⚠ Archived content — this site is no longer maintained.   Current WebKit documentation is at docs.webkit.org.

Timeline



Mar 1, 2014:

9:35 PM Changeset in webkit [164938] by timothy_horton@apple.com
  • 2 edits in trunk/Source/WebKit2

Fix the Mac build.

  • UIProcess/mac/ViewGestureController.messages.in:
9:25 PM Changeset in webkit [164937] by timothy_horton@apple.com
  • 20 edits
    3 adds in trunk/Source

WebKit2 View Gestures (Smart Magnification): Support for iOS
​https://bugs.webkit.org/show_bug.cgi?id=129146
<rdar://problem/16032668>

Reviewed by Benjamin Poulain.

Add support for double-tap smart magnification on iOS.

  • DerivedSources.make:
  • UIProcess/API/Cocoa/WKWebView.mm:

(-[WKWebView initWithFrame:configuration:]):
WKContentView is no longer API, so it can know about WKWebView
and will always have one.

(-[WKWebView contentView:didCommitLayerTree:WebKit::]):
Don't replace the zoom scale if we're currently animating.

(-[WKWebView _takeViewSnapshot]):
Rename.

(-[WKWebView _zoomToPoint:WebCore::atScale:]):
(-[WKWebView _zoomToRect:WebCore::atScale:origin:WebCore::]):
(constrainContentOffset):
(-[WKWebView _scrollToRect:WebCore::origin:WebCore::minimumScrollDistance:]):
(-[WKWebView _zoomOutWithOrigin:]):
(-[WKWebView _zoomToRect:withOrigin:fitEntireRect:minimumScale:maximumScale:minimumScrollDistance:]):
Add _zoomToRect:withOrigin:fitEntireRect:minimumScale:maximumScale:minimumScrollDistance:
and _zoomOutWithOrigin: as internal API, for use by SmartMagnificationController via WKContentView.
These will use some heuristics to generally center the target rect in the view
in a way that should match WebKit1 fairly closely.

  • UIProcess/API/Cocoa/WKWebViewInternal.h:

Don't use the WKContentViewDelegate protocol for things that don't belong there.

  • UIProcess/ios/SmartMagnificationController.h: Added.
  • UIProcess/ios/SmartMagnificationController.messages.in: Added.
  • UIProcess/ios/SmartMagnificationController.mm: Added.

(WebKit::SmartMagnificationController::SmartMagnificationController):
(WebKit::SmartMagnificationController::~SmartMagnificationController):
(WebKit::SmartMagnificationController::handleSmartMagnificationGesture):
(WebKit::SmartMagnificationController::didCollectGeometryForSmartMagnificationGesture):
SmartMagnificationController sits off of WKContentView, and sends/receives
messages to the WebProcess to do the smart magnification hit-testing.

  • UIProcess/ios/WKContentView.h:
  • UIProcess/ios/WKContentView.mm:

(-[WKContentView initWithFrame:context:WebKit::configuration:WebKit::webView:]):
As before, teach WKContentView about its owning WKWebView.

(-[WKContentView _takeViewSnapshot]):
Rename.

(-[WKContentView _zoomToRect:withOrigin:fitEntireRect:minimumScale:maximumScale:minimumScrollDistance:]):
(-[WKContentView _zoomOutWithOrigin:]):
Forward _zoomToRect:::::: and _zoomOutWithOrigin: to WKWebView.

  • UIProcess/ios/WKContentViewInteraction.h:
  • UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView setupInteraction]):
(-[WKContentView _doubleTapRecognized:]):
(-[WKContentView _twoFingerDoubleTapRecognized:]):
Forward double tap to the SmartMagnificationController.

  • UIProcess/mac/ViewGestureController.h:
  • UIProcess/mac/ViewGestureController.messages.in:
  • UIProcess/mac/ViewGestureControllerMac.mm:

(WebKit::ViewGestureController::didCollectGeometryForSmartMagnificationGesture):
Retrieve the minimum/maximum viewport scale from the WebProcess.

  • WebKit2.xcodeproj/project.pbxproj:

Add new files.

  • WebProcess/WebPage/ViewGestureGeometryCollector.cpp:

(WebKit::ViewGestureGeometryCollector::ViewGestureGeometryCollector):
(WebKit::ViewGestureGeometryCollector::dispatchDidCollectGeometryForSmartMagnificationGesture):
(WebKit::ViewGestureGeometryCollector::collectGeometryForSmartMagnificationGesture):
(WebKit::ViewGestureGeometryCollector::collectGeometryForMagnificationGesture):
Factor out code to send the geometry back to the UI process.

  • WebProcess/WebPage/ViewGestureGeometryCollector.h:
  • WebProcess/WebPage/ViewGestureGeometryCollector.messages.in:
  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::WebPage):

  • WebProcess/WebPage/WebPage.h:

We should have a ViewGestureGeometryCollector on iOS too.

  • WebCore.exp.in:

Add an export.

7:58 PM Changeset in webkit [164936] by commit-queue@webkit.org
  • 5 edits
    4 deletes in trunk

Unreviewed, rolling out r164929 and r164934.
​http://trac.webkit.org/changeset/164929
​http://trac.webkit.org/changeset/164934
​https://bugs.webkit.org/show_bug.cgi?id=129570

Caused assertions on two srcset tests (Requested by ap on
#webkit).

Source/WebCore:

  • html/parser/HTMLParserIdioms.cpp:

(WebCore::isHTMLSpaceOrComma):
(WebCore::parseImagesWithScaleFromSrcsetAttribute):
(WebCore::bestFitSourceForImageAttributes):

LayoutTests:

  • fast/hidpi/image-srcset-invalid-descriptor-expected.txt: Removed.
  • fast/hidpi/image-srcset-invalid-descriptor.html: Removed.
  • fast/hidpi/image-srcset-invalid-inputs-correct-src.html:
  • fast/hidpi/image-srcset-src-selection-1x-both-expected.txt: Removed.
  • fast/hidpi/image-srcset-src-selection-1x-both.html: Removed.
  • fast/hidpi/resources/srcset-helper.js:

(runTest):

7:45 PM Changeset in webkit [164935] by fpizlo@apple.com
  • 2 edits
    1 move in trunk

Source/JavaScriptCore: This shouldn't have been a layout test since it runs only under jsc. Moving it to JSC
stress tests.

  • tests/stress/generational-opaque-roots.js: Copied from LayoutTests/js/script-tests/generational-opaque-roots.js.

LayoutTests: This shouldn't be a layout test since it runs only under jsc. Moving it to JSC
stress tests.

  • js/script-tests/generational-opaque-roots.js: Removed.
4:35 PM Changeset in webkit [164934] by mitz@apple.com
  • 2 edits in trunk/Source/WebCore

Build fix.

  • html/parser/HTMLParserIdioms.cpp: Removed an unused function.
4:20 PM Changeset in webkit [164933] by ddkilzer@apple.com
  • 3 edits
    2 adds in trunk

Ensure keySplines is valid in SMIL animations
<​http://webkit.org/b/129547>
<rdar://problem/15676128>

Reviewed by Darin Adler.

Merged from Blink (patch by Philip Rogers):
​https://src.chromium.org/viewvc/blink?revision=156452&view=revision
​http://crbug.com/276111

This patch fixes a crash in SMIL animations when keySplines are not
specified. The SMIL spec is clear on this:
​http://www.w3.org/TR/2001/REC-smil-animation-20010904/#AnimFuncCalcMode
"If there are any errors in the keyTimes specification (bad values,
too many or too few values), the animation will have no effect."

This patch simply checks that keyTimes is not empty. Previously,
splinesCount was set to be m_keySplines.size() + 1 in
SVGAnimationElement.cpp; this patch changes splinesCount to be equal
to m_keySplines.size() to make the logic easier to follow and to
match other checks in SVGAnimationElement::startedActiveInterval.

Source/WebCore:

Test: svg/animations/animate-keysplines-crash.html

  • svg/SVGAnimationElement.cpp:

(WebCore::SVGAnimationElement::startedActiveInterval):

LayoutTests:

  • svg/animations/animate-keysplines-crash-expected.txt: Added.
  • svg/animations/animate-keysplines-crash.html: Added.
3:50 PM Changeset in webkit [164932] by benjamin@webkit.org
  • 3 edits
    2 adds in trunk

Tighten minimumRegisterRequirements()
​https://bugs.webkit.org/show_bug.cgi?id=129538

Reviewed by Andreas Kling.

Source/WebCore:

Fix small things that made minimumRegisterRequirements() a little optimistic
when dealing with attributes.

Test: fast/selectors/adjacent-descendant-tail-register-requirement.html

  • cssjit/SelectorCompiler.cpp:

(WebCore::SelectorCompiler::SelectorCodeGenerator::SelectorCodeGenerator):
Attribute Set does not do value matching, the case sensitive value matching is irrelevant
The problem is that flag is also used by minimumRegisterRequirements()
to find if one more register is needed.

Set the flag to case sensitive to avoid reserving one extra register.

(WebCore::SelectorCompiler::minimumRegisterRequirements):
Use a new backtrackingFlag to know if there is a descendant tail, thus a backtracking register
reserved.
This is better than using the backtracking action because the backtracking chain could be
an adjacent chain inside a descendant chain.

The flags are designed for that, just set one for minimumRegisterRequirements().

The 2 extra registers for the attribute count and address become limited to all attributes
except the last one. We don't keep a copy for the last matching, those registers were not needed.

(WebCore::SelectorCompiler::SelectorCodeGenerator::computeBacktrackingInformation):

LayoutTests:

  • fast/selectors/adjacent-descendant-tail-register-requirement-expected.txt: Added.
  • fast/selectors/adjacent-descendant-tail-register-requirement.html: Added.
3:31 PM Changeset in webkit [164931] by psolanki@apple.com
  • 4 edits in trunk/Source

[iOS] selectionImageForcingBlackText should return autoreleased object
​https://bugs.webkit.org/show_bug.cgi?id=129437
<rdar://problem/15810384>

Reviewed by Darin Adler.

Source/WebCore:

  • bindings/objc/DOM.mm:

(-[DOMRange renderedImageForcingBlackText:renderedImageForcingBlackText:]):

Source/WebKit/mac:

  • WebView/WebHTMLView.mm:

(-[WebHTMLView selectionImageForcingBlackText:selectionImageForcingBlackText:]):

3:30 PM Changeset in webkit [164930] by Darin Adler
  • 2 edits in trunk/Websites/webkit.org

Some small modernizing tweaks to the coding style guide
​https://bugs.webkit.org/show_bug.cgi?id=129427

Reviewed by Alexey Proskuryakov.

My intent here was originally to do the nullptr style guide update, but that was already done.
I did some other improvements. There's still quite a bit to do -- the guide is out of date in
a number of subtle ways.

  • coding/coding-style.html: Change "left side operator" example to be a more modern example

without non-WebKit-style abbreviations and such. Use auto& instead of const auto& in for loop
example, since that's normally preferred. Changed section title to say "zero" instead of "0".
Use words instead of abbreviations in code examples. Show that a modern C++ for loop is preferred
over index iteration in vector iteration example and also showed use of unsigned rather than
size_t since that is almost always what we want. Eliminated the use of PassOwnPtr in the "return
a newly created object" example. Use references instead of pointers in one example.

3:28 PM Changeset in webkit [164929] by commit-queue@webkit.org
  • 5 edits
    4 adds in trunk

Fix srcset related bugs
​https://bugs.webkit.org/show_bug.cgi?id=129539

Source/WebCore:

Fixed a bug reported against Blink's srcset implementation at ​https://code.google.com/p/chromium/issues/detail?id=347998
When both src and srcset had only 1x descriptors and the DPR was higher than 1, the src's resource was picked.

Also fixed the invalid descriptor handling, which wasn't aligned with the spec, and therefore was not as forward compatible as it should be.
Invalid descriptors caused the entire resource to be ignored, rather than just the descriptors themselves.

Patch by Yoav Weiss <​yoav@yoav.ws> on 2014-03-01
Reviewed by Andreas Kling.

Tests: fast/hidpi/image-srcset-invalid-descriptor.html

fast/hidpi/image-srcset-src-selection-1x-both.html

  • html/parser/HTMLParserIdioms.cpp:

(WebCore::parseDescriptors):
(WebCore::parseImagesWithScaleFromSrcsetAttribute):
(WebCore::bestFitSourceForImageAttributes):

LayoutTests:

These tests make sure that:

  1. Invalid srcset descriptors are handled according to the spec. The invalid descriptors are ignored, but the resource is not.
  2. When both src and srcset have only 1x descriptors and the DPR is higher than 1, the srcset's resource is picked.

I've also fixed the invalid inputs test, which was faulty.

Patch by Yoav Weiss <​yoav@yoav.ws> on 2014-03-01
Reviewed by Andreas Kling.

  • fast/hidpi/image-srcset-invalid-descriptor-expected.txt: Added.
  • fast/hidpi/image-srcset-invalid-descriptor.html: Added.
  • fast/hidpi/image-srcset-invalid-inputs-correct-src.html:
  • fast/hidpi/image-srcset-src-selection-1x-both-expected.txt: Added.
  • fast/hidpi/image-srcset-src-selection-1x-both.html: Added.
  • fast/hidpi/resources/srcset-helper.js:

(runTest):

3:22 PM Changeset in webkit [164928] by Darin Adler
  • 6 edits in trunk/Source/WebCore

Improve "bad parent" and "bad child list" assertions in line boxes
​https://bugs.webkit.org/show_bug.cgi?id=125656

Reviewed by Sam Weinig.

My previous fix for this problem was incomplete. This continuation of that fix addresses
the flaw in the original and adds additional lifetime checking so problems can be seen in
debug builds without a memory debugger.

  • rendering/InlineBox.cpp:

(WebCore::InlineBox::assertNotDeleted): Added. Poor man's memory debugging helper.
(WebCore::InlineBox::~InlineBox): Refactored body into a new function named
invalidateParentChildList. Added code to update the deletion sentinel to record
that this object is deleted.
(WebCore::InlineBox::setHasBadParent): Moved here from header since this debug-only
feature does not need to be inlined. Added a call to assertNotDeleted.
(WebCore::InlineBox::invalidateParentChildList): Added. Refactored from the destructor,
this is used by RenderTextLineBoxes.

  • rendering/InlineBox.h: Added the deletion sentinel, and called it in the parent

function. Also changed the expansion/setExpansion functions to use the type name "int",
since we don't use the type name "signed" in the WebKit coding style.

  • rendering/InlineFlowBox.cpp:

(WebCore::InlineFlowBox::~InlineFlowBox): Call setHasBadChildList rather than doing the
setHasBadParent work on children directly, to avoid code duplication.
(WebCore::InlineFlowBox::setHasBadChildList): Moved here from header. Added code to set
"has bad parent" on all children, something we previously did only on destruction. Also
added assertNotDeleted.
(WebCore::InlineFlowBox::checkConsistency): Added call to assertNotDeleted. Also tweaked
code style and variable names a little bit.

  • rendering/InlineFlowBox.h: Moved setHasBadChildList out of the header when it's on.

The empty version for ASSERT_WITH_SECURITY_IMPLICATION_DISABLED is still in the header.

  • rendering/RenderTextLineBoxes.cpp:

(WebCore::RenderTextLineBoxes::invalidateParentChildLists): Call the new
InlineBox::invalidateParentChildList function instead of calling setHasBadChildList directly.
The new function checks m_hasBadParent, something we couldn't do here.

3:05 PM Changeset in webkit [164927] by akling@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

JSCell::fastGetOwnProperty() should get the Structure more efficiently.
<​https://webkit.org/b/129560>

Now that structure() is nontrivial and we have a faster structure(VM&),
make use of that in fastGetOwnProperty() since we already have VM.

Reviewed by Sam Weinig.

  • runtime/JSCellInlines.h:

(JSC::JSCell::fastGetOwnProperty):

2:27 PM Changeset in webkit [164926] by cavalcantii@gmail.com
  • 2 edits in trunk/LayoutTests

Set css3/filter/huge-region as skipped
​https://bugs.webkit.org/show_bug.cgi?id=129552

It tries to use an insanely huge kernel for a huge blur filtered
element while it is expected to *not* execute. Thanks to the changes
introduced in #123716, this test will try to be executed and may
timeout in some machines. The proper fix will follow up in #129553.

Reviewed by Dirk Schulze.

2:21 PM Changeset in webkit [164925] by akling@apple.com
  • 3 edits in trunk/Source/JavaScriptCore

Avoid going through ExecState for VM when we already have it (in some places.)
<​https://webkit.org/b/129554>

Tweak some places that jump through unnecessary hoops to get the VM.
There are many more like this.

Reviewed by Sam Weinig.

  • runtime/JSObject.cpp:

(JSC::JSObject::putByIndexBeyondVectorLength):
(JSC::JSObject::putDirectIndexBeyondVectorLength):

  • runtime/ObjectPrototype.cpp:

(JSC::objectProtoFuncToString):

2:09 PM Changeset in webkit [164924] by benjamin@webkit.org
  • 7 edits
    2 adds in trunk

Optimized querySelector(All) when selector contains #id
​https://bugs.webkit.org/show_bug.cgi?id=116502

Reviewed by Andreas Kling.

Source/WebCore:

Test: fast/selectors/querySelector-id-filtering.html

The idea of this patch is to perform querySelector on a subtree
rooted at the last element with #id matching. For example, if we have the selector
"#foobar div a", we start by looking for the element with ID foobar in the TreeScope
cache, and start matching the children from there.

The idea comes from Rune for
​https://chromium.googlesource.com/chromium/blink/+/1cd83d3588973a02ab15d94b1b05a28620853624
but the code as diverged too much so it had to be reimplemented specifically for WebKit.

  • css/CSSSelectorList.cpp:

(WebCore::CSSSelectorList::CSSSelectorList):
(WebCore::CSSSelectorList::operator=):

  • css/CSSSelectorList.h:

(WebCore::CSSSelectorList::~CSSSelectorList):

  • css/StyleRule.h:

(WebCore::StyleRule::wrapperAdoptSelectorList):
(WebCore::StyleRulePage::wrapperAdoptSelectorList):

  • dom/SelectorQuery.cpp:

(WebCore::isSingleTagNameSelector):
(WebCore::isSingleClassNameSelector):
(WebCore::findIdMatchingType):
(WebCore::SelectorDataList::SelectorDataList):
(WebCore::filterRootById):
(WebCore::SelectorDataList::executeCompiledSimpleSelectorChecker):
(WebCore::SelectorDataList::execute):
(WebCore::SelectorQuery::SelectorQuery):
(WebCore::SelectorQueryCache::add):

  • dom/SelectorQuery.h:

(WebCore::SelectorDataList::SelectorData::SelectorData):

LayoutTests:

  • fast/selectors/querySelector-id-filtering-expected.txt: Added.
  • fast/selectors/querySelector-id-filtering.html: Added.
11:57 AM Changeset in webkit [164923] by fpizlo@apple.com
  • 10 edits
    2 adds in trunk/Source/JavaScriptCore

FTL should support PhantomArguments
​https://bugs.webkit.org/show_bug.cgi?id=113986

Reviewed by Oliver Hunt.

Adding PhantomArguments to the FTL mostly means wiring the recovery of the Arguments
object into the FTL's OSR exit compiler.

This isn't a speed-up yet, since there is still more to be done to fully support
all of the arguments craziness that our varargs benchmarks do.

  • dfg/DFGOSRExitCompiler32_64.cpp:

(JSC::DFG::OSRExitCompiler::compileExit): move the recovery code to DFGOSRExitCompilerCommon.cpp

  • dfg/DFGOSRExitCompiler64.cpp:

(JSC::DFG::OSRExitCompiler::compileExit): move the recovery code to DFGOSRExitCompilerCommon.cpp

  • dfg/DFGOSRExitCompilerCommon.cpp:

(JSC::DFG::ArgumentsRecoveryGenerator::ArgumentsRecoveryGenerator):
(JSC::DFG::ArgumentsRecoveryGenerator::~ArgumentsRecoveryGenerator):
(JSC::DFG::ArgumentsRecoveryGenerator::generateFor): this is the common place for the recovery code

  • dfg/DFGOSRExitCompilerCommon.h:
  • ftl/FTLCapabilities.cpp:

(JSC::FTL::canCompile):

  • ftl/FTLExitValue.cpp:

(JSC::FTL::ExitValue::dumpInContext):

  • ftl/FTLExitValue.h:

(JSC::FTL::ExitValue::argumentsObjectThatWasNotCreated):
(JSC::FTL::ExitValue::isArgumentsObjectThatWasNotCreated):
(JSC::FTL::ExitValue::valueFormat):

  • ftl/FTLLowerDFGToLLVM.cpp:

(JSC::FTL::LowerDFGToLLVM::compileNode):
(JSC::FTL::LowerDFGToLLVM::compilePhantomArguments):
(JSC::FTL::LowerDFGToLLVM::buildExitArguments):
(JSC::FTL::LowerDFGToLLVM::tryToSetConstantExitArgument):

  • ftl/FTLOSRExitCompiler.cpp:

(JSC::FTL::compileStub): Call into the ArgumentsRecoveryGenerator

  • tests/stress/slightly-more-difficult-to-fold-reflective-arguments-access.js: Added.
  • tests/stress/trivially-foldable-reflective-arguments-access.js: Added.
11:42 AM Changeset in webkit [164922] by fpizlo@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

Unreviewed, uncomment some code. It wasn't meant to be commented in the first place.

  • dfg/DFGCSEPhase.cpp:

(JSC::DFG::CSEPhase::getPropertyStorageLoadElimination):

10:48 AM Changeset in webkit [164921] by timothy@apple.com
  • 4 edits in trunk/Source/WebInspectorUI

Make Start Timeline Recording in the Develop menu show the Timeline view again.

This also fixes a noticeable delay between showing the sidebar and the default Timeline
view when initially opening the Web Inspector to the Timeline.

​https://bugs.webkit.org/show_bug.cgi?id=129545

Reviewed by Joseph Pecoraro.

  • UserInterface/Base/Main.js:

(WebInspector.contentLoaded): Call TimelineSidebarPanel.initialize to prime the content view.
This is needed so the view is ready in this run loop cycle, where it was delayed before.

  • UserInterface/Protocol/InspectorFrontendAPI.js:

(InspectorFrontendAPI.setTimelineProfilingEnabled): Call TimelineSidebarPanel.showTimelineOverview.

  • UserInterface/Views/TimelineSidebarPanel.js:

(WebInspector.TimelineSidebarPanel.prototype.initialize): Added.

10:48 AM Changeset in webkit [164920] by ap@apple.com
  • 9 edits in trunk

Node::compareDocumentPosition leaks memory structure
​https://bugs.webkit.org/show_bug.cgi?id=120244

Reviewed by Ryosuke Niwa.

Source/WebCore:

  • dom/Node.cpp: (WebCore::Node::compareDocumentPosition): Don't do it.

LayoutTests:

  • TestExpectations:
  • dom/xhtml/level3/core/nodecomparedocumentposition38-expected.txt:
  • fast/dom/compare-document-position-disconnected-nodes-expected.txt:
  • fast/dom/compare-document-position-disconnected-nodes.html:
  • fast/dom/shadow/compare-document-position-expected.txt:
  • fast/dom/shadow/compare-document-position.html:
10:46 AM Changeset in webkit [164919] by timothy@apple.com
  • 2 edits in trunk/Source/WebInspectorUI

Select the Resource navigation sidebar by default when there is no cookie.

​https://bugs.webkit.org/show_bug.cgi?id=129544

Reviewed by Joseph Pecoraro.

  • UserInterface/Base/Main.js:

(WebInspector._restoreInspectorViewStateFromCookie):

9:51 AM Changeset in webkit [164918] by timothy@apple.com
  • 5 edits in trunk/Source/WebInspectorUI

Label JavaScript forced layouts as such in the Timeline.

​https://bugs.webkit.org/show_bug.cgi?id=129546

Reviewed by David Kilzer.

  • Localizations/en.lproj/localizedStrings.js: Updated.
  • UserInterface/Controllers/TimelineManager.js:

(WebInspector.TimelineManager.prototype.eventRecorded.processRecord):
(WebInspector.TimelineManager.prototype.eventRecorded):

  • UserInterface/Models/LayoutTimelineRecord.js:

(WebInspector.LayoutTimelineRecord.EventType.displayName):

  • UserInterface/Views/TimelineRecordTreeElement.js:

(WebInspector.TimelineRecordTreeElement):

8:26 AM Changeset in webkit [164917] by ddkilzer@apple.com
  • 8 edits
    1 move
    2 adds in trunk

Fix lifetime handling of SVGPropertyTearOffs
<​http://webkit.org/b/129211>
<rdar://problem/15696025>

Reviewed by Maciej Stachowiak.

Merged from Blink (patch by Ojan Vafai):
​https://src.chromium.org/viewvc/blink?revision=157801&view=revision
​http://crbug.com/288771

-Replace SVGStaticPropertyWithParentTearOff with SVGMatrixTearOff
since it's only used in that one place. This means we can get rid
of the templatizing and the method pointer.
-Change SVGPropertyTearOff to keep track of it's child tearoffs
and call detachWrapper on its child tearoffs when it's destroyed
or when it's wrapper is detached.
-Have SVGPropertyTearOff hold WeakPtrs to the child tearoffs
to avoid having a cycle.

Source/WebCore:

Test: svg/transforms/svg-matrix-tearoff-crash.html

  • GNUmakefile.list.am:
  • WebCore.vcxproj/WebCore.vcxproj:
  • WebCore.vcxproj/WebCore.vcxproj.filters:
  • WebCore.xcodeproj/project.pbxproj:
  • bindings/scripts/CodeGeneratorJS.pm:

(NativeToJSValue):

  • svg/properties/SVGMatrixTearOff.h: Renamed from Source/WebCore/svg/properties/SVGStaticPropertyWithParentTearOff.h.

(WebCore::SVGMatrixTearOff::create):
(WebCore::SVGMatrixTearOff::commitChange):
(WebCore::SVGMatrixTearOff::SVGMatrixTearOff):

  • svg/properties/SVGPropertyTearOff.h:

(WebCore::SVGPropertyTearOff::addChild):
(WebCore::SVGPropertyTearOff::detachChildren):

LayoutTests:

  • svg/transforms/svg-matrix-tearoff-crash-expected.txt: Added.
  • svg/transforms/svg-matrix-tearoff-crash.html: Added.
1:35 AM Changeset in webkit [164916] by zandobersek@gmail.com
  • 2 edits in trunk/Source/WebCore

Unreviewed build fix after r164913.

  • platform/CountedUserActivity.h: Some ports don't (yet) support forwarding headers for WebCore,

so UserActivity.h should be included as non-forwarding.

1:08 AM Changeset in webkit [164915] by zandobersek@gmail.com
  • 3 edits in trunk/Source/WebKit2

Unreviewed follow-up to r164905 and r164912.

  • Platform/IPC/Connection.h: Only declare setShouldCloseConnectionOnMachExceptions() for the Mac port.
  • UIProcess/Plugins/PluginProcessProxy.cpp:

(WebKit::PluginProcessProxy::didFinishLaunching): Only call setShouldCloseConnectionOnMachExceptions() on the Mac port.

1:01 AM Changeset in webkit [164914] by barraclough@apple.com
  • 5 edits in trunk/Source/WebCore

Change PageActivityAssertionToken to use a WeakPtr
​https://bugs.webkit.org/show_bug.cgi?id=129526

Reviewed by Sam Weinig.

PageThrottler effectively implements a bespoke weak pointer mechanism; remove this & just use WeakPtr.

  • page/PageActivityAssertionToken.cpp:

(WebCore::PageActivityAssertionToken::PageActivityAssertionToken):
(WebCore::PageActivityAssertionToken::~PageActivityAssertionToken):

  • addActivityToken->incrementActivityCount, removeActivityToken->decrementActivityCount
  • page/PageActivityAssertionToken.h:
    • removed invalidate, made m_throttler a WeakPtr
  • page/PageThrottler.cpp:

(WebCore::PageThrottler::PageThrottler):

  • initialize m_weakPtrFactory, m_activityCount.

(WebCore::PageThrottler::~PageThrottler):

  • removed called to invalidate.

(WebCore::PageThrottler::startThrottleHysteresisTimer):

  • m_activityTokens.size()->m_activityCount

(WebCore::PageThrottler::throttleHysteresisTimerFired):

  • m_activityTokens.size()->m_activityCount

(WebCore::PageThrottler::incrementActivityCount):

  • m_activityTokens.add->++

(WebCore::PageThrottler::decrementActivityCount):

  • m_activityTokens.remove->--
  • page/PageThrottler.h:

(WebCore::PageThrottler::weakPtr):

  • replaced HashSet with WeakPtrFactory & count.
12:49 AM Changeset in webkit [164913] by barraclough@apple.com
  • 16 edits
    1 copy in trunk/Source

Split UserActivity/CountedUserActivity
​https://bugs.webkit.org/show_bug.cgi?id=129520

Reviewed by Sam Weinig.

UserActivity currently provides an interface allowing counted, nested calls to enable/disable.
In some cases it would be easier to use if it were a simpler boolean enabled/disabled interface.

Source/WebCore:

  • WebCore.exp.in:
    • beginActivity->start, endActivity->stop.
  • WebCore.xcodeproj/project.pbxproj:
  • page/PageThrottler.cpp:

(WebCore::PageThrottler::PageThrottler):
(WebCore::PageThrottler::~PageThrottler):
(WebCore::PageThrottler::throttlePage):
(WebCore::PageThrottler::unthrottlePage):
(WebCore::PageThrottler::setIsVisuallyIdle):

  • beginActivity->increment, endActivity->decrement.
  • page/PageThrottler.h:
    • UserActivity->CountedUserActivity for m_pageActivity
  • platform/CountedUserActivity.h: Copied from Source/WebCore/platform/UserActivity.h.

(WebCore::CountedUserActivity::CountedUserActivity):
(WebCore::CountedUserActivity::increment):
(WebCore::CountedUserActivity::decrement):

  • Added, provides counted interface to UserActivity.
  • platform/UserActivity.cpp:

(WebCore::UserActivity::UserActivity):

  • m_count removed

(WebCore::UserActivity::start):
(WebCore::UserActivity::stop):

  • beginActivity->start, endActivity->stop, implentation now empty.
  • platform/UserActivity.h:
    • beginActivity->start, endActivity->stop, m_count moved to HAVE(NS_ACTIVITY)
  • platform/mac/UserActivityMac.mm:

(WebCore::UserActivity::UserActivity):

  • m_count->m_started

(WebCore::UserActivity::isValid):

  • m_count->m_started

(WebCore::UserActivity::start):
(WebCore::UserActivity::stop):

  • beginActivity->start, endActivity->stop, no longer nest

Source/WebKit2:

  • PluginProcess/PluginProcess.h:

(WebKit::PluginProcess::connectionActivity):

  • UserActivity->CountedUserActivity
  • PluginProcess/mac/PluginProcessMac.mm:

(WebKit::PluginProcess::platformInitializeProcess):

  • beginActivity->start
  • Shared/ActivityAssertion.cpp:

(WebKit::ActivityAssertion::ActivityAssertion):
(WebKit::ActivityAssertion::~ActivityAssertion):

  • beginActivity->increment, endActivity->decrement.
  • Shared/ActivityAssertion.h:
    • UserActivity->CountedUserActivity
  • Shared/ChildProcess.cpp:

(WebKit::ChildProcess::setProcessSuppressionEnabled):

  • beginActivity->start, endActivity->stop.
  • Shared/ChildProcess.h:
    • removed processSuppressionEnabled()
  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::updatePreferences):

  • beginActivity->start, endActivity->stop.
12:35 AM Changeset in webkit [164912] by Csaba Osztrogonác
  • 2 edits in trunk/Source/WebKit2

Unreviewed trivial Linux buildfix after r164905.

setShouldCloseConnectionOnMachExceptions() is defined in the Mac only
Source/WebKit2/Platform/IPC/mac/ConnectionMac.cpp inside !PLATFORM(IOS)
which means PLATFORM(MAC), because nobody else uses this source file.

But ChildProcessProxy.cpp is used by Linux ports too, so we have to
use the most specific #if guard here, which is PLATFORM(MAC).

  • Shared/ChildProcessProxy.cpp:

(WebKit::ChildProcessProxy::didFinishLaunching):

Feb 28, 2014:

10:13 PM Changeset in webkit [164911] by benjamin@webkit.org
  • 4 edits
    4 adds in trunk

Caller saved registers can be accidentally discarded when clearing the local stack
​https://bugs.webkit.org/show_bug.cgi?id=129532

Reviewed by Andreas Kling.

Source/WebCore:

Tests: fast/selectors/tree-modifying-case-insensitive-selectors.html

fast/selectors/tree-modifying-selectors.html

StackAllocator::discard() no longer make sense now that we can use caller saved regsiter.
We should instead discard everything up to the beginning of the local stack.

  • cssjit/SelectorCompiler.cpp:

(WebCore::SelectorCompiler::SelectorCodeGenerator::generateSelectorChecker):

  • cssjit/StackAllocator.h:

(WebCore::StackAllocator::popAndDiscardUpTo):

LayoutTests:

  • fast/selectors/tree-modifying-case-insensitive-selectors.html: Added.
  • fast/selectors/tree-modifying-selectors.html: Added.
9:36 PM Changeset in webkit [164910] by Simon Fraser
  • 4 edits in trunk/Source/WebKit2

[iOS] Assertions and bad behavior when zooming into an iframe containing apple.com
​https://bugs.webkit.org/show_bug.cgi?id=129537

Reviewed by Benjamin Poulain.

When building RemoteLayerTreeTransactions, the ordering of setting created
layers and doing the recursive tree walk was wrong, such that we failed
to noticed layers created during the recursiveBuildTransaction().

Also harden the UI-side code against Obj-C exceptions thrown when layers
are missing.

  • Shared/mac/RemoteLayerTreePropertyApplier.mm:

(WebKit::RemoteLayerTreePropertyApplier::applyProperties): Assert when
children are not found, and protect against Obj-C exceptions.

  • Shared/mac/RemoteLayerTreeTransaction.mm:

(WebKit::RemoteLayerTreeTransaction::LayerCreationProperties::LayerCreationProperties):
Initialize the members.

  • WebProcess/WebPage/mac/RemoteLayerTreeContext.mm:

(WebKit::RemoteLayerTreeContext::buildTransaction): Do the recursive walk
before setting created and destroyed layers, since the walk can create
layers (especially when swapping into tiled layers).

9:34 PM Changeset in webkit [164909] by jinwoo7.song@samsung.com
  • 2 edits in trunk/Source/WebKit2

Fix WebKit2 build after r164890
​https://bugs.webkit.org/show_bug.cgi?id=129533

Reviewed by Simon Fraser.

LayerRepresentation.h is included to WebPageProxy.h in r164890
but it is not used in non Mac or iOS port.

  • UIProcess/WebPageProxy.h:
9:31 PM Changeset in webkit [164908] by mitz@apple.com
  • 4 edits
    1 delete in trunk/Source/WebKit

[Mac] Remove MailQuirksUserScript.js
​https://bugs.webkit.org/show_bug.cgi?id=129536

Reviewed by Andy Estes.

../WebKit:

  • WebKit.xcodeproj/project.pbxproj: Removed reference to the file.

../WebKit/mac:

  • Configurations/WebKit.xcconfig: Removed from EXCLUDED_SOURCE_FILE_NAMES_iphoneos.
  • Misc/MailQuirksUserScript.js: Removed. This file has been unused since r147447.
9:14 PM Changeset in webkit [164907] by aestes@apple.com
  • 2 edits in trunk/Source/WebCore

[iOS] FrameLoader has a NULL m_progressTracker when initialized with initForSynthesizedDocument()
​https://bugs.webkit.org/show_bug.cgi?id=129534

Reviewed by Simon Fraser.

No new tests because TestWebKitAPI does not yet support iOS. See <​http://webkit.org/b/129535>.

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::initForSynthesizedDocument): Create a FrameProgressTracker.

9:05 PM Changeset in webkit [164906] by ap@apple.com
  • 2 edits in trunk/Source/WebKit2

[Mac] Extract _data->_thumbnailView checks into a helper function
​https://bugs.webkit.org/show_bug.cgi?id=129517

Reviewed by Sam Weinig.

  • UIProcess/API/mac/WKView.mm:

(-[WKView shouldIgnoreMouseEvents]):
(-[WKView scrollWheel:]):
(-[WKView mouseMoved:]):
(-[WKView mouseDown:]):
(-[WKView mouseUp:]):
(-[WKView mouseDragged:]):

8:10 PM Changeset in webkit [164905] by psolanki@apple.com
  • 4 edits in trunk/Source/WebKit2

[iOS][WebKit2] Don't use any of the mach exception handling code on iOS
​https://bugs.webkit.org/show_bug.cgi?id=129516

Reviewed by Sam Weinig.

This code is not used on iOS after my fix in r164883. We can just move it all under
!PLATFORM(IOS).

  • Platform/IPC/Connection.h:
  • Platform/IPC/mac/ConnectionMac.cpp:

(IPC::Connection::platformInvalidate):
(IPC::Connection::platformInitialize):
(IPC::Connection::open):

  • Shared/ChildProcessProxy.cpp:

(WebKit::ChildProcessProxy::didFinishLaunching):

8:01 PM Changeset in webkit [164904] by akling@apple.com
  • 3 edits in trunk/Source/JavaScriptCore

JSObject::findPropertyHashEntry() should take VM instead of ExecState.
<​https://webkit.org/b/129529>

Callers already have VM in a local, and findPropertyHashEntry() only
uses the VM, no need to go all the way through ExecState.

Reviewed by Geoffrey Garen.

  • runtime/JSObject.cpp:

(JSC::JSObject::put):
(JSC::JSObject::deleteProperty):
(JSC::JSObject::findPropertyHashEntry):

  • runtime/JSObject.h:
7:11 PM Changeset in webkit [164903] by weinig@apple.com
  • 2 edits in trunk/Source/WebCore

ASSERT(isMainThread()) hit under platformUserPreferredLanguages() with WebKit1 and WebKit2 in the same process
​https://bugs.webkit.org/show_bug.cgi?id=129528

Reviewed by Andreas Kling.

  • platform/mac/Language.mm:

(WebCore::preferredLanguagesMutex):
(WebCore::preferredLanguages):
(+[WebLanguageChangeObserver languagePreferencesDidChange:]):
(WebCore::httpStyleLanguageCode):
(WebCore::platformUserPreferredLanguages):
Add a mutex and do an isolated copy of the strings to allow this function to be called from multiple threads.

6:51 PM Changeset in webkit [164902] by commit-queue@webkit.org
  • 5 edits in trunk/Source/JavaScriptCore

Deadlock remotely inspecting iOS Simulator
​https://bugs.webkit.org/show_bug.cgi?id=129511

Patch by Joseph Pecoraro <Joseph Pecoraro> on 2014-02-28
Reviewed by Timothy Hatcher.

Avoid synchronous setup. Do it asynchronously, and let
the RemoteInspector singleton know later if it failed.

  • inspector/remote/RemoteInspector.h:
  • inspector/remote/RemoteInspector.mm:

(Inspector::RemoteInspector::setupFailed):

  • inspector/remote/RemoteInspectorDebuggableConnection.h:
  • inspector/remote/RemoteInspectorDebuggableConnection.mm:

(Inspector::RemoteInspectorDebuggableConnection::setup):

6:05 PM Changeset in webkit [164901] by dino@apple.com
  • 2 edits in trunk/Source/WebCore

Crash at ebay.com when viewing auction items at com.apple.WebCore: WebCore::GraphicsContext3D::getExtensions + 10
​https://bugs.webkit.org/show_bug.cgi?id=129523

Reviewed by Simon Fraser.

  • html/canvas/WebGLRenderingContext.cpp:

(WebCore::WebGLRenderingContext::getSupportedExtensions): Early return if the
context is pending policy resolution.

5:11 PM Changeset in webkit [164900] by akling@apple.com
  • 2 edits in trunk/Source/WebCore

Micro-optimize JSNodeOwner::isReachableFromOpaqueRoots().
<​https://webkit.org/b/129518>

Only do image and audio element specific checks for element nodes.
Time spent in here goes from 0.8% to 0.5% on DYEB.

Reviewed by Benjamin Poulain.

  • bindings/js/JSNodeCustom.cpp:

(WebCore::isReachableFromDOM):

5:07 PM Changeset in webkit [164899] by bshafiei@apple.com
  • 1 copy in tags/Safari-537.75.4

New tag.

5:06 PM Changeset in webkit [164898] by bshafiei@apple.com
  • 5 edits in branches/safari-537.75-branch/Source

Versioning.

4:45 PM Changeset in webkit [164897] by ggaren@apple.com
  • 2 edits in trunk/Source/WebCore

JSC Assertion failure every time I start Safari (r164846)
​https://bugs.webkit.org/show_bug.cgi?id=129510

Reviewed by Mark Hahnenberg.

Take the lock before allocating objects because afterward is too late.

  • bindings/js/ScriptController.cpp:

(WebCore::ScriptController::evaluateInWorld):
(WebCore::ScriptController::windowScriptNPObject):
(WebCore::ScriptController::jsObjectForPluginElement):

4:42 PM Changeset in webkit [164896] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

Web Inspector: ASSERT in WebCore::JSDOMWindowBase::supportsRichSourceInfo inspecting iOS
​https://bugs.webkit.org/show_bug.cgi?id=129512

Patch by Joseph Pecoraro <Joseph Pecoraro> on 2014-02-28
Reviewed by Simon Fraser.

  • page/PageDebuggable.cpp:

(WebCore::PageDebuggable::connect):
(WebCore::PageDebuggable::disconnect):

4:40 PM Changeset in webkit [164895] by andersca@apple.com
  • 4 edits in trunk/Source/WebCore

VisitedLinkStore should be able to invalidate link styles for associated pages
​https://bugs.webkit.org/show_bug.cgi?id=129515

Reviewed by Dan Bernstein.

  • page/Page.cpp:

(WebCore::Page::Page):
Add the VisitedLinkStore if we have one.

(WebCore::Page::~Page):
Remove the VisitedLinkStore if we have one.

  • page/VisitedLinkStore.cpp:

(WebCore::VisitedLinkStore::~VisitedLinkStore):
Assert that we don't have any pages.

(WebCore::VisitedLinkStore::addPage):
Add the page to the set.

(WebCore::VisitedLinkStore::removePage):
Remove the page from the set.

(WebCore::VisitedLinkStore::invalidateStylesForAllLinks):
Traverse all the pages and invalidate the style for all links.

(WebCore::VisitedLinkStore::invalidateStylesForLink):
Traverse all the pages and invalidate the style for a given link.

  • page/VisitedLinkStore.h:

Add new members.

4:27 PM Changeset in webkit [164894] by Brent Fulgham
  • 2 edits in branches/safari-537.75-branch/WebKitLibraries

Unreviewed build fix.
<rdar://problem/16197002>

  • win/tools/vsprops/debugsuffix.props: Set OFFICIAL_BUILD=1 for DebugSuffix target, since this

is the official debug analog to the Production build target.

4:22 PM Changeset in webkit [164893] by Brent Fulgham
  • 7 edits in branches/safari-537.75-branch/Source

Unreviewed build fix.
<rdar://problem/16197002>

Source/JavaScriptCore:

slash from path names, as this was causing multiple arguments
to be concatenated together (the whitespace was getting escaped
by the slash.)

Source/WebCore:

  • WebCore.vcxproj/WebCore.proj: Remove trailing

slash from path names, as this was causing multiple arguments
to be concatenated together (the whitespace was getting escaped
by the slash.)

Source/WTF:

  • WTF.vcxproj/WTF.proj: Remove trailing

slash from path names, as this was causing multiple arguments
to be concatenated together (the whitespace was getting escaped
by the slash.)

  • WTF.vcxproj/work-around-vs-dependency-tracking-bugs.py:

Check for presence of RC_XBS in environment, and exit since
this logic is not needed for 'official' builds.

4:04 PM WebKitGTK/2.4.x edited by desrt@desrt.ca
geoclue2 is on track (diff)
3:55 PM Changeset in webkit [164892] by benjamin@webkit.org
  • 4 edits in trunk/Source

Source/WebCore: [iOS][WK2] highlight rects should never big bigger than the view
​https://bugs.webkit.org/show_bug.cgi?id=129472

Patch by Benjamin Poulain <​bpoulain@apple.com> on 2014-02-28
Reviewed by Enrica Casucci.

  • WebCore.exp.in:

Source/WebKit2: [iOS][WK2] highlight rects should never big bigger than the view
​https://bugs.webkit.org/show_bug.cgi?id=129472
<rdar://problem/15673655>

Patch by Benjamin Poulain <​bpoulain@apple.com> on 2014-02-28
Reviewed by Enrica Casucci.

The unobscuredRect on WebPageProxy is the "true" view rect as seen by the user.
Use that to constrain our highlight rects.

  • UIProcess/ios/WKContentViewInteraction.mm:

(highlightedQuadsFitInRect):
(-[WKContentView _didGetTapHighlightForRequest:color:WebCore::quads:WebCore::topLeftRadius:WebCore::topRightRadius:WebCore::bottomLeftRadius:WebCore::bottomRightRadius:WebCore::]):

3:38 PM Changeset in webkit [164891] by Beth Dakin
  • 2 edits in trunk/Source/WebCore

Turn off margin tiles on iOS
​https://bugs.webkit.org/show_bug.cgi?id=129509
-and corresponding-
<rdar://problem/16200383>

Reviewed by Simon Fraser.

  • page/FrameView.cpp:

(WebCore::FrameView::needsExtendedBackgroundRectForPainting):

3:36 PM Changeset in webkit [164890] by Simon Fraser
  • 18 edits
    2 copies
    1 move
    2 adds in trunk/Source/WebKit2

Make UIViews for compositing layers in the UI process on iOS
​https://bugs.webkit.org/show_bug.cgi?id=129508

Reviewed by Sam Weinig.

In order to parent UIScrollViews for -webkit-overflow-scrolling:touch,
we need to use UIViews in the compositing layer hierarchy on iOS.

Make much of the code that currently uses CALayers layer-or-view
agnostic by using a typedef LayerOrView.

Fix RemoteLayerTreePropertyApplier to apply properties to UIViews
or CALayers; on iOS, we need both paths because we have to set properties
on CALayer for remotely hosted content.

Add iOS-specific implementation files for RemoteLayerTreeHost and RemoteScrollingCoordinatorProxy,
since these need to see the UIView implementation. WebVideoFullscreenManagerProxy.cpp also needs
to be come a .mm file.

  • Shared/mac/RemoteLayerTreePropertyApplier.h:
  • Shared/mac/RemoteLayerTreePropertyApplier.mm:

(-[UIView setSubviews:]):
(WebKit::applyPropertiesToLayer):
(WebKit::RemoteLayerTreePropertyApplier::applyProperties):

  • UIProcess/Cocoa/LayerRepresentation.h: Copied from Source/WebKit2/Shared/mac/RemoteLayerTreePropertyApplier.h.

(asLayer):

  • UIProcess/PageClient.h:
  • UIProcess/Scrolling/RemoteScrollingCoordinatorProxy.cpp:
  • UIProcess/WebPageProxy.h:
  • UIProcess/ios/LayerRepresentation.mm: Copied from Source/WebKit2/Shared/mac/RemoteLayerTreePropertyApplier.h.

(asLayer):

  • UIProcess/ios/PageClientImplIOS.h:
  • UIProcess/ios/PageClientImplIOS.mm:

(WebKit::PageClientImpl::setAcceleratedCompositingRootLayer):
(WebKit::PageClientImpl::acceleratedCompositingRootLayer):

  • UIProcess/ios/RemoteLayerTreeHostIOS.mm: Added.

(+[WKTransformView layerClass]):
(-[WKRemoteView initWithFrame:contextID:]):
(+[WKRemoteView layerClass]):
(WebKit::RemoteLayerTreeHost::createLayer):

  • UIProcess/ios/RemoteScrollingCoordinatorProxyIOS.mm: Added.

(WebKit::layerRepresentationFromLayerOrView):
(WebKit::RemoteScrollingCoordinatorProxy::connectStateNodeLayers):

  • UIProcess/ios/WKContentView.h:
  • UIProcess/ios/WKContentView.mm:

(-[WKContentView _setAcceleratedCompositingRootView:]):

  • UIProcess/ios/WebPageProxyIOS.mm:

(WebKit::WebPageProxy::setAcceleratedCompositingRootLayer):

  • UIProcess/ios/WebVideoFullscreenManagerProxy.mm: Renamed from Source/WebKit2/UIProcess/ios/WebVideoFullscreenManagerProxy.cpp.

(WebKit::WebVideoFullscreenManagerProxy::create):
(WebKit::WebVideoFullscreenManagerProxy::WebVideoFullscreenManagerProxy):
(WebKit::WebVideoFullscreenManagerProxy::~WebVideoFullscreenManagerProxy):
(WebKit::WebVideoFullscreenManagerProxy::didCommitLayerTree):
(WebKit::WebVideoFullscreenManagerProxy::setVideoLayerID):
(WebKit::WebVideoFullscreenManagerProxy::enterFullscreen):
(WebKit::WebVideoFullscreenManagerProxy::requestExitFullScreen):
(WebKit::WebVideoFullscreenManagerProxy::play):
(WebKit::WebVideoFullscreenManagerProxy::pause):
(WebKit::WebVideoFullscreenManagerProxy::togglePlayState):
(WebKit::WebVideoFullscreenManagerProxy::seekToTime):
(WebKit::WebVideoFullscreenManagerProxy::didExitFullscreen):

  • UIProcess/mac/PageClientImpl.h:
  • UIProcess/mac/RemoteLayerTreeDrawingAreaProxy.mm:

(WebKit::RemoteLayerTreeDrawingAreaProxy::commitLayerTree):
(WebKit::RemoteLayerTreeDrawingAreaProxy::updateDebugIndicator):

  • UIProcess/mac/RemoteLayerTreeHost.h:

(WebKit::RemoteLayerTreeHost::rootLayer):

  • UIProcess/mac/RemoteLayerTreeHost.mm:

(WebKit::RemoteLayerTreeHost::updateLayerTree):
(WebKit::RemoteLayerTreeHost::getLayer):
(WebKit::RemoteLayerTreeHost::createLayer):

  • UIProcess/mac/WebPageProxyMac.mm:

(WebKit::WebPageProxy::setAcceleratedCompositingRootLayer):
(WebKit::WebPageProxy::acceleratedCompositingRootLayer):

  • WebKit2.xcodeproj/project.pbxproj:
  • WebProcess/WebPage/mac/PlatformCALayerRemote.cpp:

(PlatformCALayerRemote::recursiveBuildTransaction):

3:11 PM Changeset in webkit [164889] by andersca@apple.com
  • 2 edits in trunk/Source/WebKit2

WebKit2.TerminateTwice test is failing.

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::resetStateAfterProcessExited):
Turns out this can be called even when we're in the launching state so don't try to
remove the process from the visited link provider when in that state.

3:05 PM Changeset in webkit [164888] by Alan Bujtas
  • 3 edits in trunk/Source/WebCore

Subpixel rendering: Add subpixel support to border type of double, groove, ridge, inset and outset.
​https://bugs.webkit.org/show_bug.cgi?id=129226

Reviewed by Simon Fraser.

This is the conversion of double, inset, outset, groove and ridge border type
painting to support device pixel precision width/height.

Regression is covered by existing tests.

  • rendering/RenderObject.cpp:

(WebCore::RenderObject::drawLineForBoxSide):

  • rendering/RenderObject.h:
3:04 PM Changeset in webkit [164887] by Chris Fleizach
  • 2 edits in trunk/Source/WebKit2

AX: Crash in WebKit::WebPageProxy::registerWebProcessAccessibilityToken
​https://bugs.webkit.org/show_bug.cgi?id=127387

Reviewed by Alexey Proskuryakov.

Speculative fix for crasher. We should check if the connection is still valid before proceeding.

  • UIProcess/mac/WebPageProxyMac.mm:

(WebKit::WebPageProxy::registerWebProcessAccessibilityToken):

2:54 PM Changeset in webkit [164886] by cavalcantii@gmail.com
  • 7 edits
    4 adds in trunk

Filters should test for area instead of single dimension
​https://bugs.webkit.org/show_bug.cgi?id=123716

Reviewed by Dirk Schulze.

A filtered SVG element with a sigle dimension bigger than 5000
(counting the margin/border) will fail to render. So an element
with 4000x4000 will render fine, while another with 5000x10
won't. This patch instead tests against the total element area,
which fixes such cases.

Source/WebCore:

It also fixes some layer violations where FilterEffectRenderer and
RenderSVGResourceFilter were directly accessing the maximum filter
size in FilterEffect. Since before a somewhat bigger filter would
be aborted, it required to change the maximum kernel size
for blur filter (some filters are slower to run than others for
bigger SVGs).

Tests: svg/filters/big-height-filter-expected.svg

svg/filters/big-height-filter.svg
svg/filters/big-width-filter-expected.svg
svg/filters/big-width-filter.svg

  • platform/graphics/filters/FEGaussianBlur.cpp:
  • platform/graphics/filters/FilterEffect.cpp:

(WebCore::FilterEffect::maxFilterArea): New function, returns the
maximum allowed element area (currently is 4096 x 4096).
(WebCore::FilterEffect::isFilterSizeValid): Now it tests for
element area instead of a single dimension. This code is reused by
FilterEffectRenderer.

  • platform/graphics/filters/FilterEffect.h:
  • rendering/FilterEffectRenderer.cpp:

(WebCore::FilterEffectRenderer::updateBackingStoreRect): Uses the
new function in FilterEffect to test for valid sizes.
(WebCore::FilterEffectRendererHelper::beginFilterEffect): Same case.

  • rendering/svg/RenderSVGResourceFilter.cpp:

(WebCore::RenderSVGResourceFilter::fitsInMaximumImageSize): Test
for maximum area and properly changes the scale of transformed
element to fit.

LayoutTests:

The current patch fixes the rendering issue and at same time,
fixes some layer violations where FilterEffectRenderer and
RenderSVGResourceFilter were directly accessing the kMax value
in FilterEffect. Since before a somewhat bigger filter would
be aborted, it required to change the maximum kernel size
for blur filter (some filters are slower to run than others for
bigger SVGs).

  • svg/filters/big-height-filter-expected.svg: Added.
  • svg/filters/big-height-filter.svg: Added.
  • svg/filters/big-width-filter-expected.svg: Added.
  • svg/filters/big-width-filter.svg: Added.
2:54 PM Changeset in webkit [164885] by timothy@apple.com
  • 3 edits in trunk/Source/WebInspectorUI

Enable breakpoints when adding a new breakpoint or enabling an existing breakpoint.

This eliminates a multi-step process for the user that can be confusing.

​https://bugs.webkit.org/show_bug.cgi?id=129426

Reviewed by Joseph Pecoraro.

  • UserInterface/Controllers/DebuggerManager.js:

(WebInspector.DebuggerManager.prototype.set breakpointsEnabled):
(WebInspector.DebuggerManager.prototype._breakpointDisabledStateDidChange):

  • UserInterface/Views/DebuggerSidebarPanel.js:

(WebInspector.DebuggerSidebarPanel):
(WebInspector.DebuggerSidebarPanel.prototype._breakpointsEnabledDidChange):
(WebInspector.DebuggerSidebarPanel.prototype._breakpointsToggleButtonClicked):

2:51 PM Changeset in webkit [164884] by timothy@apple.com
  • 5 edits in trunk/Source

Properly handle when Test.html is not present in Production builds.

​https://bugs.webkit.org/show_bug.cgi?id=129506

Reviewed by Joseph Pecoraro.

Source/WebKit/mac:

  • WebCoreSupport/WebInspectorClient.mm:

(-[WebInspectorWindowController inspectorTestPagePath]):
(-[WebInspectorWindowController webView:decidePolicyForNavigationAction:request:frame:decisionListener:]):

Source/WebKit2:

  • UIProcess/WebInspectorProxy.cpp:

(WebKit::isMainOrTestInspectorPage):

  • UIProcess/mac/WebInspectorProxyMac.mm:

(WebKit::WebInspectorProxy::inspectorTestPageURL):

2:10 PM Changeset in webkit [164883] by psolanki@apple.com
  • 2 edits in trunk/Source/WebKit2

[iOS][WebKit2] Don't grab mach exception port on iOS
​https://bugs.webkit.org/show_bug.cgi?id=129505
<rdar://problem/15972749>

Reviewed by Anders Carlsson.

Don't grab the mach exception port on iOS so we get crash logs for web process and network
process.

  • Shared/ChildProcessProxy.cpp:

(WebKit::ChildProcessProxy::didFinishLaunching):

1:43 PM Changeset in webkit [164882] by Alan Bujtas
  • 11 edits
    2 adds in trunk

Add hiDPI support to DumpRenderTree/WebKitTestRunner without the need of reloading the test case.
​https://bugs.webkit.org/show_bug.cgi?id=129438

Reviewed by Simon Fraser.

'hidpi-' prefixed test cases now trigger 2x scale factor on the testing
offscreen window. It makes testing subpixel rendering and positioning possible.

Both the offscreen window's and WebKit's scaling are set accordingly.

Tools:

  • DumpRenderTree/mac/DumpRenderTree.mm:

(createWebViewAndOffscreenWindow): cleanup
(destroyWebViewAndOffscreenWindow): cleanup
(dumpRenderTree):
(changeWindowScaleIfNeeded):
(runTest):

  • DumpRenderTree/mac/DumpRenderTreeWindow.h:
  • DumpRenderTree/mac/DumpRenderTreeWindow.mm:

(-[DumpRenderTreeWindow initWithContentRect:styleMask:backing:defer:]): cleanup

  • WebKitTestRunner/PlatformWebView.h:
  • WebKitTestRunner/TestInvocation.cpp:

(WTR::changeWindowScaleIfNeeded):
(WTR::TestInvocation::invoke):

  • WebKitTestRunner/efl/PlatformWebViewEfl.cpp:

(WTR::PlatformWebView::changeWindowScaleIfNeeded):

  • WebKitTestRunner/gtk/PlatformWebViewGtk.cpp:

(WTR::PlatformWebView::changeWindowScaleIfNeeded):

  • WebKitTestRunner/ios/PlatformWebViewIOS.mm:

(WTR::PlatformWebView::changeWindowScaleIfNeeded):

  • WebKitTestRunner/mac/PlatformWebViewMac.mm:

(WTR::PlatformWebView::changeWindowScaleIfNeeded):

LayoutTests:

  • fast/borders/hidpi-simple-hairline-border-painting-expected.html: Added.
  • fast/borders/hidpi-simple-hairline-border-painting.html: Added.
1:12 PM Changeset in webkit [164881] by Brent Fulgham
  • 2 edits in trunk/Source/WebCore

[Mac, iOS] Incorrect ordering of construction arguments for out-of-band text elements
​https://bugs.webkit.org/show_bug.cgi?id=129501
<rdar://problem/16198587>

Reviewed by Eric Carlson.

  • html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::parseAttribute): Correct order of arguments
to PlatformTextTrack.

1:03 PM Changeset in webkit [164880] by oliver@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

REGRESSION(r164835): It broke 10 JSC stress test on 32 bit platforms
​https://bugs.webkit.org/show_bug.cgi?id=129488

Reviewed by Mark Lam.

Whoops, modify the right register.

  • jit/JITCall32_64.cpp:

(JSC::JIT::compileLoadVarargs):

1:00 PM Changeset in webkit [164879] by Brent Fulgham
  • 2 edits in trunk/Source/WebKit2

Unreviewed build fix after r164832.

  • UIProcess/ios/WebProcessProxyIOS.mm:

(WebKit::WebProcessProxy::updateProcessState): Switch from removed
'isValid' to 'state() != State::Running'

12:40 PM Changeset in webkit [164878] by fpizlo@apple.com
  • 3 edits in trunk/Source/JavaScriptCore

FTL should be able to call sin/cos directly on platforms where the intrinsic is busted
​https://bugs.webkit.org/show_bug.cgi?id=129503

Reviewed by Mark Lam.

  • ftl/FTLIntrinsicRepository.h:
  • ftl/FTLOutput.h:

(JSC::FTL::Output::doubleSin):
(JSC::FTL::Output::doubleCos):
(JSC::FTL::Output::intrinsicOrOperation):

11:32 AM Changeset in webkit [164877] by andersca@apple.com
  • 6 edits in trunk

Source/WebKit2: VisitedLinkProvider should keep track of processes, not pages
​https://bugs.webkit.org/show_bug.cgi?id=129497

Reviewed by Dan Bernstein.

Use a counted set of WebProcessProxy pointers instead of a set of pages.

  • UIProcess/VisitedLinkProvider.cpp:

(WebKit::VisitedLinkProvider::~VisitedLinkProvider):
Assert that m_processes is null.

(WebKit::VisitedLinkProvider::addProcess):
Add the process to the set.

(WebKit::VisitedLinkProvider::removeProcess):
Remove the process from the set.

  • UIProcess/VisitedLinkProvider.h:
  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::WebPageProxy):
If our process is currently running, add it to the visited link provider.
Otherwise it will be added in connectionWillOpen.

(WebKit::WebPageProxy::close):
If our process is running, remove it from the visited link provider.
Otherwise it's either crashed or not yet launched and will not have been added.

(WebKit::WebPageProxy::connectionWillOpen):
Add the process to the visited link provider.

(WebKit::WebPageProxy::resetStateAfterProcessExited):
Remove the process from the visited link provider.

Tools: Remove logging.

Reviewed by Dan Bernstein.

  • MiniBrowser/mac/WK2BrowserWindowController.m:

(-[WK2BrowserWindowController browsingContextController:decidePolicyForNavigationAction:decisionHandler:]):

11:19 AM Changeset in webkit [164876] by Bem Jones-Bey
  • 2 edits in trunk/Source/WebCore

Properly clear m_logicallyLastRun to remove use-after-free possibility
​https://bugs.webkit.org/show_bug.cgi?id=129489

Reviewed by David Hyatt.

A use-after-free issue was caught in Blink because m_logicallyLastRun
is not cleared when the item it points to is deleted. Clearing it
turns the use-after-free into a segfault, and prevents any future
use-after-frees from happening.

  • platform/text/BidiRunList.h:

(WebCore::BidiRunList<Run>::deleteRuns):

11:03 AM Changeset in webkit [164875] by bshafiei@apple.com
  • 1 copy in tags/Safari-537.75.3

New tag.

11:02 AM Changeset in webkit [164874] by bshafiei@apple.com
  • 5 edits in branches/safari-537.75-branch/Source

Versioning.

10:59 AM Changeset in webkit [164873] by jer.noble@apple.com
  • 2 edits in trunk/Source/WebCore

[Mac] Disambiguate calls to [[window contentView] layer] and [... setLayer:]
​https://bugs.webkit.org/show_bug.cgi?id=129491

Reviewed by Alexey Proskuryakov.

-[NSWindow contentView] returns a bare id, leading to ambiguity when calling
certain methods, including -layer and -setLayer:. Resolve the ambiguity by
explicitly casting the -contentView call to a NSView*.

  • platform/mac/WebVideoFullscreenController.mm:

(-[WebVideoFullscreenController setupVideoOverlay:]):
(-[WebVideoFullscreenController windowDidLoad]):
(-[WebVideoFullscreenController windowDidExitFullscreen]):

10:58 AM Changeset in webkit [164872] by ap@apple.com
  • 2 edits in trunk/LayoutTests

paragraphs with different directionality in textarea with unicode-bidi: plaintext are aligned the same
​https://bugs.webkit.org/show_bug.cgi?id=71194

  • platform/mac/fast/text/international/unicode-bidi-plaintext-in-textarea-expected.txt:

Updated Mac results for Mavericks (the original patch only updated Mountain Lion results).

10:44 AM Changeset in webkit [164871] by akling@apple.com
  • 3 edits
    2 adds in trunk

MouseEvent.offsetX/Y should just return 0,0 for simulated clicks.
<​https://webkit.org/b/129477>

Source/WebCore:

There's no need to compute the exact target-relative coordinates for
simulated mouse events, e.g those fired by HTMLElement.click().

The offsetX/Y properties are not supported by Firefox.

Test: fast/events/relative-offset-of-simulated-click.html

Reviewed by Alexey Proskuryakov.

  • dom/MouseRelatedEvent.cpp:

(WebCore::MouseRelatedEvent::offsetX):
(WebCore::MouseRelatedEvent::offsetY):

LayoutTests:

Add a test documenting the behavior of offsetX/Y on the simulated
mouse events that get sent by HTMLElement.click().

Reviewed by Alexey Proskuryakov.

  • fast/events/relative-offset-of-simulated-click-expected.txt: Added.
  • fast/events/relative-offset-of-simulated-click.html: Added.
10:36 AM Changeset in webkit [164870] by Brent Fulgham
  • 4 edits in branches/safari-537.75-branch

Source/WTF: Unreviewed build fix: Pass build environment to sub-processes via command line.
<rdar://problem/16197002>

  • WTF.vcxproj/build-generated-files.sh: Pass WEBKIT_LIBRARIES via argument, rather than

relying on global environment.

WebKitLibraries: Unreviewed build fix: Pass WEBKIT_LIBRARIES as argument to sub-process build steps.
<rdar://problem/16197002>

  • win/tools/vsprops/common.props: Pass WEBKIT_LIBRARIES directly to sub-processes, rather

than relying on global environment already having these values set.

10:34 AM Changeset in webkit [164869] by svillar@igalia.com
  • 7 edits
    4 adds in trunk

[CSS Grid Layout] Fix positioning grid items using named grid lines/areas
​https://bugs.webkit.org/show_bug.cgi?id=129372

Reviewed by Darin Adler.

Source/WebCore:

Our code was assuming that a <custom-ident> in
-webkit-grid-{column|row}-{start|end} and
-webkit-grid-{column|row} was always a grid area name. That's
wrong because the <custom-ident> could be also a explicitly named
grid line or the an implicitly named grid line created by a grid
area definition.

The style resolution code was not correct either. This patch fixes
it so it now matches the spec, which means that:

  • first we try to match any existing grid area.
  • then if there is a named grid line with the name

<custom-ident>-{start|end} for -webkit-grid-{column|row}-{start|end}
defined before the grid area then we use it instead of the grid
area.

  • otherwise if there is a named grid line we resolve to the first such line.
  • otherwise we treat it as 'auto'.

Fixing this uncovered a bug in GridPosition, we were not using the
name of the stored grid area to check if two GridPositions were
the same.

Tests: fast/css-grid-layout/grid-item-position-changed-dynamic.html

fast/css-grid-layout/named-grid-lines-with-named-grid-areas-resolution.html

  • css/StyleResolver.cpp:

(WebCore::gridLineDefinedBeforeGridArea): New function to check if
a given named grid line was defined before an implicit named grid
line created by a grid area definition.
(WebCore::StyleResolver::adjustNamedGridItemPosition): New
function that adjusts the position of a GridPosition parsed as a
grid area.
(WebCore::StyleResolver::adjustGridItemPosition): Use the new
function adjustNamedGridItemPosition to adjust the positions of
named grid lines.

  • css/StyleResolver.h:
  • rendering/RenderGrid.cpp:

(WebCore::RenderGrid::resolveNamedGridLinePositionFromStyle): Use GridPosition:: namespace.
(WebCore::RenderGrid::resolveGridPositionFromStyle): Ditto.
(WebCore::RenderGrid::resolveRowEndColumnEndNamedGridLinePositionAgainstOppositePosition): Ditto.

  • rendering/RenderGrid.h:
  • rendering/style/GridPosition.h:

(WebCore::GridPosition::adjustGridPositionForRowEndColumnEndSide): Moved from RenderGrid.cpp.
(WebCore::GridPosition::adjustGridPositionForSide): Ditto.
(WebCore::GridPosition::operator==): Use the named grid line to check equality.

LayoutTests:

Added a new test that checks that we correctly position grid items
using named grid lines, grid areas and also with the implicit
named grid lines created by grid areas.

I'm also importing a test from Blink that checks that we can
dynamically change the position of a grid item by changing the
name of the grid lines used to position it.

  • fast/css-grid-layout/grid-item-position-changed-dynamic-expected.txt:

Merged from Blink r153913 by <​jchaffraix@chromium.org>.

  • fast/css-grid-layout/grid-item-position-changed-dynamic.html: Ditto.
  • fast/css-grid-layout/named-grid-lines-with-named-grid-areas-resolution-expected.txt: Added.
  • fast/css-grid-layout/named-grid-lines-with-named-grid-areas-resolution.html: Added.
10:14 AM Changeset in webkit [164868] by zoltan@webkit.org
  • 4 edits in trunk/Source/WebCore

[CSS Shapes] Update line segments for ShapeInside only if the new line is wide enough
​https://bugs.webkit.org/show_bug.cgi?id=129461

Reviewed by David Hyatt.

Shape-inside can make a line only narrower than the original line width, thus we don't need
to update the line/shape segments in fitBelowFloats for every single line inside shape-inside.
This patch adds a helper function, which updates the line segments, furthermore it updates the
line segments only when the content would fit without the shape.

No new tests, no behavior change.

  • rendering/line/BreakingContextInlineHeaders.h: Use new helper.

(WebCore::updateSegmentsForShapes):

  • rendering/line/LineWidth.cpp: Use new helper.

(WebCore::LineWidth::updateLineSegment): Add new helper.
(WebCore::LineWidth::fitBelowFloats):

10:12 AM Changeset in webkit [164867] by mario.prada@samsung.com
  • 7 edits in trunk

paragraphs with different directionality in textarea with unicode-bidi: plaintext are aligned the same
​https://bugs.webkit.org/show_bug.cgi?id=71194

Reviewed by David Hyatt.

Merged from Blink r157263 by <​igor.o@sisa.samsung.com>
<​https://src.chromium.org/viewvc/blink?revision=157263&view=revision>

Source/WebCore:

In the css3-text spec (​http://www.w3.org/TR/css3-text/#bidi-linebox):

The start and end edges of a line box are determined by the inline
base direction of the line box. In most cases, this is given by
its containing block's computed ‘direction’. However if its
containing block has ‘unicode-bidi: plaintext’, the inline base
direction the line box must be determined by the base direction of
the bidi paragraph to which it belongs: that is, the bidi
paragraph for which the line box holds content.

This patch just implements the behavior described above modifying
updateLogicalWidthForAlignment and setting the bidi level of the
RootInlineBox.

  • rendering/RenderBlockFlow.h:
  • rendering/RenderBlockLineLayout.cpp:

(WebCore::RenderBlockFlow::updateLogicalWidthForAlignment):
(WebCore::RenderBlockFlow::computeInlineDirectionPositionsForSegment):
(WebCore::RenderBlockFlow::createLineBoxesFromBidiRuns):
(WebCore::RenderBlockFlow::layoutRunsAndFloatsInRange):
(WebCore::RenderBlockFlow::deleteEllipsisLineBoxes):
(WebCore::RenderBlockFlow::checkLinesForTextOverflow):
(WebCore::RenderBlockFlow::startAlignedOffsetForLine):

LayoutTests:

Modified expectations for two layout tests to make them conforming to the spec.

  • fast/text/international/unicode-bidi-plaintext-expected.html:
  • platform/gtk/fast/text/international/unicode-bidi-plaintext-in-textarea-expected.txt:
  • platform/mac-mountainlion/fast/text/international/unicode-bidi-plaintext-in-textarea-expected.txt:
10:09 AM Changeset in webkit [164866] by jochen@chromium.org
  • 3 edits
    2 adds in trunk

Update meta-referrer behavior for invalid policies
​https://bugs.webkit.org/show_bug.cgi?id=129475

Source/WebCore:

This patch aligns the behavior with the CSP 1.1 referrer directive,
where the fallback for an invalid value is the "never" policy.

Original patch from Mike West: ​https://src.chromium.org/viewvc/blink?view=rev&revision=165627

Reviewed by Alexey Proskuryakov.

Test: http/tests/security/referrer-policy-invalid.html

  • dom/Document.cpp:

(WebCore::Document::processReferrerPolicy):

LayoutTests:

Reviewed by Alexey Proskuryakov.

  • http/tests/security/referrer-policy-invalid-expected.txt: Added.
  • http/tests/security/referrer-policy-invalid.html: Added.
9:53 AM Changeset in webkit [164865] by dbates@webkit.org
  • 2 edits in trunk/Tools

[iOS] DumpRenderTree Perl Support may build against wrong SDK and toolchain
​https://bugs.webkit.org/show_bug.cgi?id=129163

Reviewed by David Kilzer.

Fixes an issue where DumpRenderTree Perl Support may build with the specified
iOS Simulator SDK and toolchain instead of using the OS X SDK and toolchain.
This issue manifests itself in a build failure when building the IPhoneSimulatorNotification
bundle.

Currently we build DumpRenderTree Perl Support with respect to the variable SDKROOT,
which is specified either explicitly when building with MAKE(1) or implicitly by
build-{dumprendertree, webkit}. Instead we should always build Perl Support using
the OS X SDK and toolchain since the built products are only applicable to OS X.

  • DumpRenderTree/ios/PerlSupport/Makefile:
9:46 AM Changeset in webkit [164864] by mhahnenberg@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

Fix !ENABLE(GGC) builds

  • heap/Heap.cpp:

(JSC::Heap::markRoots):
(JSC::Heap::gatherJSStackRoots): Also fix one of the names of the GC phases.

9:30 AM Changeset in webkit [164863] by dbates@webkit.org
  • 3 edits
    2 adds in trunk

SubresourceLoader::didFinishLoading() should not assert when a decode error occurs
​https://bugs.webkit.org/show_bug.cgi?id=127029

Reviewed by Darin Adler.

Source/WebCore:

SubresourceLoader::didFinishLoading() can be called for a resource (e.g. an image) that
failed to be decoded or, in the case of an image, whose estimated decoded size exceeds
the maximum decoded size (Settings::maximumDecodedImageSize()).

Test: fast/images/decoded-size-exceeds-max-decoded-size.html

  • loader/SubresourceLoader.cpp:

(WebCore::SubresourceLoader::didFinishLoading):

LayoutTests:

Added a test to ensure that we don't cause an assertion failure when an image fails
to load because of a decode error. In particular, the estimated decoded image size
exceeds the maximum decoded image size.

  • fast/images/decoded-size-exceeds-max-decoded-size-expected.txt: Added.
  • fast/images/decoded-size-exceeds-max-decoded-size.html: Added.
8:56 AM Changeset in webkit [164862] by mhahnenberg@apple.com
  • 3 edits in trunk/Source/JavaScriptCore

Clean up Heap::collect and Heap::markRoots
​https://bugs.webkit.org/show_bug.cgi?id=129464

Reviewed by Geoffrey Garen.

These functions have built up a lot of cruft recently.
We should do a bit of cleanup to make them easier to grok.

  • heap/Heap.cpp:

(JSC::Heap::finalizeUnconditionalFinalizers):
(JSC::Heap::gatherStackRoots):
(JSC::Heap::gatherJSStackRoots):
(JSC::Heap::gatherScratchBufferRoots):
(JSC::Heap::clearLivenessData):
(JSC::Heap::visitSmallStrings):
(JSC::Heap::visitConservativeRoots):
(JSC::Heap::visitCompilerWorklists):
(JSC::Heap::markProtectedObjects):
(JSC::Heap::markTempSortVectors):
(JSC::Heap::markArgumentBuffers):
(JSC::Heap::visitException):
(JSC::Heap::visitStrongHandles):
(JSC::Heap::visitHandleStack):
(JSC::Heap::traceCodeBlocksAndJITStubRoutines):
(JSC::Heap::converge):
(JSC::Heap::visitWeakHandles):
(JSC::Heap::clearRememberedSet):
(JSC::Heap::updateObjectCounts):
(JSC::Heap::resetVisitors):
(JSC::Heap::markRoots):
(JSC::Heap::copyBackingStores):
(JSC::Heap::deleteUnmarkedCompiledCode):
(JSC::Heap::collect):
(JSC::Heap::collectIfNecessaryOrDefer):
(JSC::Heap::suspendCompilerThreads):
(JSC::Heap::willStartCollection):
(JSC::Heap::deleteOldCode):
(JSC::Heap::flushOldStructureIDTables):
(JSC::Heap::flushWriteBarrierBuffer):
(JSC::Heap::stopAllocation):
(JSC::Heap::reapWeakHandles):
(JSC::Heap::sweepArrayBuffers):
(JSC::Heap::snapshotMarkedSpace):
(JSC::Heap::deleteSourceProviderCaches):
(JSC::Heap::notifyIncrementalSweeper):
(JSC::Heap::rememberCurrentlyExecutingCodeBlocks):
(JSC::Heap::resetAllocators):
(JSC::Heap::updateAllocationLimits):
(JSC::Heap::didFinishCollection):
(JSC::Heap::resumeCompilerThreads):

  • heap/Heap.h:
8:35 AM Changeset in webkit [164861] by commit-queue@webkit.org
  • 3 edits
    2 adds in trunk

ASSERTION FAILED: roundedIntPoint(rendererMappedResult) == roundedIntPoint(result) in WebCore::RenderGeometryMap::mapToContainer
​https://bugs.webkit.org/show_bug.cgi?id=119626

Source/WebCore:

SVGRenderSupport::mapLocalToContainer() was trying to apply transforms
in the incorrect order. Specifically, it would attempt to apply its
localToParentTransform before its localToBorderBoxTransform. This
was causing an ASSERT to fail when the computed transforms didn't
match up to those computed by RenderGeometryMap.

Backported from Blink: ​https://codereview.chromium.org/143363004

Patch by Martin Hodovan <​mhodovan@inf.u-szeged.hu> on 2014-02-28
Reviewed by Simon Fraser.

Test: svg/transforms/svg-geometry-crash.html

  • rendering/svg/SVGRenderSupport.cpp:

(WebCore::SVGRenderSupport::mapLocalToContainer):

LayoutTests:

Backported from Blink: ​https://codereview.chromium.org/143363004

Patch by Martin Hodovan <​mhodovan@inf.u-szeged.hu> on 2014-02-28
Reviewed by Simon Fraser.

  • svg/transforms/svg-geometry-crash-expected.txt: Added.
  • svg/transforms/svg-geometry-crash.html: Added.
7:52 AM Changeset in webkit [164860] by commit-queue@webkit.org
  • 11 edits
    2 deletes in trunk

Unreviewed, rolling out r164859.
​http://trac.webkit.org/changeset/164859
​https://bugs.webkit.org/show_bug.cgi?id=129483

caused WK1 crashes (DumpRenderTree) (Requested by zalan on
#webkit).

Tools:

  • DumpRenderTree/mac/DumpRenderTree.mm:

(createWebViewAndOffscreenWindow):
(dumpRenderTree):
(runTest):

  • DumpRenderTree/mac/DumpRenderTreeWindow.h:
  • DumpRenderTree/mac/DumpRenderTreeWindow.mm:

(-[DumpRenderTreeWindow initWithContentRect:styleMask:backing:defer:]):

  • WebKitTestRunner/PlatformWebView.h:
  • WebKitTestRunner/TestInvocation.cpp:

(WTR::TestInvocation::invoke):

  • WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
  • WebKitTestRunner/gtk/PlatformWebViewGtk.cpp:
  • WebKitTestRunner/ios/PlatformWebViewIOS.mm:
  • WebKitTestRunner/mac/PlatformWebViewMac.mm:

(-[WebKitTestRunnerWindow frameRespectingFakeOrigin]):

LayoutTests:

  • fast/borders/hidpi-simple-hairline-border-painting-expected.html: Removed.
  • fast/borders/hidpi-simple-hairline-border-painting.html: Removed.
7:24 AM Changeset in webkit [164859] by Alan Bujtas
  • 11 edits
    2 adds in trunk

Add hiDPI support to DumpRenderTree/WebKitTestRunner without the need of reloading the test case.
​https://bugs.webkit.org/show_bug.cgi?id=129438

Reviewed by Simon Fraser.

'hidpi-' prefixed test cases now trigger 2x scale factor on the testing
offscreen window. It makes testing subpixel rendering and positioning possible.

Both the offscreen window's and WebKit's scaling are set accordingly.

Tools:

  • DumpRenderTree/mac/DumpRenderTree.mm:

(createWebViewAndOffscreenWindow): unrelated cleanup.
(destroyWebViewAndOffscreenWindow): unrelated cleanup.
(dumpRenderTree):
(changeOffscreenWindowScaleIfNeeded):
(runTest):

  • DumpRenderTree/mac/DumpRenderTreeWindow.h:
  • DumpRenderTree/mac/DumpRenderTreeWindow.mm:

(-[DumpRenderTreeWindow initWithContentRect:styleMask:backing:defer:]): cleanup.

  • WebKitTestRunner/PlatformWebView.h:
  • WebKitTestRunner/TestInvocation.cpp:

(WTR::changeOffscreenWindowScaleIfNeeded):
(WTR::TestInvocation::invoke):

  • WebKitTestRunner/mac/PlatformWebViewMac.mm:

(WTR::PlatformWebView::changeOffscreenWindowScaleIfNeeded):

LayoutTests:

  • fast/borders/hidpi-simple-hairline-border-painting-expected.html: Added.
  • fast/borders/hidpi-simple-hairline-border-painting.html: Added.
6:34 AM Changeset in webkit [164858] by mihnea@adobe.com
  • 4 edits
    2 adds in trunk

[CSSRegions] ASSERTION FAILED: !m_regionsInvalidated in RenderFlowThread::regionAtBlockOffset
​https://bugs.webkit.org/show_bug.cgi?id=129371

Source/WebCore:

Reviewed by Andrei Bucur.

regionAtBlockOffset method is meant to be used after the flow thread validated
its region chain, otherwise we cannot rely on regionAtBlockOffset to determine correctly
the target region based on the offset in flow thread coordinates.

When the video element is collected in a named flow and displayed in a region, we ensure
that the region's decorations are taken into account to properly position the video.
However, we have to do that only if the named flow regions are validated.

I changed the method cachedRegionForCompositedLayer to check only the "cached" region
for a layer and not update the layer to region mappings before returning the region.
Because method cachedRegionForCompositedLayer does not update the region - layer mappings,
the code does not call regionAtBlockOffset in situations where the region chain is not validated yet.

Test: compositing/regions/video-in-overflow-region.html

  • rendering/RenderFlowThread.cpp:

(WebCore::RenderFlowThread::getLayerListForRegion): Make it a getter that relies on
cached information and refrain from updating the mappings. Make it const too.
(WebCore::RenderFlowThread::cachedRegionForCompositedLayer):

  • rendering/RenderFlowThread.h:

LayoutTests:

Reviewed Andrei Bucur.

  • compositing/regions/video-in-overflow-region-expected.txt: Added.
  • compositing/regions/video-in-overflow-region.html: Added.
2:06 AM Changeset in webkit [164857] by calvaris@igalia.com
  • 4 edits in trunk/Source/WebCore

[GTK] Finetune captions menu as per design team
​https://bugs.webkit.org/show_bug.cgi?id=129432

Reviewed by Eric Carlson.

Some design fine tuning of the captions dialog was required.

No new tests, current suffice.

  • Modules/mediacontrols/mediaControlsApple.js:

Added out class.

  • Modules/mediacontrols/mediaControlsGtk.js:

(ControllerGtk.prototype.buildCaptionMenu): Setting out class when
menu is going to show overlap the panel right border.

  • css/mediaControlsGtk.css:

(video::-webkit-media-controls-closed-captions-container.out):
Setting all borders to 5px;
(video::-webkit-media-controls-closed-captions-container h3):
Reduced font size and increasing top padding.
(video::-webkit-media-controls-closed-captions-container ul):
Increased top padding.

Note: See TracTimeline for information about the timeline view.