77982>

Reviewed by Geoff Garen.

Source/WebCore:

Refactor EventListenerMap to store pair<AtomicString, EventListenerVector> in a Vector
instead of using key/value HashMap stores. This is much more space efficient and actually
faster since we were spending more time/effort managing the hash map than it costs us
to iterate over and compare a couple of pointers. (It's very rare to have more than
4 different event types registered in a single EventListenerMap.)

This gets rid of the slightly hacky optimization for nodes with listeners of a single type,
reducing the complexity of EventListenerMap greatly.

~1.1MB progression on Membuster. Also strong (20+ MB on larger patches) savings on WebKit
bugzilla review pages, though they don't necessarily represent a common usecase.

* dom/EventListenerMap.cpp:
(WebCore::EventListenerMap::contains):
(WebCore::EventListenerMap::clear):
(WebCore::EventListenerMap::eventTypes):
(WebCore::EventListenerMap::add):
(WebCore::EventListenerMap::remove):
(WebCore::EventListenerMap::find):
(WebCore::EventListenerMap::removeFirstEventListenerCreatedFromMarkup):
(WebCore::EventListenerMap::copyEventListenersNotCreatedFromMarkupToTarget):
(WebCore::EventListenerIterator::EventListenerIterator):
(WebCore::EventListenerIterator::nextListener):
* dom/EventListenerMap.h:
(WebCore::EventListenerMap::isEmpty):
(WebCore::EventListenerMapEntry::EventListenerMapEntry):
(EventListenerMapEntry):
(EventListenerMap):
(EventListenerIterator):

LayoutTests:

Rebaseline inspector test whose output depended on the internal ordering of event
listeners (changed as listeners are no longer stored in an unordered HashMap.)

* inspector/console/command-line-api-getEventListeners-expected.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@128002 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      80eb5f60
    • lforschler@apple.com's avatar
      Build Fix. <rdar://problem/12260893> · f2820a8e
      lforschler@apple.com authored 
              
* platform/graphics/mac/WebGLLayer.mm:



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@127998 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      f2820a8e
    • commit-queue@webkit.org's avatar
      Unreviewed, rolling out r127979. · fc65bcd7
      commit-queue@webkit.org authored 
      http://trac.webkit.org/changeset/127979
https://bugs.webkit.org/show_bug.cgi?id=96217

Appears to have broken the windows and linux builds (Requested
by jamesr_ on #webkit).

Patch by Sheriff Bot <webkit.review.bot@gmail.com> on 2012-09-09

* bindings/scripts/CodeGeneratorV8.pm:
(NativeToJSValue):
* bindings/v8/V8DOMWrapper.h:
(WebCore::V8DOMWrapper::convertEventTargetToV8Object):
(V8DOMWrapper):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@127997 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      fc65bcd7
    • jamesr@google.com's avatar
      [chromium] Export WebFilterOperation(FilterType, SkScalar[20]) constructor · bf53f98b
      jamesr@google.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=96215

Reviewed by Adrienne Walker.

Unlike the other WebFilterOperation constructors this one is defined out-of-line and so must be exported to be
usable outside of WebKit.dll

* chromium/public/WebFilterOperation.h:
(WebFilterOperation):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@127996 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      bf53f98b
    • jamesr@google.com's avatar
      [chromium] webkit_platform should only be linked into the webkit target · 1977bb15
      jamesr@google.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=96213

Reviewed by Adrienne Walker.

The Platform.gyp:webkit_platform target provides the implementation of several WEBKIT_EXPORTed symbols. It only
makes sense as a target of WebKit.gyp:webkit, other targets shouldn't directly depend on it.

Source/WebCore:

* WebCore.gyp/WebCore.gyp:

Source/WebKit/chromium:

* WebKit.gyp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@127995 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      1977bb15