Timeline



Sep 1, 2012:

10:51 PM Changeset in webkit [127386] by weinig@apple.com
  • 2 edits in trunk/Source/WebKit2

Remove unused member variable from WebProcess.

Rubber-stamped by Dan Bernstein.

  • WebProcess/WebProcess.h:

(WebProcess):
Remove m_applicationCachePathExtension.

9:31 PM Changeset in webkit [127385] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/efl

[EFL] Refactoring the DumpRenderTreeSupportEfl
​https://bugs.webkit.org/show_bug.cgi?id=95544

Patch by Jinwoo Song <​jinwoo7.song@samsung.com> on 2012-09-01
Reviewed by Gyuyoung Kim.

There are lots of repeated codes for getting the WebCore::Frame
and WebCore::Page. To reduce these duplications, simple macros
are added.

No new tests, just refactoring.

  • WebCoreSupport/DumpRenderTreeSupportEfl.cpp:

(DumpRenderTreeSupportEfl::activeAnimationsCount):
(DumpRenderTreeSupportEfl::callShouldCloseOnWebView):
(DumpRenderTreeSupportEfl::clearFrameName):
(DumpRenderTreeSupportEfl::clearOpener):
(DumpRenderTreeSupportEfl::elementDoesAutoCompleteForElementWithId):
(DumpRenderTreeSupportEfl::frameChildren):
(DumpRenderTreeSupportEfl::frameParent):
(DumpRenderTreeSupportEfl::layoutFrame):
(DumpRenderTreeSupportEfl::numberOfPages):
(DumpRenderTreeSupportEfl::pageSizeAndMarginsInPixels):
(DumpRenderTreeSupportEfl::pageProperty):
(DumpRenderTreeSupportEfl::pauseAnimation):
(DumpRenderTreeSupportEfl::pauseTransition):
(DumpRenderTreeSupportEfl::pendingUnloadEventCount):
(DumpRenderTreeSupportEfl::renderTreeDump):
(DumpRenderTreeSupportEfl::responseMimeType):
(DumpRenderTreeSupportEfl::selectionRectangle):
(DumpRenderTreeSupportEfl::suitableDRTFrameName):
(DumpRenderTreeSupportEfl::setDefersLoading):
(DumpRenderTreeSupportEfl::setLoadsSiteIconsIgnoringImageLoadingSetting):
(DumpRenderTreeSupportEfl::addUserScript):
(DumpRenderTreeSupportEfl::clearUserScripts):
(DumpRenderTreeSupportEfl::addUserStyleSheet):
(DumpRenderTreeSupportEfl::clearUserStyleSheets):
(DumpRenderTreeSupportEfl::executeCoreCommandByName):
(DumpRenderTreeSupportEfl::findString):
(DumpRenderTreeSupportEfl::setCSSGridLayoutEnabled):
(DumpRenderTreeSupportEfl::isCommandEnabled):
(DumpRenderTreeSupportEfl::setSmartInsertDeleteEnabled):
(DumpRenderTreeSupportEfl::setSelectTrailingWhitespaceEnabled):
(DumpRenderTreeSupportEfl::evaluateInWebInspector):
(DumpRenderTreeSupportEfl::evaluateScriptInIsolatedWorld):
(DumpRenderTreeSupportEfl::globalContextRefForFrame):
(DumpRenderTreeSupportEfl::setInteractiveFormValidationEnabled):
(DumpRenderTreeSupportEfl::setValidationMessageTimerMagnification):
(DumpRenderTreeSupportEfl::setAuthorAndUserStylesEnabled):
(DumpRenderTreeSupportEfl::setComposition):
(DumpRenderTreeSupportEfl::hasComposition):
(DumpRenderTreeSupportEfl::compositionRange):
(DumpRenderTreeSupportEfl::confirmComposition):
(DumpRenderTreeSupportEfl::firstRectForCharacterRange):
(DumpRenderTreeSupportEfl::selectedRange):

8:08 PM Changeset in webkit [127384] by weinig@apple.com
  • 10 edits in trunk/Source/WebKit2

Use dynamic sandbox extension rather than sandbox parameters for paths provided by the UIProcess
​https://bugs.webkit.org/show_bug.cgi?id=95651

Reviewed by Dan Bernstein.

This is the first step to not relying on the initialization message to initialize
the sandbox.

  • Shared/SandboxExtension.h:

(SandboxExtension):

  • Shared/mac/SandboxExtensionMac.mm:

(WebKit::SandboxExtension::consumePermanently):
Add static version of SandboxExtension::consumePermanently which takes a
SandboxExtension::Handle. There is no reason to create a SandboxExtension
just to consume it forever.

  • Shared/WebProcessCreationParameters.cpp:

(WebKit::WebProcessCreationParameters::encode):
(WebKit::WebProcessCreationParameters::decode):

  • Shared/WebProcessCreationParameters.h:

(WebProcessCreationParameters):
Add SandboxExtension::Handles for paths we need access to, and remove
webInspectorBaseDirectory, which is unused.

  • UIProcess/WebContext.cpp:

(WebKit::WebContext::createNewWebProcess):

  • UIProcess/mac/WebContextMac.mm:

(WebKit::WebContext::platformInitializeWebProcess):
Create SandboxExtension::Handles for paths we need access to, and move
getting the presenterApplicationPid to platformInitializeWebProcess, since
it is Mac only.

  • WebProcess/WebProcess.cpp:

(WebKit::WebProcess::initializeWebProcess):
Move setting the presenterApplicationPid to platformInitializeWebProcess.

  • WebProcess/mac/WebProcessMac.mm:

(WebKit::initializeSandbox):
(WebKit::WebProcess::platformInitializeWebProcess):
Replace the use of sandbox parameters for opening access for necessary paths
with permanent consumption of sandbox extensions.

  • WebProcess/com.apple.WebProcess.sb.in:

Remove sandbox rules that are now handled via extensions.

5:15 PM Changeset in webkit [127383] by krit@webkit.org
  • 5 edits
    18 adds in trunk

Use -webkit-clip-path shapes to clip SVG elements
​https://bugs.webkit.org/show_bug.cgi?id=95620

Reviewed by Rob Buis.

Source/WebCore:

This patch adds a path segment for a BasicShape to a given Path object. This
path and it's wind rule are used to clip the context of the SVG element.

Tests: svg/clip-path/clip-path-shape-circle-1-expected.svg

svg/clip-path/clip-path-shape-circle-1.svg
svg/clip-path/clip-path-shape-circle-2-expected.svg
svg/clip-path/clip-path-shape-circle-2.svg
svg/clip-path/clip-path-shape-ellipse-1-expected.svg
svg/clip-path/clip-path-shape-ellipse-1.svg
svg/clip-path/clip-path-shape-ellipse-2-expected.svg
svg/clip-path/clip-path-shape-ellipse-2.svg
svg/clip-path/clip-path-shape-polygon-1-expected.svg
svg/clip-path/clip-path-shape-polygon-1.svg
svg/clip-path/clip-path-shape-polygon-2-expected.svg
svg/clip-path/clip-path-shape-polygon-2.svg
svg/clip-path/clip-path-shape-polygon-3-expected.svg
svg/clip-path/clip-path-shape-polygon-3.svg
svg/clip-path/clip-path-shape-rounded-rect-1-expected.svg
svg/clip-path/clip-path-shape-rounded-rect-1.svg
svg/clip-path/clip-path-shape-rounded-rect-2-expected.svg
svg/clip-path/clip-path-shape-rounded-rect-2.svg

  • rendering/style/BasicShapes.cpp: Added helper functions that apply path segments to a given path.

(WebCore::BasicShapeRectangle::path):
(WebCore::BasicShapeCircle::path):
(WebCore::BasicShapeEllipse::path):
(WebCore::BasicShapePolygon::path):

  • rendering/style/BasicShapes.h: Make BasicShape virtualized again, since new virtual functions were added.

(WebCore::BasicShape::~BasicShape):
(BasicShape):
(WebCore::BasicShape::windRule): Will return the wind rule of the shape - nonzero by default.
(WebCore::BasicShape::BasicShape):
(BasicShapeRectangle):
(WebCore::BasicShapeRectangle::type): Removed member variable and return type per inheriting class directly.
(WebCore::BasicShapeRectangle::BasicShapeRectangle):
(BasicShapeCircle):
(WebCore::BasicShapeCircle::type): Ditto.
(WebCore::BasicShapeCircle::BasicShapeCircle):
(BasicShapeEllipse):
(WebCore::BasicShapeEllipse::type): Ditto.
(WebCore::BasicShapeEllipse::BasicShapeEllipse):
(BasicShapePolygon):
(WebCore::BasicShapePolygon::windRule):
(WebCore::BasicShapePolygon::type): Ditto.
(WebCore::BasicShapePolygon::BasicShapePolygon):

  • rendering/svg/SVGRenderingContext.cpp: If -webkit-clip-path was defined, clip the context to the shape.

Right now -webkit-clip-path overrides clip-path, so that people don't use both at the same time. Current
clip-path property will be replaced, once -webkit-clip-path gets unprefixed.
(WebCore::SVGRenderingContext::prepareToRenderSVGContent):

LayoutTests:

New tests to check behavior of -webkit-clip-path on SVG elements.

  • svg/clip-path/clip-path-shape-circle-1-expected.svg: Added.
  • svg/clip-path/clip-path-shape-circle-1.svg: Added.
  • svg/clip-path/clip-path-shape-circle-2-expected.svg: Added.
  • svg/clip-path/clip-path-shape-circle-2.svg: Added.
  • svg/clip-path/clip-path-shape-ellipse-1-expected.svg: Added.
  • svg/clip-path/clip-path-shape-ellipse-1.svg: Added.
  • svg/clip-path/clip-path-shape-ellipse-2-expected.svg: Added.
  • svg/clip-path/clip-path-shape-ellipse-2.svg: Added.
  • svg/clip-path/clip-path-shape-polygon-1-expected.svg: Added.
  • svg/clip-path/clip-path-shape-polygon-1.svg: Added.
  • svg/clip-path/clip-path-shape-polygon-2-expected.svg: Added.
  • svg/clip-path/clip-path-shape-polygon-2.svg: Added.
  • svg/clip-path/clip-path-shape-polygon-3-expected.svg: Added.
  • svg/clip-path/clip-path-shape-polygon-3.svg: Added.
  • svg/clip-path/clip-path-shape-rounded-rect-1-expected.svg: Added.
  • svg/clip-path/clip-path-shape-rounded-rect-1.svg: Added.
  • svg/clip-path/clip-path-shape-rounded-rect-2-expected.svg: Added.
  • svg/clip-path/clip-path-shape-rounded-rect-2.svg: Added.
1:29 PM Changeset in webkit [127382] by wangxianzhu@chromium.org
  • 4 edits
    1 add
    1 delete in trunk

[Chromium-Android] Remove hard-coded skipped directories in chromium_android.py
​https://bugs.webkit.org/show_bug.cgi?id=95605

Reviewed by Dirk Pranke.

Tools:

Now skipping rules are put in LayoutTests/platform/chromium-android/TestExpectations.

  • Scripts/webkitpy/layout_tests/port/chromium_android.py:

(ChromiumAndroidPort.check_sys_deps):
(ChromiumAndroidPort.expectations_files):
(ChromiumAndroidPort.clean_up_test_run):

LayoutTests:

Moved the skipping rules for ANDROID into platform/chromium-android/TestExpectations.

  • platform/chromium/TestExpectations:
  • platform/chromium/test_expectations_android.txt: Removed.
  • platform/chromium-android/TestExpectations: Added.
12:46 PM Changeset in webkit [127381] by commit-queue@webkit.org
  • 8 edits
    2 adds in trunk

Regression(r126763): Heap-use-after-free in WebCore::nextBreakablePosition
​https://bugs.webkit.org/show_bug.cgi?id=95229

Patch by Ned Holbrook <​nholbrook@apple.com> on 2012-09-01
Reviewed by Dan Bernstein.

Source/WebCore:

TextLayout and LazyLineBreakIterator cache a RenderText's string during line breaking but RenderCounter
and RenderQuote can replace that string during preferred width calculation. This patch adds a non-virtual member
function to RenderText named updateTextIfNeeded() that triggers immediate string replacement by calling
the new virtual updateText() if necessary, which in turn calls computePreferredLogicalWidths(). In this way
RenderBlock::LineBreaker::nextLineBreak() can ensure a RenderText's string is current before caching it.

Test: fast/css/content/content-quotes-crash.html

  • rendering/RenderBlockLineLayout.cpp:

(WebCore::dirtyLineBoxesForRenderer): Drive-by: replace existing code with the equivalent updateTextIfNeeded().
(WebCore::RenderBlock::LineBreaker::nextLineBreak): Use updateTextIfNeeded() prior to caching RenderText string.

  • rendering/RenderCounter.cpp:

(WebCore::RenderCounter::updateText): Call computePreferredLogicalWidths().
(WebCore):

  • rendering/RenderCounter.h:

(RenderCounter):

  • rendering/RenderQuote.cpp:

(WebCore::RenderQuote::updateText): Call computePreferredLogicalWidths().
(WebCore):

  • rendering/RenderQuote.h:

(RenderQuote):

  • rendering/RenderText.h:

(WebCore::RenderText::updateTextIfNeeded): Only call virtual updateText() if necessary.
(RenderText):
(WebCore::RenderText::updateText): Add no-op default implementation for new virtual member function.

LayoutTests:

Ensure line breaking doesn't crash due to text of RenderText being changed.

  • fast/css/content/content-quotes-crash-expected.txt: Added.
  • fast/css/content/content-quotes-crash.html: Added.
12:42 PM Changeset in webkit [127380] by commit-queue@webkit.org
  • 9 edits in trunk

fast/events/message-port-clone.html hits ASSERT in Debug (usually in later tests)
​https://bugs.webkit.org/show_bug.cgi?id=85811

Patch by Li Yin <​li.yin@intel.com> on 2012-09-01
Reviewed by Kenneth Rohde Christiansen.

Source/WebCore:

When MessagePort invoked close function, the close status is needed to set,
regardless of whether the port is cloned or not.

Tests: fast/events/message-port-clone.html

fast/events/message-port-close.html

  • dom/MessagePort.cpp:

(WebCore::MessagePort::close): Must set m_closed to be true.

LayoutTests:

  • platform/gtk/TestExpectations:
  • platform/mac/Skipped:
  • platform/qt-4.8/TestExpectations:
  • platform/qt-mac/TestExpectations:
  • platform/qt/TestExpectations:
  • platform/win/Skipped:
12:26 PM Changeset in webkit [127379] by wangxianzhu@chromium.org
  • 2 edits
    2 adds
    2 deletes in trunk/LayoutTests

[Chromium-Android] Upstream layout test expectations (Part 4)
​https://bugs.webkit.org/show_bug.cgi?id=95594

Reviewed by Adam Barth.

Remove a font boosting test which has been removed in downstream.
Update canvas test expectations.
Rebaselined one canvas test (only one space diff).

  • platform/chromium-android/fast/canvas/fallback-content-expected.txt: Added.
  • platform/chromium-android/fast/font-boosting/various-tricky-cases-expected.txt: Removed.
  • platform/chromium-android/fast/font-boosting/various-tricky-cases.html: Removed.
  • platform/chromium/TestExpectations:
12:06 PM Changeset in webkit [127378] by jamesr@google.com
  • 2 edits in trunk/Source/WebKit/chromium

Unreviewed chromium buildfix when use_libcc_for_compositor=1

This gyp target was renamed downstream.

  • WebKit.gyp:
11:16 AM Changeset in webkit [127377] by commit-queue@webkit.org
  • 2 edits in trunk/LayoutTests

[Gtk] accessibility/replaced-objects-in-anonymous-blocks layout test failure
​https://bugs.webkit.org/show_bug.cgi?id=95643

Unreviewed GTK gardening, rebaselining an accessibility test.

Patch by Joanmarie Diggs <​jdiggs@igalia.com> on 2012-09-01

  • platform/gtk/accessibility/replaced-objects-in-anonymous-blocks-expected.txt: Corrected.
10:49 AM Changeset in webkit [127376] by ggaren@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

2012-09-01 Geoffrey Garen <​ggaren@apple.com>

Rolled back out a piece of <​http://trac.webkit.org/changeset/127293>
because it broke inspector tests on Windows.

Shrink activation objects by half
​https://bugs.webkit.org/show_bug.cgi?id=95591

Reviewed by Sam Weinig.

10:47 AM Changeset in webkit [127375] by kling@webkit.org
  • 4 edits in trunk/Source/WebCore

Share inline style between cloned Nodes (copy on write.)
<​http://webkit.org/b/95451>

Reviewed by Antti Koivisto.

When cloning a Node, use an immutable StylePropertySet for the new Node's inline style.
If the old Node already had an immutable inline style, we now reuse that, avoiding a copy.
Copying is deferred until mutation (either via CSSOM or setting of the style attribute.)

  • dom/ElementAttributeData.cpp:

(WebCore::ElementAttributeData::cloneDataFrom):

  • css/StylePropertySet.h:
  • css/StylePropertySet.cpp:

(WebCore::StylePropertySet::immutableCopyIfNeeded):

Added. Simply returns 'this' if the object is already immutable, otherwise creates a new one.

10:36 AM Changeset in webkit [127374] by commit-queue@webkit.org
  • 25 edits
    1 copy
    2 adds in trunk/Source

LLInt C loop backend.
​https://bugs.webkit.org/show_bug.cgi?id=91052.

Patch by Mark Lam <​mark.lam@apple.com> on 2012-09-01
Reviewed by Filip Pizlo.

Source/JavaScriptCore:

(JSC::CodeBlock::dump):
(JSC::CodeBlock::bytecodeOffset):

  • interpreter/Interpreter.cpp:

(JSC::Interpreter::execute):
(JSC::Interpreter::executeCall):
(JSC::Interpreter::executeConstruct):
(JSC):

  • interpreter/Interpreter.h:
  • jit/JITStubs.h:

(JITStackFrame):
(JSC):

  • llint/LLIntCLoop.cpp: Added.

(JSC):
(LLInt):
(JSC::LLInt::CLoop::initialize):
(JSC::LLInt::CLoop::catchRoutineFor):
(JSC::LLInt::CLoop::hostCodeEntryFor):
(JSC::LLInt::CLoop::jsCodeEntryWithArityCheckFor):
(JSC::LLInt::CLoop::jsCodeEntryFor):

  • llint/LLIntCLoop.h: Added.

(JSC):
(LLInt):
(CLoop):

  • llint/LLIntData.cpp:

(JSC::LLInt::initialize):

  • llint/LLIntData.h:

(JSC):

  • llint/LLIntOfflineAsmConfig.h:
  • llint/LLIntOpcode.h:
  • llint/LLIntThunks.cpp:

(LLInt):

  • llint/LowLevelInterpreter.asm:
  • llint/LowLevelInterpreter.cpp:

(LLInt):
(JSC::LLInt::Ints2Double):
(JSC):
(JSC::CLoop::execute):

  • llint/LowLevelInterpreter.h:

(JSC):

  • llint/LowLevelInterpreter32_64.asm:
  • llint/LowLevelInterpreter64.asm:
  • offlineasm/asm.rb:
  • offlineasm/backends.rb:
  • offlineasm/cloop.rb: Added.
  • offlineasm/instructions.rb:
  • runtime/Executable.h:

(ExecutableBase):
(JSC::ExecutableBase::hostCodeEntryFor):
(JSC::ExecutableBase::jsCodeEntryFor):
(JSC::ExecutableBase::jsCodeWithArityCheckEntryFor):
(JSC::ExecutableBase::catchRoutineFor):
(NativeExecutable):

  • runtime/JSValue.h:

(JSC):
(LLInt):
(JSValue):

  • runtime/JSValueInlineMethods.h:

(JSC):
(JSC::JSValue::JSValue):

  • runtime/Options.cpp:

(JSC::Options::initialize):

Source/WTF:

Added configs for the llint C loop backend.

  • wtf/Platform.h:
10:22 AM Changeset in webkit [127373] by ggaren@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

Rolled back in a piece of <​http://trac.webkit.org/changeset/127293>.

Shrink activation objects by half
​https://bugs.webkit.org/show_bug.cgi?id=95591

Reviewed by Sam Weinig.

9:14 AM Changeset in webkit [127372] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/mac

Added #include <JavaScriptCore/Identifier.h>.
​https://bugs.webkit.org/show_bug.cgi?id=95641.

Patch by Mark Lam <​mark.lam@apple.com> on 2012-09-01
Reviewed by Dan Bernstein.

The code references JSC::Indentifier but was previously depending on
someone else to #include it. This results in build problems when
building for the llint C loop configuration.

  • Plugins/Hosted/NetscapePluginInstanceProxy.h:
7:42 AM Changeset in webkit [127371] by krit@webkit.org
  • 2 edits in trunk/Source/WebCore

[Qt] Fix the --minimal build after r127327
​https://bugs.webkit.org/show_bug.cgi?id=95639

Unreviewed attempt to fix Qt minimal build.

  • css/StyleBuilder.cpp:
7:02 AM WebKitGTK/1.8.x edited by Martin Robinson
(diff)
7:01 AM WebKitGTK/1.8.x edited by Martin Robinson
(diff)
3:51 AM Changeset in webkit [127370] by commit-queue@webkit.org
  • 5 edits
    2 adds in trunk

[Gtk] No accessible caret-moved events found in certain content
​https://bugs.webkit.org/show_bug.cgi?id=72811

Part of the bug was due to expected accessible objects of DivRole
and ParagraphRole being ignored, in favor of including child blocks.

Patch by Joanmarie Diggs <​jdiggs@igalia.com> on 2012-09-01
Reviewed by Chris Fleizach.

Source/WebCore:

Test: platform/gtk/accessibility/spans-paragraphs-and-divs.html

  • accessibility/AccessibilityObject.cpp:

(WebCore::AccessibilityObject::firstAnonymousBlockChild):
(WebCore):

  • accessibility/AccessibilityObject.h:

(AccessibilityObject):
New method to return the first child which is an anonymous block.

  • accessibility/gtk/AccessibilityObjectAtk.cpp:

(WebCore::AccessibilityObject::accessibilityPlatformIncludesObject):
Include paragraphs and divs which contain a non-nested anonymous block.

LayoutTests:

Added new test to verify that paragraphs and divs which contain anonymous
blocks are included in the accessible object hierarchy and have the correct
number of children with the expected role.

  • platform/gtk/accessibility/spans-paragraphs-and-divs-expected.txt: Added.
  • platform/gtk/accessibility/spans-paragraphs-and-divs.html: Added.
3:47 AM Changeset in webkit [127369] by zandobersek@gmail.com
  • 2 edits in trunk/LayoutTests

Unreviewed GTK gardening, rebaselining an accessibility test.

  • platform/gtk/accessibility/replaced-objects-in-anonymous-blocks-expected.txt:
1:53 AM Changeset in webkit [127368] by zandobersek@gmail.com
  • 8 edits
    2 adds in trunk

[Gtk] No accessible caret-moved events found in certain content
​https://bugs.webkit.org/show_bug.cgi?id=72811

Part of the bug was due to extraneous accessible objects resulting
from unignored inline and block spans.

Patch by Joanmarie Diggs <​jdiggs@igalia.com> on 2012-09-01
Reviewed by Chris Fleizach.

Source/WebCore:

Test: platform/gtk/accessibility/spans.html

  • accessibility/AccessibilityRenderObject.cpp:

(WebCore::AccessibilityRenderObject::accessibilityIsIgnored): Ignore objects that have spanTag tag name.

  • accessibility/gtk/AccessibilityObjectAtk.cpp:

(WebCore::AccessibilityObject::accessibilityPlatformIncludesObject): Ignore most anonymous blocks.

  • accessibility/gtk/WebKitAccessibleWrapperAtk.cpp:

(roleIsTextType): Add ListItem to the roles which should implement AtkText.

LayoutTests:

Added new test, updated the results of one test to reflect the fix,
corrected a test with a mismatched element tag.

  • platform/gtk/accessibility/aria-roles-unignored-expected.txt: Corrected results having fixed tag.
  • platform/gtk/accessibility/aria-roles-unignored.html: Fixed mismatched element tag.
  • platform/gtk/accessibility/media-element-expected.txt: Updated to reflect removal of extraneous object.
  • platform/gtk/accessibility/spans-expected.txt: Added.
  • platform/gtk/accessibility/spans.html: Added.
1:43 AM Changeset in webkit [127367] by zandobersek@gmail.com
  • 8 edits
    4 adds in trunk

[Gtk] Incorrect/unexpected characters in the text of certain accessibles
​https://bugs.webkit.org/show_bug.cgi?id=95180

Patch by Joanmarie Diggs <​jdiggs@igalia.com> on 2012-09-01
Reviewed by Chris Fleizach.

Source/WebCore:

The bug was caused by failing to properly handle anonymous block text
which had object replacement characters (multibyte) in it. Calculating
the string length based on the UTF-8 string meant that we were returning
more characters than were there and in danger of splitting a multibyte
character.

Tests: platform/gtk/accessibility/entry-and-password.html

platform/gtk/accessibility/replaced-objects-in-anonymous-blocks.html

  • accessibility/gtk/WebKitAccessibleInterfaceText.cpp:

(webkitAccessibleTextGetText): Convert the text returned by textForObject()
to Unicode before calculating its length.

Source/WebKit/gtk:

Corrected a unit test in which the expected accessible text was wrong as
a result of this bug. In particular, the AtkText inserted into an empty
text field is expected to be the same text atk_text_get_text() returns.
That was not happening -- and presumably not noticed as a result of the
hard to read textual representation of the multibyte password field
bullets.

  • tests/testatk.c:

(testWebkitAtkTextChangedNotifications): Corrected the test and added a
comment so that one knows what the multibyte character is.

Tools:

The bug that was fixed stood in the way of fully implementing stringValue().
Testing that the bug is fixed requires stringValue() to be fully implemented
and object replacement characters to be printable.

  • DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:

(replaceCharactersForResults): New method which turns object replacement
characters into "<obj>" so that the characters can be properly shown in
Layout Test results. Also turns "\n" into "<
n>" so that printing the
accessible text of a single object in the accessible tree doesn't mess up
the readibility of the results.
(AccessibilityUIElement::stringValue): Remove the code that immediately
returned upon encountering an object of ATK_ROLE_PANEL and call the new
replaceCharactersForResults() prior to returning the accessible string
value.

LayoutTests:

Two new layout tests, plus one updated one.

  • platform/gtk/accessibility/deleting-iframe-destroys-axcache-expected.txt: Indicated replaced objects.
  • platform/gtk/accessibility/entry-and-password-expected.txt: Added.
  • platform/gtk/accessibility/entry-and-password.html: Added.
  • platform/gtk/accessibility/replaced-objects-in-anonymous-blocks-expected.txt: Added.
  • platform/gtk/accessibility/replaced-objects-in-anonymous-blocks.html: Added.
1:20 AM Changeset in webkit [127366] by abarth@webkit.org
  • 17 edits in trunk/Source/WebCore

Remove all-but-one use of WTF::String::operator+= from WebCore
​https://bugs.webkit.org/show_bug.cgi?id=95508

Reviewed by Benjamin Poulain.

This patch removes all the uses of WTF::String::operator+= from
WebCore, except those in WorkerScriptLoader (which need a more delicate
patch). There were actually a handful of legitimate use cases for += in
this group. I've replaced them with calls to String::append rather than
+= so that we can remove += and encourage most contributors to use
more efficient string idioms.

(There are likely some more uses in WebCore hiding in port-specific
code---this patch covers only those call sites found when compiling the
chromium-mac port.)

  • inspector/InspectorStyleTextEditor.cpp:

(WebCore::InspectorStyleTextEditor::insertProperty):

  • This complicated function looks really inefficient, but I didn't have the heart to rewrite it.
  • inspector/NetworkResourcesData.cpp:

(WebCore::NetworkResourcesData::ResourceData::decodeDataToContent):

  • loader/cache/CachedCSSStyleSheet.cpp:

(WebCore::CachedCSSStyleSheet::sheetText):
(WebCore::CachedCSSStyleSheet::data):

  • loader/cache/CachedFont.cpp:

(WebCore::CachedFont::ensureSVGFontData):

  • loader/cache/CachedScript.cpp:

(WebCore::CachedScript::script):

  • loader/cache/CachedXSLStyleSheet.cpp:

(WebCore::CachedXSLStyleSheet::data):

  • This decoder/flush pattern can probably be improved by combining the decode and flush operations, but I didn't do that in this patch.
  • page/FrameTree.cpp:

(WebCore::FrameTree::uniqueChildName):

  • I found this code very amusing. It's worried enough about efficiency to give a big speech about why snprintf is safe, but then it implicitly performs a large number of mallocs and memcpy operations.
  • page/Page.cpp:

(WebCore::Page::userStyleSheet):

  • platform/chromium/support/WebHTTPLoadInfo.cpp:

(WebKit::addHeader):

  • platform/chromium/support/WebURLResponse.cpp:

(WebKit::WebURLResponse::addHTTPHeaderField):

  • This header-appending idiom looks like a reasonable use case for String::append.
  • xml/XMLHttpRequest.cpp:

(WebCore::XMLHttpRequest::send):
(WebCore::XMLHttpRequest::setRequestHeaderInternal):

  • xml/XPathFunctions.cpp:

(WebCore::XPath::FunTranslate::evaluate):

  • Fixes 6 year old FIXME.
  • xml/parser/XMLDocumentParser.cpp:

(WebCore::XMLDocumentParser::append):

  • xml/parser/XMLDocumentParser.h:

(XMLDocumentParser):

  • xml/parser/XMLDocumentParserLibxml2.cpp:

(WebCore::XMLDocumentParser::doEnd):

  • xml/parser/XMLDocumentParserQt.cpp:

(WebCore::XMLDocumentParser::doEnd):

  • Changed m_originalSourceForTransform to a SegmentedString so that we don't need to malloc and copy the entire document every time we get more data from the network.
12:54 AM Changeset in webkit [127365] by tommyw@google.com
  • 22 edits
    2 adds in trunk

MediaStream API: Add MediaStream management to RTCPeerConnection
​https://bugs.webkit.org/show_bug.cgi?id=95543

Reviewed by Adam Barth.

Source/Platform:

  • chromium/public/WebMediaStreamCenter.h:

(WebMediaStreamCenter):

  • chromium/public/WebRTCPeerConnectionHandler.h:

(WebKit):
(WebRTCPeerConnectionHandler):

  • chromium/public/WebRTCPeerConnectionHandlerClient.h:

(WebKit):
(WebRTCPeerConnectionHandlerClient):

Source/WebCore:

This includes addStream/removeStream to RTCPeerConnection,
as well as didAddRemoteStream/didRemoveRemoteStream to RTCPeerConnectionHandlerClient.

Test: fast/mediastream/RTCPeerConnection-AddRemoveStream.html

  • Modules/mediastream/RTCPeerConnection.cpp:

(WebCore::RTCPeerConnection::parseConfiguration):
(WebCore::RTCPeerConnection::RTCPeerConnection):
(WebCore::RTCPeerConnection::addStream):
(WebCore):
(WebCore::RTCPeerConnection::removeStream):
(WebCore::RTCPeerConnection::localStreams):
(WebCore::RTCPeerConnection::remoteStreams):
(WebCore::RTCPeerConnection::didAddRemoteStream):
(WebCore::RTCPeerConnection::didRemoveRemoteStream):

  • Modules/mediastream/RTCPeerConnection.h:

(RTCPeerConnection):

  • Modules/mediastream/RTCPeerConnection.idl:
  • bindings/js/Dictionary.cpp:

(WebCore::Dictionary::getOwnPropertiesAsStringHashMap):
(WebCore::Dictionary::getOwnPropertyNames):

  • bindings/v8/Dictionary.cpp:

(WebCore::Dictionary::getOwnPropertiesAsStringHashMap):
(WebCore::Dictionary::getOwnPropertyNames):

  • bindings/v8/Dictionary.h:

(Dictionary):

  • platform/mediastream/RTCPeerConnectionHandler.cpp:

(RTCPeerConnectionHandlerDummy):
(WebCore::RTCPeerConnectionHandlerDummy::addStream):
(WebCore):
(WebCore::RTCPeerConnectionHandlerDummy::removeStream):

  • platform/mediastream/RTCPeerConnectionHandler.h:

(RTCPeerConnectionHandler):

  • platform/mediastream/RTCPeerConnectionHandlerClient.h:

(WebCore):
(RTCPeerConnectionHandlerClient):

  • platform/mediastream/chromium/RTCPeerConnectionHandlerChromium.cpp:

(WebCore::RTCPeerConnectionHandlerChromium::addStream):
(WebCore):
(WebCore::RTCPeerConnectionHandlerChromium::removeStream):
(WebCore::RTCPeerConnectionHandlerChromium::didAddRemoteStream):
(WebCore::RTCPeerConnectionHandlerChromium::didRemoveRemoteStream):

  • platform/mediastream/chromium/RTCPeerConnectionHandlerChromium.h:

(WebKit):
(RTCPeerConnectionHandlerChromium):

Source/WebKit/chromium:

  • src/AssertMatchingEnums.cpp:

Tools:

Extending MockWebRTCPeerConnectionHandler to handle addStream/removeStream.

  • DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:

(MockWebRTCPeerConnectionHandler::MockWebRTCPeerConnectionHandler):
(MockWebRTCPeerConnectionHandler::addStream):
(MockWebRTCPeerConnectionHandler::removeStream):

  • DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:

(MockWebRTCPeerConnectionHandler):

LayoutTests:

  • fast/mediastream/RTCPeerConnection-AddRemoveStream-expected.txt: Added.
  • fast/mediastream/RTCPeerConnection-AddRemoveStream.html: Added.
12:49 AM Changeset in webkit [127364] by jonlee@apple.com
  • 1 edit
    4 adds in trunk/LayoutTests

[Tests] Add a test for all notification events
​https://bugs.webkit.org/show_bug.cgi?id=95609
<rdar://problem/12218309>

Reviewed by Jessie Berlin.

This test tests for the onshow, onclick, and onclose events. onerror is called only if the
origin did not have permission to show the notification, and that is covered in
http/tests/notifications/{legacy,}/show.html.

  • http/tests/notifications/events-expected.txt: Added.
  • http/tests/notifications/events.html: Added.
  • http/tests/notifications/legacy/events-expected.txt: Added.
  • http/tests/notifications/legacy/events.html: Added.
12:45 AM Changeset in webkit [127363] by ggaren@apple.com
  • 12 edits in trunk/Source/JavaScriptCore

2012-09-01 Geoffrey Garen <​ggaren@apple.com>

Rolled back in a piece of <​http://trac.webkit.org/changeset/127293>.

Shrink activation objects by half
​https://bugs.webkit.org/show_bug.cgi?id=95591

Reviewed by Sam Weinig.

  • runtime/JSActivation.cpp: (JSC::JSActivation::JSActivation):
  • runtime/JSGlobalObject.cpp: (JSC::JSGlobalObject::JSGlobalObject): (JSC::JSGlobalObject::setGlobalThis): (JSC): (JSC::JSGlobalObject::visitChildren):
  • runtime/JSGlobalObject.h: (JSGlobalObject): (JSC::JSScope::globalThis): (JSC): (JSC::JSGlobalObject::globalThis):
  • runtime/JSNameScope.h: (JSC::JSNameScope::JSNameScope):
  • runtime/JSScope.cpp: (JSC::JSScope::visitChildren):
  • runtime/JSScope.h: (JSScope): (JSC::JSScope::JSScope): (JSC::JSScope::globalObject): (JSC::JSScope::globalData):
  • runtime/JSSegmentedVariableObject.h: (JSC::JSSegmentedVariableObject::JSSegmentedVariableObject):
  • runtime/JSSymbolTableObject.h: (JSC::JSSymbolTableObject::JSSymbolTableObject):
  • runtime/JSVariableObject.h: (JSC::JSVariableObject::JSVariableObject):
  • runtime/JSWithScope.h: (JSC::JSWithScope::JSWithScope):
  • runtime/StrictEvalActivation.cpp: (JSC::StrictEvalActivation::StrictEvalActivation):
12:38 AM Changeset in webkit [127362] by jonlee@apple.com
  • 1 edit
    2 adds
    2 deletes in trunk/LayoutTests

[Tests] Move fast/notifications/notifications-event-listener-crash.html to http/tests/notifications
​https://bugs.webkit.org/show_bug.cgi?id=95535
<rdar://problem/12213781>

Reviewed by Jessie Berlin.

This test is among a few that are being migrated from fast/notifications to http/tests/notifications because of permissions.
Since most notifications tests will involve permissions, it is easier to have all of them consolidated in the same directory.

This test has been slightly modified to take advantage of the new testRunner notifications API. The test itself also does not
need to be protected within the window.testRunner check since it does not use that object.

  • fast/notifications/notifications-event-listener-crash-expected.txt: Removed.
  • fast/notifications/notifications-event-listener-crash.html: Removed.
  • http/tests/notifications/event-listener-crash-expected.txt: Added.
  • http/tests/notifications/event-listener-crash.html: Added.
12:29 AM Changeset in webkit [127361] by tkent@chromium.org
  • 2 edits in trunk/Source/WebCore

Touch a file so that Chromium-win bots correctly re-build for OPENTYPE_VERTICAL
​https://bugs.webkit.org/show_bug.cgi?id=51450

  • platform/graphics/opentype/OpenTypeVerticalData.cpp:
12:15 AM Changeset in webkit [127360] by ggaren@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

2012-09-01 Geoffrey Garen <​ggaren@apple.com>

Rolled back out a piece of <​http://trac.webkit.org/changeset/127293>
because it broke Window inspector tests.

Shrink activation objects by half
​https://bugs.webkit.org/show_bug.cgi?id=95591

Reviewed by Sam Weinig.

  • runtime/JSActivation.h: (JSActivation):

Aug 31, 2012:

10:38 PM Changeset in webkit [127359] by tkent@chromium.org
  • 2 edits in trunk/LayoutTests

[Chromium] Text expectation update
​https://bugs.webkit.org/show_bug.cgi?id=51450

  • platform/chromium/TestExpectations:

Fix incorrect entry for track-cue-rendering-vertical.html

10:31 PM Changeset in webkit [127358] by tkent@chromium.org
  • 2 edits in trunk/LayoutTests

[Chromium] Test expectation update.
​https://bugs.webkit.org/show_bug.cgi?id=51450

  • platform/chromium/TestExpectations:
  • r127354 wrongly removed failure expectation for MAC.
  • Also, fix expectation of vertical-font-fallback.html on WIN.
9:59 PM LineBreakingCSS3Mapping edited by glenn@skynav.com
Fix anchor. (diff)
9:57 PM LineBreakingCSS3Mapping edited by glenn@skynav.com
Add anchor. (diff)
9:53 PM LineBreakingCSS3Mapping edited by glenn@skynav.com
Editorial update. (diff)
9:44 PM LineBreakingCSS3Mapping edited by glenn@skynav.com
Add content on after-white-space and migrating to unprefixed line-break. (diff)
9:31 PM Changeset in webkit [127357] by pdr@google.com
  • 8 edits in trunk/LayoutTests

Fix 6 broken SVG animation tests

Unreviewed test update.

Six SVG animation tests were marked as TIMEOUT because they never ran. Each
of these tests required a click event to start and the event was not being fired.
This patch updates these tests to correctly start, fixing 6 TIMEOUTS.

  • platform/chromium/TestExpectations:
  • svg/animations/script-tests/animate-path-animation-qQ-tT-inverse.js:
  • svg/animations/script-tests/animate-path-nested-transforms.js:
  • svg/animations/script-tests/svgnumberoptionalnumber-animation-1.js:
  • svg/animations/script-tests/svgnumberoptionalnumber-animation-2.js:
  • svg/animations/script-tests/svgnumberoptionalnumber-animation-3.js:
  • svg/animations/script-tests/svgnumberoptionalnumber-animation-4.js:
9:24 PM Changeset in webkit [127356] by tkent@chromium.org
  • 2 edits in trunk/LayoutTests

[Chromium] Unreviewed, test expectation update

  • platform/chromium/TestExpectations:
    • Skip removing-underline-after-accepting-autocorrection-using-punctuation.html. Chromium doesn't support autocorrection yet.
    • Remove ANDROID entry for http/tests/notifications
9:13 PM Changeset in webkit [127355] by tkent@chromium.org
  • 4 edits in trunk/Source/WebCore

Touch some files so that Chromium-win bots correctly re-build
files for OPENTYPE_VERTICAL.
​https://bugs.webkit.org/show_bug.cgi?id=51450

  • platform/graphics/FontCache.h:
  • platform/graphics/SimpleFontData.h:
  • platform/graphics/opentype/OpenTypeTypes.h:
9:06 PM Changeset in webkit [127354] by tkent@chromium.org
  • 10 edits in trunk

Glyphs in vertical text tests are rotated 90 degrees clockwise on Chromium Windows
​https://bugs.webkit.org/show_bug.cgi?id=51450

Patch by Koji Ishii <Koji Ishii> on 2012-08-31
Reviewed by Tony Chang.

When writing-mode is vertical-*, Chromium Win does not render glyphs correctly.
​http://dev.w3.org/csswg/css3-writing-modes/#writing-mode
Glyphs are now drawn in the correct orientation, regardless of the leading '@' in the font name,
and vertical alternate glyphs (OpenType 'vert' feature) is not enabled.
CJK characters drawn at wrong baseline position is also fixed.
In addition, fontMetrics().unitsPerEm() and SimpleFontData::platformBoundsForGlyph() were empty
in Chromium Win but they were implemented because they were needed to draw vertical flow correctly.

Note that this patch does not address issues in vertical flow for complex code path.
The issue is under discussion with Google people and it will not be fixed in short term.

Source/WebKit/chromium:

Existing tests in fast/writing-mode/* are enabled in LayoutTests/platform/chromium/TestExpectations.

  • features.gypi: ENABLE_OPENTYPE_VERTICAL=1.

LayoutTests:

  • fast/text/international/text-spliced-font.html: Added "MS Gothic" to the font list so that Japanese font is chosen on Windows.
  • fast/writing-mode/border-vertical-lr.html: Ditto.
  • fast/writing-mode/japanese-lr-selection.html: Ditto.
  • fast/writing-mode/japanese-lr-text.html: Ditto.
  • fast/writing-mode/japanese-rl-selection.html: Ditto.
  • fast/writing-mode/japanese-rl-text.html: Ditto.
  • platform/chromium/TestExpectations: writing-mode and vertical tests are separated to rebase later.
8:46 PM Changeset in webkit [127353] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit2

Build warning : -Wformat on WebMemorySampler.cpp.
​https://bugs.webkit.org/show_bug.cgi?id=95550

Patch by Byungwoo Lee <​bw80.lee@samsung.com> on 2012-08-31
Reviewed by Darin Adler.

Fix build warning.
Use String::number instead of String::format.

  • Shared/WebMemorySampler.cpp:

(WebKit::WebMemorySampler::appendCurrentMemoryUsageToFile):

8:08 PM Changeset in webkit [127352] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

-webkit-flex: 0 in inspector.css is confusing
​https://bugs.webkit.org/show_bug.cgi?id=95568

Patch by Chris Hutten-Czapski <​chutten@rim.com> on 2012-08-31
Reviewed by Tony Chang.

The initial value of -webkit-flex results in the same layout as
-webkit-flex: 0; but does not depend on the new initial value of auto
for min-width.

This increases the support for the new Web Inspector style to renderers
built before rev 122264 but after 117385.

No new tests (has no effect except to out-of-date renderers)

  • inspector/front-end/inspector.css:

(.status-bar-item):

7:52 PM LineBreakingCSS3Mapping edited by glenn@skynav.com
Add missing description of IS class. (diff)
7:34 PM LineBreakingCSS3Mapping edited by glenn@skynav.com
Fix table order. (diff)
7:14 PM Changeset in webkit [127351] by fpizlo@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

Unreviewed, attempt to fix Windows, take two.

6:51 PM Changeset in webkit [127350] by fpizlo@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

Unreviewed, attempt to fix Windows.

6:50 PM Changeset in webkit [127349] by fpizlo@apple.com
  • 12 edits in trunk/Source

JSArray::putDirectIndex should by default behave like JSObject::putDirect
​https://bugs.webkit.org/show_bug.cgi?id=95630

Reviewed by Gavin Barraclough.

Source/JavaScriptCore:

  • interpreter/Interpreter.cpp:

(JSC::Interpreter::privateExecute):

  • jit/JITStubs.cpp:

(JSC::DEFINE_STUB_FUNCTION):

  • jsc.cpp:

(GlobalObject::finishCreation):

  • llint/LLIntSlowPaths.cpp:

(JSC::LLInt::LLINT_SLOW_PATH_DECL):

  • runtime/JSArray.cpp:

(JSC::SparseArrayValueMap::putDirect):
(JSC::JSArray::defineOwnNumericProperty):
(JSC::JSArray::putDirectIndexBeyondVectorLength):

  • runtime/JSArray.h:

(SparseArrayValueMap):
(JSArray):
(JSC::JSArray::putDirectIndex):

  • runtime/JSONObject.cpp:

(JSC::Walker::walk):

  • runtime/RegExpMatchesArray.cpp:

(JSC::RegExpMatchesArray::reifyAllProperties):
(JSC::RegExpMatchesArray::reifyMatchProperty):

  • runtime/StringPrototype.cpp:

(JSC::splitStringByOneCharacterImpl):
(JSC::stringProtoFuncSplit):

Source/WebCore:

No new tests because no change in behavior.

  • bindings/js/SerializedScriptValue.cpp:

(WebCore::CloneDeserializer::putProperty):

6:39 PM Changeset in webkit [127348] by ggaren@apple.com
  • 7 edits in trunk/Source/JavaScriptCore

Rolled back in a piece of <​http://trac.webkit.org/changeset/127293>.

Shrink activation objects by half
​https://bugs.webkit.org/show_bug.cgi?id=95591

Reviewed by Sam Weinig.

  • runtime/JSGlobalData.cpp:

(JSC::JSGlobalData::JSGlobalData):

  • runtime/JSGlobalData.h:

(JSGlobalData):

  • runtime/JSNameScope.h:

(JSC::JSNameScope::JSNameScope):

  • runtime/JSWithScope.h:

(JSC::JSWithScope::JSWithScope):

  • runtime/StrictEvalActivation.cpp:

(JSC::StrictEvalActivation::StrictEvalActivation):

6:09 PM Changeset in webkit [127347] by Nate Chapin
  • 5 edits in trunk

Source/WebCore: fast/loader/document-destruction-within-unload.html causes assertion failures on mac and qt.
​https://bugs.webkit.org/show_bug.cgi?id=66783

Reviewed by Darin Adler.

This was a not-quite-regression from trac.webkit.org/changeset/93521, in that we hit asserts
in a case where we previously were use-after-freeing. Tweak how we handle cases where a Document
is cleared from within an unload handler.

No new tests, fixing fast/loader/document-destruction-within-unload.html on mac and qt.

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::setDocumentLoader): Instead of trying to reattach a partially removed

DocumentLoader if it is detached before being fully added, leave the old one in place, completed.

(WebCore::FrameLoader::transitionToCommitted):

LayoutTests: Unskip fast/loader/document-destruction-within-unload.html
on mac and qt.
​https://bugs.webkit.org/show_bug.cgi?id=66783

Reviewed by Darin Adler.

  • platform/mac/Skipped:
  • platform/qt/Skipped:
6:00 PM Changeset in webkit [127346] by tony@chromium.org
  • 6 edits in trunk/Source/WebCore

Make computeBlockDirectionMargins const
​https://bugs.webkit.org/show_bug.cgi?id=95595

Reviewed by Ojan Vafai.

This is a step in making computeLogicalHeight const.
Refactor RenderBox::computeBlockDirectionMargins and make it const.
Add a helper method computeAndSetBlockDirectionMargins for the callers outside of
computeLogicalHeight that are calling this.

No new tests, this is already covered by tests in fast/writing-mode and fast/multicol.

  • rendering/RenderBlock.cpp:

(WebCore::RenderBlock::marginBeforeEstimateForChild): Switch to computeAndSetBlockDirectionMargins
(WebCore::RenderBlock::layoutBlockChild): Switch to computeAndSetBlockDirectionMargins
(WebCore::RenderBlock::insertFloatingObject): Switch to computeAndSetBlockDirectionMargins

  • rendering/RenderBox.cpp:

(WebCore::RenderBox::computeLogicalHeight): Use const version and fill in a MarginValues struct (to be merged in the next patch).
(WebCore::RenderBox::computeBlockDirectionMargins): Pass in out params.
(WebCore):
(WebCore::RenderBox::computeAndSetBlockDirectionMargins): Does the same as the old computeBlockDirectionMargins.

  • rendering/RenderBox.h:

(RenderBox):

  • rendering/RenderDeprecatedFlexibleBox.cpp:

(WebCore::RenderDeprecatedFlexibleBox::layoutHorizontalBox): Switch to computeAndSetBlockDirectionMargins
(WebCore::RenderDeprecatedFlexibleBox::layoutVerticalBox): Switch to computeAndSetBlockDirectionMargins

  • rendering/RenderTableRow.cpp:

(WebCore::RenderTableRow::layout): Switch to computeAndSetBlockDirectionMargins

5:46 PM Changeset in webkit [127345] by ggaren@apple.com
  • 10 edits in trunk/Source/JavaScriptCore

Rolled back in a piece of <​http://trac.webkit.org/changeset/127293>.

Shrink activation objects by half
​https://bugs.webkit.org/show_bug.cgi?id=95591

Reviewed by Sam Weinig.

  • dfg/DFGAbstractState.cpp:

(JSC::DFG::AbstractState::execute):

  • jit/JITOpcodes.cpp:

(JSC::JIT::emit_op_resolve_global_dynamic):

  • llint/LowLevelInterpreter32_64.asm:
  • llint/LowLevelInterpreter64.asm:
  • runtime/JSActivation.cpp:

(JSC::JSActivation::JSActivation):

  • runtime/JSGlobalData.cpp:

(JSC::JSGlobalData::JSGlobalData):

  • runtime/JSGlobalData.h:

(JSGlobalData):

  • runtime/JSGlobalObject.cpp:

(JSC::JSGlobalObject::reset):
(JSC::JSGlobalObject::visitChildren):

  • runtime/JSGlobalObject.h:

(JSGlobalObject):
(JSC::JSGlobalObject::withScopeStructure):
(JSC::JSGlobalObject::strictEvalActivationStructure):
(JSC::JSGlobalObject::activationStructure):
(JSC::JSGlobalObject::nameScopeStructure):

5:40 PM Changeset in webkit [127344] by roger_fong@apple.com
  • 1 edit
    2 adds in trunk/LayoutTests

Unreviewed. CSS_VARIABLES are not enabled on Windows. Adding Windows specific results.

  • platform/win/fast/css/variables: Added.
  • platform/win/fast/css/variables/build-supports-variables-expected.txt: Added.
5:38 PM Changeset in webkit [127343] by mhahnenberg@apple.com
  • 5 edits in trunk/Source/JavaScriptCore

Remove use of ClassInfo in SpeculativeJIT::emitBranch
​https://bugs.webkit.org/show_bug.cgi?id=95623

Reviewed by Filip Pizlo.

  • dfg/DFGAbstractState.cpp:

(JSC::DFG::AbstractState::execute):

  • dfg/DFGSpeculativeJIT.h:

(SpeculativeJIT):

  • dfg/DFGSpeculativeJIT32_64.cpp:

(JSC::DFG::SpeculativeJIT::emitNonStringCellOrOtherBranch):
(JSC::DFG::SpeculativeJIT::emitBranch):

  • dfg/DFGSpeculativeJIT64.cpp:

(JSC::DFG::SpeculativeJIT::emitNonStringCellOrOtherBranch):
(JSC::DFG::SpeculativeJIT::emitBranch):

5:33 PM Changeset in webkit [127342] by wjmaclean@chromium.org
  • 5 edits
    2 copies
    4 adds in trunk

[chromium] Make link highlighter use same node selection criteria as Chromium for Android.
​https://bugs.webkit.org/show_bug.cgi?id=95371

Reviewed by James Robinson.

This CL brings the selection criteria for link highlight targets into alignment with Chromium for Android.

Source/WebKit/chromium:

  • src/WebViewImpl.cpp:

(WebKit::highlightConditions):
(WebKit::WebViewImpl::bestTouchLinkNode):

  • tests/data/test_touch_link_highlight.html:

LayoutTests:

  • platform/chromium-linux/compositing/gestures/gesture-tapHighlight-simple-ZIndex.html: Copied from LayoutTests/platform/chromium-linux/compositing/gestures/gesture-tapHighlight-simple-background.html.
  • platform/chromium-linux/compositing/gestures/gesture-tapHighlight-simple-background.html:
  • platform/chromium-linux/compositing/gestures/gesture-tapHighlight-simple-body.html: Copied from LayoutTests/platform/chromium-linux/compositing/gestures/gesture-tapHighlight-simple-background.html.
  • platform/chromium-linux/platform/chromium-linux/compositing/gestures/gesture-tapHighlight-simple-ZIndex-expected.png: Added.
  • platform/chromium-linux/platform/chromium-linux/compositing/gestures/gesture-tapHighlight-simple-ZIndex-expected.txt: Added.
  • platform/chromium-linux/platform/chromium-linux/compositing/gestures/gesture-tapHighlight-simple-body-expected.png: Added.
  • platform/chromium-linux/platform/chromium-linux/compositing/gestures/gesture-tapHighlight-simple-body-expected.txt: Added.
5:30 PM Changeset in webkit [127341] by roger_fong@apple.com
  • 1 edit
    8 adds in trunk/LayoutTests

Unreviewed. Adding expected failing results for some fast/css(3) tests.
​https://bugs.webkit.org/show_bug.cgi?id=95627.

  • platform/win/fast/css/image-resolution: Added.
  • platform/win/fast/css/image-resolution/image-resolution-expected.txt: Added.
  • platform/win/fast/css/image-set-setting-expected.txt: Added.
  • platform/win/fast/css/parse-border-image-repeat-null-crash-expected.txt: Added.
  • platform/win/fast/css3-text-decoration: Added.
  • platform/win/fast/css3-text-decoration/getComputedStyle: Added.
  • platform/win/fast/css3-text-decoration/getComputedStyle/getComputedStyle-text-decoration-line-expected.txt: Added.
  • platform/win/fast/css3-text-decoration/getComputedStyle/getComputedStyle-text-decoration-style-expected.txt: Added.
5:20 PM Changeset in webkit [127340] by jamesr@google.com
  • 13 edits
    1 move
    1 add in trunk/Source/WebKit/chromium

[chromium] Consolidate geometry unit testing functions for cc
​https://bugs.webkit.org/show_bug.cgi?id=95526

Reviewed by Julien Chaffraix.

This consolidates geometry equality (rects, matrices) in CCGeometryUtils and removes a few duplicate macros for
testing rect equality.

  • WebKit.gypi:
  • tests/CCDamageTrackerTest.cpp:

(WebKitTests::TEST_F):

  • tests/CCGeometryTestUtils.cpp: Added.

(WebKitTests):
(WebKitTests::ExpectTransformationMatrixEq):

  • tests/CCGeometryTestUtils.h: Renamed from Source/WebKit/chromium/tests/CCLayerTreeTestCommon.h.

(WebKit):
(WebKitTests):

  • tests/CCLayerTreeHostTest.cpp:
  • tests/CCMathUtilTest.cpp:

(WebKit::TEST):

  • tests/TiledLayerChromiumTest.cpp:
5:12 PM Changeset in webkit [127339] by ojan@chromium.org
  • 2 edits in trunk/LayoutTests

Mark some indexeddb tests as SLOW. They consistently time out for me
when I run with --fully-parallel.

  • platform/chromium/TestExpectations:
5:08 PM Changeset in webkit [127338] by ggaren@apple.com
  • 6 edits in trunk/Source/JavaScriptCore

Rolled back in a piece of <​http://trac.webkit.org/changeset/127293>.

Shrink activation objects by half
​https://bugs.webkit.org/show_bug.cgi?id=95591

Reviewed by Sam Weinig.

  • heap/MarkedBlock.cpp:

(JSC::MarkedBlock::MarkedBlock):

  • heap/MarkedBlock.h:

(MarkedBlock):
(JSC::MarkedBlock::globalData):
(JSC):

  • heap/WeakSet.cpp:

(JSC::WeakSet::addAllocator):

  • heap/WeakSet.h:

(WeakSet):
(JSC::WeakSet::WeakSet):
(JSC::WeakSet::globalData):

  • runtime/JSGlobalData.h:

(JSC::WeakSet::heap):
(JSC):

4:49 PM Changeset in webkit [127337] by ap@apple.com
  • 3 edits in trunk/Source/WebKit2

[WK2] Use initial process in multi-process mode, too
​https://bugs.webkit.org/show_bug.cgi?id=95612

Reviewed by Sam Weinig.

  • UIProcess/WebContext.cpp: (WebKit::WebContext::WebContext): (WebKit::WebContext::warmInitialProcess): (WebKit::WebContext::disconnectProcess): (WebKit::WebContext::createWebPage):
  • UIProcess/WebContext.h: Added a boolean telling whether the last (normally, the only) process in m_processes is available for use in a new page.
4:46 PM Changeset in webkit [127336] by jamesr@google.com
  • 21 edits in trunk/Source/WebKit/chromium

[chromium] Clean up Web*Layer initialization paths
​https://bugs.webkit.org/show_bug.cgi?id=95523

Reviewed by Darin Fisher.

Constructing a Web*LayerImpl (or other compositor type) wrapper shouldn't require knowledge of the underyling
implementation. Also normalizes naming conventions in these files.

  • src/WebAnimationImpl.cpp:

(WebKit::WebAnimation::create):
(WebKit::WebAnimationImpl::WebAnimationImpl):

  • src/WebAnimationImpl.h:

(WebAnimationImpl):

  • src/WebContentLayerImpl.cpp:

(WebKit::WebContentLayerImpl::WebContentLayerImpl):
(WebKit::WebContentLayerImpl::~WebContentLayerImpl):
(WebKit::WebContentLayerImpl::layer):
(WebKit::WebContentLayerImpl::setDoubleSided):
(WebKit::WebContentLayerImpl::setContentsScale):
(WebKit::WebContentLayerImpl::setUseLCDText):
(WebKit::WebContentLayerImpl::setDrawCheckerboardForMissingTiles):

  • src/WebContentLayerImpl.h:

(WebContentLayerImpl):

  • src/WebFloatAnimationCurveImpl.cpp:

(WebKit::WebFloatAnimationCurve::create):
(WebKit::WebFloatAnimationCurveImpl::WebFloatAnimationCurveImpl):

  • src/WebFloatAnimationCurveImpl.h:

(WebFloatAnimationCurveImpl):

  • src/WebIOSurfaceLayerImpl.cpp:

(WebKit::WebIOSurfaceLayer::create):
(WebKit::WebIOSurfaceLayerImpl::WebIOSurfaceLayerImpl):

  • src/WebIOSurfaceLayerImpl.h:

(WebIOSurfaceLayerImpl):

  • src/WebImageLayerImpl.cpp:

(WebKit::WebImageLayer::create):
(WebKit::WebImageLayerImpl::WebImageLayerImpl):

  • src/WebImageLayerImpl.h:

(WebImageLayerImpl):

  • src/WebLayerImpl.cpp:

(WebKit::WebLayer::create):
(WebKit):
(WebKit::WebLayerImpl::WebLayerImpl):

  • src/WebLayerImpl.h:

(WebLayerImpl):

  • src/WebScrollbarLayerImpl.cpp:

(WebKit::WebScrollbarLayer::create):
(WebKit::WebScrollbarLayerImpl::WebScrollbarLayerImpl):

  • src/WebScrollbarLayerImpl.h:

(WebScrollbarLayerImpl):

  • src/WebSolidColorLayerImpl.cpp:

(WebKit::WebSolidColorLayer::create):
(WebKit::WebSolidColorLayerImpl::WebSolidColorLayerImpl):

  • src/WebSolidColorLayerImpl.h:

(WebSolidColorLayerImpl):

  • src/WebTransformAnimationCurveImpl.cpp:

(WebKit::WebTransformAnimationCurve::create):
(WebKit::WebTransformAnimationCurveImpl::WebTransformAnimationCurveImpl):

  • src/WebTransformAnimationCurveImpl.h:

(WebTransformAnimationCurveImpl):

  • src/WebVideoLayerImpl.cpp:

(WebKit::WebVideoLayer::create):
(WebKit::WebVideoLayerImpl::WebVideoLayerImpl):

  • src/WebVideoLayerImpl.h:

(WebVideoLayerImpl):

4:42 PM Changeset in webkit [127335] by commit-queue@webkit.org
  • 2 edits in trunk/LayoutTests

Unreviewed, rolling out r127309.
​http://trac.webkit.org/changeset/127309
​https://bugs.webkit.org/show_bug.cgi?id=95624

The Windows tests weren't actually fixed. (Requested by
mhahnenberg on #webkit).

Patch by Sheriff Bot <​webkit.review.bot@gmail.com> on 2012-08-31

  • platform/win/Skipped:
4:27 PM Changeset in webkit [127334] by ap@apple.com
  • 11 edits in trunk/Source/WebKit2

[WK2] WebProcesses should not wait 60 seconds to close in multi-process mode
​https://bugs.webkit.org/show_bug.cgi?id=95616

Reviewed by Darin Adler.

  • Shared/ChildProcess.cpp: (WebKit::ChildProcess::ChildProcess):
  • Shared/ChildProcess.h: (WebKit::ChildProcess::setTerminationTimeout): Expose a setter for timeout instead of taking it at construction time. If a derived class doesn't call the setter, default to 0.
  • PluginProcess/PluginProcess.cpp: (WebKit::PluginProcess::PluginProcess): (WebKit::PluginProcess::initializePluginProcess): Use timeout from initialization message.
  • WebProcess/WebProcess.cpp: (WebKit::WebProcess::WebProcess): (WebKit::WebProcess::initializeWebProcess): Ditto.
  • Shared/Plugins/PluginProcessCreationParameters.cpp: (WebKit::PluginProcessCreationParameters::encode): (WebKit::PluginProcessCreationParameters::decode):
  • Shared/Plugins/PluginProcessCreationParameters.h:
  • Shared/WebProcessCreationParameters.cpp: (WebKit::WebProcessCreationParameters::encode): (WebKit::WebProcessCreationParameters::decode):
  • Shared/WebProcessCreationParameters.h: Funnel timeout across process boundary.
  • UIProcess/Plugins/PluginProcessProxy.cpp: (WebKit::PluginProcessProxy::didFinishLaunching): Pass timeout as initialization message after launch.
  • UIProcess/WebContext.cpp: (WebKit::WebContext::createNewWebProcess): Pass timeout as initialization message (unlike plugin process proxy, web process proxy doesn't wait, and relies on the message being queued). (WebKit::WebContext::disconnectProcess): Skip invalidating global managers when one process quits. We still need to do something, but running this code would just result in assertion failures any time a page was closed.
4:25 PM Changeset in webkit [127333] by commit-queue@webkit.org
  • 15 edits in trunk/Source/JavaScriptCore

Refactor LLInt and supporting code in preparation for the C Loop backend.
​https://bugs.webkit.org/show_bug.cgi?id=95531.

Patch by Mark Lam <​mark.lam@apple.com> on 2012-08-31
Reviewed by Filip Pizlo.

  • bytecode/GetByIdStatus.cpp:

(JSC::GetByIdStatus::computeFromLLInt):

  • bytecode/PutByIdStatus.cpp:

(JSC::PutByIdStatus::computeFromLLInt):

  • jit/JITExceptions.cpp:

(JSC::genericThrow): Use ExecutableBase::catchRoutineFor() to fetch

fetch the catch routine for a thrown exception. This will allow
us to redefine that for the C loop later, and still keep this
code readable.

  • llint/LLIntOfflineAsmConfig.h: Moved ASM macros to

LowLevelInterpreter.cpp which is the only place they are used. This
will make it more convenient to redefine them for the C loop later.

  • llint/LLIntSlowPaths.cpp:

(JSC::LLInt::setUpCall): Use ExecutableBase's hostCodeEntry()

jsCodeEntryFor(), and jsCodeWithArityCheckEntryFor() for computing
the entry points to functions being called.

  • llint/LLIntSlowPaths.h:

(SlowPathReturnType):
(JSC::LLInt::encodeResult):
(LLInt):
(JSC::LLInt::decodeResult): Added. Needed by LLInt C Loop later.

  • llint/LowLevelInterpreter.asm:
  • llint/LowLevelInterpreter.cpp:
  • llint/LowLevelInterpreter32_64.asm:
  • llint/LowLevelInterpreter64.asm:
  • offlineasm/asm.rb: Disambiguate between opcodes and other labels.
  • offlineasm/config.rb:
  • runtime/Executable.h:

(JSC::ExecutableBase::hostCodeEntryFor): Added.
(ExecutableBase):
(JSC::ExecutableBase::jsCodeEntryFor): Added.
(JSC::ExecutableBase::jsCodeWithArityCheckEntryFor): Added.
(JSC::ExecutableBase::catchRoutineFor): Added.

  • runtime/JSValueInlineMethods.h:

(JSC):

4:16 PM Changeset in webkit [127332] by jberlin@webkit.org
  • 2 edits
    1 copy
    1 add in trunk/LayoutTests

[Lion] css3/filters/composited-during-animation-layertree.html produces empty results
​https://bugs.webkit.org/show_bug.cgi?id=95622

Move the empty results to mac-lion and add the ML results to mac to get the bots greener.

  • platform/mac-lion/css3/filters/composited-during-animation-layertree-expected.txt: Copied from LayoutTests/platform/mac/css3/filters/composited-during-animation-layertree-expected.txt.
  • platform/mac/css3/filters/composited-during-animation-layertree-expected.txt:
4:08 PM Changeset in webkit [127331] by roger_fong@apple.com
  • 1 edit
    2 deletes in trunk/LayoutTests

Unreviewed. Getting rid of Windows specific results because CSS3 flexbox has been re-enabled here: ​https://bugs.webkit.org/show_bug.cgi?id=95382

  • platform/win/fast/css/getComputedStyle/computed-style-expected.txt: Removed.
  • platform/win/fast/css/getComputedStyle/computed-style-without-renderer-expected.txt: Removed.
4:00 PM Changeset in webkit [127330] by jberlin@webkit.org
  • 2 edits
    1 copy
    1 move
    1 add in trunk/LayoutTests

[Mountain Lion] platform/mac/editing/spelling/removing-underline-after-accepting-autocorrection-using-punctuation.html failing.
​https://bugs.webkit.org/show_bug.cgi?id=95477

Add Mountain Lion expected (failing?) results and move the ones previously in mac to mac-lion to get the bots greener.

  • platform/mac-lion/editing/spelling/removing-underline-after-accepting-autocorrection-using-punctuation-expected.png: Renamed from LayoutTests/platform/mac/editing/spelling/removing-underline-after-accepting-autocorrection-using-punctuation-expected.png.
  • platform/mac-lion/editing/spelling/removing-underline-after-accepting-autocorrection-using-punctuation-expected.txt: Copied from LayoutTests/platform/mac/editing/spelling/removing-underline-after-accepting-autocorrection-using-punctuation-expected.txt.
  • platform/mac/editing/spelling/removing-underline-after-accepting-autocorrection-using-punctuation-expected.txt:
3:56 PM Changeset in webkit [127329] by pdr@google.com
  • 2 edits in trunk/LayoutTests

Unreviewed update of fast/js/function-constructor-this-value.html expectations on Chromium
​https://bugs.webkit.org/show_bug.cgi?id=95617

This test was added in r127323 but fails on Chromium ports.

  • platform/chromium/TestExpectations:
3:51 PM Changeset in webkit [127328] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/gtk

[GTK] Assert on ChromeClientGtk::scroll with delta (0, -1).
​https://bugs.webkit.org/show_bug.cgi?id=95590

Change the assert to avoid hitting when the delta does not have any
value > 0.

Patch by José Dapena Paz <​jdapena@igalia.com> on 2012-08-31
Reviewed by Martin Robinson.

  • WebCoreSupport/ChromeClientGtk.cpp:

(WebKit::ChromeClient::scroll):

3:29 PM Changeset in webkit [127327] by krit@webkit.org
  • 12 edits
    3 adds in trunk

Introduce new CSS property for clip-path
​https://bugs.webkit.org/show_bug.cgi?id=95474

Reviewed by Tim Horton.

Source/WebCore:

This path introduces the new, prefixed CSS property 'clip-path' from the
CSS Masking specification. In a first step the property just accepts
'none' and the the basic shapes from CSS Exclusion. Later it will also be
possible to reference SVG 'clipPath' elements - like Firefox already does.
To enable the parsing of the shapes, the exclusion compiler flags
around the shape parser were removed.

​http://dvcs.w3.org/hg/FXTF/raw-file/tip/masking/index.html

Test: fast/masking/parsing-clip-path-shape.html

  • css/CSSComputedStyleDeclaration.cpp:

(WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue): Return CSSValue for CSSPropertyWebkitClipPath.

  • css/CSSParser.cpp:

(WebCore::CSSParser::parseValue): Add CSSPropertyWebkitClipPath.
(WebCore::CSSParser::parseBasicShape): Modify parser function to accept CSSPropertyWebkitClipPath as well.

  • css/CSSParser.h:
  • css/CSSProperty.cpp:

(WebCore::CSSProperty::isInheritedProperty):

  • css/CSSPropertyNames.in: Added property name -webkit-clip-path.
  • css/StyleBuilder.cpp:

(ApplyPropertyClipPath): New property applier just for -webkit-clip-path.
(WebCore::ApplyPropertyClipPath::setValue):
(WebCore::ApplyPropertyClipPath::applyValue):
(WebCore::ApplyPropertyClipPath::createHandler):
(WebCore::StyleBuilder::StyleBuilder):

  • rendering/style/RenderStyle.cpp: Repaint if there is a difference between two BasicShapes.

(WebCore::RenderStyle::diff):

  • rendering/style/RenderStyle.h:
  • rendering/style/StyleRareNonInheritedData.cpp:

(WebCore::StyleRareNonInheritedData::StyleRareNonInheritedData):
(WebCore::StyleRareNonInheritedData::operator==):
(WebCore::StyleRareNonInheritedData::reportMemoryUsage):

  • rendering/style/StyleRareNonInheritedData.h: Add new member variable for clip shape.

(StyleRareNonInheritedData):

LayoutTests:

Test parsing behavior of basic shapes on new introduced '-webkit-clip-path' CSS property.

  • fast/masking/parsing-clip-path-shape-expected.txt: Added.
  • fast/masking/parsing-clip-path-shape.html: Added.
3:17 PM Changeset in webkit [127326] by zhajiang@rim.com
  • 2 edits in trunk/Source/WebKit/blackberry

[BlackBerry] ASSERT failure in RenderBlock::checkPositionedObjectsNeedLayout
​https://bugs.webkit.org/show_bug.cgi?id=95611

Reviewed by Antonio Gomes.
Patch by Jacky Jiang <​zhajiang@rim.com>

PR:200127
Remove r118405 now as we already have a new fix webkit/43f8394b for
that issue and it is not reproducible anymore after the removal.
In this way, we can get rid of this ASSERT failure as well.
Thanks Antonio for the suggestion.

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::setViewportSize):

3:09 PM Changeset in webkit [127325] by rwlbuis@webkit.org
  • 4 edits in trunk/Source

[BlackBerry] Eliminate WebGL IT policy
​https://bugs.webkit.org/show_bug.cgi?id=95604
Internal RIM PR #198764

Patch by Joshua Netterfield <​jnetterfield@rim.com> on 2012-08-31
Reviewed by Rob Buis.

This is no longer a requirement for the BlackBerry port.

Source/WebCore:

No new tests, because this introduces no new functionality.

  • html/canvas/WebGLRenderingContext.cpp:

(WebCore):
(WebCore::WebGLRenderingContext::create):

Source/WebKit/blackberry:

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::init):
(BlackBerry::WebKit::WebPage::setWebGLEnabled):

  • WebKitSupport/AboutData.cpp:

(BlackBerry::WebKit::configPage):

2:49 PM Changeset in webkit [127324] by ggaren@apple.com
  • 2 edits in trunk/LayoutTests

Not reviewed.

Removed some tabs that snuck into my last patch.

  • fast/js/function-constructor-this-value.html:
2:47 PM Changeset in webkit [127323] by ggaren@apple.com
  • 1 edit
    2 adds in trunk/LayoutTests

Use one object instead of two for closures, eliminating ScopeChainNode
​https://bugs.webkit.org/show_bug.cgi?id=95501

Reviewed by Sam Weinig.

Adding a test to demonstrate the behavior change in this patch.

  • fast/js/function-constructor-this-value-expected.txt: Added.
  • fast/js/function-constructor-this-value.html: Added.
2:34 PM Changeset in webkit [127322] by timothy@apple.com
  • 2 edits in branches/safari-536.26-branch/Source/WebKit2

Merge r127319 for <rdar://problem/12217897>.

2:26 PM Changeset in webkit [127321] by jonlee@apple.com
  • 5 edits in trunk

ondisplay event handlers are not called
​https://bugs.webkit.org/show_bug.cgi?id=95263
<rdar://problem/12193359>

Reviewed by Alexey Proskuryakov.

Source/WebCore:

To support the legacy API, we conflate ondisplay and onshow.

  • Modules/notifications/Notification.h:

(WebCore::Notification::ondisplay): Use display event as an alias for the show event.
(WebCore::Notification::setOndisplay):

LayoutTests:

Adjust the test to use the ondisplay handler, to fully test legacy API.

  • http/tests/notifications/legacy/show-expected.txt: Verify that event listener for ondisplay is

the same as that for onshow.

  • http/tests/notifications/legacy/show.html:
2:24 PM Changeset in webkit [127320] by roger_fong@apple.com
  • 1 edit
    1 add in trunk/LayoutTests

Unreviewed. Adding passing test results for fast/canvas/canvas-strokeRect-alpha-shadow.
Test passes on Windows port now but Mac port currently uses expected failing results.

  • platform/win/fast/canvas/canvas-strokeRect-alpha-shadow-expected.txt: Added.
2:18 PM Changeset in webkit [127319] by beidson@apple.com
  • 2 edits in trunk/Source/WebKit2

REGRESSION (r127252): incomplete repaint on Flash element after pinch to zoom
<rdar://problem/12217897> and ​https://bugs.webkit.org/show_bug.cgi?id=95607

Reviewed by Simon Fraser.

  • WebProcess/Plugins/PluginView.cpp: (WebKit::PluginView::viewGeometryDidChange): Only give the window coordinate clip rect a chance

to override the boundsRect() if the page scale factor is 1.

2:17 PM Changeset in webkit [127318] by tony@chromium.org
  • 37 edits in trunk

Remove ENABLE_CSS3_FLEXBOX compile time flag
​https://bugs.webkit.org/show_bug.cgi?id=95382

Reviewed by Ojan Vafai.

Everyone is already enabling this by default and the spec has stablized.

.:

  • Source/cmake/WebKitFeatures.cmake:
  • Source/cmakeconfig.h.cmake:
  • configure.ac:

Source/JavaScriptCore:

  • Configurations/FeatureDefines.xcconfig:

Source/WebCore:

No new tests, shouldn't change anything.

  • Configurations/FeatureDefines.xcconfig:
  • GNUmakefile.am:
  • GNUmakefile.features.am:
  • css/CSSComputedStyleDeclaration.cpp:

(WebCore):
(WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):

  • css/CSSParser.cpp:

(WebCore::isValidKeywordPropertyAndValue): Expand check range to include flex and inline-flex.
(WebCore::isKeywordPropertyID):
(WebCore::CSSParser::parseValue):

  • css/CSSPrimitiveValueMappings.h:

(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):

  • css/CSSProperty.cpp:

(WebCore::CSSProperty::isInheritedProperty):

  • css/CSSPropertyNames.in:
  • css/CSSValueKeywords.in:
  • css/StyleBuilder.cpp:

(WebCore::StyleBuilder::StyleBuilder):

  • css/StylePropertySet.cpp:

(WebCore::StylePropertySet::getPropertyValue):
(WebCore::StylePropertySet::asText):

  • css/StylePropertyShorthand.cpp:

(WebCore::webkitFlexShorthand):
(WebCore::shorthandForProperty):

  • css/StylePropertyShorthand.h:

(WebCore):

  • css/StyleResolver.cpp:

(WebCore::StyleResolver::collectMatchingRulesForList):

  • page/animation/CSSPropertyAnimation.cpp:

(WebCore::CSSPropertyAnimation::ensurePropertyMap):

  • rendering/RenderObject.cpp:

(WebCore::RenderObject::createObject):

  • rendering/style/RenderStyle.h:
  • rendering/style/RenderStyleConstants.h:

Source/WebKit/chromium:

  • features.gypi:

Source/WebKit/mac:

  • Configurations/FeatureDefines.xcconfig:

Source/WebKit2:

  • Configurations/FeatureDefines.xcconfig:

Tools:

  • Scripts/webkitperl/FeatureList.pm:
  • qmake/mkspecs/features/features.pri:

WebKitLibraries:

  • win/tools/vsprops/FeatureDefines.vsprops:
  • win/tools/vsprops/FeatureDefinesCairo.vsprops:
2:11 PM Changeset in webkit [127317] by commit-queue@webkit.org
  • 7 edits in trunk/Source

[chromium] Add main versus impl thread asserts
​https://bugs.webkit.org/show_bug.cgi?id=95596

Patch by Christopher Cameron <​ccameron@chromium.org> on 2012-08-31
Reviewed by James Robinson.

Add asserts to prioritized texture manager that specify that a
function be called on the main or impl thread.

Add debug signs that the main thread is blocked to the threaded
proxy.

Add a mechanism to simultaneously set the thread to the impl thread
and indicate that the main thread is blocked.

Update tests that hit asserts in the prioritized texture manager to
set their impl thread and main thread blocked state correctly.

Source/WebCore:

  • platform/graphics/chromium/cc/CCPrioritizedTextureManager.cpp:

(WebCore::CCPrioritizedTextureManager::prioritizeTextures):
(WebCore::CCPrioritizedTextureManager::clearPriorities):
(WebCore::CCPrioritizedTextureManager::requestLate):
(WebCore::CCPrioritizedTextureManager::acquireBackingTextureIfNeeded):
(WebCore::CCPrioritizedTextureManager::reduceMemory):
(WebCore::CCPrioritizedTextureManager::registerTexture):
(WebCore::CCPrioritizedTextureManager::unregisterTexture):
(WebCore::CCPrioritizedTextureManager::returnBackingTexture):
(WebCore::CCPrioritizedTextureManager::createBacking):
(WebCore::CCPrioritizedTextureManager::assertInvariants):

  • platform/graphics/chromium/cc/CCSingleThreadProxy.h:

(WebCore):
(DebugScopedSetImplThreadAndMainThreadBlocked):

  • platform/graphics/chromium/cc/CCThreadProxy.cpp:

(WebCore::CCThreadProxy::compositeAndReadback):
(WebCore::CCThreadProxy::finishAllRendering):
(WebCore::CCThreadProxy::setVisible):
(WebCore::CCThreadProxy::initializeRenderer):
(WebCore::CCThreadProxy::recreateContext):
(WebCore::CCThreadProxy::implSideRenderingStats):
(WebCore::CCThreadProxy::start):
(WebCore::CCThreadProxy::forceSerializeOnSwapBuffers):
(WebCore::CCThreadProxy::acquireLayerTextures):

Source/WebKit/chromium:

  • tests/CCPrioritizedTextureTest.cpp:

(WTF::CCPrioritizedTextureTest::CCPrioritizedTextureTest):
(WTF::CCPrioritizedTextureTest::~CCPrioritizedTextureTest):
(WTF::CCPrioritizedTextureTest::validateTexture):
(WTF::TEST_F):

  • tests/TiledLayerChromiumTest.cpp:
2:03 PM Changeset in webkit [127316] by pdr@google.com
  • 2 edits in trunk/LayoutTests

Unreviewed update of fast/js/function-apply-aliased.html expectations.

fast/js/function-apply-aliased.html has been failing for TEXT for some time but
the chromium-mac TestExpectations did not have an entry for mac. This patch
updates the expectations to include all platforms.

The bug (crbug.com/48239) has been updated to reflect this.

  • platform/chromium/TestExpectations:
1:55 PM Changeset in webkit [127315] by jberlin@webkit.org
  • 4 edits in trunk/LayoutTests

LayoutTests/http/tests/websocket/tests/hixie76/frame-lengths.html was removed. Remove it from the Skipped list.

  • platform/mac-wk2/Skipped:
  • platform/qt-arm/Skipped:
  • platform/wincairo/Skipped:
1:55 PM Changeset in webkit [127314] by jberlin@webkit.org
  • 4 edits in trunk/LayoutTests

compositing/tiling/huge-layer* flakey
​https://bugs.webkit.org/show_bug.cgi?id=48454

These tests were disabled in r89631. Remove them from the Skipped list so that
run-webkit-tests doesn't try to find them.

  • platform/win/Skipped:
  • platform/wincairo/Skipped:
  • platform/wk2/Skipped:
1:47 PM Changeset in webkit [127313] by psolanki@apple.com
  • 2 edits in trunk/Source/WebKit2

objc_msgSend and IMP should be cast appropriately before using
​https://bugs.webkit.org/show_bug.cgi?id=95242

Reviewed by Benjamin Poulain.

Follow on fix for Mac EWS - use objc-runtime.h instead of runtime.h so
that it finds the declaration for objc_msgSend.

  • UIProcess/API/mac/PDFViewController.mm:
1:41 PM Changeset in webkit [127312] by jberlin@webkit.org
  • 2 edits in trunk/LayoutTests

The fast/forms/ file tests were moved in the file sub-directory in r96918. Remove the
old entries from the Skipped list.

  • platform/mac-wk2/Skipped:
1:40 PM Changeset in webkit [127311] by jamesr@google.com
  • 16 edits
    1 copy
    2 adds in trunk/Source

[chromium] Put compositor client interfaces in separate headers, normalize naming
​https://bugs.webkit.org/show_bug.cgi?id=95522

Reviewed by Adam Barth.

Source/WebCore:

Client interfaces should be in their own headers so code implementing these interfaces do not have to
transitively #include all of the compositor internals.

We call things 'clients' not 'delegates' in the compositor.

  • WebCore.gypi:
  • platform/graphics/chromium/ContentLayerChromium.cpp:

(WebCore::ContentLayerPainter::ContentLayerPainter):
(WebCore::ContentLayerPainter::create):
(WebCore::ContentLayerPainter::paint):
(WebCore::ContentLayerChromium::create):
(WebCore::ContentLayerChromium::ContentLayerChromium):
(WebCore::ContentLayerChromium::drawsContent):
(WebCore::ContentLayerChromium::createTextureUpdaterIfNeeded):

  • platform/graphics/chromium/ContentLayerChromium.h:

(WebCore):
(ContentLayerPainter):
(ContentLayerChromium):
(WebCore::ContentLayerChromium::clearClient):

  • platform/graphics/chromium/TextureLayerChromium.cpp:
  • platform/graphics/chromium/TextureLayerChromium.h:

(WebCore):

  • platform/graphics/chromium/cc/CCLayerTreeHost.cpp:
  • platform/graphics/chromium/cc/CCLayerTreeHost.h:

(WebCore):

Source/WebKit/chromium:

  • src/WebContentLayerImpl.cpp:

(WebKit::WebContentLayerImpl::~WebContentLayerImpl):

  • src/WebContentLayerImpl.h:

(WebContentLayerImpl):

  • src/WebExternalTextureLayerImpl.h:
  • src/WebLayerTreeViewImpl.h:

(WebCore):

  • tests/CCLayerTreeHostCommonTest.cpp:
  • tests/CCLayerTreeHostTest.cpp:

(WebKitTests::ContentLayerChromiumWithUpdateTracking::create):
(WebKitTests::ContentLayerChromiumWithUpdateTracking::ContentLayerChromiumWithUpdateTracking):
(WebKitTests::CCLayerTreeHostTestDeviceScaleFactorScalesViewportAndLayers::CCLayerTreeHostTestDeviceScaleFactorScalesViewportAndLayers):
(WebKitTests::CCLayerTreeHostTestAtomicCommit::CCLayerTreeHostTestAtomicCommit):
(WebKitTests::CCLayerTreeHostTestAtomicCommitWithPartialUpdate::CCLayerTreeHostTestAtomicCommitWithPartialUpdate):
(WebKitTests::CCLayerTreeHostTestLostContextWhileUpdatingResources::CCLayerTreeHostTestLostContextWhileUpdatingResources):
(CCLayerTreeHostTestLostContextWhileUpdatingResources):

  • tests/ContentLayerChromiumTest.cpp:

(WebKit::MockContentLayerChromiumClient::MockContentLayerChromiumClient):
(WebKit::TEST):

1:36 PM Changeset in webkit [127310] by roger_fong@apple.com
  • 2 edits in trunk/LayoutTests

Unreviewed gardening. Adding fast/css/nested-layers-with-over.html to Windows skip list.
​https://bugs.webkit.org/show_bug.cgi?id=95388

  • platform/win/Skipped:
1:32 PM Changeset in webkit [127309] by mhahnenberg@apple.com
  • 2 edits in trunk/LayoutTests

Unskip plugins/npruntime/delete-plugin-within-* tests on Windows
​https://bugs.webkit.org/show_bug.cgi?id=95597

Reviewed by Darin Adler.

  • platform/win/Skipped:
1:28 PM Changeset in webkit [127308] by jonlee@apple.com
  • 2 edits in trunk/LayoutTests

Unreviewed. Skip http/tests/notifications on Lion, since there is no support.

  • platform/mac-lion/Skipped:
1:15 PM Changeset in webkit [127307] by roger_fong@apple.com
  • 2 edits
    1 add in trunk/LayoutTests

Unreviewed. Adding expected failing results for accessibility/canvas tests.
​https://bugs.webkit.org/show_bug.cgi?id=95598

Windows description and role strings don't match up with what is expected on Mac.
Particularly, they do not include the "AX" at the beginning of the descriptions/roles of canvas fallback content.

  • platform/win/accessibility/canvas-description-and-role-expected.txt: Added.
  • platform/win/accessibility/canvas-fallback-content-expected.txt:
1:11 PM Changeset in webkit [127306] by psolanki@apple.com
  • 9 edits in trunk

objc_msgSend and IMP should be cast appropriately before using
​https://bugs.webkit.org/show_bug.cgi?id=95242

Reviewed by Benjamin Poulain.

Fix for older compilers. Pass id as the return type to the template
instead of relying on default type.

Source/WebKit/mac:

  • WebView/WebDelegateImplementationCaching.mm:

(CallDelegate):
(CallFormDelegate):

  • WebView/WebHTMLView.mm:

(setCursor):
(setNeedsDisplayInRect):

Source/WTF:

  • wtf/Functional.h:

(WTF::R):

  • wtf/ObjcRuntimeExtras.h:

Tools:

  • DumpRenderTree/mac/DumpRenderTree.mm:

(drt_NSFontManager_availableFontFamilies):

  • WebKitTestRunner/InjectedBundle/mac/ActivateFonts.mm:

(WTR::wtr_NSFontManager_availableFontFamilies):

1:10 PM Changeset in webkit [127305] by gavinp@chromium.org
  • 4 edits in trunk

[chromium] DEPS roll with a test skip.
​https://bugs.webkit.org/show_bug.cgi?id=95601

Unreviewed gardening & DEPS roll.

Source/WebKit/chromium:

fast/frames/cached-frame-counter.html was never the most stable test, and now it's failing on more platforms. Marking it flaky, along with a DEPS roll (earlier attempts at the DEPS roll failed on this)

  • DEPS:

LayoutTests:

fast/frames/cached-frame-counter.html was never the most stable test, and now it's failing on more platforms. Marking it flaky, along with a DEPS roll (earlier attempts at the DEPS roll failed on this)

  • platform/chromium/TestExpectations:
1:10 PM Changeset in webkit [127304] by ggaren@apple.com
  • 23 edits in trunk/Source/JavaScriptCore

Not reviewed.

Rolled out ​http://trac.webkit.org/changeset/127293 because it broke
inspector tests on Windows.

Shrink activation objects by half
​https://bugs.webkit.org/show_bug.cgi?id=95591

Reviewed by Sam Weinig.

12:51 PM Changeset in webkit [127303] by benjamin@webkit.org
  • 2 edits in trunk/Source/WebCore

Fix the Debug builds after r127277

Unreviewed.

One assertion was not updated in r127277. The encapsulation of CSSParserString's length
has changed and now use an accessor.

Patch by Benjamin Poulain <​bpoulain@apple.com> on 2012-08-31

  • css/CSSParser.cpp:

(WebCore::CSSParser::storeVariableDeclaration):

12:49 PM Changeset in webkit [127302] by zandobersek@gmail.com
  • 7 edits in trunk/Tools

nrwt: use scm instead of calling svn directly to get the revision in json results generator
​https://bugs.webkit.org/show_bug.cgi?id=89616

Reviewed by Eric Seidel.

Add the new 'executive' optional function argument to the SCM.in_working_directory
class method. This way we get to reuse an Executive instance which is necessary in testing
when we're operating with a mock version of the object.

In JSONResultsGeneratorBase, it's now possible to get the current SVN revision by using
the SCMDetector with the generator's FileSystem and Executive instances.

  • Scripts/webkitpy/common/checkout/scm/detection.py:

(SCMDetector.detect_scm_system):

  • Scripts/webkitpy/common/checkout/scm/detection_unittest.py:

(SCMDetectorTest.test_detect_scm_system): Update the unit test to capture the mock
Executive output and test it against expected output.

  • Scripts/webkitpy/common/checkout/scm/git.py:

(Git.in_working_directory):

  • Scripts/webkitpy/common/checkout/scm/scm.py:

(SCM.in_working_directory):

  • Scripts/webkitpy/common/checkout/scm/svn.py:

(SVN.in_working_directory):

  • Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:

(JSONResultsGeneratorBase.init):
(JSONResultsGeneratorBase._get_svn_revision):

12:38 PM Changeset in webkit [127301] by Simon Fraser
  • 6 edits
    2 adds in trunk

(Regression: r126774): Fix crash when scrolling after removing inline sticky element
​https://bugs.webkit.org/show_bug.cgi?id=95584

Reviewed by Dave Hyatt.

Source/WebCore:

Move fixed/sticky registration and unregistration with the FrameView from
RenderBox and RenderInline into RenderBoxModelObject, which also
fixes the issue that inlines didn't unregister themselves on destruction.

Test: fast/css/sticky/remove-inline-sticky-crash.html

  • rendering/RenderBox.cpp:

(WebCore::RenderBox::willBeDestroyed): Code moved to RenderBoxModelObject. No need
to null-check style.
(WebCore::RenderBox::styleWillChange): Code moved to RenderBoxModelObject.

  • rendering/RenderBoxModelObject.cpp:

(WebCore::RenderBoxModelObject::willBeDestroyed): Remove fixed objects. Check
isPositioned() to avoid this work for most non-positioned renderers.
(WebCore::RenderBoxModelObject::styleWillChange): Register and unregister fixed
and sticky objects with the FrameView.

  • rendering/RenderInline.cpp: styleWillChange() is no longer needed.
  • rendering/RenderInline.h: Ditto.

LayoutTests:

Testcase removes an inline sticky, then scrolls.

  • fast/css/sticky/remove-inline-sticky-crash-expected.txt: Added.
  • fast/css/sticky/remove-inline-sticky-crash.html: Added.
12:11 PM Changeset in webkit [127300] by commit-queue@webkit.org
  • 35 edits in trunk/Source/WebCore

Remove extraneous includes (Node.h, Document.h, Element.h, RenderObject.h)
​https://bugs.webkit.org/show_bug.cgi?id=95259

Patch by Nikhil Bhargava <​nbhargava@google.com> on 2012-08-31
Reviewed by Sam Weinig.

Node.h, Document.h, Element.h, and RenderObject.h are four of the most
expensive headers to include. This patch removes them from many of the
places they are not needed. There is a minor compile-time performance
boost.

  • dom/ElementShadow.h:
  • dom/Range.h:

(WebCore):

  • dom/TreeScopeAdopter.h:

(WebCore):

  • editing/VisiblePosition.h:

(WebCore):

  • html/HTMLCollection.h:

(WebCore):

  • html/parser/HTMLElementStack.h:
  • html/parser/HTMLTreeBuilder.h:

(WebCore):

  • html/track/WebVTTParser.h:

(WebCore):

  • inspector/InspectorCSSAgent.h:

(WebCore):

  • inspector/InspectorInstrumentation.h:

(WebCore):

  • inspector/InspectorOverlay.h:
  • loader/ImageLoader.h:
  • loader/TextTrackLoader.h:

(WebCore):

  • platform/chromium/PopupListBox.cpp:
  • platform/chromium/PopupListBox.h:

(WebCore):

  • platform/graphics/MediaPlayer.h:

(WebCore):

  • rendering/FilterEffectRenderer.h:
  • rendering/InlineFlowBox.cpp:
  • rendering/RenderCounter.cpp:
  • rendering/RenderGeometryMap.h:

(WebCore):

  • rendering/RenderInputSpeech.h:

(WebCore):

  • rendering/RenderMediaControls.h:

(WebCore):

  • rendering/RenderMediaControlsChromium.h:

(WebCore):

  • rendering/RenderQuote.cpp:
  • rendering/RenderQuote.h:

(WebCore):

  • rendering/RenderScrollbar.h:

(WebCore):

  • rendering/RenderText.h:

(WebCore):

  • rendering/RenderTheme.h:

(WebCore):

  • rendering/RenderThemeChromiumSkia.h:

(WebCore):

  • rendering/svg/RenderSVGResourceMarker.h:

(WebCore):

  • svg/graphics/filters/SVGFilterBuilder.h:

(WebCore):

  • xml/NativeXPathNSResolver.h:

(WebCore):
(NativeXPathNSResolver):
(WebCore::NativeXPathNSResolver::create):

  • xml/XPathStep.h:

(WebCore):
(XPath):
(Step):
(NodeTest):
(WebCore::XPath::Step::NodeTest::NodeTest):
(WebCore::XPath::Step::NodeTest::kind):
(WebCore::XPath::Step::NodeTest::data):
(WebCore::XPath::Step::NodeTest::namespaceURI):
(WebCore::XPath::Step::NodeTest::mergedPredicates):
(WebCore::XPath::Step::axis):
(WebCore::XPath::Step::nodeTest):

  • xml/parser/NewXMLDocumentParser.cpp:
  • xml/parser/NewXMLDocumentParser.h:

(WebCore):

  • xml/parser/XMLToken.h:
  • xml/parser/XMLTreeBuilder.cpp:
  • xml/parser/XMLTreeBuilder.h:

(WebCore):

12:08 PM Changeset in webkit [127299] by jonlee@apple.com
  • 19 edits
    11 adds in trunk

[Tests] Add basic tests to http/tests/notifications
​https://bugs.webkit.org/show_bug.cgi?id=95493
<rdar://problem/12209303>

Reviewed by Alexey Proskuryakov.

Source/WebKit2:

Simplify the passing of the permission policy to just a boolean instead of converting it
back and forth from the NotificationClient::Permission enum.

  • WebProcess/InjectedBundle/API/c/WKBundlePrivate.h: Move the declaration to the section

of functions that are TestRunner SPI.

  • WebProcess/InjectedBundle/InjectedBundle.cpp:

(WebKit::InjectedBundle::setWebNotificationPermission): Use a boolean.

  • WebProcess/Notifications/NotificationPermissionRequestManager.cpp:

(WebKit::NotificationPermissionRequestManager::setPermissionLevelForTesting):

  • WebProcess/Notifications/NotificationPermissionRequestManager.h:

(NotificationPermissionRequestManager):

  • win/WebKit2.vcproj: Add WKNotificationPermissionRequest.{h,cpp}.
  • win/WebKit2Generated.make: Export WKNotificationPermissionRequest.h.

Tools:

Add permission request support for WTR.

  • WebKitTestRunner/TestController.cpp:

(WTR::TestController::initialize):
(WTR::TestController::decidePolicyForNotificationPermissionRequest): Always allow.

  • WebKitTestRunner/TestController.h:

(TestController):

LayoutTests:

  • http/tests/notifications/request-expected.txt: Added.
  • http/tests/notifications/request.html: Added. This test requests permission. DRT/WTR

is expected to accept it, and we verify that both the parameter in the callback function,
and Notification.permission, are both set to "granted".

  • http/tests/notifications/show-expected.txt: Added.
  • http/tests/notifications/show.html: Added. This test checks that the various permission

APIs on TestRunner work.

Legacy analog to above tests.

  • http/tests/notifications/legacy/request-expected.txt: Added.
  • http/tests/notifications/legacy/request.html: Added.
  • http/tests/notifications/legacy/show-expected.txt: Added.
  • http/tests/notifications/legacy/show.html: Added.
  • http/tests/resources/js-test-post-async.js: Added. This will be used for most tests since

the events are called asynchronously.

  • http/tests/resources/js-test-pre.js: Updated from fast/js/resources.
  • http/tests/resources/notifications-test-pre.js: Added. Adds a helper function to complete

the test, and defines the domain of http/tests.
(testCompleted):

Skip.

  • platform/chromium/TestExpectations: Skip on Android. Filed 95506 to bring Chromium support.
  • platform/efl/TestExpectations:
  • platform/gtk/TestExpectations:
  • platform/mac/Skipped:
  • platform/qt/Skipped: Filed 95507 to bring Qt support.
  • platform/win/Skipped:
  • platform/wincairo/Skipped:
11:55 AM Changeset in webkit [127298] by jonlee@apple.com
  • 5 edits in trunk/Tools

[DRT] Make simulating a web click on a notification a queued task
​https://bugs.webkit.org/show_bug.cgi?id=95546
<rdar://problem/12214170>

Reviewed by Alexey Proskuryakov.

Making the web click happen asynchronously better mimics user interaction with the platform.

  • DumpRenderTree/TestRunner.h: Add a flag that determines whether we have a pending notification click.

(TestRunner::hasPendingWebNotificationClick): Exposed so that dump() can check that the flag is
not set.

  • DumpRenderTree/mac/DumpRenderTree.mm:

(dump): Assert that the flag is not set.

  • DumpRenderTree/mac/TestRunnerMac.mm:

(TestRunner::simulateWebNotificationClick): Set the flag, then add a task to click the
notification. In the case where the block gets executed after the flag has been unset, we avoid
performing the click.

11:51 AM Changeset in webkit [127297] by hyatt@apple.com
  • 3 edits in trunk/Source/WebCore

[New Multicolumn] Implement column repainting.
​https://bugs.webkit.org/show_bug.cgi?id=95593

Reviewed by Dan Bernstein.

Make the new columns repaint properly. Note this code is a significant improvement over the
old multicolumn code in that repaints are properly issued per-column, i.e., no more fuzzy
uniting of rects.

  • rendering/RenderMultiColumnSet.cpp:

(WebCore::RenderMultiColumnSet::columnIndexAtOffset):
(WebCore):
(WebCore::RenderMultiColumnSet::repaintFlowThreadContent):

  • rendering/RenderMultiColumnSet.h:
11:45 AM Changeset in webkit [127296] by pdr@google.com
  • 2 edits in trunk/LayoutTests

Unreviewed update of test expectations.

Marking media/track/track-cue-rendering-snap-to-lines-not-set.html as
flaky again until WK89167 is fixed.

  • platform/chromium/TestExpectations:
11:25 AM Changeset in webkit [127295] by roger_fong@apple.com
  • 2 edits in trunk/LayoutTests

Unreviewed Gardening. inspector/profiler/webgl/webgl-profiler-api-changes.html fails on Windows.
WebGL not enabled on Windows.

  • platform/win/Skipped:
11:24 AM Changeset in webkit [127294] by alokp@chromium.org
  • 2 edits in trunk/Source/WebCore

[chromium] Layout tests svg-filters.html and alpha.html are crashing intermittently
​https://bugs.webkit.org/show_bug.cgi?id=92660

Reviewed by James Robinson.

We use two GL contexts to create and upload tile textures.
One context creates the texture, while another uploads it.
Even though the texture is shared by the two contexts, it is not
available to the uploading context immediately after creation,
unless the creating context is flushed.

  • platform/graphics/chromium/FrameBufferSkPictureCanvasLayerTextureUpdater.cpp:

(WebCore::createAcceleratedCanvas):
(WebCore::FrameBufferSkPictureCanvasLayerTextureUpdater::updateTextureRect):

11:22 AM Changeset in webkit [127293] by ggaren@apple.com
  • 23 edits in trunk/Source/JavaScriptCore

Shrink activation objects by half
​https://bugs.webkit.org/show_bug.cgi?id=95591

Reviewed by Sam Weinig.

Removed the global object, global data, and global this pointers from
JSScope, and changed an int to a bitfield. This gets the JSActivation
class down to 64 bytes, which in practice cuts it in half by getting it
out of the 128 byte size class.

Now, it's one extra indirection to get these pointers. These pointers
aren't accessed by JIT code, so I thought there would be no cost to the
extra indirection. However, some C++-heavy SunSpider tests regressed a
bit in an early version of the patch, which added even more indirection.
This suggests that calls to exec->globalData() and/or exec->lexicalGlobalObject()
are common and probably duplicated in lots of places, and could stand
further optimization in C++.

  • dfg/DFGAbstractState.cpp:

(JSC::DFG::AbstractState::execute): Test against the specific activation
for our global object, since there's no VM-shared activation structure
anymore. This is guaranteed to have the same success rate as the old test
because activation scope is fixed at compile time.

  • heap/MarkedBlock.cpp:

(JSC::MarkedBlock::MarkedBlock):

  • heap/MarkedBlock.h:

(JSC::MarkedBlock::globalData):

  • heap/WeakSet.cpp:

(JSC::WeakSet::addAllocator):

  • heap/WeakSet.h:

(WeakSet):
(JSC::WeakSet::WeakSet):
(JSC::WeakSet::globalData): Store a JSGlobalData* instead of a Heap*
because JSGlobalData->Heap is just a constant fold in the addressing
mode, while Heap->JSGlobalData is an extra pointer dereference. (These
objects should eventually just merge.)

  • jit/JITOpcodes.cpp:

(JSC::JIT::emit_op_resolve_global_dynamic): See DFGAbstractState.cpp.

  • llint/LowLevelInterpreter32_64.asm:
  • llint/LowLevelInterpreter64.asm: Load the activation structure from

the code block instead of the global data because the structure is not
VM-shared anymore. (See DFGAbstractState.cpp.)

  • runtime/JSActivation.cpp:

(JSC::JSActivation::JSActivation):

  • runtime/JSActivation.h:

(JSActivation): This is the point of the patch: Remove the data.

  • runtime/JSGlobalData.cpp:

(JSC::JSGlobalData::JSGlobalData):

  • runtime/JSGlobalData.h:

(JSGlobalData): No longer VM-shared. (See DFGAbstractState.cpp.)

(JSC::WeakSet::heap): (See WeakSet.h.)

  • runtime/JSGlobalObject.cpp:

(JSC::JSGlobalObject::JSGlobalObject):
(JSC::JSGlobalObject::setGlobalThis):
(JSC::JSGlobalObject::reset):
(JSC::JSGlobalObject::visitChildren):

  • runtime/JSGlobalObject.h:

(JSGlobalObject):
(JSC::JSGlobalObject::withScopeStructure):
(JSC::JSGlobalObject::strictEvalActivationStructure):
(JSC::JSGlobalObject::activationStructure):
(JSC::JSGlobalObject::nameScopeStructure):
(JSC::JSScope::globalThis):
(JSC::JSGlobalObject::globalThis): Data that used to be in the JSScope
class goes here now, so it's not duplicated across all activations.

  • runtime/JSNameScope.h:

(JSC::JSNameScope::JSNameScope):

  • runtime/JSScope.cpp:

(JSC::JSScope::visitChildren): This is the point of the patch: Remove the data.

  • runtime/JSScope.h:

(JSScope):
(JSC::JSScope::JSScope):
(JSC::JSScope::globalObject):
(JSC::JSScope::globalData):

  • runtime/JSSegmentedVariableObject.h:

(JSC::JSSegmentedVariableObject::JSSegmentedVariableObject):

  • runtime/JSSymbolTableObject.h:

(JSC::JSSymbolTableObject::JSSymbolTableObject):

  • runtime/JSVariableObject.h:

(JSC::JSVariableObject::JSVariableObject):

  • runtime/JSWithScope.h:

(JSC::JSWithScope::JSWithScope):

  • runtime/StrictEvalActivation.cpp:

(JSC::StrictEvalActivation::StrictEvalActivation): Simplified now that
we don't need to pass so much data to JSScope.

11:12 AM Changeset in webkit [127292] by jberlin@webkit.org
  • 2 edits in trunk/LayoutTests

Mark http/tests/security/inactive-document-with-empty-security-origin.html while Geoff Garen
investigates in ​https://bugs.webkit.org/show_bug.cgi?id=95501 in order to get the bots green.

  • platform/mac/TestExpectations:
11:07 AM Changeset in webkit [127291] by gavinp@chromium.org
  • 2 edits in trunk/LayoutTests

[chromium] Fix platform/chromium/TestExpectations typo
​https://bugs.webkit.org/show_bug.cgi?id=95592

Unreviewed gardening.

Unfortunately r127174 contained a typo that broke webkit-lint on our build. See ​https://bugs.webkit.org/show_bug.cgi?id=95581 .

  • platform/chromium/TestExpectations:
11:00 AM Changeset in webkit [127290] by jberlin@webkit.org
  • 3 edits in trunk/LayoutTests

fast/css/align-positioned-object-on-resize.html skipped on Mac platforms
​https://bugs.webkit.org/show_bug.cgi?id=83489

This test was removed in r126911. Remove it from the Skipped list.

  • platform/mac/Skipped:
  • platform/win/Skipped:
11:00 AM Changeset in webkit [127289] by jberlin@webkit.org
  • 2 edits in trunk/LayoutTests

http/tests/cache/cancel-in-progress-load.html fails on several platforms
​https://bugs.webkit.org/show_bug.cgi?id=87728

The test was rolled out in r119918. Remove it from the Skipped list so that run-webkit-tests
doesn't complain about not being able to find the test.

  • platform/mac/Skipped:
11:00 AM Changeset in webkit [127288] by jberlin@webkit.org
  • 2 edits in trunk/LayoutTests

plugins/npruntime/leak-window-scriptable-object.html is failing
​https://bugs.webkit.org/show_bug.cgi?id=90710

This test was removed in r91637. Remove it from the Skipped list so that run-webkit-tests
doesn't complain about not being able to find the test.

  • platform/mac/Skipped:
11:00 AM Changeset in webkit [127287] by jberlin@webkit.org
  • 3 edits in trunk/LayoutTests

[Qt] http/tests/cookies/js-get-and-set-http-only-cookie.html failing since introduction in r119947
​https://bugs.webkit.org/show_bug.cgi?id=87208

The .php version was removed in r118166, but the .html version it was replaced with in r119947
was separately marked as failing on Qt. Update the Skipped lists so that run-webkit-tests
doesn't complain about not being able to find the test on Mac (and so that the comment in the
Qt Skipped list is correct).

  • platform/mac/Skipped:
  • platform/qt/Skipped:
11:00 AM Changeset in webkit [127286] by jberlin@webkit.org
  • 3 edits in trunk/LayoutTests

fast/workers/storage/interrupt-database-sync.html fails intermittently
​https://bugs.webkit.org/show_bug.cgi?id=49906

The test was disabled in r81333, so remove it from the Skipped lists so that run-webkit-tests
doesn't complain about not being able to find the test.

  • platform/win-xp/Skipped:
  • platform/wk2/Skipped:
10:58 AM Changeset in webkit [127285] by kareng@chromium.org
  • 1 edit in branches/chromium/1132/Source/WebCore/bindings/v8/V8DOMWindowShell.cpp

Merge 124689 - Disabling eval changes the timing of DidCreateScriptContext
​https://bugs.webkit.org/show_bug.cgi?id=92189

Reviewed by Eric Seidel.

When we implemented Content-Security-Policy, we added the ability to
disable eval in the JavaScript engine. However, when we process the
Content-Security-Policy header, we might not have initialized the
script context for the given frame.

Previously, we would initialize the context, but that generates a
DidCreateScriptContext callback to the embedder earlier in the Document's
lifetime that before. A natural thing to do in this callback is to run
script to customize the script context, but Document isn't fully
initialized yet, which leads to odd bugs and general confusion.

In this patch, we delay actually disabling eval until we would have
created the scripting context previously. From the perspective of the
web platform, this has the same behavior. The only difference is that
now the DidCreateScriptContext notification occurs at the same time
regardless of whether Content-Security-Policy disables eval.

I tried to write a test for this change, but it was unclear to me how
to write a good test. I tried writing a Chromium WebKit unit test to no
avail. The good news is that this patch will be covered by the
PlatformAppBrowserTest.Iframes test in Chromium once
​https://bugs.webkit.org/show_bug.cgi?id=93079 lands. That's not the
best way to test this change, but it might be sufficient.

  • bindings/js/ScriptController.cpp:

(WebCore::ScriptController::initScript):
(WebCore::ScriptController::disableEval):

  • bindings/v8/ScriptController.cpp:

(WebCore::ScriptController::enableEval):
(WebCore::ScriptController::disableEval):

  • bindings/v8/V8DOMWindowShell.cpp:

(WebCore::V8DOMWindowShell::initContextIfNeeded):

TBR=​abarth@webkit.org
Review URL: ​https://chromiumcodereview.appspot.com/10919034

10:50 AM Changeset in webkit [127284] by timothy@apple.com
  • 2 edits in branches/safari-536.26-branch/LayoutTests

Rename testRunner to layoutTestController for a test on the branch.

Patch by Brady Eidson <​beidson@apple.com> on 2012-08-31
Rubberstamped by Tim Hatcher.

  • platform/mac-wk2/plugins/npp-setwindow-called-on-scroll.html:
10:50 AM Changeset in webkit [127283] by kareng@chromium.org
  • 1 edit in branches/chromium/1132/Source/WebCore/bindings/v8/V8DOMWindowShell.cpp

Merge 124689 - Disabling eval changes the timing of DidCreateScriptContext
​https://bugs.webkit.org/show_bug.cgi?id=92189

Reviewed by Eric Seidel.

When we implemented Content-Security-Policy, we added the ability to
disable eval in the JavaScript engine. However, when we process the
Content-Security-Policy header, we might not have initialized the
script context for the given frame.

Previously, we would initialize the context, but that generates a
DidCreateScriptContext callback to the embedder earlier in the Document's
lifetime that before. A natural thing to do in this callback is to run
script to customize the script context, but Document isn't fully
initialized yet, which leads to odd bugs and general confusion.

In this patch, we delay actually disabling eval until we would have
created the scripting context previously. From the perspective of the
web platform, this has the same behavior. The only difference is that
now the DidCreateScriptContext notification occurs at the same time
regardless of whether Content-Security-Policy disables eval.

I tried to write a test for this change, but it was unclear to me how
to write a good test. I tried writing a Chromium WebKit unit test to no
avail. The good news is that this patch will be covered by the
PlatformAppBrowserTest.Iframes test in Chromium once
​https://bugs.webkit.org/show_bug.cgi?id=93079 lands. That's not the
best way to test this change, but it might be sufficient.

  • bindings/js/ScriptController.cpp:

(WebCore::ScriptController::initScript):
(WebCore::ScriptController::disableEval):

  • bindings/v8/ScriptController.cpp:

(WebCore::ScriptController::enableEval):
(WebCore::ScriptController::disableEval):

  • bindings/v8/V8DOMWindowShell.cpp:

(WebCore::V8DOMWindowShell::initContextIfNeeded):

TBR=​abarth@webkit.org

10:48 AM Changeset in webkit [127282] by kareng@chromium.org
  • 4 edits in branches/chromium/1132/Source/WebCore

Merge 124689 - Disabling eval changes the timing of DidCreateScriptContext
​https://bugs.webkit.org/show_bug.cgi?id=92189

Reviewed by Eric Seidel.

When we implemented Content-Security-Policy, we added the ability to
disable eval in the JavaScript engine. However, when we process the
Content-Security-Policy header, we might not have initialized the
script context for the given frame.

Previously, we would initialize the context, but that generates a
DidCreateScriptContext callback to the embedder earlier in the Document's
lifetime that before. A natural thing to do in this callback is to run
script to customize the script context, but Document isn't fully
initialized yet, which leads to odd bugs and general confusion.

In this patch, we delay actually disabling eval until we would have
created the scripting context previously. From the perspective of the
web platform, this has the same behavior. The only difference is that
now the DidCreateScriptContext notification occurs at the same time
regardless of whether Content-Security-Policy disables eval.

I tried to write a test for this change, but it was unclear to me how
to write a good test. I tried writing a Chromium WebKit unit test to no
avail. The good news is that this patch will be covered by the
PlatformAppBrowserTest.Iframes test in Chromium once
​https://bugs.webkit.org/show_bug.cgi?id=93079 lands. That's not the
best way to test this change, but it might be sufficient.

  • bindings/js/ScriptController.cpp:

(WebCore::ScriptController::initScript):
(WebCore::ScriptController::disableEval):

  • bindings/v8/ScriptController.cpp:

(WebCore::ScriptController::enableEval):
(WebCore::ScriptController::disableEval):

  • bindings/v8/V8DOMWindowShell.cpp:

(WebCore::V8DOMWindowShell::initContextIfNeeded):

TBR=​abarth@webkit.org

10:45 AM Changeset in webkit [127281] by pdr@google.com
  • 2 edits in trunk/LayoutTests

Mark fast/dom/shadow/shadowdom-for-textarea-complex-shadow.html as failing on mac
​https://bugs.webkit.org/show_bug.cgi?id=95588

Unreviewed gardening.

This test is failing on mac due to the sizing of the textbox: in the
expected results there is no scrollbar but in the actual results there
is one. This patch marks this test as failing temporarily.

  • platform/chromium/TestExpectations:
10:37 AM Changeset in webkit [127280] by hyatt@apple.com
  • 7 edits in trunk/Source/WebCore

[New Multicolumn] Refactor flow thread repainting.
​https://bugs.webkit.org/show_bug.cgi?id=95586

Reviewed by Simon Fraser.

Refactor flow thread repainting in regions to delegate the work to the region. This allows subclasses
to customize how the repainting happens so that individual repaints can be issued for columns and pages
inside sets.

repaintFlowThreadContent is the new virtual function that can be subclassed by sets to customize
the repaint process.

repaintFlowThreadContentRectangle is a parameterized helper that will be called for each individual
column and page in a set.

  • WebCore.exp.in:

Change repaintRectangle's signature to be const.

  • rendering/RenderFlowThread.cpp:

(WebCore::RenderFlowThread::repaintRectangleInRegions):
Now delegates the work for each region to RenderRegion::repaintFlowThreadContent.

  • rendering/RenderObject.cpp:

(WebCore::RenderObject::repaintUsingContainer):
(WebCore::RenderObject::repaint):
(WebCore::RenderObject::repaintRectangle):
(WebCore::RenderObject::isRooted):

  • rendering/RenderObject.h:

Change a bunch of functions to be const.

  • rendering/RenderRegion.cpp:

(WebCore::RenderRegion::repaintFlowThreadContent):
(WebCore):
(WebCore::RenderRegion::repaintFlowThreadContentRectangle):

  • rendering/RenderRegion.h:

(RenderRegion):
Add the two new functions (the virtual repaintFlowThreadContent and the helper it
uses, repaintFlowThreadContentRectangle).

10:34 AM Changeset in webkit [127279] by commit-queue@webkit.org
  • 1 edit
    2 deletes in trunk/LayoutTests

Unreviewed, rolling out r127182.
​http://trac.webkit.org/changeset/127182
​https://bugs.webkit.org/show_bug.cgi?id=95587

Reftests shouldn't be rebaselined (Requested by pdr on
#webkit).

Patch by Sheriff Bot <​webkit.review.bot@gmail.com> on 2012-08-31

  • platform/chromium-mac-snowleopard/fast/dom/shadow/shadowdom-for-textarea-complex-shadow-expected.png: Removed.
  • platform/chromium-mac/fast/dom/shadow/shadowdom-for-textarea-complex-shadow-expected.png: Removed.
10:28 AM Changeset in webkit [127278] by yoli@rim.com
  • 3 edits in trunk/Source/WebCore

[BlackBerry] Disable GCController timer for PLATFORM(BLACKBERRY)
​https://bugs.webkit.org/show_bug.cgi?id=93007

Reviewed by Rob Buis.

Disable GCController timer.

  • bindings/js/GCController.cpp:

(WebCore::GCController::GCController):
(WebCore::GCController::garbageCollectSoon):
(WebCore):

  • bindings/js/GCController.h:

(GCController):

10:26 AM Changeset in webkit [127277] by msaboff@apple.com
  • 15 edits
    2 adds in trunk

CSS Parser should directly parse 8 bit source strings
​https://bugs.webkit.org/show_bug.cgi?id=95207

Reviewed by Geoffrey Garen.

Source/WebCore:

Changed CSSParser to handle 8 bit strings directly. Templatized most private parsing methods
and helper methods by character type. Introduced several templatized helper methods for
parsing methods. Some of these helper methods have independent source and destination template
types to handle that an 8 bit source string when parsed produces a 16 bit result. This is due
to unicode escapes. Renamed the lex method "realLex" and made it a templated method. The lex
method that is called by cssyylex(), which is now inlined, invokes the correct instance of realLex
via a member function pointer.

Split out the encoding part of parseEscape() into a new templated UnicodeToChars() method.

Added 8/16 bit switching to CSSParserString as well as added initialization, setters and getters.
Note that CSSParserString must have a default constructor as it is part of the YYLVAL union.
Therefore created init() methods instead of constructors. Many of the parser function now directly
modify the instance in yylval. Made supporting changes in other webcore classes that are used
directly by CSSParser.

Test: fast/css/unicode-escape-in-8bit-string.html

  • css/CSSGrammar.y: Removed cssyylex() and changed manipulation of CSSParserString objects to

use newly written methods instead of direct manipulation of attributes.

  • css/CSSParser.cpp:

Templatized new or existing helper methods
(WebCore::makeLower):
(WebCore::isCSSLetter):
(WebCore::isCSSEscape):
(WebCore::isURILetter):
(WebCore::isIdentifierStartAfterDash):
(WebCore::isEqualToCSSIdentifier):
(WebCore::checkAndSkipEscape):
(WebCore::checkForValidDouble):
(WebCore::parseDouble):
(WebCore::parseColorIntOrPercentage):
(WebCore::isTenthAlpha):
(WebCore::parseAlphaValue):
(WebCore::mightBeRGBA):
(WebCore::mightBeRGB):
(WebCore::fastParseColorInternal):
(WebCore::CSSParser::fastParseColor):
(WebCore::CSSParser::parseImageSet):
(WebCore::skipWhiteSpace):
(WebCore::CSSParser::currentCharacter16):
(WebCore::CSSParser::isIdentifierStart):
(WebCore::CSSParser::checkAndSkipString):
(WebCore::CSSParser::parseUnicodeRange):
(WebCore::CSSParser::parseNthChild):
(WebCore::CSSParser::parseNthChildExtra):
(WebCore::CSSParser::detectFunctionTypeToken):
(WebCore::CSSParser::detectMediaQueryToken):
(WebCore::CSSParser::detectNumberToken):
(WebCore::CSSParser::detectDashToken):
(WebCore::CSSParser::detectAtToken):
(WebCore::CSSParser::setRuleHeaderEnd):
(WebCore::cssPropertyID):
(WebCore::isCSSTokenizerIdentifier):
(WebCore::isCSSTokenizerURL):

(WebCore::CSSParser::CSSParser): Updated constructor for initializing new attributes.

(WebCore::CSSParserString::lower): Calls new makeLower() template helper.

(WebCore::CSSParser::setupParser): Added 8 bit path and set up parser appropriate to
string type.

(WebCore::equal): Updated to use CSSParserString getters and StringImpl::equal
(WebCore::equalIgnoringCase): Updated to use CSSParserString getters and
StringImpl::equalIgnoringCase
(WebCore::parseColorValue): Updated to use new CSSParserString::init()
(WebCore::parseSimpleLength):
(WebCore::parseKeywordValue): Updated to use new CSSParserString::init()
(WebCore::parseTransformArguments):
(WebCore::CSSParser::parseSystemColor): Updated to use new CSSParserString::init()
(WebCore::CSSParser::storeVariableDeclaration): Create right character sized AtomicString.
(WebCore::FontFamilyValueBuilder::add):
(WebCore::TransformOperationInfo::TransformOperationInfo):
(WebCore::CSSParser::parseFontFeatureTag):

(WebCore::CSSParser::parseEscape): Now only parses an escape and returns the resulting
Unicode value as an unsigned.
(WebCore::CSSParser::UnicodeToChars): New method to write out unicode character(s)
based on character type.
(WebCore::CSSParser::parseIdentifierInternal): Templated helper.
(WebCore::CSSParser::parseIdentifier): Parses identifiers, possibly upconverting
the result when a unicode escape is found in an 8 bit source.
(WebCore::CSSParser::parseStringInternal): Templated helper.
(WebCore::CSSParser::parseString): Parses strings, possibly upconverting
the result when a unicode escape is found in an 8 bit source.
(WebCore::CSSParser::parseURIInternal): Templated helper.
(WebCore::CSSParser::parseURI): Templated and split logic out to parseURIInternal.
Reparses a parsed identifier to see if it is a URI. This method doesn't need to up
convert from 8 to 16 bits since that would have been done in parseIdentfier().
(WebCore::CSSParser::realLex): Lot of template related changes. Removed yylval->string
ASSERT checks at the end, since they are too naive to handle upconverted identifier,
string and URIs. They compared source pointers (m_tokenStart) with resulting strings.
(WebCore::fixUnparsedProperties):
(WebCore::CSSParser::fixUnparsedPropertyRanges): Added new template helper with same name.
(WebCore::CSSParser::markRuleHeaderStart):
(WebCore::CSSParser::markRuleHeaderEnd):
(WebCore::CSSParser::markRuleBodyStart):
(WebCore::CSSParser::markRuleBodyEnd):
(WebCore::CSSParser::markPropertyStart):
(WebCore::CSSParser::markPropertyEnd):
(WebCore::cssValueKeywordID):

  • css/CSSParser.h:

(CSSParser):
(WebCore::CSSParser::lex): Made this inline helper that calls correct realLex().
(WebCore::CSSParser::is8BitSource):
(WebCore::CSSParser::tokenStartOffset): Templated method returning offset in source
of the current token start.
(WebCore::CSSParser::tokenStartChar): Templated access to m_tokenStart.
(WebCore::cssyylex): Made this inline to avoid call.

  • css/CSSParserValues.h:

(WebCore::CSSParserString::init): Added three new initialization methods.

New setters, getters and helper methods.
(WebCore::CSSParserString::clear):
(WebCore::CSSParserString::is8Bit):
(WebCore::CSSParserString::characters8):
(WebCore::CSSParserString::characters16):
(WebCore::CSSParserString::length):
(WebCore::CSSParserString::setLength):
(WebCore::CSSParserString::operator[]):
(WebCore::CSSParserString::equalIgnoringCase):
(WebCore::CSSParserString::operator String):
(WebCore::CSSParserString::operator AtomicString):

Other methods templatized to support changes in CSSParser.

  • platform/graphics/Color.cpp:

(WebCore::parseHexColorInternal):
(WebCore::Color::parseHexColor): New LChar flavor
(WebCore::Color::Color):

  • platform/graphics/Color.h:

(Color):

  • platform/text/TextEncoding.cpp:

(WebCore::TextEncoding::TextEncoding):

  • platform/text/TextEncodingRegistry.cpp:

(WebCore::atomicCanonicalTextEncodingName):

  • platform/text/TextEncodingRegistry.h:
  • svg/SVGParserUtilities.cpp:

(WebCore::genericParseNumber):
(WebCore::parseSVGNumber):

  • svg/SVGParserUtilities.h:

(WebCore::isSVGSpace):
(WebCore::skipOptionalSVGSpaces):
(WebCore::skipOptionalSVGSpacesOrDelimiter):

Source/WTF:

  • wtf/text/StringImpl.h:

(WTF::equalIgnoringCase): Added (const char*, const LChar*, unsigned) flavor to support
8 bit CSS parser work.

LayoutTests:

New test that checks the case that a 16 bit unicode escape in an 8 bit CSS
source is properly handled.

  • fast/css/unicode-escape-in-8bit-string-expected.html: Added.
  • fast/css/unicode-escape-in-8bit-string.html: Added.
10:16 AM Changeset in webkit [127276] by timothy@apple.com
  • 2 edits in branches/safari-536.26-branch/Source/WebKit2

Merge r127272 for <rdar://problem/12211180>.

10:08 AM Changeset in webkit [127275] by kbalazs@webkit.org
  • 2 edits in trunk/Source/WebKit2

[Qt] ASSERT(m_useFixedLayout) in WebPage when using QRawWebView without setting fixed layout
​https://bugs.webkit.org/show_bug.cgi?id=95557

Reviewed by Noam Rosenthal.

Let QRawWebView work with non-fixed layout size.

  • UIProcess/API/qt/raw/qrawwebview.cpp:

(QRawWebView::setSize): Also put the member assigment into
the point where the new size is ensured to be committed.

10:07 AM Changeset in webkit [127274] by dominik.rottsches@intel.com
  • 2 edits in trunk/LayoutTests

[Chromium] javascript-url-stop-loaders.html and zero-length-response.html are flaky on Chromium EWS
​https://bugs.webkit.org/show_bug.cgi?id=95581

Unreviewed, Chromium gardening.

Tests mentioned in the title seem to cause flakiness on the Chromium EWS.

  • platform/chromium/TestExpectations: Mark as TIMEOUT PASS for now.
9:54 AM Changeset in webkit [127273] by kbalazs@webkit.org
  • 3 edits in trunk/Source/WebKit2

[Qt] LayerTreeRenderer::callOnMainTread method name is misleading
​https://bugs.webkit.org/show_bug.cgi?id=95558

Reviewed by Kenneth Rohde Christiansen.

Fix the typo in the function name so callers will really call this
and not WTF::callOnMainThread. Renamed to dispatchOnMainThread to
make it clear that these things are not the same.

  • UIProcess/CoordinatedGraphics/LayerTreeRenderer.cpp:

(WebKit::LayerTreeRenderer::dispatchOnMainThread):
(WebKit::LayerTreeRenderer::flushLayerChanges):
(WebKit::LayerTreeRenderer::purgeGLResources):

  • UIProcess/CoordinatedGraphics/LayerTreeRenderer.h:

(LayerTreeRenderer):

9:20 AM Changeset in webkit [127272] by beidson@apple.com
  • 2 edits in trunk/Source/WebKit2

REGRESSION (127047): Crash in PluginProxy::windowVisibilityChanged when changing tabs
<rdar://problem/12211180> and ​https://bugs.webkit.org/show_bug.cgi?id=95515

Reviewed by Dan Bernstein.

Plug-ins created in background tabs don't get initialized until they first become visible,
so telling them they are now invisible leads to messaging on behalf of an uninitialized plug-in.

Many other methods related to visibility and focus only call through if the plug-in is initialized,
so setWindowIsVisible should do the same.

  • WebProcess/Plugins/PluginView.cpp:

(WebKit::PluginView::setWindowIsVisible): Only tell initialized plug-ins of window visibility changes.

9:03 AM Changeset in webkit [127271] by jberlin@webkit.org
  • 2 edits
    1 delete in trunk/LayoutTests

Web Inspector: sort object properties when dumping them in tests
​https://bugs.webkit.org/show_bug.cgi?id=95459

Fix the expected results for timeline-paint to actually be in order.

  • inspector/timeline/timeline-paint-expected.txt:

'height' comes before 'width' alphabetically.

  • platform/chromium/inspector/timeline/timeline-paint-expected.txt: Removed.

Once the cross-platform results were fixed, this file was identical to the cross-platform
file. Remove it.

9:03 AM Changeset in webkit [127270] by hyatt@apple.com
  • 2 edits in trunk/Source/WebCore

[New Multicolumn] Make RenderFlowThreads into selection roots.
​https://bugs.webkit.org/show_bug.cgi?id=95580

Reviewed by Dan Bernstein.

Make RenderFlowThreads selection roots so that they paint their own selection gaps.
This fixes multicolumn selection rendering so that the gaps paint correctly and stop
at the edge of the flow thread instead of going out to the edges of enclosing blocks.

  • rendering/RenderBlock.cpp:

(WebCore::RenderBlock::isSelectionRoot):

8:50 AM Changeset in webkit [127269] by commit-queue@webkit.org
  • 8 edits
    2 adds in trunk

[CSS Regions] It should be possible to specify ::BEFORE/AFTER as regions
​https://bugs.webkit.org/show_bug.cgi?id=90079

Patch by Andrei Bucur <​abucur@adobe.com> on 2012-08-31
Reviewed by David Hyatt.

Source/WebCore:

The calls for RenderRegion::node() have been replaced with RenderRegion::generatingNode() to allow pseudo-elements to have RenderRegion renderers.
The compare function that establishes the region order was adapted to take into account pseudo-regions.
A renderer for the pseudo-element is now generated if it's a region even though there's no content property specified.

Test: fast/regions/content-flowed-into-pseudo-regions.html

  • rendering/RenderFlowThread.cpp:

(WebCore::RenderFlowThread::computeOverflowStateForRegions):

  • rendering/RenderNamedFlowThread.cpp:

(WebCore::compareRenderRegions):

  • rendering/RenderObjectChildList.cpp:

(WebCore::ensureBeforeAfterContainer):
(WebCore):
(WebCore::RenderObjectChildList::updateBeforeAfterContent):

  • rendering/RenderRegion.cpp:

(WebCore::RenderRegion::nodeAtPoint):
(WebCore::RenderRegion::checkRegionStyle):
(WebCore::RenderRegion::computeStyleInRegion):

  • rendering/RenderTreeAsText.cpp:

(WebCore::writeRenderNamedFlowThreads):

LayoutTests:

The new ref-test has a block with both a :before and :after pseudo-elements marked as regions.
In the block there's another region.

The test verifies the content is correctly flowed (the region order is respected) in the region chain:
:before region -> child region -> :after region

As a region a pseudo-element should also be displayed if there's no content property specified in the style sheet.

  • fast/regions/content-flowed-into-pseudo-regions-expected.html: Added.
  • fast/regions/content-flowed-into-pseudo-regions.html: Added.
  • fast/regions/resources/region-style.css:

(.regionArticleAfter::after, .regionArticleBefore::before):
(.borderBoxNoFloat):

8:49 AM Changeset in webkit [127268] by kbalazs@webkit.org
  • 2 edits in trunk/Tools

[Qt] ImageDiff output is not in the expected form if image dimensions differ
​https://bugs.webkit.org/show_bug.cgi?id=95575

Reviewed by Csaba Osztrogonác.

Output 100% difference in the case of differing image dimensions so
the test harness knows that we won't produce any output further.
This matches the behavior of other ports.

  • DumpRenderTree/qt/ImageDiff.cpp:

(main):

8:47 AM Changeset in webkit [127267] by hyatt@apple.com
  • 14 edits in trunk/Source/WebCore

[New Multicolumn] Implement unforced breaking in the new column layout.
​https://bugs.webkit.org/show_bug.cgi?id=95498

Reviewed by Dan Bernstein.

This patch implements unforced breaking for lines and objects inside columns. Forced breaks are
not yet handled properly and will be covered in a separate patch.

  • rendering/LayoutState.cpp:

(WebCore::LayoutState::LayoutState):
Eliminate the flow thread constructor for LayoutState. Flow threads were doing a double push of
layout states, once in RenderFlowThread::layout() and then again in RenderBlock::layout(). This
double push was unnecessary, so I refactored the code to ensure the RenderBlock push does the
right thing for RenderFlowThreads.

(WebCore):

  • rendering/LayoutState.h:

(LayoutState):
Eliminate the flow thread constructor for LayoutState.

  • rendering/RenderBlock.cpp:

(WebCore::RenderBlock::checkForPaginationLogicalHeightChange):
This function now handles RenderFlowThreads and properly sets a non-zero page height as well
as querying the flow thread to ask if the pagination logical height has changed. The flow thread
has a new member that it sets in layout() so that it can answer this question.

(WebCore::RenderBlock::hasNextPage):
Make sure RenderRegionSets always claim to have a next page when they are the last region, since
we know columns/pages can always generate additional boxes as needed.

  • rendering/RenderFlowThread.cpp:

(WebCore::RenderFlowThread::RenderFlowThread):
Add the m_pageLogicalHeightChanged member variable for tracking whether regions changed dimensions,
thus forcing us to repaginate everything.

(WebCore::RenderFlowThread::layout):
Remove the redundant push of layout state now that RenderBlock's covers it.

(WebCore::RenderFlowThread::regionAtBlockOffset):
Always return the last region if it's a set, regardless of the extendLastRegion boolean.
This reflects the fact that the last region set can always generate new page/column boxes.

(WebCore::RenderFlowThread::pageLogicalTopForOffset):
Refactored to call into the region it finds for a given offset, so that the region itself
can answer the question of where the top of the page is. This enables sets to return the top of
specific columns/pages within the set.

(WebCore::RenderFlowThread::pageRemainingLogicalHeightForOffset):
Changed to be written in terms of pageLogicalTop and pageLogicalHeight so that it correctly uses
only the current column/page when determining remaining height.

  • rendering/RenderFlowThread.h:

Added pageLogicalHeightChanged() so that RenderBlock can ask the flow thread if its regions changed
dimensions.

  • rendering/RenderMultiColumnBlock.cpp:

(WebCore::RenderMultiColumnBlock::checkForPaginationLogicalHeightChange):
Tweaked to no longer turn on pagination, since the RenderMultiColumnFlowThread does that and we
want to avoid a double push of paginated states. Instead it now just sets the column height ahead
of time so that the RenderMultiColumnSets know how big their columns are.

  • rendering/RenderMultiColumnSet.cpp:

(WebCore::RenderMultiColumnSet::pageLogicalTopForOffset):
Subclassed to return the top of a specific column.

(WebCore):

  • rendering/RenderMultiColumnSet.h:

Add the subclassed method for pageLogicalTopForOffset.

  • rendering/RenderRegion.cpp:

(WebCore::RenderRegion::pageLogicalTopForOffset):
The code from RenderFlow::pageLogicalTopForOffset moved here. Just returns the top of the flow
thread portion rect in the region (same as before).

(WebCore::RenderRegion::layout):
Changed to check only the page logical width and height, since we only consider a region as
invalidating how you break if there is an actual page height change.

  • rendering/RenderRegion.h:

(WebCore::RenderRegion::isRenderRegionSet):
Added an isRenderRegionSet() virtual function so that we can test for sets.

  • rendering/RenderRegionSet.h:

(WebCore::RenderRegionSet::isRenderRegionSet):
Subclassed isRenderRegionSet() to return true.

  • rendering/RenderView.cpp:
  • rendering/RenderView.h:

(LayoutStateMaintainer):
(WebCore::LayoutStateMaintainer::push):
Remove the pushes of LayoutState that were specific to RenderFlowThreads.

8:40 AM Changeset in webkit [127266] by pfeldman@chromium.org
  • 7 edits in trunk

Web Inspector: Update $ to alias to querySelector rather than getElementById
​https://bugs.webkit.org/show_bug.cgi?id=92648

Reviewed by Yury Semikhatsky.

Source/WebCore:

Completed implementation started by Addy Osmani. Updates $ to alias to querySelector
with more flexible handling of selectors. Also adds a warning about the change (as per Firebug).

  • inspector/InjectedScriptSource.js:

(.):

LayoutTests:

  • inspector/console/command-line-api-expected.txt:
  • inspector/console/command-line-api-inspect.html:
  • inspector/console/command-line-api.html:
8:36 AM Changeset in webkit [127265] by Csaba Osztrogonác
  • 1 edit
    2 adds in trunk/LayoutTests

[Qt][WK2] Unreviewed gardening, add Qt-WK2 specific expected
file, because this doesn't fail as expected, but passes.

  • platform/qt-5.0-wk2/css3/compositing/should-have-compositing-layer-expected.txt: Added.
8:04 AM Changeset in webkit [127264] by commit-queue@webkit.org
  • 2 edits in trunk/Tools

Unreviewed, rolling out r127261.
​http://trac.webkit.org/changeset/127261
​https://bugs.webkit.org/show_bug.cgi?id=95572

This patch depends on other build fixes (see blockers of bug
95466) which should land first (Requested by milian on
#webkit).

Patch by Sheriff Bot <​webkit.review.bot@gmail.com> on 2012-08-31

  • qmake/mkspecs/features/features.prf:
7:59 AM Changeset in webkit [127263] by gavinp@chromium.org
  • 2 edits in trunk/Source/WebKit/chromium

[chromium] Disable WebViewTest.AutoResizeMinimumSize
​https://bugs.webkit.org/show_bug.cgi?id=95571

Unreviewed gardening.

Disabling this test since it started failing in build 17225, see ​http://build.chromium.org/p/chromium.webkit/builders/Webkit%20Win/builds/17247 for instance. Will shortly file bug to re-enable this.

  • tests/WebViewTest.cpp:
7:54 AM Changeset in webkit [127262] by caseq@chromium.org
  • 2 edits in trunk/Source/WebCore

Web Inspector: disable popover on click outside of popover anchor
​https://bugs.webkit.org/show_bug.cgi?id=95564

Reviewed by Yury Semikhatsky.

Immediately hide popover when receiving a click outside of popover anchor.

  • inspector/front-end/Popover.js:

(WebInspector.PopoverHelper.prototype._mouseDown):

7:47 AM Changeset in webkit [127261] by commit-queue@webkit.org
  • 2 edits in trunk/Tools

[Qt] Support WebKit2/QML on QNX
​https://bugs.webkit.org/show_bug.cgi?id=95466

Patch by Milian Wolff <​milian.wolff@kdab.com> on 2012-08-31
Reviewed by Noam Rosenthal.

Enable USE(3D_GRAPHICS) on QNX again, paired with the other bug fixes
this should finally work now.

  • qmake/mkspecs/features/features.prf:
7:38 AM Changeset in webkit [127260] by timothy@apple.com
  • 1 edit in branches/safari-536.26-branch/Tools/DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj

Fix build from another bad project merge.

7:35 AM Changeset in webkit [127259] by gavinp@chromium.org
  • 2 edits
    2 adds in trunk/LayoutTests

Unreviewed rebaseline after r127226

The windows results needed this rebaseline, apparently the original commit omitted these windows diffs per ​yosin@chromium.org.

  • platform/chromium-win-xp/fast/forms/time-multiple-fields/time-multiple-fields-appearance-basic-expected.png: Added.
  • platform/chromium-win/fast/forms/time-multiple-fields/time-multiple-fields-appearance-basic-expected.png:
7:35 AM Changeset in webkit [127258] by commit-queue@webkit.org
  • 7 edits
    2 adds in trunk/Source

[EFL] Add proper support for navigator.onLine and associated events
​https://bugs.webkit.org/show_bug.cgi?id=95532

Patch by Christophe Dumez <Christophe Dumez> on 2012-08-31
Reviewed by Kenneth Rohde Christiansen.

Source/WebCore:

Add support for detecting online state in EFL port and
monitor state changes. The EFL port now has proper
support for navigator.onLine and associated events.

The implementation is based on Eeze library for
detecting network interfaces and their configuration.
Detection of online state changes relies on rtnetlink.

No new tests, no behavior change for layout tests.

  • PlatformEfl.cmake: Link unconditionally against Eeze

library now.

  • platform/network/NetworkStateNotifier.cpp:

(WebCore):

  • platform/network/NetworkStateNotifier.h:

(NetworkStateNotifier):
(WebCore):

  • platform/network/efl/NetworkStateNotifierEfl.cpp: Added.

(WebCore):
(WebCore::NetworkStateNotifier::updateState):
(WebCore::NetworkStateNotifier::networkInterfaceChanged):
(WebCore::readSocketCallback):
(WebCore::NetworkStateNotifier::~NetworkStateNotifier):
(WebCore::NetworkStateNotifier::NetworkStateNotifier):

Source/WebKit/efl:

Remove ewk_network_state_notifier_online_set() API since
the EFL port no longer relies on the client application to
provide it with the online state.

  • ewk/ewk_network.cpp:
  • ewk/ewk_network.h:
7:29 AM Changeset in webkit [127257] by Csaba Osztrogonác
  • 2 edits in trunk/LayoutTests

[Qt][WK2] REGRESSION(r127091): It made layout tests extremely slow
​https://bugs.webkit.org/show_bug.cgi?id=95450

Unreviewed gardening, skip http/tests/appcache tests until proper fix.

  • platform/qt-5.0-wk2/Skipped:
7:13 AM Changeset in webkit [127256] by commit-queue@webkit.org
  • 4 edits in trunk/LayoutTests

Remove http/tests/loading/gmail-assert-on-load.html test case from Skipped.
​https://bugs.webkit.org/show_bug.cgi?id=95562

Unreviewed gardening.

Removed http/tests/loading/gmail-assert-on-load.html test case from Skipped files since it has been moved out from http/loading to http/misc, for details see bug 88923.

Patch by Kangil Han <​kangil.han@samsung.com> on 2012-08-31

  • platform/win/Skipped:
  • platform/wincairo/Skipped:
  • platform/wk2/Skipped:
7:04 AM Changeset in webkit [127255] by zandobersek@gmail.com
  • 2 edits in trunk/LayoutTests

Unreviewed GTK gardening.

Adding failure expectations for two HTTP security tests.

  • platform/gtk/TestExpectations:
6:39 AM Changeset in webkit [127254] by timothy@apple.com
  • 1 edit in branches/safari-536.26-branch/Tools/DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj

Fix build from bad merge.

6:34 AM LineBreakingCSS3Mapping edited by glenn@skynav.com
Fix typos. (diff)
6:33 AM Changeset in webkit [127253] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit2

[Qt] QNX does not support SOCK_SEQPACKET socket type for socketpair in ProcessLauncherQt
​https://bugs.webkit.org/show_bug.cgi?id=95553

Patch by Milian Wolff <​milian.wolff@kdab.com> on 2012-08-31
Reviewed by Simon Hausmann.

Calling socketpair with SOCK_SEQPACKET socket type fails on QNX with an error 241 (EPROTOTYPE),
as can be seen by the following debug output: "Creation of socket failed with errno: 241".
A test app shows that SOCK_DGRAM works properly on QNX, thus use it instead of SOCK_SEQPACKET.

  • UIProcess/Launcher/qt/ProcessLauncherQt.cpp:
6:33 AM LineBreakingCSS3Mapping edited by glenn@skynav.com
Minor fix. (diff)
6:31 AM LineBreakingCSS3Mapping edited by glenn@skynav.com
Fix typo. (diff)
6:28 AM LineBreaking edited by glenn@skynav.com
Fix typo. (diff)
6:21 AM WikiStart edited by glenn@skynav.com
Add link to new CSS3 Text (webkit) wiki page. (diff)
6:20 AM CSS3Text created by glenn@skynav.com
Initial content.
6:17 AM Changeset in webkit [127252] by timothy@apple.com
  • 9 edits
    3 adds in branches/safari-536.26-branch

Merge r127047 for <rdar://problem/12133021>.

6:14 AM LineBreaking edited by glenn@skynav.com
Elaborate content. (diff)
5:58 AM Changeset in webkit [127251] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

[Qt] QNX build fails due to missing OpenGL ES defines
​https://bugs.webkit.org/show_bug.cgi?id=95554

Patch by Milian Wolff <​milian.wolff@kdab.com> on 2012-08-31
Reviewed by Simon Hausmann.

Apply the fix of bug 91030 to all platforms running on QNX.

  • platform/graphics/opengl/Extensions3DOpenGLES.h:
5:53 AM LineBreakingCSS3Mapping edited by glenn@skynav.com
Minor fixes. (diff)
5:52 AM Changeset in webkit [127250] by Csaba Osztrogonác
  • 2 edits in trunk/LayoutTests

[Qt] Unreviewed gardening, skip a new failing test.

  • platform/qt/Skipped:
5:50 AM LineBreakingCSS3Mapping edited by glenn@skynav.com
Elaborate content. (diff)
5:46 AM Changeset in webkit [127249] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebCore

[CMake] Changes on the bindings generator does not trigger generation of the bindings
​https://bugs.webkit.org/show_bug.cgi?id=95552

Patch by Thiago Marcos P. Santos <​thiago.santos@intel.com> on 2012-08-31
Reviewed by Kenneth Rohde Christiansen.

Added the engine especific part of the generator to the dependency
list.

  • UseJSC.cmake:
  • UseV8.cmake:
5:30 AM LineBreakingCSS3Mapping edited by glenn@skynav.com
Add links. (diff)
5:28 AM LineBreakingCSS3Mapping edited by glenn@skynav.com
Fix typo. (diff)
5:28 AM Changeset in webkit [127248] by Patrick Gansterer
  • 2 edits in trunk/Source/JavaScriptCore

Build fix for WinCE after r127191.

  • bytecode/JumpTable.h:
5:27 AM LineBreakingCSS3Mapping edited by glenn@skynav.com
Elaborate content. (diff)
5:15 AM LineBreakingCSS3Mapping edited by glenn@skynav.com
Elaborate content. (diff)
5:07 AM Changeset in webkit [127247] by dominik.rottsches@intel.com
  • 2 edits
    1 add in trunk/LayoutTests

[EFL] Gardening of failing tests and new passes
​https://bugs.webkit.org/show_bug.cgi?id=95555

Unreviewed EFL gardening of failing tests and new passes.

Patch by Sudarsana Nagineni <​sudarsana.nagineni@linux.intel.com> on 2012-08-31

  • platform/efl/TestExpectations:
  • platform/efl/http/tests/misc/willCacheResponse-delegate-callback-expected.txt: Added.
5:05 AM Changeset in webkit [127246] by Csaba Osztrogonác
  • 2 edits in trunk/LayoutTests

[Qt] Gardening after r127140. Skip a failing reftest.

Patch by János Badics <János Badics> on 2012-08-31
Reviewed by Csaba Osztrogonác.

  • platform/qt/Skipped:
4:59 AM Changeset in webkit [127245] by Csaba Osztrogonác
  • 27 edits
    14 adds
    4 deletes in trunk/LayoutTests

[Qt] Unreviewed gardening, unskip now passing tests, update expected files.

  • platform/qt-5.0/css2.1/t0805-c5519-brdr-r-01-e-expected.png: Removed.
  • platform/qt-5.0/css2.1/t0805-c5519-brdr-r-01-e-expected.txt: Removed.
  • platform/qt-5.0/fast/inline/absolute-positioned-inline-in-centred-block-expected.png: Removed.
  • platform/qt-5.0/fast/inline/absolute-positioned-inline-in-centred-block-expected.txt: Removed.
  • platform/qt/Skipped:
  • platform/qt/TestExpectations:
  • platform/qt/css2.1/t0805-c5519-brdr-r-00-a-expected.txt:
  • platform/qt/css2.1/t0805-c5519-brdr-r-01-e-expected.png:
  • platform/qt/css2.1/t0805-c5519-brdr-r-01-e-expected.txt:
  • platform/qt/css2.1/t0805-c5520-brdr-b-01-e-expected.txt:
  • platform/qt/css2.1/t0805-c5521-brdr-l-00-a-expected.txt:
  • platform/qt/css2.1/t0805-c5521-brdr-l-01-e-expected.txt:
  • platform/qt/css2.1/t0805-c5521-ibrdr-l-00-a-expected.txt:
  • platform/qt/css3/filters/blur-filter-page-scroll-expected.png: Added.
  • platform/qt/css3/filters/blur-filter-page-scroll-expected.txt: Added.
  • platform/qt/css3/filters/blur-filter-page-scroll-parents-expected.png: Added.
  • platform/qt/css3/filters/blur-filter-page-scroll-parents-expected.txt: Added.
  • platform/qt/css3/filters/effect-reference-expected.png: Added.
  • platform/qt/css3/filters/effect-reference-expected.txt: Added.
  • platform/qt/css3/filters/effect-reference-external-expected.png: Added.
  • platform/qt/css3/filters/effect-reference-external-expected.txt: Added.
  • platform/qt/css3/filters/effect-reference-hw-expected.txt: Added.
  • platform/qt/css3/filters/effect-reference-ordering-expected.png: Added.
  • platform/qt/css3/filters/effect-reference-ordering-expected.txt: Added.
  • platform/qt/fast/block/float/024-expected.png:
  • platform/qt/fast/block/float/024-expected.txt:
  • platform/qt/fast/block/margin-collapse/025-expected.png:
  • platform/qt/fast/block/margin-collapse/025-expected.txt:
  • platform/qt/fast/block/margin-collapse/block-inside-inline/025-expected.png:
  • platform/qt/fast/block/margin-collapse/block-inside-inline/025-expected.txt:
  • platform/qt/fast/block/margin-collapse/empty-clear-blocks-expected.png:
  • platform/qt/fast/block/margin-collapse/empty-clear-blocks-expected.txt:
  • platform/qt/fast/css/absolute-child-with-percent-height-inside-relative-parent-expected.png:
  • platform/qt/fast/css/absolute-child-with-percent-height-inside-relative-parent-expected.txt:
  • platform/qt/fast/css/nested-layers-with-hover-expected.txt:
  • platform/qt/fast/inline/absolute-positioned-inline-in-centred-block-expected.png:
  • platform/qt/fast/inline/absolute-positioned-inline-in-centred-block-expected.txt:
  • platform/qt/fast/inline/left-right-center-inline-alignment-in-ltr-and-rtl-blocks-expected.png:
  • platform/qt/fast/inline/left-right-center-inline-alignment-in-ltr-and-rtl-blocks-expected.txt:
  • platform/qt/svg/filters/color-interpolation-filters-expected.png: Added.
  • platform/qt/tables/mozilla_expected_failures/marvin/table_overflow_hidden_tbody-expected.png: Added.
  • platform/qt/tables/mozilla_expected_failures/marvin/table_overflow_hidden_tbody-expected.txt:
  • platform/qt/tables/mozilla_expected_failures/marvin/table_overflow_hidden_tr-expected.png: Added.
  • platform/qt/tables/mozilla_expected_failures/marvin/table_overflow_hidden_tr-expected.txt:
4:35 AM Changeset in webkit [127244] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit/efl

[EFL] Simplify FrameLoaderClinetEfl by adding a private method.
​https://bugs.webkit.org/show_bug.cgi?id=95444

Patch by Jinwoo Song <​jinwoo7.song@samsung.com> on 2012-08-31
Reviewed by Gyuyoung Kim.

In the FrameLoaderClientEfl, the codes which checks that
if current frame is main frame are frequently used.
This patch adds a private method and replaces those codes
with the method to simplify.

  • WebCoreSupport/FrameLoaderClientEfl.cpp:

(WebCore::FrameLoaderClientEfl::dispatchDidChangeLocationWithinPage):
(WebCore::FrameLoaderClientEfl::dispatchDidReceiveIcon):
(WebCore::FrameLoaderClientEfl::dispatchDidStartProvisionalLoad):
(WebCore::FrameLoaderClientEfl::dispatchDidReceiveTitle):
(WebCore::FrameLoaderClientEfl::dispatchDidCommitLoad):
(WebCore::FrameLoaderClientEfl::dispatchDidFailProvisionalLoad):
(WebCore::FrameLoaderClientEfl::transitionToCommittedForNewPage):

  • WebCoreSupport/FrameLoaderClientEfl.h:

(WebCore::FrameLoaderClientEfl::isLoadingMainFrame):

3:41 AM LineBreakingCSS3Mapping created by glenn@skynav.com
Initial content.
3:33 AM LineBreaking edited by glenn@skynav.com
Elaborate content. (diff)
3:17 AM LineBreaking edited by glenn@skynav.com
Elaborate content (diff)
3:01 AM EFLWebKitCodingStyle edited by gyuyoung.kim@samsung.com
(diff)
2:56 AM Changeset in webkit [127243] by commit-queue@webkit.org
  • 5 edits in trunk/Source/WebCore

Web Inspector: [WebGL] Link WebGL calls to the source code where it was made
​https://bugs.webkit.org/show_bug.cgi?id=95426

Patch by Andrey Adaikin <​aandrey@chromium.org> on 2012-08-31
Reviewed by Vsevolod Vlasov.

Save a stack trace for every call we capture and report to the frontend sourceURL,lineNumber,columnNumber were the call was made from.
Right now, this will work only with V8.

  • inspector/InjectedScriptExterns.js:

(CallSite.prototype.getFileName):
(CallSite.prototype.getLineNumber):
(CallSite.prototype.getColumnNumber):

  • inspector/InjectedScriptWebGLModuleSource.js:

(.):

  • inspector/Inspector.json:
  • inspector/front-end/WebGLProfileView.js:

(WebInspector.WebGLProfileView):
(WebInspector.WebGLProfileView.prototype._showTraceLog):

2:43 AM Changeset in webkit [127242] by pfeldman@chromium.org
  • 6 edits
    1 add in trunk/Source/WebCore

Web Inspector: Add high-DPI images for status bar glyphs
​https://bugs.webkit.org/show_bug.cgi?id=95471

Patch by Patrick Dubroy <​dubroy@chromium.org> on 2012-08-31
Reviewed by Pavel Feldman.

Add the new 2x resolution image for the status bar glyphs, and add
media queries to use the images on high-DPI devices.

  • inspector/front-end/Images/statusbarButtonGlyphs2x.png: Added.
  • inspector/front-end/WebKit.qrc:
  • inspector/front-end/inspector.css:

(.long-click-glyph):
(@media (-webkit-min-device-pixel-ratio: 1.5)):
(.status-bar-item > .glyph):

  • inspector/front-end/scriptsPanel.css:

(.toggle-breakpoints .glyph):
(@media (-webkit-min-device-pixel-ratio: 1.5)):
(.status-bar-item.scripts-navigator-show-hide-button > .glyph):
(.status-bar-item.scripts-debugger-show-hide-button > .glyph):

  • WebCore.gypi
  • inspector/front-end/timelinePanel.css:

(#timeline-overview-sidebar .icon):
(@media (-webkit-min-device-pixel-ratio: 1.5)):

2:42 AM LineBreaking edited by glenn@skynav.com
Elaborate content. (diff)
2:40 AM Changeset in webkit [127241] by commit-queue@webkit.org
  • 14 edits in trunk/LayoutTests

[CSS Exclusions] Exclusions tests should run without errors in standalone browser
​https://bugs.webkit.org/show_bug.cgi?id=95369

Patch by Bear Travis <​betravis@adobe.com> on 2012-08-31
Reviewed by Andreas Kling.

Adding a check to make sure window.internals exists before accessing it.
Without the check, a reference error would occur when trying to run these
tests from the browser (as opposed to the test script).

  • fast/exclusions/css-exclusions-disabled.html:
  • fast/exclusions/script-tests/parsing-wrap-shape-inside.js:
  • fast/exclusions/script-tests/parsing-wrap-shape-outside.js:
  • fast/exclusions/script-tests/wrap-flow-parsing.js:
  • fast/exclusions/script-tests/wrap-margin-parsing.js:
  • fast/exclusions/script-tests/wrap-padding-parsing.js:
  • fast/exclusions/script-tests/wrap-parsing.js:
  • fast/exclusions/script-tests/wrap-through-parsing.js:
  • fast/exclusions/shape-inside/shape-inside-floats-simple.html:
  • fast/exclusions/shape-inside/shape-inside-inline-elements.html:
  • fast/exclusions/shape-inside/shape-inside-overflow.html:
  • fast/exclusions/shape-inside/shape-inside-percentage-auto.html:
  • fast/exclusions/shape-inside/shape-inside-text.html:
2:35 AM Changeset in webkit [127240] by pfeldman@chromium.org
  • 9 edits in trunk/Source/WebCore

Web Inspector: render box model elements and grid on inspector overlay
​https://bugs.webkit.org/show_bug.cgi?id=95456

Reviewed by Vsevolod Vlasov.

This change migrates native rendering for box model to the InspectorOverlay.

  • inspector/InspectorInstrumentation.cpp:

(WebCore):
(WebCore::InspectorInstrumentation::didScrollImpl):

  • inspector/InspectorInstrumentation.h:

(InspectorInstrumentation):
(WebCore::InspectorInstrumentation::didScroll):
(WebCore):

  • inspector/InspectorOverlay.cpp:

(WebCore::InspectorOverlay::drawNodeHighlight):
(WebCore::InspectorOverlay::drawRectHighlight):
(WebCore::InspectorOverlay::overlayPage):
(WebCore::InspectorOverlay::evaluateInOverlay):

  • inspector/InspectorOverlay.h:

(WebCore::Highlight::setColors):
(Highlight):
(InspectorOverlay):

  • inspector/InspectorOverlayPage.html:
  • inspector/InspectorPageAgent.cpp:

(WebCore::InspectorPageAgent::didLayout):
(WebCore):
(WebCore::InspectorPageAgent::didScroll):

  • inspector/InspectorPageAgent.h:
  • page/Chrome.cpp:

(WebCore::Chrome::scroll):

2:23 AM LineBreaking edited by glenn@skynav.com
Add content. (diff)
2:14 AM Changeset in webkit [127239] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

Web Inspector: [WebGL] Make wrap functions initialization lazy
​https://bugs.webkit.org/show_bug.cgi?id=95533

Patch by Andrey Adaikin <​aandrey@chromium.org> on 2012-08-31
Reviewed by Vsevolod Vlasov.

WebGL wrap functions initialization is not always needed to be executed: for example, when we will instrument a 2D canvas.
Drive-by: fixed variable namings, moved duplicated code into a small auxiliary function.

  • inspector/InjectedScriptWebGLModuleSource.js:

(.):

1:34 AM LineBreaking created by glenn@skynav.com
Outline new page describing line breaking.
1:07 AM Changeset in webkit [127238] by Simon Hausmann
  • 8 edits in trunk/Source

[Qt] Port convertValueToQVariant to use the JSC C API
​https://bugs.webkit.org/show_bug.cgi?id=94695

Reviewed by Kenneth Rohde Christiansen.

Source/WebCore:

Most of the code is portable to the C API, with a few exceptions for internal
types such as byte arrays, dates or when unboxing primitive types.

  • bridge/qt/qt_instance.cpp:

(JSC::Bindings::QtField::setValueToInstance):

  • bridge/qt/qt_runtime.cpp:

(JSC::Bindings::operator<<):
(JSC::Bindings::isJSUint8ClampedArray):
(JSC::Bindings::isJSArray):
(Bindings):
(JSC::Bindings::isJSDate):
(JSC::Bindings::isQtObject):
(JSC::Bindings::valueRealType):
(JSC::Bindings::toString):
(JSC::Bindings::unwrapBoxedPrimitive):
(JSC::Bindings::convertValueToQVariantMap):
(JSC::Bindings::convertToList):
(JSC::Bindings::toQString):
(JSC::Bindings::getGregorianDateTimeUTC):
(JSC::Bindings::toQDateTimeUTC):
(JSC::Bindings::convertValueToQVariant):
(JSC::Bindings::findMethodIndex):
(JSC::Bindings::QtConnectionObject::execute):

  • bridge/qt/qt_runtime.h:

(Bindings):

Source/WebKit/qt:

Minor adaptations are needed here to accomodate the changed signature of
convertValueToQVariant.

  • Api/qwebelement.cpp:

(QWebElement::evaluateJavaScript):
(convertJSValueToWebElementVariant):

  • Api/qwebframe.cpp:

(QWebFrame::evaluateJavaScript):

  • WebCoreSupport/DumpRenderTreeSupportQt.cpp:

(convertJSValueToNodeVariant):

12:58 AM Changeset in webkit [127237] by pfeldman@chromium.org
  • 4 edits in trunk/Source/WebCore

Web Inspector: render path to file and line number as subtitles in selector dialog.
​https://bugs.webkit.org/show_bug.cgi?id=95481

Reviewed by Vsevolod Vlasov.

In case several files have matching names, we should render the path next to it.
Using the same subtitle field, we now also render line number for JS and Style outlines.

  • inspector/front-end/FilteredItemSelectionDialog.js:

(WebInspector.FilteredItemSelectionDialog.prototype.get _itemsLoaded):
(WebInspector.FilteredItemSelectionDialog.prototype._createItemElement):
(WebInspector.FilteredItemSelectionDialog.prototype._filterItems):
(WebInspector.SelectionDialogContentProvider.prototype.itemSuffixAt):
(WebInspector.JavaScriptOutlineDialog.prototype.itemSuffixAt):
(WebInspector.JavaScriptOutlineDialog.prototype.itemSubtitleAt):
(WebInspector.OpenResourceDialog.prototype.itemSuffixAt):
(WebInspector.OpenResourceDialog.prototype.itemSubtitleAt):

  • inspector/front-end/StyleSheetOutlineDialog.js:

(WebInspector.StyleSheetOutlineDialog.prototype.itemSuffixAt):
(WebInspector.StyleSheetOutlineDialog.prototype.itemSubtitleAt):

  • inspector/front-end/filteredItemSelectionDialog.css:

(.js-outline-dialog > .container > div.item):
(.js-outline-dialog span.subtitle):

12:36 AM Changeset in webkit [127236] by commit-queue@webkit.org
  • 2 edits
    2 moves
    1 add in trunk/LayoutTests

[Qt] New inspector/profiler/webgl-profiler-api-changes.html times out
​https://bugs.webkit.org/show_bug.cgi?id=95183

Patch by Andrey Adaikin <​aandrey@chromium.org> on 2012-08-31
Reviewed by Yury Semikhatsky.

Move inspector WebGL test(s) to its own folder so that they are easy to blacklist. Skip them for QT where WebGL is disabled at the moment.
Drive-by: make the test fail instead of timing out when WebGL is disabled.

  • inspector/profiler/webgl/webgl-profiler-api-changes-expected.txt: Renamed from LayoutTests/inspector/profiler/webgl-profiler-api-changes-expected.txt.
  • inspector/profiler/webgl/webgl-profiler-api-changes.html: Renamed from LayoutTests/inspector/profiler/webgl-profiler-api-changes.html.
  • platform/qt/Skipped:
12:09 AM Changeset in webkit [127235] by yosin@chromium.org
  • 6 edits in trunk

[Forms] Left/Right keys in multiple fields time input UI should move focus physical left/right instead of logical left/right regardless text direction.
​https://bugs.webkit.org/show_bug.cgi?id=95434

Reviewed by Kent Tamura.

Source/WebCore:

This patch also changes handling of Left/Right keys to aware text
direction ("dir" attribute"). Left/Right keys should move focus
physical right/left instead of logical right/left.

This patch affects ports which enable both ENABLE_INPUT_TYPE_TIME and
ENABLE_INPUT_TYPE_TIME_MULTIPLE_FIELDS.

No new tests. This patch adds a test case into fast/forms/time-multiple-fields/time-multiple-fields-keyboard-events.html

  • html/shadow/DateTimeFieldElement.cpp:

(WebCore::DateTimeFieldElement::defaultKeyboardEventHandler): Changed to check isRTL() for Left/Right keys then we call
appropriate focus motion function.
(WebCore::DateTimeFieldElement::isRTL): Added for checking text direction is right-to-left or not.

  • html/shadow/DateTimeFieldElement.h:

(DateTimeFieldElement): Added declaration of isRTL().

LayoutTests:

This patch adds a new test case of Left/Right keys navigation on
right-to-left text direction for multiple field time input UI.

This patch affects ports which enable both ENABLE_INPUT_TYPE_TIME and
ENABLE_INPUT_TYPE_TIME_MULTIPLE_FIELDS.

  • fast/forms/time-multiple-fields/time-multiple-fields-keyboard-events.html: Changed to have Left/Right keys navigation in

right-to-left text direction.

Note: See TracTimeline for information about the timeline view.