WebKit-https.git
4 years agoClean up how we reason about the states of AccessCases
fpizlo@apple.com [Mon, 11 Apr 2016 18:20:59 +0000 (18:20 +0000)]
Clean up how we reason about the states of AccessCases
https://bugs.webkit.org/show_bug.cgi?id=156454

Reviewed by Mark Lam.

Currently when we add an AccessCase to a PolymorphicAccess stub, we regenerate the stub.
That means that as we grow a stub to have N cases, we will do O(N^2) generation work. I want
to explore buffering AccessCases so that we can do O(N) generation work instead. But to
before I go there, I want to make sure that the statefulness of AccessCase makes sense. So,
I broke it down into three different states and added assertions about the transitions. I
also broke out a separate operation called AccessCase::commit(), which is the work that
cannot be buffered since there cannot be any JS effects between when the AccessCase was
created and when we do the work in commit().

This opens up a fairly obvious path to buffering AccessCases: add them to the list without
regenerating. Then when we do eventually trigger regeneration, those cases will get cloned
and generated automagically. This patch doesn't implement this technique yet, but gives us
an opportunity to independently test the scaffolding necessary to do it.

This is perf-neutral on lots of tests.

* bytecode/PolymorphicAccess.cpp:
(JSC::AccessGenerationResult::dump):
(JSC::AccessCase::clone):
(JSC::AccessCase::commit):
(JSC::AccessCase::guardedByStructureCheck):
(JSC::AccessCase::dump):
(JSC::AccessCase::generateWithGuard):
(JSC::AccessCase::generate):
(JSC::AccessCase::generateImpl):
(JSC::PolymorphicAccess::regenerateWithCases):
(JSC::PolymorphicAccess::regenerate):
(WTF::printInternal):
* bytecode/PolymorphicAccess.h:
(JSC::AccessCase::type):
(JSC::AccessCase::state):
(JSC::AccessCase::offset):
(JSC::AccessCase::viaProxy):
(JSC::AccessCase::callLinkInfo):
* bytecode/StructureStubInfo.cpp:
(JSC::StructureStubInfo::addAccessCase):
* bytecode/Watchpoint.h:
* dfg/DFGOperations.cpp:
* jit/Repatch.cpp:
(JSC::repatchGetByID):
(JSC::repatchPutByID):
(JSC::repatchIn):
* runtime/VM.cpp:
(JSC::VM::dumpRegExpTrace):
(JSC::VM::ensureWatchpointSetForImpureProperty):
(JSC::VM::registerWatchpointForImpureProperty):
(JSC::VM::addImpureProperty):
* runtime/VM.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199297 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDOMTokenList.contains() should not throw
cdumez@apple.com [Mon, 11 Apr 2016 16:50:44 +0000 (16:50 +0000)]
DOMTokenList.contains() should not throw
https://bugs.webkit.org/show_bug.cgi?id=156453

Reviewed by Ryosuke Niwa.

LayoutTests/imported/w3c:

Re-sync dom/nodes/Element-classlist.html with upstream @26308720.

* web-platform-tests/dom/nodes/Element-classlist-expected.txt:
* web-platform-tests/dom/nodes/Element-classlist.html:

Source/WebCore:

DOMTokenList.contains() should not throw if the input token is invalid:
https://github.com/whatwg/dom/commit/6d3076e3cbcba662489b272a718bc6b8c0082a74

We now return false in such cases, instead of throwing, which should be
safe with regards to backward compatibility.

No new tests, already covered by existing tests.

* html/DOMTokenList.cpp:
(WebCore::DOMTokenList::contains):
* html/DOMTokenList.h:
* html/DOMTokenList.idl:

LayoutTests:

Update existing layout tests now that DOMTokenList.contains() no longer
throws when called with an invalid token.

* fast/dom/HTMLElement/class-list-expected.txt:
* fast/dom/HTMLElement/class-list-quirks-expected.txt:
* fast/dom/HTMLElement/script-tests/class-list.js:
(shouldThrowDOMException): Deleted.
* fast/dom/HTMLOutputElement/dom-settable-token-list-expected.txt:
* fast/dom/HTMLOutputElement/script-tests/dom-settable-token-list.js:
(shouldThrowDOMException): Deleted.
* fast/dom/rel-list-expected.txt:
* fast/dom/rel-list.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199296 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRefactor RenderMathMLFraction layout to avoid using flexbox
fred.wang@free.fr [Mon, 11 Apr 2016 15:38:46 +0000 (15:38 +0000)]
Refactor RenderMathMLFraction layout to avoid using flexbox
https://bugs.webkit.org/show_bug.cgi?id=153917

Patch by Frederic Wang <fwang@igalia.com> on 2016-04-11
Reviewed by Sergio Villar Senin.

Source/WebCore:

Based on a patch by Alejandro G. Castro <alex@igalia.com>

Implement the layoutBlock method to handle the layout calculations
directly in the class. This also fixes parsing of absolute values for
linethickness attribute (e.g. 10px) and adds support for the AxisHeight
and FractionRuleThickness MATH parameters.

Test: mathml/opentype/fraction-line.html

* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::mathLineThickness): Use the thickness relative to the
default line thickness since that's really what is expected by mathml-line-fraction.html
* css/mathml.css: Remove flexbox properties for mfrac.
(mfrac): Deleted.
(mfrac > *): Deleted.
(mfrac[numalign="left"] > :first-child): Deleted.
(mfrac[numalign="right"] > :first-child): Deleted.
(mfrac[denomalign="left"] > :last-child): Deleted.
(mfrac[denomalign="right"] > :last-child): Deleted.
(mfrac > :first-child): Deleted.
(mfrac > :last-child): Deleted.
(mfrac): Deleted.
* rendering/mathml/RenderMathMLBlock.cpp: Introduce a helper function to retrieve the math
axis height.
(WebCore::RenderMathMLBlock::mathAxisHeight):
* rendering/mathml/RenderMathMLBlock.h: Declare mathAxisHeight.
* rendering/mathml/RenderMathMLFraction.cpp:
(WebCore::RenderMathMLFraction::RenderMathMLFraction):
(WebCore::RenderMathMLFraction::parseAlignmentAttribute): Helper function to parse the align
attribute.
(WebCore::RenderMathMLFraction::isValid): Helper function to verify whether the child list
is valid with respect to the MathML specificitation.
(WebCore::RenderMathMLFraction::numerator): Helper function to retrieve the numerator.
(WebCore::RenderMathMLFraction::denominator): Helper function to retrieve the denominator.
(WebCore::RenderMathMLFraction::updateFromElement): Use the FractionRuleThickness parameter
when avaiable to calculate the default linethickness.
Fix computation of linethickness for absolute values (e.g. 10px), the default linethickness
must not be involved for such values.
We no longer need to manage style of anonymous wrappers.
(WebCore::RenderMathMLFraction::unembellishedOperator): Use the helper function and we no
longer care about anonymous wrappers.
(WebCore::RenderMathMLFraction::computePreferredLogicalWidths): Implement this function
without using flexbox.
(WebCore::RenderMathMLFraction::horizontalOffset): Helper function to get the horizontal
offsets of children depending of the alignment.
(WebCore::RenderMathMLFraction::layoutBlock): Implement this function without using flexbox.
(WebCore::RenderMathMLFraction::paint): Do not paint if the fraction is invalid. Use helper
function. Use the width of the renderer (instead of the one of the denominator) as the
length of the fraction bar.
(WebCore::RenderMathMLFraction::firstLineBaseline): Use the helper functions to get children
and axis height.
(WebCore::RenderMathMLFraction::paintChildren): Temporary function to remove in a
follow-up patch.
(WebCore::RenderMathMLFraction::fixChildStyle): Deleted. We no longer need to manage style
of anonymous wrappers.
(WebCore::RenderMathMLFraction::addChild): Deleted. We no longer need to manage
anonymous wrappers.
(WebCore::RenderMathMLFraction::styleDidChange): We no longer need to manage style of
anonymous wrappers.
(WebCore::RenderMathMLFraction::layout): Deleted.
* rendering/mathml/RenderMathMLFraction.h: Replace lineThickness with relativeLineThickness,
as needed by the accessibility code. Update function and members declarations.

LayoutTests:

* TestExpectations: No longer skip mathml/presentation/fractions-positions.html
* mathml/opentype/fraction-line-expected.html: Added. New test to verify AxisHeight and
FractionRuleThickness parameters.
* mathml/opentype/fraction-line.html: Added. New test to verify axis height and rule
thickness parameters.
* mathml/presentation/fractions-linethickness-expected.html: Adjust the test to be sure that
the default rule thickness is 1px.
* mathml/presentation/fractions-linethickness.html: Adjust the test to be sure that the
default rule thickness is 1px.
* platform/gtk/mathml/presentation/roots-expected.txt: Update reference to take into account
changes in the render tree.
* platform/ios-simulator/mathml/presentation/roots-expected.txt: Ditto
* platform/mac/TestExpectations: Mark fraction-line and fractions-linethickness as
possibly failing since these tests require Latin Modern Math to work reliably.
* platform/ios-simulator/TestExpectations: Ditto

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199295 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r199290.
commit-queue@webkit.org [Mon, 11 Apr 2016 12:46:30 +0000 (12:46 +0000)]
Unreviewed, rolling out r199290.
https://bugs.webkit.org/show_bug.cgi?id=156465

broke 300 tests (Requested by mcatanzaro on #webkit).

Reverted changeset:

"Merge CG ImageSource and non CG ImageSource implementation in
one file"
https://bugs.webkit.org/show_bug.cgi?id=155456
http://trac.webkit.org/changeset/199290

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199294 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRefactor RenderMathMLUnderOver layout functions to avoid using flexbox
fred.wang@free.fr [Mon, 11 Apr 2016 12:11:18 +0000 (12:11 +0000)]
Refactor RenderMathMLUnderOver layout functions to avoid using flexbox
https://bugs.webkit.org/show_bug.cgi?id=153742

Patch by Frederic Wang <fwang@igalia.com> on 2016-04-11
Reviewed by Sergio Villar Senin.

Source/WebCore:

Based on a patch by Javier Fernandez <jfernandez@igalia.com>

Refactor the UnderOver renderer to use its own layoutBlock method that
does all the layout calculations without considering the flexbox
restrictions.

* css/mathml.css:
(mo, mfrac, munder, mover, munderover): Delete the underover elements from the line defining
the column direction.
(munder, mover, munderover): Deleted. This flexbox property is no longer needed.
(mover > :last-child, munderover > :last-child): Deleted. This flexbox property is no longer
needed.
* rendering/mathml/RenderMathMLUnderOver.cpp:
(WebCore::RenderMathMLUnderOver::firstLineBaseline): Use ascentForChild.
(WebCore::RenderMathMLUnderOver::computeOperatorsHorizontalStretch): Avoid stretching
operators that are not stretchy.
(WebCore::RenderMathMLUnderOver::isValid): Helper function to ensure that the child list is
valid with respect to the MathML specification.
(WebCore::RenderMathMLUnderOver::base): Added. Helper function.
(WebCore::RenderMathMLUnderOver::under): Added. Helper function.
(WebCore::RenderMathMLUnderOver::over): Added. Helper function.
(WebCore::RenderMathMLUnderOver::computePreferredLogicalWidths): Added.
The preferred width is the maximum preferred width of the base, under and over scripts.
(WebCore::RenderMathMLUnderOver::horizontalOffset): Added, helper to calculate the
horizontal position of children (horizontally centered).
(WebCore::RenderMathMLUnderOver::layoutBlock): Added, it lays out the base, underscript and
overscript. It calculates the exact logical width, which may differ from the preferred width when
one child contains stretchy operators. It later sets the locations of children accordingly
and sets the heigth of the render element.
(WebCore::RenderMathMLUnderOver::paintChildren): Added, we have to use the usual traverse
instead of the one that comes from the flexbox. This will be removed in a follow-up patch.
(WebCore::RenderMathMLUnderOver::layout): Deleted.
* rendering/mathml/RenderMathMLUnderOver.h: Added new functions definitions.

LayoutTests:

Apply some small adjustments to the expectations of MathML tests after
the refactoring of RenderMathMLUnderOver. We also add a test for
non-stretchy horizontal operators in underover.

* platform/gtk/mathml/opentype/horizontal-expected.png:
* platform/gtk/mathml/opentype/horizontal-expected.txt:
* platform/gtk/mathml/opentype/opentype-stretchy-horizontal-expected.png:
* platform/gtk/mathml/opentype/opentype-stretchy-horizontal-expected.txt:
* platform/mac/mathml/opentype/opentype-stretchy-horizontal-expected.png:
* platform/mac/mathml/opentype/opentype-stretchy-horizontal-expected.txt:
* mathml/mn-as-list-item-assert.html: Move the test description out of the invalid munderover
so that it is still displayed.
* mathml/mn-as-list-item-assert-expected.txt: Update the text expectation.
* mathml/presentation/underover-nonstretchy-horizontal.html: Ensure that nonstretchy horizontal operators are not stretched in munderover.
* mathml/presentation/underover-nonstretchy-horizontal-expected.html: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199293 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Rework the theming code for GTK+ 3.20
carlosgc@webkit.org [Mon, 11 Apr 2016 10:59:03 +0000 (10:59 +0000)]
[GTK] Rework the theming code for GTK+ 3.20
https://bugs.webkit.org/show_bug.cgi?id=156333

Reviewed by Michael Catanzaro.

.:

Add a manual test to check how themed elements are rendered.

* ManualTests/gtk/theme.html: Added.

Source/WebCore:

During the 3.19 GTK+ release cycle, the GTK+ css system was reworked, making themes and programs rendering
themed widgets, incompatible with the new system. We were trying to fix our rendering every time GTK+ broke
something, but we were just changing whatever it was needed to make our rendering look like current GTK+ with
the default theme Adwaita. This means that our rendering will be broken for other themes or that changes in
Adwaita can break our rendering. This solution was good enough to ensure WebKitGTK+ 2.12 looked good with GTK+
3.20, but it doesn't work in the long term. We need to ensure that our theming code honors the new GTK+ CSS
properties (max-width, min-width, margin, padding, border, ...) in all the cases, not only the cases where
Adwaita uses them like we currently do.
This patch splits all rendering methods to keep the current code for previous GTK+ versions and adds new code
for GTK+ >= 3.20 using the new RenderThemeGadget classes. This makes the code easier to read, since there aren't
ifdef blocks in the functions, and we ensure we don't break previous rendering.

* PlatformGTK.cmake: Add new files to compilation.
* html/shadow/SpinButtonElement.cpp:
(WebCore::SpinButtonElement::defaultEventHandler): Check the button layout used by the theme to decide the
current buttons state.
* platform/gtk/RenderThemeGadget.cpp: Added.
(WebCore::RenderThemeGadget::create):
(WebCore::createStyleContext):
(WebCore::appendElementToPath):
(WebCore::RenderThemeGadget::RenderThemeGadget):
(WebCore::RenderThemeGadget::~RenderThemeGadget):
(WebCore::RenderThemeGadget::marginBox):
(WebCore::RenderThemeGadget::borderBox):
(WebCore::RenderThemeGadget::paddingBox):
(WebCore::RenderThemeGadget::contentsBox):
(WebCore::RenderThemeGadget::color):
(WebCore::RenderThemeGadget::backgroundColor):
(WebCore::RenderThemeGadget::minimumSize):
(WebCore::RenderThemeGadget::preferredSize):
(WebCore::RenderThemeGadget::render):
(WebCore::RenderThemeGadget::renderFocus):
(WebCore::RenderThemeBoxGadget::RenderThemeBoxGadget):
(WebCore::RenderThemeTextFieldGadget::RenderThemeTextFieldGadget):
(WebCore::RenderThemeTextFieldGadget::minimumSize):
(WebCore::RenderThemeToggleGadget::RenderThemeToggleGadget):
(WebCore::RenderThemeToggleGadget::render):
(WebCore::RenderThemeArrowGadget::RenderThemeArrowGadget):
(WebCore::RenderThemeArrowGadget::render):
(WebCore::RenderThemeIconGadget::RenderThemeIconGadget):
(WebCore::RenderThemeIconGadget::gtkIconSizeForPixelSize):
(WebCore::RenderThemeIconGadget::render):
(WebCore::RenderThemeIconGadget::minimumSize):
* platform/gtk/RenderThemeGadget.h: Added.
(WebCore::RenderThemeGadget::context):
* rendering/RenderTheme.h:
(WebCore::RenderTheme::innerSpinButtonLayout): Added this method to allow themes use a different layout for the
buttons.
* rendering/RenderThemeGtk.cpp:
(WebCore::themeChangedCallback): Just moved this code to a common place.
(WebCore::RenderThemeGtk::RenderThemeGtk): Initialize the theme monitor in the constructor.
(WebCore::createStyleContext): Remove the render parts that are specific to GTK+ 3.20.
(WebCore::RenderThemeGtk::adjustRepaintRect): Moved inside a GTK+ < 3.20 ifdef block.
(WebCore::themePartStateFlags): Helper function to get the GtkStateFlags of a theme part for a given RenderObject.
(WebCore::shrinkToMinimumSizeAndCenterRectangle): Move this common code to a helper function.
(WebCore::setToggleSize):
(WebCore::paintToggle):
(WebCore::RenderThemeGtk::paintButton):
(WebCore::RenderThemeGtk::popupInternalPaddingBox):
(WebCore::RenderThemeGtk::paintMenuList):
(WebCore::RenderThemeGtk::adjustTextFieldStyle): For GTK+ 3.20 we need to ensure a minimum size for spin buttons,
so if the text field is for a spin button, we adjust the desired size here.
(WebCore::RenderThemeGtk::paintTextField): In GTK+ 3.20 the CSS gadgets used to render spin buttons are
different, so we check here if this is the entry of a spin button to use the right gadgets.
(WebCore::adjustSearchFieldIconStyle):
(WebCore::RenderThemeGtk::paintTextArea):
(WebCore::RenderThemeGtk::adjustSearchFieldResultsButtonStyle):
(WebCore::RenderThemeGtk::paintSearchFieldResultsButton):
(WebCore::RenderThemeGtk::adjustSearchFieldResultsDecorationPartStyle):
(WebCore::RenderThemeGtk::adjustSearchFieldCancelButtonStyle):
(WebCore::paintSearchFieldIcon):
(WebCore::RenderThemeGtk::paintSearchFieldResultsDecorationPart):
(WebCore::RenderThemeGtk::paintSearchFieldCancelButton):
(WebCore::centerRectVerticallyInParentInputElement): Moved inside a GTK+ < 3.20 ifdef block.
(WebCore::RenderThemeGtk::paintSliderTrack):
(WebCore::RenderThemeGtk::adjustSliderThumbSize):
(WebCore::RenderThemeGtk::paintSliderThumb):
(WebCore::RenderThemeGtk::progressBarRectForBounds): Ensure a minimum size of progress bars in GTK+ 3.20.
(WebCore::RenderThemeGtk::paintProgressBar):
(WebCore::RenderThemeGtk::innerSpinButtonLayout): Use an horizontal layout for spin buttons.
(WebCore::RenderThemeGtk::adjustInnerSpinButtonStyle):
(WebCore::RenderThemeGtk::paintInnerSpinButton):
(WebCore::styleColor):
(WebCore::RenderThemeGtk::paintMediaButton):
* rendering/RenderThemeGtk.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199292 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoImplement functional :host() pseudo class
antti@apple.com [Mon, 11 Apr 2016 09:13:06 +0000 (09:13 +0000)]
Implement functional :host() pseudo class
https://bugs.webkit.org/show_bug.cgi?id=156397
<rdar://problem/25621445>

Reviewed by Darin Adler.

Source/WebCore:

We already support :host. Add functional syntax too.

* css/CSSGrammar.y.in:

    Parse functional :host().

* css/CSSParser.cpp:
(WebCore::CSSParser::detectFunctionTypeToken):
* css/CSSParserValues.cpp:
(WebCore::CSSParserSelector::parsePseudoClassHostFunctionSelector):
* css/CSSParserValues.h:
* css/ElementRuleCollector.cpp:
(WebCore::ElementRuleCollector::matchedRuleList):
(WebCore::ElementRuleCollector::addMatchedRule):

    Factor some shared code here.

(WebCore::ElementRuleCollector::matchHostPseudoClassRules):

    Instead of using the generic paths use a :host specific code path for matching.
    This makes it easier to avoid :host matching when it shouldn't.

(WebCore::ElementRuleCollector::collectMatchingRulesForList):
* css/ElementRuleCollector.h:
* css/RuleSet.cpp:
(WebCore::computeMatchBasedOnRuleHash):

    :host is always handled by the special matching path.

* css/SelectorChecker.cpp:
(WebCore::SelectorChecker::match):
(WebCore::SelectorChecker::matchHostPseudoClass):

    Add a function specifically for checking :host. In always fails on the normal code paths.
    Check the argument selector if provided.

(WebCore::hasScrollbarPseudoElement):
* css/SelectorChecker.h:

LayoutTests:

Enable, fix and expand the test.

* fast/shadow-dom/css-scoping-shadow-host-functional-rule.html:
* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199291 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMerge CG ImageSource and non CG ImageSource implementation in one file
commit-queue@webkit.org [Mon, 11 Apr 2016 07:29:33 +0000 (07:29 +0000)]
Merge CG ImageSource and non CG ImageSource implementation in one file
https://bugs.webkit.org/show_bug.cgi?id=155456

Patch by Said Abou-Hallawa <sabouhallawa@apple,com> on 2016-04-11
Reviewed by Darin Adler.
Source/WebCore:

ImageSource for CG and CG code paths look very similar. All the platform
specific code can be moved to ImageDecoder classes for CG and non CG. And
we can have the ImageSource be platform independent and we get rid of
ImageSourceCG.cpp.

Test: fast/images/image-subsampling.html

* CMakeLists.txt:
* PlatformAppleWin.cmake:
* PlatformMac.cmake:
* WebCore.xcodeproj/project.pbxproj:
Delete ImageSourceCG.cpp form all make files and add ImageSource.cpp to
CMakeLists.txt.

* platform/Cursor.cpp:
(WebCore::determineHotSpot):
* platform/graphics/BitmapImage.cpp:
(WebCore::BitmapImage::hotSpot):
(WebCore::BitmapImage::getHotSpot): Deleted.
* platform/graphics/BitmapImage.h:
* platform/graphics/Image.h:
(WebCore::Image::hotSpot):
(WebCore::Image::getHotSpot): Deleted.
Rename getHotSpot() to hotSpot() and change it to return Optional<IntPoint>.

* platform/graphics/ImageSource.cpp:
(WebCore::ImageSource::~ImageSource): Remove clear(true) call. It does nothing.
(WebCore::ImageSource::clearFrameBufferCache): A wrapper which calls ImageDecoder::clearFrameBufferCache().
(WebCore::ImageSource::clear): Calls clearFrameBufferCache() which will do nothing for CG.

(WebCore::ImageSource::ensureDecoderIsCreated): Change SharedBuffer* to
const SharedBuffer& and remove the call to ImageDecoder::setMaxNumPixels().
The value of const static int CG ImageDecoder::m_maxNumPixels will be set
based on IMAGE_DECODER_DOWN_SAMPLING.

(WebCore::ImageSource::setData): Pass SharedBuffer& to the underlying functions.

(WebCore::ImageSource::calculateMaximumSubsamplingLevel): Returns the maximum
subsampling level allowed for an image.

(WebCore::ImageSource::subsamplingLevelForScale): Converts from a scale to
SubsamplingLevel taking into consideration the maximumSubsamplingLevel for
a particular image.

(WebCore::ImageSource::bytesDecodedToDetermineProperties): Returns the number
of encoded bytes which can determine the image properties. For non CG it's
zero. For CG it is a maximum value which can be corrected later.

(WebCore::ImageSource::isSizeAvailable):
(WebCore::ImageSource::sizeRespectingOrientation):
(WebCore::ImageSource::frameCount):
(WebCore::ImageSource::repetitionCount):
(WebCore::ImageSource::filenameExtension):
(WebCore::ImageSource::getHotSpot):
(WebCore::ImageSource::frameIsCompleteAtIndex):
(WebCore::ImageSource::frameHasAlphaAtIndex):
(WebCore::ImageSource::allowSubsamplingOfFrameAtIndex):
(WebCore::ImageSource::frameSizeAtIndex):
(WebCore::ImageSource::frameBytesAtIndex):
(WebCore::ImageSource::frameDurationAtIndex):
(WebCore::ImageSource::orientationAtIndex):
(WebCore::ImageSource::createFrameImageAtIndex):
These are wrappers for the ImageDecoder APIs. The purpose of these functions
is to ensure the ImageDecoder is created.

(WebCore::ImageSource::dump): Called from BitmapImage::dump().

(WebCore::ImageSource::getHotSpot): Deleted.

* platform/graphics/ImageSource.h:
(WebCore::ImageSource::setAllowSubsampling): Called from BitmapImage::setAllowSubsampling().

(WebCore::ImageSource::maxPixelsPerDecodedImage): Deleted.
(WebCore::ImageSource::setMaxPixelsPerDecodedImage): Deleted.
Setting maxPixelsPerDecodedImage was moved to the non CG ImageDecoder.

* platform/graphics/cg/ImageDecoderCG.cpp:
(WebCore::ImageDecoder::setData): Change SharedBuffer* to SharedBuffer&.

(WebCore::ImageDecoder::subsamplingLevelForScale): Deleted.
The code was moved to ImageSource::subsamplingLevelForScale().

* platform/graphics/cg/ImageDecoderCG.h:
(WebCore::ImageDecoder::create): Make the prototype of this function
suitable for CG and non CG cases.
(WebCore::ImageDecoder::clearFrameBufferCache): Empty functions for CG.

* platform/graphics/cg/ImageSourceCG.cpp: Removed.

* platform/image-decoders/ImageDecoder.cpp:
(WebCore::ImageDecoder::frameIsCompleteAtIndex): A mew function to return
whether the frame decoding is complete or not.

(WebCore::ImageDecoder::frameHasAlphaAtIndex): Simplify the logic.

(WebCore::ImageDecoder::frameDurationAtIndex): The code was moved from
ImageSource::frameDurationAtIndex() in ImageSource.cpp.

(WebCore::ImageDecoder::createFrameImageAtIndex): The code was moved from
ImageSource::createFrameImageAtIndex() in ImageSource.cpp.

* platform/image-decoders/ImageDecoder.h:
(WebCore::ImageDecoder::ImageDecoder): Initialize the members in class.
(WebCore::ImageDecoder::~ImageDecoder): Fix the braces style.
(WebCore::ImageDecoder::setData): Change the type of the argument from
SharedBuffer* to SharedBuffer&.
(WebCore::ImageDecoder::frameSizeAtIndex): Add the argument SubsamplingLevel
so it can have the same prototype as CG.
(WebCore::ImageDecoder::orientationAtIndex): Rename it to the same of CG.

(WebCore::ImageDecoder::allowSubsamplingOfFrameAtIndex):
(WebCore::ImageDecoder::bytesDecodedToDetermineProperties):
(WebCore::ImageDecoder::subsamplingLevelForScale): Add these functions
and return the default values so we do not have to add directive compiled
non CG blocks in ImageSource.cpp.

(WebCore::ImageDecoder::hotSpot): Return Optional<IntPoint>.

(WebCore::ImageDecoder::orientation): Deleted.
(WebCore::ImageDecoder::setMaxNumPixels): Deleted.

* platform/image-decoders/bmp/BMPImageDecoder.cpp:
(WebCore::BMPImageDecoder::setData):
* platform/image-decoders/bmp/BMPImageDecoder.h:
* platform/image-decoders/gif/GIFImageDecoder.cpp:
(WebCore::GIFImageDecoder::setData):
(WebCore::GIFImageDecoder::decode):
* platform/image-decoders/gif/GIFImageDecoder.h:
* platform/image-decoders/gif/GIFImageReader.h:
(GIFImageReader::setData):
* platform/image-decoders/ico/ICOImageDecoder.cpp:
(WebCore::ICOImageDecoder::setData):
Use reference SharedBuffer instead of pointer SharedBuffer.

(WebCore::ICOImageDecoder::hotSpot):
(WebCore::ICOImageDecoder::hotSpotAtIndex):
Change hotSpot() to return Optional<IntPoint>.
* platform/image-decoders/ico/ICOImageDecoder.h:

(WebCore::ICOImageDecoder::setDataForPNGDecoderAtIndex):
Pass reference SharedBuffer instead of pointer SharedBuffer.

Source/WebKit2:

* UIProcess/API/efl/EwkView.cpp:
(EwkView::setCursor):
Replace the call to Image::getHotSpot() by Image::hotSpot().

LayoutTests:

Add a test for image sub-sampling. The image subsampling is enabled by
default for iOS platform only. But it can be explicitly enabled through
the setting ImageSubsamplingEnabled.

* fast/images/image-subsampling-expected.html: Added.
* fast/images/image-subsampling.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199290 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[CMake] Make FOLDER property INHERITED
commit-queue@webkit.org [Mon, 11 Apr 2016 07:28:42 +0000 (07:28 +0000)]
[CMake] Make FOLDER property INHERITED
https://bugs.webkit.org/show_bug.cgi?id=156460

Patch by Fujii Hironori <Hironori.Fujii@jp.sony.com> on 2016-04-11
Reviewed by Brent Fulgham.

.:

Some CMake targets are not setting the FOLDER property. This causes the
generated projects to be displayed in the top-level folder of the solution.

Making the FOLDER property INHERITED ensures that all the targets
are placed in their proper directories.

* Source/cmake/OptionsCommon.cmake:
Define FOLDER property as a inherited property.
* Source/cmake/WebKitMacros.cmake:
Do not set FOLDER target property.

Source/bmalloc:

* CMakeLists.txt:
Set FOLDER property as a directory property not a target property

Source/JavaScriptCore:

* CMakeLists.txt:
* shell/CMakeLists.txt:
* shell/PlatformWin.cmake:
Set FOLDER property as a directory property not a target property

Source/ThirdParty/ANGLE:

* CMakeLists.txt:
Set FOLDER property as a directory property not a target property

Source/WebCore:

* CMakeLists.txt:
Set FOLDER property as a directory property not a target property

Source/WebKit:

* CMakeLists.txt:
* PlatformWin.cmake:
Set FOLDER property as a directory property not a target property

Source/WebKit2:

* CMakeLists.txt:
Set FOLDER property as a directory property not a target property

Source/WTF:

* CMakeLists.txt:
Set FOLDER directory property

Tools:

* CMakeLists.txt:
* DumpRenderTree/CMakeLists.txt:
* DumpRenderTree/PlatformWin.cmake:
* ImageDiff/CMakeLists.txt:
* MiniBrowser/efl/CMakeLists.txt:
* MiniBrowser/gtk/CMakeLists.txt:
* MiniBrowser/win/CMakeLists.txt:
Set FOLDER property as a directory property not a target property

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199289 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Tools] correctly check for braces in multiline branches in macro definition
commit-queue@webkit.org [Mon, 11 Apr 2016 00:10:17 +0000 (00:10 +0000)]
[Tools] correctly check for braces in multiline branches in macro definition
https://bugs.webkit.org/show_bug.cgi?id=156441

Patch by Caitlin Potter <caitp@igalia.com> on 2016-04-10
Reviewed by Darin Adler.

Prevents emitting whitespace/braces warning for code like the
following:

```
#define MACRO(x) \
    if (x) { \
        doTheThing(); \
        continue; \
    }
```

* Scripts/webkitpy/style/checkers/cpp.py:
(check_braces):
* Scripts/webkitpy/style/checkers/cpp_unittest.py:
(WebKitStyleTest.test_line_breaking):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199288 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix the build.
weinig@apple.com [Sun, 10 Apr 2016 19:04:41 +0000 (19:04 +0000)]
Fix the build.

* platform/graphics/avfoundation/objc/MediaPlaybackTargetPickerMac.mm:
(WebCore::MediaPlaybackTargetPickerMac::showPlaybackTargetPicker):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199287 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove support for custom target picker actions
weinig@apple.com [Sun, 10 Apr 2016 18:48:37 +0000 (18:48 +0000)]
Remove support for custom target picker actions
<rdar://problem/24987783>
https://bugs.webkit.org/show_bug.cgi?id=156434

Reviewed by Eric Carlson.

This mostly entailed rolling out r197429 and r197569.

Source/WebCore:

* Modules/mediasession/WebMediaSessionManager.cpp:
(WebCore::WebMediaSessionManager::removeAllPlaybackTargetPickerClients):
(WebCore::WebMediaSessionManager::showPlaybackTargetPicker):
(WebCore::WebMediaSessionManager::clientStateDidChange):
(WebCore::WebMediaSessionManager::externalOutputDeviceAvailableDidChange):
(WebCore::WebMediaSessionManager::configureNewClients):
(WebCore::WebMediaSessionManager::customPlaybackActionSelected): Deleted.
* Modules/mediasession/WebMediaSessionManager.h:
* Modules/mediasession/WebMediaSessionManagerClient.h:
* dom/Document.cpp:
(WebCore::Document::removePlaybackTargetPickerClient):
(WebCore::Document::showPlaybackTargetPicker):
(WebCore::Document::playbackTargetPickerClientStateDidChange):
(WebCore::Document::setShouldPlayToPlaybackTarget):
(WebCore::Document::customPlaybackActionSelected): Deleted.
* dom/Document.h:
* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::enqueuePlaybackTargetAvailabilityChangedEvent):
(WebCore::HTMLMediaElement::setShouldPlayToPlaybackTarget):
(WebCore::HTMLMediaElement::webkitCurrentPlaybackTargetIsWireless):
(WebCore::HTMLMediaElement::customPlaybackActionSelected): Deleted.
(WebCore::HTMLMediaElement::playbackTargetPickerCustomActionName): Deleted.
* html/HTMLMediaElement.h:
* html/MediaElementSession.cpp:
(WebCore::MediaElementSession::showPlaybackTargetPicker):
(WebCore::MediaElementSession::hasWirelessPlaybackTargets):
(WebCore::MediaElementSession::setShouldPlayToPlaybackTarget):
(WebCore::MediaElementSession::mediaStateDidChange):
(WebCore::MediaElementSession::customPlaybackActionSelected): Deleted.
* html/MediaElementSession.h:
* page/ChromeClient.h:
* page/Page.cpp:
(WebCore::Page::removePlaybackTargetPickerClient):
(WebCore::Page::showPlaybackTargetPicker):
(WebCore::Page::setShouldPlayToPlaybackTarget):
(WebCore::Page::ensureTestTrigger):
(WebCore::Page::customPlaybackActionSelected): Deleted.
* page/Page.h:
(WebCore::Page::testTrigger):
* platform/audio/PlatformMediaSession.h:
(WebCore::PlatformMediaSessionClient::canPlayToWirelessPlaybackTarget):
(WebCore::PlatformMediaSessionClient::isPlayingToWirelessPlaybackTarget):
(WebCore::PlatformMediaSessionClient::setShouldPlayToPlaybackTarget):
(WebCore::PlatformMediaSessionClient::customPlaybackActionSelected): Deleted.
* platform/graphics/MediaPlaybackTargetClient.h:
* platform/graphics/MediaPlaybackTargetPicker.cpp:
(WebCore::MediaPlaybackTargetPicker::pendingActionTimerFired):
(WebCore::MediaPlaybackTargetPicker::addPendingAction):
(WebCore::MediaPlaybackTargetPicker::showPlaybackTargetPicker):
* platform/graphics/MediaPlaybackTargetPicker.h:
(WebCore::MediaPlaybackTargetPicker::availableDevicesDidChange):
(WebCore::MediaPlaybackTargetPicker::currentDeviceDidChange):
(WebCore::MediaPlaybackTargetPicker::Client::customPlaybackActionSelected): Deleted.
(WebCore::MediaPlaybackTargetPicker::customPlaybackActionSelected): Deleted.
* platform/graphics/avfoundation/objc/MediaPlaybackTargetPickerMac.h:
* platform/graphics/avfoundation/objc/MediaPlaybackTargetPickerMac.mm:
(WebCore::MediaPlaybackTargetPickerMac::devicePicker):
(WebCore::MediaPlaybackTargetPickerMac::showPlaybackTargetPicker):
* platform/mac/WebVideoFullscreenInterfaceMac.h:
* platform/mac/WebVideoFullscreenInterfaceMac.mm:
(WebCore::WebVideoFullscreenInterfaceMac::preparedToReturnToInline):
(WebCore::WebVideoFullscreenInterfaceMac::setVideoDimensions):
(WebCore::WebVideoFullscreenInterfaceMac::setExternalPlayback): Deleted.
* platform/mock/MediaPlaybackTargetPickerMock.cpp:
(WebCore::MediaPlaybackTargetPickerMock::timerFired):
(WebCore::MediaPlaybackTargetPickerMock::showPlaybackTargetPicker):
* platform/mock/MediaPlaybackTargetPickerMock.h:
* platform/spi/cocoa/AVKitSPI.h:

Source/WebKit/mac:

* WebCoreSupport/WebChromeClient.h:
* WebCoreSupport/WebChromeClient.mm:
(WebChromeClient::removePlaybackTargetPickerClient):
(WebChromeClient::showPlaybackTargetPicker):
* WebView/WebMediaPlaybackTargetPicker.h:
* WebView/WebMediaPlaybackTargetPicker.mm:
(WebMediaPlaybackTargetPicker::removePlaybackTargetPickerClient):
(WebMediaPlaybackTargetPicker::showPlaybackTargetPicker):
(WebMediaPlaybackTargetPicker::playbackTargetPickerClientStateDidChange):
(WebMediaPlaybackTargetPicker::setShouldPlayToPlaybackTarget):
(WebMediaPlaybackTargetPicker::invalidate):
(WebMediaPlaybackTargetPicker::customPlaybackActionSelected): Deleted.
* WebView/WebView.mm:
(-[WebView _showPlaybackTargetPicker:location:hasVideo:]):
(-[WebView _playbackTargetPickerClientStateDidChange:state:]):

Source/WebKit2:

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::removePlaybackTargetPickerClient):
(WebKit::WebPageProxy::showPlaybackTargetPicker):
(WebKit::WebPageProxy::playbackTargetPickerClientStateDidChange):
(WebKit::WebPageProxy::setShouldPlayToPlaybackTarget):
(WebKit::WebPageProxy::didChangeBackgroundColor):
(WebKit::WebPageProxy::customPlaybackActionSelected): Deleted.
* UIProcess/WebPageProxy.h:
* UIProcess/WebPageProxy.messages.in:
* WebProcess/WebCoreSupport/WebChromeClient.cpp:
(WebKit::WebChromeClient::removePlaybackTargetPickerClient):
(WebKit::WebChromeClient::showPlaybackTargetPicker):
(WebKit::WebChromeClient::playbackTargetPickerClientStateDidChange):
* WebProcess/WebCoreSupport/WebChromeClient.h:
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:
* WebProcess/WebPage/mac/WebPageMac.mm:
(WebKit::WebPage::setShouldPlayToPlaybackTarget):
(WebKit::WebPage::customPlaybackActionSelected): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199286 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agotryGetById should be supported by the DFG/FTL
keith_miller@apple.com [Sun, 10 Apr 2016 03:38:44 +0000 (03:38 +0000)]
tryGetById should be supported by the DFG/FTL
https://bugs.webkit.org/show_bug.cgi?id=156378

Reviewed by Filip Pizlo.

This patch adds support for tryGetById in the DFG/FTL. It adds a new DFG node
TryGetById, which acts similarly to the normal GetById DFG node. One key
difference between GetById and TryGetById is that in the LLInt and Baseline
we do not profile the result type. This profiling is unnessary for the current
use case of tryGetById, which is expected to be a strict equality comparision
against a specific object or undefined. In either case other DFG optimizations
will make this equally fast with or without the profiling information.

Additionally, this patch adds new reuse modes for JSValueRegsTemporary that take
an operand and attempt to reuse the registers for that operand if they are free
after the current DFG node.

* bytecode/GetByIdStatus.cpp:
(JSC::GetByIdStatus::computeFromLLInt):
(JSC::GetByIdStatus::computeForStubInfoWithoutExitSiteFeedback):
* dfg/DFGAbstractInterpreterInlines.h:
(JSC::DFG::AbstractInterpreter<AbstractStateType>::executeEffects):
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::handleGetById):
(JSC::DFG::ByteCodeParser::parseBlock):
* dfg/DFGCapabilities.cpp:
(JSC::DFG::capabilityLevel):
* dfg/DFGClobberize.h:
(JSC::DFG::clobberize):
* dfg/DFGDoesGC.cpp:
(JSC::DFG::doesGC):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
* dfg/DFGNode.h:
(JSC::DFG::Node::hasIdentifier):
* dfg/DFGNodeType.h:
* dfg/DFGPredictionPropagationPhase.cpp:
(JSC::DFG::PredictionPropagationPhase::propagate):
* dfg/DFGSafeToExecute.h:
(JSC::DFG::safeToExecute):
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileTryGetById):
(JSC::DFG::JSValueRegsTemporary::JSValueRegsTemporary):
* dfg/DFGSpeculativeJIT.h:
(JSC::DFG::GPRTemporary::operator=):
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::cachedGetById):
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::cachedGetById):
(JSC::DFG::SpeculativeJIT::compile):
* ftl/FTLCapabilities.cpp:
(JSC::FTL::canCompile):
* ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::LowerDFGToB3::compileNode):
(JSC::FTL::DFG::LowerDFGToB3::compileGetById):
(JSC::FTL::DFG::LowerDFGToB3::getById):
* jit/JITOperations.cpp:
* jit/JITOperations.h:
* tests/stress/try-get-by-id.js:
(tryGetByIdTextStrict):
(get let):
(let.get createBuiltin):
(get throw):
(getCaller.obj.1.throw.new.Error): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199279 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFixed compilation of JPEGImageDecoder with libjpeg v9.
commit-queue@webkit.org [Sun, 10 Apr 2016 01:46:27 +0000 (01:46 +0000)]
Fixed compilation of JPEGImageDecoder with libjpeg v9.
https://bugs.webkit.org/show_bug.cgi?id=156445

Patch by Konstantin Tokarev <annulen@yandex.ru> on 2016-04-09
Reviewed by Michael Catanzaro.

ICU defines TRUE and FALSE macros, breaking libjpeg v9 headers.

No new tests needed.

* platform/image-decoders/jpeg/JPEGImageDecoder.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199278 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAllocation sinking SSA Defs are allowed to have replacements
sbarati@apple.com [Sun, 10 Apr 2016 00:26:25 +0000 (00:26 +0000)]
Allocation sinking SSA Defs are allowed to have replacements
https://bugs.webkit.org/show_bug.cgi?id=156444

Reviewed by Filip Pizlo.

Consider the following program and the annotations that explain why
the SSA defs we create in allocation sinking can have replacements.

function foo(a1) {
    let o1 = {x: 20, y: 50};
    let o2 = {y: 40, o1: o1};
    let o3 = {};

    // We're Defing a new variable here, call it o3_field.
    // o3_field is defing the value that is the result of
    // a GetByOffset that gets eliminated through allocation sinking.
    o3.field = o1.y;

    dontCSE();

    // This control flow is here to not allow the phase to consult
    // its local SSA mapping (which properly handles replacements)
    // for the value of o3_field.
    if (a1) {
        a1 = true;
    } else {
        a1 = false;
    }

    // Here, we ask for the reaching def of o3_field, and assert
    // it doesn't have a replacement. It does have a replacement
    // though. The original Def was the GetByOffset. We replaced
    // that GetByOffset with the value of the o1_y variable.
    let value = o3.field;
    assert(value === 50);
}

* dfg/DFGObjectAllocationSinkingPhase.cpp:
* tests/stress/allocation-sinking-defs-may-have-replacements.js: Added.
(dontCSE):
(assert):
(foo):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199277 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r199242.
commit-queue@webkit.org [Sat, 9 Apr 2016 20:54:57 +0000 (20:54 +0000)]
Unreviewed, rolling out r199242.
https://bugs.webkit.org/show_bug.cgi?id=156442

Caused many many leaks (Requested by ap on #webkit).

Reverted changeset:

"Web Inspector: get rid of InspectorBasicValue and
InspectorString subclasses"
https://bugs.webkit.org/show_bug.cgi?id=156407
http://trac.webkit.org/changeset/199242

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199276 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDebug JSC test failure: stress/multi-put-by-offset-reallocation-butterfly-cse.js...
fpizlo@apple.com [Sat, 9 Apr 2016 20:41:04 +0000 (20:41 +0000)]
Debug JSC test failure: stress/multi-put-by-offset-reallocation-butterfly-cse.js.ftl-no-cjit-small-pool
https://bugs.webkit.org/show_bug.cgi?id=156406

Reviewed by Saam Barati.

The failure was because the GC ran from within the butterfly allocation call in a put_by_id
transition AccessCase that had to deal with indexing storage. When the GC runs in a call from a stub,
then we need to be extra careful:

1) The GC may reset the IC and delete the stub. So, the stub needs to tell the GC that it might be on
   the stack during GC, so that the GC keeps it alive if it's currently running.

2) If the stub uses (dereferences or stores) some object after the call, then we need to ensure that
   the stub routine knows about that object independently of the IC.

In the case of put_by_id transitions that use a helper to allocate the butterfly, we have both
issues. A long time ago, we had to deal with (2), and we still had code to handle that case, although
it appears to be dead. This change revives that code and glues it together with PolymorphicAccess.

* bytecode/PolymorphicAccess.cpp:
(JSC::AccessCase::alternateBase):
(JSC::AccessCase::doesCalls):
(JSC::AccessCase::couldStillSucceed):
(JSC::AccessCase::generate):
(JSC::PolymorphicAccess::regenerate):
* bytecode/PolymorphicAccess.h:
(JSC::AccessCase::customSlotBase):
(JSC::AccessCase::isGetter):
(JSC::AccessCase::doesCalls): Deleted.
* jit/GCAwareJITStubRoutine.cpp:
(JSC::GCAwareJITStubRoutine::markRequiredObjectsInternal):
(JSC::MarkingGCAwareJITStubRoutine::MarkingGCAwareJITStubRoutine):
(JSC::MarkingGCAwareJITStubRoutine::~MarkingGCAwareJITStubRoutine):
(JSC::MarkingGCAwareJITStubRoutine::markRequiredObjectsInternal):
(JSC::GCAwareJITStubRoutineWithExceptionHandler::GCAwareJITStubRoutineWithExceptionHandler):
(JSC::createJITStubRoutine):
(JSC::MarkingGCAwareJITStubRoutineWithOneObject::MarkingGCAwareJITStubRoutineWithOneObject): Deleted.
(JSC::MarkingGCAwareJITStubRoutineWithOneObject::~MarkingGCAwareJITStubRoutineWithOneObject): Deleted.
(JSC::MarkingGCAwareJITStubRoutineWithOneObject::markRequiredObjectsInternal): Deleted.
* jit/GCAwareJITStubRoutine.h:
(JSC::createJITStubRoutine):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199275 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r199268.
commit-queue@webkit.org [Sat, 9 Apr 2016 20:13:12 +0000 (20:13 +0000)]
Unreviewed, rolling out r199268.
https://bugs.webkit.org/show_bug.cgi?id=156440

Broke Windows build (Requested by ap on #webkit).

Reverted changeset:

"Implement functional :host() pseudo class"
https://bugs.webkit.org/show_bug.cgi?id=156397
http://trac.webkit.org/changeset/199268

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199274 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Update another GStreamer test expectation
mcatanzaro@igalia.com [Sat, 9 Apr 2016 18:16:14 +0000 (18:16 +0000)]
[GTK] Update another GStreamer test expectation

Unreviewed.

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199273 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Update some more IndexedDB test expectations.
mcatanzaro@igalia.com [Sat, 9 Apr 2016 18:11:52 +0000 (18:11 +0000)]
[GTK] Update some more IndexedDB test expectations.

Unreviewed.

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199272 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Remove failure expectation from storage/indexeddb/connection-leak.html
mcatanzaro@igalia.com [Sat, 9 Apr 2016 17:58:42 +0000 (17:58 +0000)]
[GTK] Remove failure expectation from storage/indexeddb/connection-leak.html

It's skipped in the global TestExpectations, see bug #152643.

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199271 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Gardening unexpected passes and IndexedDB tests.
mcatanzaro@igalia.com [Sat, 9 Apr 2016 17:29:46 +0000 (17:29 +0000)]
[GTK] Gardening unexpected passes and IndexedDB tests.

Unreviewed gardening.

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199270 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix. Don't treat a build number 0 as a pending build.
rniwa@webkit.org [Sat, 9 Apr 2016 07:40:21 +0000 (07:40 +0000)]
Build fix. Don't treat a build number 0 as a pending build.

* tools/js/buildbot-syncer.js:
(BuildbotBuildEntry.prototype.isPending):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199269 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoImplement functional :host() pseudo class
antti@apple.com [Sat, 9 Apr 2016 07:38:32 +0000 (07:38 +0000)]
Implement functional :host() pseudo class
https://bugs.webkit.org/show_bug.cgi?id=156397
<rdar://problem/25621445>

Reviewed by Darin Adler.

Source/WebCore:

We already support :host. Add functional syntax too.

* css/CSSGrammar.y.in:

    Parse functional :host().

* css/CSSParser.cpp:
(WebCore::CSSParser::detectFunctionTypeToken):
* css/CSSParserValues.cpp:
(WebCore::CSSParserSelector::parsePseudoClassHostFunctionSelector):
* css/CSSParserValues.h:
* css/ElementRuleCollector.cpp:
(WebCore::ElementRuleCollector::matchedRuleList):
(WebCore::ElementRuleCollector::addMatchedRule):

    Factor some shared code here.

(WebCore::ElementRuleCollector::matchHostPseudoClassRules):

    Instead of using the generic paths use a :host specific code path for matching.
    This makes it easier to avoid :host matching when it shouldn't.

(WebCore::ElementRuleCollector::collectMatchingRulesForList):
* css/ElementRuleCollector.h:
* css/RuleSet.cpp:
(WebCore::computeMatchBasedOnRuleHash):

    :host is always handled by the special matching path.

* css/SelectorChecker.cpp:
(WebCore::SelectorChecker::match):
(WebCore::SelectorChecker::matchHostPseudoClass):

    Add a function specifically for checking :host. In always fails on the normal code paths.
    Check the argument selector if provided.

(WebCore::hasScrollbarPseudoElement):
* css/SelectorChecker.h:

LayoutTests:

Enable, fix and expand the test.

* fast/shadow-dom/css-scoping-shadow-host-functional-rule.html:
* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199268 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoHave Animometer benchmark always start with complexity of 1
jonlee@apple.com [Sat, 9 Apr 2016 05:16:59 +0000 (05:16 +0000)]
Have Animometer benchmark always start with complexity of 1
https://bugs.webkit.org/show_bug.cgi?id=156432

Reviewed by Ryosuke Niwa.

* Animometer/tests/resources/main.js: Update the default Controller and RampController to
set its minimum complexities to 1 instead of 0.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199267 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoEscape builder names in url* and pathFor* methods of BuildbotSyncer
rniwa@webkit.org [Sat, 9 Apr 2016 04:56:42 +0000 (04:56 +0000)]
Escape builder names in url* and pathFor* methods of BuildbotSyncer
https://bugs.webkit.org/show_bug.cgi?id=156427

Reviewed by Darin Adler.

The build fix in r199251 breaks other usage of RemoteAPI. Fix it properly by escaping builder names in
various methods of BuildbotSyncer.

Also fixed a typo in the logging and a bug that the new syncing script never updated "scheduled" to "running".

* server-tests/resources/mock-data.js:
(MockData.mockTestSyncConfigWithTwoBuilders): Renamed "some-builder-2" to "some builder 2" to test the
new escaping behavior in tools-buildbot-triggerable-tests.js and buildbot-syncer-tests.js.

* server-tests/tools-buildbot-triggerable-tests.js: Added tests for status url, and added a new test case
for updating "scheduled" to "running".

* tools/js/buildbot-syncer.js:
(BuildbotBuildEntry.buildRequestStatusIfUpdateIsNeeded): Update the status to "running" when the request's
status is "scheduled" and the buildbot's build is currently in progress.
(BuildbotSyncer.prototype.pathForPendingBuildsJSON): Escape the builder name.
(BuildbotSyncer.prototype.pathForBuildJSON): Ditto.
(BuildbotSyncer.prototype.pathForForceBuild): Ditto.
(BuildbotSyncer.prototype.url): Ditto.
(BuildbotSyncer.prototype.urlForBuildNumber): Ditto.

* tools/js/buildbot-triggerable.js:
(BuildbotTriggerable.prototype._pullBuildbotOnAllSyncers):
(BuildbotTriggerable.prototype._scheduleNextRequestInGroupIfSlaveIsAvailable): Fixed a typo. We are
scheduling new build requests, not syncing them.
* tools/js/remote.js:
(RemoteAPI.sendHttpRequest): Reverted r199251.
* unit-tests/buildbot-syncer-tests.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199266 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoImprove IDL support for object arguments that are neither optional nor nullable
darin@apple.com [Sat, 9 Apr 2016 03:46:25 +0000 (03:46 +0000)]
Improve IDL support for object arguments that are neither optional nor nullable
https://bugs.webkit.org/show_bug.cgi?id=156149

Reviewed by Chris Dumez.

Source/WebCore:

After this patch, we are almost ready to change some more DOM functions to
use references instead of pointers. Remaining blocking issue is lack of support
for ShouldPassWrapperByReference in the gobject bindings.

* bindings/objc/ExceptionHandlers.h: Add NO_RETURN to raiseDOMException.
Added a new raiseTypeErrorException. Re-indented header and removed unneeded
include and forward declarations.

* bindings/objc/ExceptionHandlers.mm:
(WebCore::raiseDOMException): Added RELEASE_ASSERT_NOT_REACHED so the compiler
will understand this is NO_RETURN. Also updated FIXME comment.
(WebCore::raiseTypeErrorException): Added.

* bindings/scripts/CodeGenerator.pm: Removed unneeded code that allows the type
"AtomicString" in IDL files.
(ShouldPassWrapperByReference): Added. Contains the logic from the function in
the JavaScript code generator that was named IsPointerParameterPassedByReference,
minus a couple checks that are unneeded. For use in other code generators so they
are all consistent about how they call the DOM implementation.

* bindings/scripts/CodeGeneratorGObject.pm:
(SkipFunction): Removed support for unused CustomBinding extended attribute.

* bindings/scripts/CodeGeneratorJS.pm:
(GenerateHeader): Removed support for unused CustomBinding extended attribute.
(GenerateImplementation): Ditto. Also changed type checking code to throw a
type error in a more efficient way, using throwVMTypeError directly.
(GenerateParametersCheck): Rearranged code a bit so that arguments that need to
be passed in unusual ways are handled all in one place. Use WTFMove for newly
created NodeFilter objects. Simplified the reference logic so it doesn't need
to do an additional check to see if a type is a callback. Also changed type
checking code to throw a type error in a more efficient way, using throwVMTypeError
directly. Also corrected mistake where null checking code was throwing
TYPE_MISMATCH_ERR instead of a type error.
(GetNativeType): Coding style tweak.
(ShouldPassWrapperByReference): Renamed from IsPointerParameterPassedByReference.
Changed to call underlying ShouldPassWrapperByReference function in the language-
independent code generator.
(GenerateConstructorDefinition): Updated for name change.

* bindings/scripts/CodeGeneratorObjC.pm:
(SkipFunction): Removed support for unused CustomBinding extended attribute.
(GenerateImplementation): Added code to null check and pass a reference when
ShouldPassWrapperByReference returns true.

* bindings/scripts/IDLAttributes.txt: Sorted in the AppleCopyright and
UsePointersEvenForNonNullableObjectArguments arguments. Removed the unused
CPPPureInterface and CustomBinding attributes.

* bindings/scripts/test/JS/JSTestActiveDOMObject.cpp: Regenerated test results.
* bindings/scripts/test/JS/JSTestInterface.cpp: Ditto.
* bindings/scripts/test/JS/JSTestMediaQueryListListener.cpp: Ditto.
* bindings/scripts/test/JS/JSTestObj.cpp: Ditto.
* bindings/scripts/test/JS/JSTestObj.h: Ditto.
* bindings/scripts/test/JS/JSTestOverloadedConstructors.cpp: Ditto.
* bindings/scripts/test/JS/JSTestTypedefs.cpp: Ditto.
* bindings/scripts/test/ObjC/DOMTestActiveDOMObject.mm: Ditto.
* bindings/scripts/test/ObjC/DOMTestCallback.mm: Ditto.
* bindings/scripts/test/ObjC/DOMTestCallbackFunction.mm: Ditto.
* bindings/scripts/test/ObjC/DOMTestInterface.mm: Ditto.
* bindings/scripts/test/ObjC/DOMTestMediaQueryListListener.mm: Ditto.
* bindings/scripts/test/ObjC/DOMTestObj.mm: Ditto.

* bindings/scripts/test/TestObj.idl: Removed test for CustomBinding.

* dom/DOMImplementation.idl: Fixed #if so that only the return type is different
between JavaScript and the other bindings. Without this change, the different
bindings got different results for ShouldPassWrapperByReference. Also formatted
functions all on a single line.

* dom/EventListener.idl: Removed CPPPureInterface, since it had no effect.
* dom/EventTarget.idl: Ditto.

LayoutTests:

* fast/canvas/canvas-path-addPath-expected.txt: Updated expected result to expect
TypeError rather than TYPE_MISMATCH_ERR. A progression.
* fast/text/font-face-set-javascript-expected.txt: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199265 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WebIDL] Add support for [ExportMacro=XXX] IDL extended attribute
cdumez@apple.com [Sat, 9 Apr 2016 03:07:43 +0000 (03:07 +0000)]
[WebIDL] Add support for [ExportMacro=XXX] IDL extended attribute
https://bugs.webkit.org/show_bug.cgi?id=156428

Reviewed by Ryosuke Niwa.

Add support for [ExportMacro=XXX] IDL extended attribute (e.g. [ExportMacro=WEBCORE_EXPORT])
so developers can indicate in the IDL which macro to use to export the generated JS bindings
class.

We previously supported this by hard-coding JS class names in the bindings generator which
was ugly.

* Modules/mediasession/MediaSession.idl:
* Modules/mediasource/SourceBuffer.idl:
* Modules/notifications/Notification.idl:
* Modules/webaudio/AudioContext.idl:
* bindings/scripts/CodeGeneratorJS.pm:
(GetExportMacroForJSClass):
(GenerateHeader):
(AddIncludesForType): Deleted.
(AddToImplIncludes): Deleted.
* bindings/scripts/IDLAttributes.txt:
* bindings/scripts/test/TestInterface.idl:
* bindings/scripts/test/TestNode.idl:
* css/CSSStyleDeclaration.idl:
* dom/ClientRect.idl:
* dom/ClientRectList.idl:
* dom/Document.idl:
* dom/Element.idl:
* dom/Node.idl:
* dom/Range.idl:
* fileapi/File.idl:
* html/DOMURL.idl:
* html/HTMLElement.idl:
* html/HTMLMediaElement.idl:
* html/TimeRanges.idl:
* html/canvas/DOMPath.idl:
* inspector/ScriptProfile.idl:
* inspector/ScriptProfileNode.idl:
* page/DOMWindow.idl:
* page/make_settings.pl:
(generateInternalSettingsIdlFile):
* testing/InternalSettings.idl:
* testing/Internals.idl:
* testing/MallocStatistics.idl:
* testing/MemoryInfo.idl:
* testing/TypeConversions.idl:
* xml/XMLHttpRequest.idl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199264 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: XHRs and Web Worker scripts are not searchable
commit-queue@webkit.org [Sat, 9 Apr 2016 02:37:19 +0000 (02:37 +0000)]
Web Inspector: XHRs and Web Worker scripts are not searchable
https://bugs.webkit.org/show_bug.cgi?id=154214
<rdar://problem/24643587>

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-04-08
Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

* inspector/protocol/Page.json:
Add optional requestId to search results properties and search
parameters for when the frameId and url are not enough. XHR
resources, and "Other" resources will use this.

Source/WebCore:

Test: inspector/page/searchInResources.html

* inspector/InspectorPageAgent.h:
* inspector/InspectorPageAgent.cpp:
(WebCore::InspectorPageAgent::searchInResource):
(WebCore::InspectorPageAgent::searchInResources):
Let the NetworkAgent handle individual search requests
with a requestId. And provide global search results for
"other" resources and will include requestId properties.

* inspector/InspectorNetworkAgent.h:
* inspector/InspectorNetworkAgent.cpp:
(WebCore::InspectorNetworkAgent::didFinishXHRLoading):
(WebCore::buildObjectForSearchResult):
(WebCore::InspectorNetworkAgent::searchOtherRequests):
(WebCore::InspectorNetworkAgent::searchInRequest):
Search saved "other" resource data content.

* inspector/NetworkResourcesData.h:
* inspector/NetworkResourcesData.cpp:
(WebCore::NetworkResourcesData::resources):
Expose the resources for iteration by the NetworkAgent.

Source/WebInspectorUI:

* UserInterface/Views/SearchSidebarPanel.js:
(WebInspector.SearchSidebarPanel.prototype.performSearch.resourceCallback):
(WebInspector.SearchSidebarPanel.prototype.performSearch.resourcesCallback):
Carry forward the requestId property if it is available.

LayoutTests:

* inspector/page/resources/search-script.js: Added.
* inspector/page/resources/search-stylesheet.css: Added.
* inspector/page/resources/search-worker.js: Added.
* inspector/page/resources/search-xhr.txt: Added.
* inspector/page/searchInResources-expected.txt: Added.
* inspector/page/searchInResources.html: Added.
Test for the Page domain's search commands.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199263 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Allocation snapshot hover persists after switching tabs
commit-queue@webkit.org [Sat, 9 Apr 2016 02:32:16 +0000 (02:32 +0000)]
Web Inspector: Allocation snapshot hover persists after switching tabs
https://bugs.webkit.org/show_bug.cgi?id=156430
<rdar://problem/25633800>

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-04-08
Reviewed by Timothy Hatcher.

* UserInterface/Views/HeapSnapshotInstanceDataGridNode.js:
(WebInspector.HeapSnapshotInstanceDataGridNode.prototype._mouseoverHandler):
Don't show the popover if the tree is no longer visible.

* UserInterface/Views/HeapSnapshotInstancesContentView.js:
(WebInspector.HeapSnapshotInstancesContentView.prototype.shown):
* UserInterface/Views/HeapSnapshotInstancesDataGridTree.js:
(WebInspector.HeapSnapshotInstancesDataGridTree):
(WebInspector.HeapSnapshotInstancesDataGridTree.prototype.get visible):
(WebInspector.HeapSnapshotInstancesDataGridTree.prototype.shown):
(WebInspector.HeapSnapshotInstancesDataGridTree.prototype.hidden):
Give the tree a visible state and have its containing ContentView
update it with normal ContentView shown/hidden.

* UserInterface/Views/Popover.js:
We are presenting while we were dismissing, so completely clear the
dismissing state.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199262 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMIPS: support Signed cond in branchTest32()
commit-queue@webkit.org [Sat, 9 Apr 2016 01:20:29 +0000 (01:20 +0000)]
MIPS: support Signed cond in branchTest32()
https://bugs.webkit.org/show_bug.cgi?id=156260

This is needed since r197688 makes use of it.

Patch by Guillaume Emont <guijemont@igalia.com> on 2016-04-08
Reviewed by Mark Lam.

* assembler/MacroAssemblerMIPS.h:
(JSC::MacroAssemblerMIPS::branchTest32):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199261 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAX: "AXLandmarkApplication" is an inappropriate subrole for ARIA "application" since...
jdiggs@igalia.com [Sat, 9 Apr 2016 01:19:07 +0000 (01:19 +0000)]
AX: "AXLandmarkApplication" is an inappropriate subrole for ARIA "application" since it's no longer a landmark
https://bugs.webkit.org/show_bug.cgi?id=155403

Reviewed by Chris Fleizach.

The new subrole is AXWebApplication and the new role description is "web application".
As part of the fix, the WebCore AccessibilityRole for ARIA's "application" role was
renamed from LandmarkApplicationRole to WebApplicationRole.

The roles-exposed.html and aria-grouping-roles.html test expectations were also updated.

Source/WebCore:

* English.lproj/Localizable.strings:
* accessibility/AccessibilityObject.cpp:
(WebCore::AccessibilityObject::accessibleNameDerivesFromContent):
(WebCore::AccessibilityObject::isLandmark):
(WebCore::initializeRoleMap):
* accessibility/AccessibilityObject.h:
* accessibility/atk/WebKitAccessibleWrapperAtk.cpp:
(atkRole):
* accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
(-[WebAccessibilityObjectWrapper determineIsAccessibilityElement]):
(-[WebAccessibilityObjectWrapper _accessibilityIsLandmarkRole:]):
* accessibility/mac/WebAccessibilityObjectWrapperBase.mm:
(-[WebAccessibilityObjectWrapperBase ariaLandmarkRoleDescription]):
* accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(createAccessibilityRoleMap):
(-[WebAccessibilityObjectWrapper subrole]):
* platform/LocalizedStrings.cpp:
(WebCore::AXARIAContentGroupText):

Source/WebKit/win:

* AccessibleBase.cpp: Update the rolename
(MSAARole):

LayoutTests:

* accessibility/mac/aria-grouping-roles-expected.txt:
* accessibility/mac/aria-grouping-roles.html:
* platform/mac/accessibility/roles-exposed-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199260 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS WK2] WKWebViews should consult ancestor UIScrollViews to determine tiling area
simon.fraser@apple.com [Sat, 9 Apr 2016 01:18:46 +0000 (01:18 +0000)]
[iOS WK2] WKWebViews should consult ancestor UIScrollViews to determine tiling area
https://bugs.webkit.org/show_bug.cgi?id=156429
rdar://problem/25455111

Reviewed by Tim Horton.

When a WKWebView is expanded to full size, then embedded in UIScrollView, it would
create huge tiles that cover the entire view area (since it considered itself non-scrollable).

Fix to always use 512x512 tiles in this configuration, and to adjust the tile coverage
for the area exposed through the enclosing UIScrollView.

Source/WebCore:

* loader/HistoryController.cpp:
(WebCore::HistoryController::saveScrollPositionAndViewStateToItem): setObscuredInset()
moved from FrameView to Page.
* page/FrameView.cpp:
(WebCore::FrameView::adjustTiledBackingScrollability): If we're clipped by an ancestor scrollView,
just assume we're scrollable on both axes.
* page/Page.h:
(WebCore::Page::obscuredInset):
(WebCore::Page::setObscuredInset):
(WebCore::Page::enclosedInScrollView):
(WebCore::Page::setEnclosedInScrollView):
* platform/ScrollView.h:
(WebCore::ScrollView::platformObscuredInset): Deleted.
(WebCore::ScrollView::platformSetObscuredInset): Deleted.

Source/WebKit2:

* Shared/VisibleContentRectUpdateInfo.cpp: Add enclosedInScrollView(), which is used to
trigger normal-sized tiles.
(WebKit::VisibleContentRectUpdateInfo::encode):
(WebKit::VisibleContentRectUpdateInfo::decode):
* Shared/VisibleContentRectUpdateInfo.h:
(WebKit::VisibleContentRectUpdateInfo::VisibleContentRectUpdateInfo):
(WebKit::VisibleContentRectUpdateInfo::enclosedInScrollView):
(WebKit::operator==):
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _didInvokeUIScrollViewDelegateCallback]): Pass our scrollView.
(-[WKWebView _didFinishScrolling]):
(-[WKWebView scrollViewDidScroll:]):
(-[WKWebView scrollViewDidZoom:]):
(-[WKWebView scrollViewDidEndZooming:withView:atScale:]):
(-[WKWebView _scrollViewDidInterruptDecelerating:]):
(-[WKWebView _visibleRectInEnclosingScrollView:]):
(-[WKWebView _visibleContentRect]): Compute the exposed part of the content relative
to the WKWebView, then intersect with the exposed part via any ancestor UIScrollView.
(-[WKWebView _didScroll]): This is called by UIKit when some ancestor UIScrollView scrolls.
However, we don't get all the UIScrollView delegate callbacks, so have to use a timer to
trigger a call to -_updateVisibleContentRects when we're in a stable state.
(-[WKWebView _enclosingScrollerScrollingEnded:]):
(-[WKWebView _frameOrBoundsChanged]):
(-[WKWebView _updateVisibleContentRects]):
(-[WKWebView _updateVisibleContentRectAfterScrollInView:]): Get the stable state from the
scroll view that the user is interacting with.
(-[WKWebView _updateContentRectsWithState:]):
* UIProcess/API/Cocoa/WKWebViewInternal.h:
* UIProcess/WebPageProxy.h: Rather than pass a bazillion arguments through updateVisibleContentRects(), just
pass the VisibleContentRectUpdateInfo struct.
* UIProcess/ios/WKContentView.h:
* UIProcess/ios/WKContentView.mm:
(-[WKContentView didUpdateVisibleRect:unobscuredRect:unobscuredRectInScrollViewCoordinates:obscuredInset:scale:minimumScale:inStableState:isChangingObscuredInsetsInteractively:enclosedInScrollView:]):
(-[WKContentView didUpdateVisibleRect:unobscuredRect:unobscuredRectInScrollViewCoordinates:obscuredInset:scale:minimumScale:inStableState:isChangingObscuredInsetsInteractively:]): Deleted.
* UIProcess/ios/WebPageProxyIOS.mm:
(WebKit::WebPageProxy::updateVisibleContentRects):
* UIProcess/mac/RemoteLayerTreeDrawingAreaProxy.mm:
(WebKit::RemoteLayerTreeDrawingAreaProxy::RemoteLayerTreeDrawingAreaProxy):
(WebKit::RemoteLayerTreeDrawingAreaProxy::indicatorLocation):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::updatePreferences):
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::updateVisibleContentRects):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199259 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS Simulator] Build failure (property 'contentsFormat' not found on object of type...
commit-queue@webkit.org [Sat, 9 Apr 2016 00:42:53 +0000 (00:42 +0000)]
[iOS Simulator] Build failure (property 'contentsFormat' not found on object of type 'LegacyTileLayer *')
https://bugs.webkit.org/show_bug.cgi?id=156415

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-04-08
Reviewed by Simon Fraser.

Source/WebCore:

* platform/spi/cocoa/QuartzCoreSPI.h:
Provide SPI forward declaration of the CALayer contentsFormat property.

Source/WebKit2:

* UIProcess/API/Cocoa/_WKElementAction.mm:
(-[_WKElementAction runActionWithElementInfo:]):
Use WeakObjCPtr instead of __weak to avoid build errors when not under ARC.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199258 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoProgress towards running CMake WebKit2 on Mac
achristensen@apple.com [Sat, 9 Apr 2016 00:36:06 +0000 (00:36 +0000)]
Progress towards running CMake WebKit2 on Mac
https://bugs.webkit.org/show_bug.cgi?id=156426

Reviewed by Tim Horton.

.:

* Source/cmake/OptionsMac.cmake:
FTL works on Mac, so let's use it.
* Source/cmake/WebKitMacros.cmake:

Source/JavaScriptCore:

* PlatformMac.cmake:

Source/WebCore:

* CMakeLists.txt:
* PlatformGTK.cmake:
* PlatformMac.cmake:
* PlatformWin.cmake:
On Mac, WTF is a static library that is linked only with JavaScriptCore.

Source/WebKit:

* CMakeLists.txt:
* PlatformMac.cmake:
* PlatformWin.cmake:

Source/WebKit2:

* CMakeLists.txt:
* PlatformMac.cmake:
Put the xpc service binaries in the right place.

Source/WTF:

* wtf/PlatformMac.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199257 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix with IndexedDB disabled but DatabaseProcess enabled after r199230
achristensen@apple.com [Sat, 9 Apr 2016 00:33:12 +0000 (00:33 +0000)]
Build fix with IndexedDB disabled but DatabaseProcess enabled after r199230
https://bugs.webkit.org/show_bug.cgi?id=156321

Rubber-stamped by Brady Eidson.

* DatabaseProcess/DatabaseProcess.cpp:
(WebKit::DatabaseProcess::deleteWebsiteDataForOrigins):
(WebKit::DatabaseProcess::grantSandboxExtensionsForBlobs):
(WebKit::DatabaseProcess::accessToTemporaryFileComplete):
(WebKit::DatabaseProcess::indexedDatabaseOrigins):
* DatabaseProcess/DatabaseProcess.h:
Add some more guards.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199256 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed 32-bit build fix; make type of std::min<> explicit.
jer.noble@apple.com [Fri, 8 Apr 2016 22:54:50 +0000 (22:54 +0000)]
Unreviewed 32-bit build fix; make type of std::min<> explicit.

* platform/audio/ios/AudioDestinationIOS.cpp:
(WebCore::AudioDestinationIOS::render):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199253 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoCRASH in AudioDestinationNode::render()
jer.noble@apple.com [Fri, 8 Apr 2016 22:41:46 +0000 (22:41 +0000)]
CRASH in AudioDestinationNode::render()
https://bugs.webkit.org/show_bug.cgi?id=156308

Reviewed by Eric Carlson.

Yet another math error in AudioDestinationIOS::render(). It is possible for the difference between
m_startSpareFrame and m_endSpareFrame to be greater than the numberOfFrames to be rendered. Protect
against this case by taking the min() of those two values and only advancing m_startSpareFrame by
that amount.  This guarantees that framesThisTime will never underflow, and that data will not be
written past the end of the ioData parameter.

* platform/audio/ios/AudioDestinationIOS.cpp:
(WebCore::AudioDestinationIOS::render):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199252 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix. We need to escape the path or http.request would fail.
rniwa@webkit.org [Fri, 8 Apr 2016 22:37:34 +0000 (22:37 +0000)]
Build fix. We need to escape the path or http.request would fail.

* tools/js/remote.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199251 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoModern IDB: Use more IDBValue and IDBGetResult in IDBBackingStore.
beidson@apple.com [Fri, 8 Apr 2016 22:01:10 +0000 (22:01 +0000)]
Modern IDB: Use more IDBValue and IDBGetResult in IDBBackingStore.
https://bugs.webkit.org/show_bug.cgi?id=156418

Reviewed by Alex Christensen.

No new tests (Refactor, no change in behavior).

* Modules/indexeddb/IDBValue.cpp:
(WebCore::IDBValue::IDBValue):
* Modules/indexeddb/IDBValue.h:

* Modules/indexeddb/server/IDBBackingStore.h:

* Modules/indexeddb/server/MemoryBackingStoreTransaction.cpp:
(WebCore::IDBServer::MemoryBackingStoreTransaction::abort):

* Modules/indexeddb/server/MemoryIDBBackingStore.cpp:
(WebCore::IDBServer::MemoryIDBBackingStore::addRecord):
(WebCore::IDBServer::MemoryIDBBackingStore::getRecord):
* Modules/indexeddb/server/MemoryIDBBackingStore.h:

* Modules/indexeddb/server/MemoryObjectStore.cpp:
(WebCore::IDBServer::MemoryObjectStore::addRecord):
* Modules/indexeddb/server/MemoryObjectStore.h:

* Modules/indexeddb/server/SQLiteIDBBackingStore.cpp:
(WebCore::IDBServer::SQLiteIDBBackingStore::addRecord):
(WebCore::IDBServer::SQLiteIDBBackingStore::getRecord):
* Modules/indexeddb/server/SQLiteIDBBackingStore.h:

* Modules/indexeddb/server/UniqueIDBDatabase.cpp:
(WebCore::IDBServer::UniqueIDBDatabase::performPutOrAdd):
(WebCore::IDBServer::UniqueIDBDatabase::performGetRecord):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199250 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDebugger may dereference m_currentCallFrame even after the VM has gone idle
sbarati@apple.com [Fri, 8 Apr 2016 21:21:25 +0000 (21:21 +0000)]
Debugger may dereference m_currentCallFrame even after the VM has gone idle
https://bugs.webkit.org/show_bug.cgi?id=156413

Reviewed by Mark Lam.

There is a bug where the debugger may dereference its m_currentCallFrame
pointer after that pointer becomes invalid to read from. This happens like so:

We may step over an instruction which causes the end of execution for the
current program. This causes the VM to exit. Then, we perform a GC which
causes us to collect the global object. The global object being collected
causes us to detach the debugger. In detaching, we think we still have a
valid m_currentCallFrame, we dereference it, and crash. The solution is to
make sure we're paused when dereferencing this pointer inside ::detach().

* debugger/Debugger.cpp:
(JSC::Debugger::detach):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199249 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoModern IDB: Make IDBGetResult contain an IDBValue instead of a buffer, and remove...
beidson@apple.com [Fri, 8 Apr 2016 21:11:37 +0000 (21:11 +0000)]
Modern IDB: Make IDBGetResult contain an IDBValue instead of a buffer, and remove unused methods.
https://bugs.webkit.org/show_bug.cgi?id=156416

Reviewed by Alex Christensen.

No new tests (Refactor, no change in behavior).

* Modules/indexeddb/IDBCursor.cpp:
(WebCore::IDBCursor::setGetResult):

* Modules/indexeddb/IDBGetResult.cpp:
(WebCore::IDBGetResult::dataFromBuffer):
(WebCore::IDBGetResult::isolatedCopy):
* Modules/indexeddb/IDBGetResult.h:
(WebCore::IDBGetResult::IDBGetResult):
(WebCore::IDBGetResult::value):
(WebCore::IDBGetResult::encode):
(WebCore::IDBGetResult::decode):
(WebCore::IDBGetResult::valueBuffer): Deleted.
(WebCore::IDBGetResult::setValueBuffer): Deleted.
(WebCore::IDBGetResult::setKeyData): Deleted.
(WebCore::IDBGetResult::setPrimaryKeyData): Deleted.
(WebCore::IDBGetResult::setKeyPath): Deleted.

* Modules/indexeddb/IDBTransaction.cpp:
(WebCore::IDBTransaction::didGetRecordOnServer):

* Modules/indexeddb/IDBValue.cpp:
(WebCore::IDBValue::IDBValue):
* Modules/indexeddb/IDBValue.h:

* Modules/indexeddb/server/SQLiteIDBBackingStore.cpp:
(WebCore::IDBServer::SQLiteIDBBackingStore::getIndexRecord):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199248 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFocus ring drawn at incorrect location on image map with CSS transform.
zalan@apple.com [Fri, 8 Apr 2016 21:01:50 +0000 (21:01 +0000)]
Focus ring drawn at incorrect location on image map with CSS transform.
https://bugs.webkit.org/show_bug.cgi?id=143527
<rdar://problem/21908735>

Reviewed by Simon Fraser.

Source/WebCore:

Implement pathForFocusRing for HTMLAreaElement. It follows the logic of RenderObject::addFocusRingRects().

Tests: fast/images/image-map-outline-in-positioned-container.html
       fast/images/image-map-outline-with-paint-root-offset.html
       fast/images/image-map-outline-with-scale-transform.html
       fast/images/image-map-outline.html

* html/HTMLAreaElement.cpp:
(WebCore::HTMLAreaElement::pathForFocusRing):
* html/HTMLAreaElement.h:
* rendering/RenderElement.cpp:
(WebCore::RenderElement::paintFocusRing): Move addFocusRingRects() out of focus ring painting.
(WebCore::RenderElement::paintOutline):
* rendering/RenderElement.h:
* rendering/RenderImage.cpp:
(WebCore::RenderImage::paint):
(WebCore::RenderImage::paintAreaElementFocusRing):
* rendering/RenderImage.h:
* rendering/RenderInline.cpp:
(WebCore::RenderInline::paintOutline):

LayoutTests:

Implement pathForFocusRing for HTMLAreaElement.

* fast/images/image-map-outline-in-positioned-container-expected.html: Added.
* fast/images/image-map-outline-in-positioned-container.html: Added.
* fast/images/image-map-outline-with-paint-root-offset-expected.html: Added.
* fast/images/image-map-outline-with-paint-root-offset.html: Added.
* fast/images/image-map-outline-with-scale-transform-expected.html: Added.
* fast/images/image-map-outline-with-scale-transform.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199247 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agobmalloc: stress_aligned test fails if you increase smallMax
ggaren@apple.com [Fri, 8 Apr 2016 20:59:51 +0000 (20:59 +0000)]
bmalloc: stress_aligned test fails if you increase smallMax
https://bugs.webkit.org/show_bug.cgi?id=156414

Reviewed by Oliver Hunt.

When size exceeds alignment and is a multiple of alignment and is not
a power of two, such as 24kB with 8kB alignment, the small allocator
did not always guarantee alignment. Let's fix that.

* bmalloc/Algorithm.h:
(bmalloc::divideRoundingUp): Math is hard.

* bmalloc/Allocator.cpp:
(bmalloc::Allocator::allocate): Align to the page size unconditionally.
Even if the page size is not a power of two, it might be a multiple of
a power of two, and we want alignment to that smaller power of two to
be guaranteed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199246 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK1] Wheel event callback removing the window causes crash in WebCore
bfulgham@apple.com [Fri, 8 Apr 2016 20:46:25 +0000 (20:46 +0000)]
[WK1] Wheel event callback removing the window causes crash in WebCore
https://bugs.webkit.org/show_bug.cgi?id=156409
<rdar://problem/25631267>

Reviewed by Simon Fraser.

Null check the Widget before using it, since the iframe may have been removed
from its parent document inside the event handler.

This is the WK1 fix for https://bugs.webkit.org/show_bug.cgi?id=150871.

Tested by fast/events/wheel-event-destroys-frame.html

* page/EventHandler.cpp:
(WebCore::widgetForElement): Added.
(WebCore::EventHandler::handleWheelEvent): Use new helper function to
clean up the code, and allow us to check that the Widget has not been
destroyed during the event handler.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199245 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix SVG benchmark test
jonlee@apple.com [Fri, 8 Apr 2016 20:19:26 +0000 (20:19 +0000)]
Fix SVG benchmark test
https://bugs.webkit.org/show_bug.cgi?id=156410

Reviewed by Dean Jackson.

* Animometer/resources/extensions.js: Update Point.zero to be a static Point.
* Animometer/tests/simple/resources/tiled-canvas-image.js:
(Stage.call._setupTiles): Refactor.
* Animometer/tests/master/resources/particles.js:
(Particle.prototype.reset): Use Point.center.
(complexity): We are not using a gradient background anymore, so remove the +1.
* Animometer/tests/master/resources/svg-particles.js: Update to use SVG transform
instead of CSS transform.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199244 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoTiming attack on SVG feComposite filter circumvents same-origin policy
said@apple.com [Fri, 8 Apr 2016 20:07:51 +0000 (20:07 +0000)]
Timing attack on SVG feComposite filter circumvents same-origin policy
https://bugs.webkit.org/show_bug.cgi?id=154338

Patch by Said Abou-Hallawa <sabouhallawa@apple,com> on 2016-04-08
Reviewed by Oliver Hunt.

Ensure the FEComposite arithmetic filter is clamping the resulted color
components in a constant time.

* platform/graphics/filters/FEComposite.cpp:
(WebCore::clampByte):
(WebCore::computeArithmeticPixels):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199243 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: get rid of InspectorBasicValue and InspectorString subclasses
bburg@apple.com [Fri, 8 Apr 2016 19:59:25 +0000 (19:59 +0000)]
Web Inspector: get rid of InspectorBasicValue and InspectorString subclasses
https://bugs.webkit.org/show_bug.cgi?id=156407
<rdar://problem/25627659>

Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

There's no point having these subclasses as they don't save any space.
Add m_stringValue to the union and merge some implementations of writeJSON.
Move uses of the subclass to InspectorValue and delete redundant methods.
Now, most InspectorValue methods are non-virtual so they can be templated.

* bindings/ScriptValue.cpp:
(Deprecated::jsToInspectorValue):
* inspector/InjectedScriptBase.cpp:
(Inspector::InjectedScriptBase::makeCall):
Don't used deleted subclasses.

* inspector/InspectorValues.cpp:
(Inspector::InspectorValue::null):
(Inspector::InspectorValue::create):
(Inspector::InspectorValue::asValue):
(Inspector::InspectorValue::asBoolean):
(Inspector::InspectorValue::asDouble):
(Inspector::InspectorValue::asInteger):
(Inspector::InspectorValue::asString):
These only need one implementation now.

(Inspector::InspectorValue::writeJSON):
Still a virtual method since Object and Array need their members.

(Inspector::InspectorObjectBase::InspectorObjectBase):
(Inspector::InspectorBasicValue::asBoolean): Deleted.
(Inspector::InspectorBasicValue::asDouble): Deleted.
(Inspector::InspectorBasicValue::asInteger): Deleted.
(Inspector::InspectorBasicValue::writeJSON): Deleted.
(Inspector::InspectorString::asString): Deleted.
(Inspector::InspectorString::writeJSON): Deleted.
(Inspector::InspectorString::create): Deleted.
(Inspector::InspectorBasicValue::create): Deleted.

* inspector/InspectorValues.h:
(Inspector::InspectorObjectBase::setBoolean):
(Inspector::InspectorObjectBase::setInteger):
(Inspector::InspectorObjectBase::setDouble):
(Inspector::InspectorObjectBase::setString):
(Inspector::InspectorArrayBase::pushBoolean):
(Inspector::InspectorArrayBase::pushInteger):
(Inspector::InspectorArrayBase::pushDouble):
(Inspector::InspectorArrayBase::pushString):
Use new factory methods.

* replay/EncodedValue.cpp:
(JSC::ScalarEncodingTraits<bool>::encodeValue):
(JSC::ScalarEncodingTraits<double>::encodeValue):
(JSC::ScalarEncodingTraits<float>::encodeValue):
(JSC::ScalarEncodingTraits<int32_t>::encodeValue):
(JSC::ScalarEncodingTraits<int64_t>::encodeValue):
(JSC::ScalarEncodingTraits<uint32_t>::encodeValue):
(JSC::ScalarEncodingTraits<uint64_t>::encodeValue):
* replay/EncodedValue.h:
Use new factory methods.

Source/WebCore:

* inspector/InspectorDatabaseAgent.cpp: Don't use deleted subclasses.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199242 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoJSContext Inspector: Fix asserts and uncaught exception showing Timeline Tab
commit-queue@webkit.org [Fri, 8 Apr 2016 19:40:48 +0000 (19:40 +0000)]
JSContext Inspector: Fix asserts and uncaught exception showing Timeline Tab
https://bugs.webkit.org/show_bug.cgi?id=156411

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-04-08
Reviewed by Timothy Hatcher.

* UserInterface/Views/OverviewTimelineView.js:
(WebInspector.OverviewTimelineView):
(WebInspector.OverviewTimelineView.prototype.closed):
Gracefully handle if we do not have a Network Timeline.

* UserInterface/Views/TimelineTabContentView.js:
(WebInspector.TimelineTabContentView.prototype._changeViewMode):
This function is always called by the constructor, so the assert
is not useful since it can be called when FPS is not supported.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199241 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd IC support for arguments.length
fpizlo@apple.com [Fri, 8 Apr 2016 19:37:04 +0000 (19:37 +0000)]
Add IC support for arguments.length
https://bugs.webkit.org/show_bug.cgi?id=156389

Reviewed by Geoffrey Garen.
Source/JavaScriptCore:

This adds support for caching accesses to arguments.length for both DirectArguments and
ScopedArguments. In strict mode, we already cached these accesses since they were just
normal properties.

Amazingly, we also already supported caching of overridden arguments.length in both
DirectArguments and ScopedArguments. This is because when you override, the property gets
materialized as a normal JS property and the structure is changed.

This patch painstakingly preserves our previous caching of overridden length while
introducing caching of non-overridden length (i.e. the common case). In fact, we even cache
the case where it could either be overridden or not, since we just end up with an AccessCase
for each and they cascade to each other.

This is a >3x speed-up on microbenchmarks that do arguments.length in a polymorphic context.
Entirely monomorphic accesses were already handled by the DFG.

* bytecode/PolymorphicAccess.cpp:
(JSC::AccessGenerationState::calculateLiveRegistersForCallAndExceptionHandling):
(JSC::AccessCase::guardedByStructureCheck):
(JSC::AccessCase::generateWithGuard):
(JSC::AccessCase::generate):
(WTF::printInternal):
* bytecode/PolymorphicAccess.h:
* jit/ICStats.h:
* jit/JITOperations.cpp:
* jit/Repatch.cpp:
(JSC::tryCacheGetByID):
(JSC::tryCachePutByID):
(JSC::tryRepatchIn):
* tests/stress/direct-arguments-override-length-then-access-normal-length.js: Added.
(args):
(foo):
(result.foo):

LayoutTests:

* js/regress/direct-arguments-length-expected.txt: Added.
* js/regress/direct-arguments-length.html: Added.
* js/regress/direct-arguments-overridden-length-expected.txt: Added.
* js/regress/direct-arguments-overridden-length.html: Added.
* js/regress/direct-arguments-possibly-overridden-length-expected.txt: Added.
* js/regress/direct-arguments-possibly-overridden-length.html: Added.
* js/regress/scoped-arguments-length-expected.txt: Added.
* js/regress/scoped-arguments-length.html: Added.
* js/regress/scoped-arguments-overridden-length-expected.txt: Added.
* js/regress/scoped-arguments-overridden-length.html: Added.
* js/regress/scoped-arguments-possibly-overridden-length-expected.txt: Added.
* js/regress/scoped-arguments-possibly-overridden-length.html: Added.
* js/regress/script-tests/direct-arguments-length.js: Added.
(args):
* js/regress/script-tests/direct-arguments-overridden-length.js: Added.
(args):
* js/regress/script-tests/direct-arguments-possibly-overridden-length.js: Added.
(args1):
(args2):
* js/regress/script-tests/scoped-arguments-length.js: Added.
(args):
* js/regress/script-tests/scoped-arguments-overridden-length.js: Added.
(args):
* js/regress/script-tests/scoped-arguments-possibly-overridden-length.js: Added.
(args1):
(args2):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199240 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix leaks in WebAVMediaSelectionOptionMac and WebPlaybackControlsManager
bdakin@apple.com [Fri, 8 Apr 2016 18:32:19 +0000 (18:32 +0000)]
Fix leaks in WebAVMediaSelectionOptionMac and WebPlaybackControlsManager
https://bugs.webkit.org/show_bug.cgi?id=156379

Reviewed by Tim Horton.

These classes should use RetainPtrs.
* platform/mac/WebVideoFullscreenInterfaceMac.mm:
(-[WebAVMediaSelectionOptionMac localizedDisplayName]):
(-[WebAVMediaSelectionOptionMac setLocalizedDisplayName:]):
(-[WebPlaybackControlsManager timing]):
(-[WebPlaybackControlsManager setTiming:]):
(-[WebPlaybackControlsManager seekableTimeRanges]):
(-[WebPlaybackControlsManager setSeekableTimeRanges:]):
(-[WebPlaybackControlsManager audioMediaSelectionOptions]):
(-[WebPlaybackControlsManager setAudioMediaSelectionOptions:]):
(-[WebPlaybackControlsManager currentAudioMediaSelectionOption]):
(-[WebPlaybackControlsManager setCurrentAudioMediaSelectionOption:]):
(-[WebPlaybackControlsManager legibleMediaSelectionOptions]):
(-[WebPlaybackControlsManager setLegibleMediaSelectionOptions:]):
(-[WebPlaybackControlsManager currentLegibleMediaSelectionOption]):
(-[WebPlaybackControlsManager setCurrentLegibleMediaSelectionOption:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199238 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoTouching any IDL files rebuilds all bindings in CMake Ninja build
commit-queue@webkit.org [Fri, 8 Apr 2016 18:28:27 +0000 (18:28 +0000)]
Touching any IDL files rebuilds all bindings in CMake Ninja build
https://bugs.webkit.org/show_bug.cgi?id=156400

Patch by Fujii Hironori <Hironori.Fujii@jp.sony.com> on 2016-04-08
Reviewed by Brent Fulgham.

* bindings/scripts/preprocess-idls.pl:
(GenerateConstructorAttribute):
WriteFileIfChanged does not work due to flaky results of 'keys'.
Sort results of 'keys'.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199237 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRedefining a method of the same name hits an assertion
joepeck@webkit.org [Fri, 8 Apr 2016 18:18:19 +0000 (18:18 +0000)]
Redefining a method of the same name hits an assertion
https://bugs.webkit.org/show_bug.cgi?id=144258

Reviewed by Ryosuke Niwa.

* TestExpectations:
* js/script-tests/class-syntax-semicolon.js:
This test no longer asserts.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199236 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUInt32ToNumber should have an Int52 path
commit-queue@webkit.org [Fri, 8 Apr 2016 18:07:25 +0000 (18:07 +0000)]
UInt32ToNumber should have an Int52 path
https://bugs.webkit.org/show_bug.cgi?id=125704

Patch by Benjamin Poulain <bpoulain@apple.com> on 2016-04-08
Reviewed by Filip Pizlo.

When dealing with big numbers, fall back to Int52 instead
of double when possible.

* dfg/DFGAbstractInterpreterInlines.h:
(JSC::DFG::AbstractInterpreter<AbstractStateType>::executeEffects):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
* dfg/DFGPredictionPropagationPhase.cpp:
(JSC::DFG::PredictionPropagationPhase::propagate):
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileUInt32ToNumber):
* ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::LowerDFGToB3::compileUInt32ToNumber):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199235 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[cmake] Use ICU include dirs in WebKit.
commit-queue@webkit.org [Fri, 8 Apr 2016 17:50:22 +0000 (17:50 +0000)]
[cmake] Use ICU include dirs in WebKit.
https://bugs.webkit.org/show_bug.cgi?id=156402

Patch by Konstantin Tokarev <annulen@yandex.ru> on 2016-04-08
Reviewed by Brent Fulgham.

* CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199234 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS WK2] Stop using exposedContentRect for history scroll state restoration
simon.fraser@apple.com [Fri, 8 Apr 2016 17:22:19 +0000 (17:22 +0000)]
[iOS WK2] Stop using exposedContentRect for history scroll state restoration
https://bugs.webkit.org/show_bug.cgi?id=156392

Reviewed by Tim Horton.

A future commit will alter the meaning of exposedContentRect on iOS to take into
account clipped out parts of the WKWebView. To achieve this, wean history restoration
off of using exposedContentRect for scroll state restoration. It did this to restore
the page to the same position relative to the view's top-left (to avoid jiggles caused
by changing obscured insets).

Do this by pushing the left/top obscured insets down with visible content rects updates,
storing them on ScrollView, and adding them to HistoryItem. Those insets are then used
for scroll state restoration in WKWebView.

Source/WebCore:

* history/HistoryItem.cpp:
(WebCore::HistoryItem::HistoryItem):
* history/HistoryItem.h:
(WebCore::HistoryItem::obscuredInset):
(WebCore::HistoryItem::setObscuredInset):
* loader/HistoryController.cpp:
(WebCore::HistoryController::saveScrollPositionAndViewStateToItem):
* platform/ScrollView.h:
(WebCore::ScrollView::platformObscuredInset):
(WebCore::ScrollView::platformSetObscuredInset):

Source/WebKit2:

* Shared/VisibleContentRectUpdateInfo.cpp: Add FloatSize for obscuredInset.
(WebKit::VisibleContentRectUpdateInfo::encode):
(WebKit::VisibleContentRectUpdateInfo::decode):
* Shared/VisibleContentRectUpdateInfo.h:
(WebKit::VisibleContentRectUpdateInfo::VisibleContentRectUpdateInfo):
(WebKit::VisibleContentRectUpdateInfo::obscuredInset):
(WebKit::operator==):
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _processDidExit]): Rename _needsToRestoreExposedRect to _needsToRestoreScrollPosition
(-[WKWebView _didCommitLayerTree:]): Restore the scroll position using the scaled scrollOffset minus
the old obscuredInset.
(-[WKWebView _layerTreeCommitComplete]):
(-[WKWebView _restorePageScrollPosition:scrollOrigin:previousObscuredInset:scale:]):
(-[WKWebView _restorePageStateToUnobscuredCenter:scale:]):
(-[WKWebView _scrollToContentScrollPosition:scrollOrigin:]):
(-[WKWebView _updateVisibleContentRects]):
(-[WKWebView _restorePageStateToExposedRect:scrollOrigin:scale:]): Deleted.
* UIProcess/API/Cocoa/WKWebViewInternal.h:
* UIProcess/PageClient.h:
* UIProcess/WebPageProxy.h:
* UIProcess/WebPageProxy.messages.in:
* UIProcess/ios/PageClientImplIOS.h:
* UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::restorePageState):
* UIProcess/ios/WKContentView.h:
* UIProcess/ios/WKContentView.mm:
(-[WKContentView didUpdateVisibleRect:unobscuredRect:unobscuredRectInScrollViewCoordinates:obscuredInset:scale:minimumScale:inStableState:isChangingObscuredInsetsInteractively:]):
(-[WKContentView didUpdateVisibleRect:unobscuredRect:unobscuredRectInScrollViewCoordinates:scale:minimumScale:inStableState:isChangingObscuredInsetsInteractively:]): Deleted.
* UIProcess/ios/WebPageProxyIOS.mm:
(WebKit::WebPageProxy::updateVisibleContentRects):
(WebKit::WebPageProxy::restorePageState):
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::restorePageState):
(WebKit::WebPage::updateVisibleContentRects):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199233 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix followup to http://trac.webkit.org/changeset/199230
beidson@apple.com [Fri, 8 Apr 2016 17:21:32 +0000 (17:21 +0000)]
Build fix followup to trac.webkit.org/changeset/199230

Unreviewed.

* platform/posix/FileSystemPOSIX.cpp:
(WebCore::hardLinkOrCopyFile): Stricter POSIX systems require a umask for O_CREAT opens,
  so let's provide one.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199232 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove 14 more unnecessary uses of UsePointersEvenForNonNullableObjectArguments
darin@apple.com [Fri, 8 Apr 2016 16:59:35 +0000 (16:59 +0000)]
Remove 14 more unnecessary uses of UsePointersEvenForNonNullableObjectArguments
https://bugs.webkit.org/show_bug.cgi?id=156405

Reviewed by Chris Dumez.

* Modules/encryptedmedia/MediaKeySession.idl:
* Modules/encryptedmedia/MediaKeys.idl:
* dom/Element.idl:
* dom/NamedNodeMap.idl:
* html/HTMLElement.idl:
* html/canvas/OESVertexArrayObject.idl:
* html/canvas/WebGLRenderingContext.idl:
* page/DOMSelection.idl:
* storage/StorageEvent.idl:
* svg/SVGSVGElement.idl:
* xml/XMLSerializer.idl:
* xml/XPathEvaluator.idl:
* xml/XPathExpression.idl:
* xml/XSLTProcessor.idl:
Removed UsePointersEvenForNonNullableObjectArguments, which was having no effect
in any of these classes. Also tweaked formatting of some of the IDL, merging things
onto single lines, changing paragraphing and indenting a bit, and fixing some typos.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199231 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoModern IDB (Blob support): Write blobs to temporary files and move them to the correc...
beidson@apple.com [Fri, 8 Apr 2016 16:57:29 +0000 (16:57 +0000)]
Modern IDB (Blob support): Write blobs to temporary files and move them to the correct location when storing them.
https://bugs.webkit.org/show_bug.cgi?id=156321

Reviewed by Alex Christensen, Andy Estes, and Darin Adler.

Source/WebCore:

No new tests (No testable change in behavior yet, current tests pass).

When asked to store a Blob (including Files) in IndexedDB, the Blob is written out to a temporary file.

Then when the putOrAdd request is received by IDBServer it includes a list of blobURLs and their mappings
to temporary files.

Finally, as part of storing the Blob value in the database, those temporary files are moved in to place
under the IndexedDB directory for storage and later retrieval.

* Modules/indexeddb/IDBValue.cpp:
(WebCore::IDBValue::IDBValue):

* Modules/indexeddb/server/IDBBackingStore.h:
(WebCore::IDBServer::IDBBackingStoreTemporaryFileHandler::~IDBBackingStoreTemporaryFileHandler):

* Modules/indexeddb/server/IDBServer.cpp:
(WebCore::IDBServer::IDBServer::create):
(WebCore::IDBServer::IDBServer::IDBServer):
(WebCore::IDBServer::IDBServer::createBackingStore):
* Modules/indexeddb/server/IDBServer.h:

* Modules/indexeddb/server/SQLiteIDBBackingStore.cpp:
(WebCore::IDBServer::blobRecordsTableSchema):
(WebCore::IDBServer::blobRecordsTableSchemaAlternate):
(WebCore::IDBServer::blobFilesTableSchema):
(WebCore::IDBServer::blobFilesTableSchemaAlternate):
(WebCore::IDBServer::SQLiteIDBBackingStore::SQLiteIDBBackingStore):
(WebCore::IDBServer::SQLiteIDBBackingStore::ensureValidBlobTables):
(WebCore::IDBServer::SQLiteIDBBackingStore::getOrEstablishDatabaseInfo):
(WebCore::IDBServer::SQLiteIDBBackingStore::addRecord):
* Modules/indexeddb/server/SQLiteIDBBackingStore.h:
(WebCore::IDBServer::SQLiteIDBBackingStore::temporaryFileHandler):

* Modules/indexeddb/server/SQLiteIDBTransaction.cpp:
(WebCore::IDBServer::SQLiteIDBTransaction::commit):
(WebCore::IDBServer::SQLiteIDBTransaction::moveBlobFilesIfNecessary):
(WebCore::IDBServer::SQLiteIDBTransaction::abort):
(WebCore::IDBServer::SQLiteIDBTransaction::reset):
(WebCore::IDBServer::SQLiteIDBTransaction::addBlobFile):
* Modules/indexeddb/server/SQLiteIDBTransaction.h:

* Modules/indexeddb/shared/InProcessIDBServer.cpp:
(WebCore::InProcessIDBServer::InProcessIDBServer):
(WebCore::InProcessIDBServer::accessToTemporaryFileComplete):
* Modules/indexeddb/shared/InProcessIDBServer.h:

* bindings/js/SerializedScriptValue.cpp:
(WebCore::SerializedScriptValue::blobURLsIsolatedCopy):
* bindings/js/SerializedScriptValue.h:

* platform/FileSystem.h:
* platform/gtk/FileSystemGtk.cpp:
(WebCore::hardLinkOrCopyFile):
* platform/posix/FileSystemPOSIX.cpp:
(WebCore::hardLinkOrCopyFile):

Source/WebKit2:

The NetworkProcess writes a blob to a temporary file, then tells the UIProcess to grant the DatabaseProcess
a Sandbox Extension to that path.

It then tells the WebProcess the paths for the temporary files, which then tells the DatabaseProcess to store
the contents of those files as blob references in the database.

Since the UIProcess had already granted it a Sandbox Extension, it is able to do so.

* DatabaseProcess/DatabaseProcess.cpp:
(WebKit::DatabaseProcess::idbServer):
(WebKit::DatabaseProcess::grantSandboxExtensionsForBlobs):
(WebKit::DatabaseProcess::prepareForAccessToTemporaryFile):
(WebKit::DatabaseProcess::accessToTemporaryFileComplete):
* DatabaseProcess/DatabaseProcess.h:
* DatabaseProcess/DatabaseProcess.messages.in:

* NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::writeBlobsToTemporaryFiles):

* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::grantSandboxExtensionsToDatabaseProcessForBlobs):
(WebKit::NetworkProcess::didGrantSandboxExtensionsToDatabaseProcessForBlobs):
* NetworkProcess/NetworkProcess.h:
* NetworkProcess/NetworkProcess.messages.in:

* UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::grantSandboxExtensionsToDatabaseProcessForBlobs):
* UIProcess/Network/NetworkProcessProxy.h:
* UIProcess/Network/NetworkProcessProxy.messages.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199230 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAX: [ATK] Crash getting text under element in CSS table
jdiggs@igalia.com [Fri, 8 Apr 2016 16:36:01 +0000 (16:36 +0000)]
AX: [ATK] Crash getting text under element in CSS table
https://bugs.webkit.org/show_bug.cgi?id=156328

Reviewed by Chris Fleizach.

Source/WebCore:

AccessibilityRenderObject::textUnderElement() assumes (and asserts) that
the first and last child of an anonymous block will each have nodes with
which to define positions. This is not the case for CSS Tables and their
anonymous descendants. AccessibilityNodeObject:textUnderElement() is our
fallback for the instances where a text range cannot be created based on
positions, so let it handle anonymous RenderTable parts.

Test: accessibility/generated-content-with-display-table-crash.html

* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::textUnderElement):
(WebCore::AccessibilityRenderObject::shouldGetTextFromNode):
* accessibility/AccessibilityRenderObject.h:

LayoutTests:

While this crash is currently seen only for ATK, there is nothing to
prevent another port from attempting to get all the text under a CSS
RenderTable. Hence the shared test.

* accessibility/generated-content-with-display-table-crash.html: Added.
* platform/gtk/accessibility/generated-content-with-display-table-crash-expected.txt: Added.
* platform/mac/accessibility/generated-content-with-display-table-crash-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199229 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed.
youenn.fablet@crf.canon.fr [Fri, 8 Apr 2016 16:11:00 +0000 (16:11 +0000)]
Unreviewed.
Rebasing LayoutTests/imported/w3c/web-platform-tests/dom/nodes/MutationObserver-childList.html expectation after https://trac.webkit.org/changeset/199225.
Removing its Timeout expectation.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199228 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Attempting to dismiss a popover that is already being dismissed causes...
mattbaker@apple.com [Fri, 8 Apr 2016 16:05:20 +0000 (16:05 +0000)]
Web Inspector: Attempting to dismiss a popover that is already being dismissed causes an error
https://bugs.webkit.org/show_bug.cgi?id=156385
<rdar://problem/25617962>

Reviewed by Timothy Hatcher.

The Popover element is removed from the DOM once it's fade-out transition
completes. Since Popover.dismiss proceeds as long as it's element has a
parent, successive calls to dismiss can run before the popover is removed.

Rather than rely on the presence of the popover in the DOM, set a "dismissing"
flag the first time dismiss is called, before the fade-out animation begins.

* UserInterface/Controllers/BreakpointPopoverController.js:
(WebInspector.BreakpointPopoverController.prototype._conditionCodeMirrorEscapeOrEnterKey):
Check for null popover.

* UserInterface/Views/Popover.js:
(WebInspector.Popover):
(WebInspector.Popover.prototype.dismiss):
Do nothing if already dismissing.

(WebInspector.Popover.prototype.handleEvent):
Reset dismissing flag after style transition completes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199227 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Quick Open fails to match pattern "bB" in file "abBc"
mattbaker@apple.com [Fri, 8 Apr 2016 16:04:24 +0000 (16:04 +0000)]
Web Inspector: Quick Open fails to match pattern "bB" in file "abBc"
https://bugs.webkit.org/show_bug.cgi?id=156398

Reviewed by Timothy Hatcher.

Source/WebInspectorUI:

Correct an off-by-one error in the backtrack routine that set the dead
branch index to the character just before the match that was popped.
The dead branch index should equal the index of the popped match.

* UserInterface/Controllers/ResourceQueryController.js:
(WebInspector.ResourceQueryController.prototype._findQueryMatches.backtrack):
(WebInspector.ResourceQueryController.prototype._findQueryMatches):

LayoutTests:

* inspector/unit-tests/resource-query-controller-expected.txt:
* inspector/unit-tests/resource-query-controller.html:
Test that two repeated characters in the search string are correctly
matched when the first character is lowercase and the second is uppercase.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199226 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoLayoutTests/imported/w3c:
youenn.fablet@crf.canon.fr [Fri, 8 Apr 2016 14:45:24 +0000 (14:45 +0000)]
LayoutTests/imported/w3c:
Testharness-based tests that time out should be able to produce detailed output
https://bugs.webkit.org/show_bug.cgi?id=145313

Reviewed by Xabier Rodriguez-Calvar.

Rebasing tests that produce output after testharness timeout() is called.

* web-platform-tests/fetch/api/request/request-cache-expected.txt:
* web-platform-tests/html/semantics/document-metadata/the-link-element/link-style-error-01-expected.txt:
* web-platform-tests/html/semantics/document-metadata/the-style-element/style-error-01-expected.txt:
* web-platform-tests/html/semantics/embedded-content/the-img-element/environment-changes/viewport-change-expected.txt:

Tools:
Testharness-based tests that time out should be able to produce a detailed output
https://bugs.webkit.org/show_bug.cgi?id=145313

Reviewed by Xabier Rodriguez-Calvar.

Adding timeout readonly accessor to TestRunner for both WK1 and WK2.

* DumpRenderTree/TestRunner.cpp:
(getTimeoutCallback): The js "timeout" property getter.
(TestRunner::staticValues): Adding "timeout" property to DumpRenderTree so that testRunner.timeout called from JS returns the timeout value.
* DumpRenderTree/TestRunner.h:
(TestRunner::timeout): Adding access to DRT m_timeout private value.
* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl: Adding timeout readonly attribute so that testRunner.timeout can be called from JS.
* WebKitTestRunner/InjectedBundle/TestRunner.h: Adding DOM timeout getter to implement timeout IDL definition.
(WTR::TestRunner::timeout):

LayoutTests:
Testharness-based tests that time out should be able to produce a detailled output
https://bugs.webkit.org/show_bug.cgi?id=145313

Reviewed by Xabier Rodriguez-Calvar.

* TestExpectations: Removed TIMEOUT for some tests for which testharness.timeout will be called just before WTR times out.
* platform/gtk/imported/w3c/web-platform-tests/fetch/api/request/request-cache-expected.txt: GTK specific baseline.
* resources/testharnessreport.js:
(add_completion_callback): Improving error logging message. Dumping of the tests status in error case.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199225 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove unneeded UsePointersEvenForNonNullableObjectArguments from event classes
darin@apple.com [Fri, 8 Apr 2016 09:04:17 +0000 (09:04 +0000)]
Remove unneeded UsePointersEvenForNonNullableObjectArguments from event classes
https://bugs.webkit.org/show_bug.cgi?id=156396

Reviewed by Youenn Fablet.

* dom/CompositionEvent.idl:
* dom/KeyboardEvent.idl:
* dom/MouseEvent.idl:
* dom/MutationEvent.idl:
* dom/TextEvent.idl:
* dom/TouchEvent.idl:
* dom/UIEvent.idl:
* dom/WheelEvent.idl:
Removed UsePointersEvenForNonNullableObjectArguments, which was having no effect.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199224 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[css-grid] Fix positioned items with grid gaps
rego@igalia.com [Fri, 8 Apr 2016 08:52:17 +0000 (08:52 +0000)]
[css-grid] Fix positioned items with grid gaps
https://bugs.webkit.org/show_bug.cgi?id=156288

Reviewed by Darin Adler.

Source/WebCore:

When we place a positioned items in a grid with gaps,
we were not taking into accounts the gutter size.
We've to use that size to properly place and size the item.

Tests: fast/css-grid-layout/grid-positioned-items-gaps-rtl.html
       fast/css-grid-layout/grid-positioned-items-gaps.html

* rendering/RenderGrid.cpp:
(WebCore::RenderGrid::offsetAndBreadthForPositionedChild):

LayoutTests:

Added new tests checking the right behavior.

* fast/css-grid-layout/grid-positioned-items-gaps-expected.txt: Added.
* fast/css-grid-layout/grid-positioned-items-gaps-rtl-expected.txt: Added.
* fast/css-grid-layout/grid-positioned-items-gaps-rtl.html: Added.
* fast/css-grid-layout/grid-positioned-items-gaps.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199223 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[css-grid] Remove unnecessary iteration in populateGridPositions loop
jfernandez@igalia.com [Fri, 8 Apr 2016 08:01:04 +0000 (08:01 +0000)]
[css-grid] Remove unnecessary iteration in populateGridPositions loop
https://bugs.webkit.org/show_bug.cgi?id=156376

Reviewed by Darin Adler.

The populateGridPositions loop limit was set to 'lastLine'. However, the
the position of last track's start line is updated after the loop, since
it does not follow the same pattern; it does not have a content
distribution offset.

So, since we are essentially overwriting the value stored in the last
iteration, we can just lower the loop limit.

No new tests added, because there is no change in the functionality.

* rendering/RenderGrid.cpp:
(WebCore::RenderGrid::populateGridPositions):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199222 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoCSP: Block XHR when calling XMLHttpRequest.send() and throw network error.
commit-queue@webkit.org [Fri, 8 Apr 2016 07:17:50 +0000 (07:17 +0000)]
CSP: Block XHR when calling XMLHttpRequest.send() and throw network error.
https://bugs.webkit.org/show_bug.cgi?id=153598
<rdar://problem/24391483>

Patch by John Wilander <wilander@apple.com> on 2016-04-08
Reviewed by Darin Adler.

Source/WebCore:

No new tests. Changes to existing tests are sufficient.

* xml/XMLHttpRequest.cpp:
(WebCore::XMLHttpRequest::open):
(WebCore::XMLHttpRequest::initSend):
    Moved the CSP check from XMLHttpRequest::open() to XMLHttpRequest::initSend().
    Changed the thrown error type from Security to Network for synchronous requests.
    Changed from throwing an error to firing an error event for asynchronous requests.
    These changes are in conformance with connect-src of Content Security Policy Level 2.
    https://www.w3.org/TR/CSP2/#directive-connect-src (W3C Candidate Recommendation, 21 July 2015)

LayoutTests:

* fast/workers/resources/worker-inherits-csp-blocks-xhr.js:
(catch):
* fast/workers/worker-inherits-csp-blocks-xhr-expected.txt:
    Changed expected error from DOMException.SECURITY_ERR to DOMException.NETWORK_ERR.
* http/tests/security/contentSecurityPolicy/connect-src-xmlhttprequest-blocked-expected.txt:
* http/tests/security/contentSecurityPolicy/connect-src-xmlhttprequest-blocked.html:
    Now tests that XMLHttpRequest.send() is blocked if the URL voilates the connect-src directive in CSP.
* http/tests/security/contentSecurityPolicy/resources/worker.php:
    Added two additional calls to XMLHttpRequest.send() and switched to receiving an error event to make
    existing tests work with code changes.
* http/tests/security/contentSecurityPolicy/source-list-parsing-malformed-meta.html:
    Added an additional call to XMLHttpRequest.send() and switched to receiving an error event to make
    existing test work with code changes.
* http/tests/security/isolatedWorld/bypass-main-world-csp-for-xhr-expected.txt:
* http/tests/security/isolatedWorld/bypass-main-world-csp-for-xhr.html:
    Added an additional call to XMLHttpRequest.send() and switched to receiving an error event to make
    existing tests work with code changes.
    Refactored test mechnism with additional parameters to cover synchronous/asynchronous as well as
    same-origin/cross-origin in isolated worlds.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199221 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix various bugs in the new syncing script
rniwa@webkit.org [Fri, 8 Apr 2016 07:15:43 +0000 (07:15 +0000)]
Fix various bugs in the new syncing script
https://bugs.webkit.org/show_bug.cgi?id=156393

Reviewed by Darin Adler.

* server-tests/resources/common-operations.js: Added. This file was supposed to be added in r199191.
(addBuilderForReport):
(addSlaveForReport):
(connectToDatabaseInEveryTest):
(submitReport):
* tools/js/buildbot-triggerable.js:
(BuildbotTriggerable.prototype._pullBuildbotOnAllSyncers): Don't log every time we pull from buildbot
builder as this dramatically increases the amount of log we generate.
* tools/js/parse-arguments.js:
(parseArguments): Fixed a typo. This should be parseArgument*s*, not parseArgument.
* tools/js/remote.js:
(RemoteAPI.prototype.url): Fixed a bug that portSuffix wasn't being expanded in the template literal.
(RemoteAPI.prototype.configure): Added more validations with nice error messages.
(RemoteAPI.prototype.sendHttpRequest): Falling back to port 80 isn't right when scheme is https. Compute
the right port in configure instead based on the scheme.
* tools/sync-buildbot.js:
(syncLoop): Fixed the bug that syncing multiple times fail because Manifest.fetch() create new Platform
and Test objects. This results in various references in BuildRequest objects to get outdated. Fixing this
properly in Manifest.fetch() because we do need to "forget" about some tests and platforms in some cases.
For now, delete all v3 model objects and start over in each syncing cycle.
* unit-tests/tools-js-remote-tests.js: Added. Unit tests for the aforementioned changes to RemoteAPI.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199220 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: protocol generator should emit an error when 'type' is used instead...
bburg@apple.com [Fri, 8 Apr 2016 07:13:28 +0000 (07:13 +0000)]
Web Inspector: protocol generator should emit an error when 'type' is used instead of '$ref'
https://bugs.webkit.org/show_bug.cgi?id=156275
<rdar://problem/25569331>

Reviewed by Darin Adler.

* inspector/protocol/Heap.json: Fix a mistake that's now caught by the protocol generator.

* inspector/scripts/codegen/models.py:
(TypeReference.__init__): Check here if type_kind is on a whitelist of primitive types.
(TypeReference.referenced_name): Update comment.

Add a new test specifically for the case when the type would otherwise be resolved. Rebaseline.

* inspector/scripts/tests/expected/fail-on-type-reference-as-primitive-type.json-error: Added.
* inspector/scripts/tests/expected/fail-on-unknown-type-reference-in-type-declaration.json-error:
* inspector/scripts/tests/fail-on-type-reference-as-primitive-type.json: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199219 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[JSC] Enable Concurrent JIT by default
utatane.tea@gmail.com [Fri, 8 Apr 2016 07:08:19 +0000 (07:08 +0000)]
[JSC] Enable Concurrent JIT by default
https://bugs.webkit.org/show_bug.cgi?id=156341

Reviewed by Filip Pizlo.

We enable Concurrent JIT by default when DFG JIT and JSVALUE64 are enabled.
This change offers Concurrent JIT to the JSCOnly port.

* wtf/Platform.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199218 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agostylebot should know about TestWebKitAPI FeatureDefines.xcconfig
commit-queue@webkit.org [Fri, 8 Apr 2016 05:04:51 +0000 (05:04 +0000)]
stylebot should know about TestWebKitAPI FeatureDefines.xcconfig
https://bugs.webkit.org/show_bug.cgi?id=156387

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-04-07
Rubber-stamped by Dan Bernstein.

* Scripts/webkitpy/style/checkers/featuredefines.py:
(FeatureDefinesChecker.check):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199217 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFontFaceSet binding does not handle null correctly
darin@apple.com [Fri, 8 Apr 2016 04:26:27 +0000 (04:26 +0000)]
FontFaceSet binding does not handle null correctly
https://bugs.webkit.org/show_bug.cgi?id=156141

Reviewed by Youenn Fablet.

Source/WebCore:

* css/FontFaceSet.cpp:
(WebCore::FontFaceSet::FontFaceSet): Pass a reference to add rather than a pointer.
(WebCore::FontFaceSet::has): Take a reference rather than a pointer.
(WebCore::FontFaceSet::add): Ditto.
(WebCore::FontFaceSet::remove): Ditto.
(WebCore::FontFaceSet::load): Initialize ec since we check it. Caller is not required
to do this, nor is the matchingFaces function. Rearranged function to avoid needless
creation/destruction of PendingPromise for the immediate failure case. Removed some
unneeded type casts and local variables.
(WebCore::FontFaceSet::status): Use ASCIILiteral instead of ConstructFromLiteral.
No reason to use the more aggressive optimization.
(WebCore::FontFaceSet::faceFinished): Factored out a common hasReachedTerminalState
check to streamline the logic a bit.
(WebCore::FontFaceSet::load): Moved overload without a string in here; not critical
to inline it.
(WebCore::FontFaceSet::check): Ditto.

* css/FontFaceSet.h: Removed many unneeded includes and forward declarations.
Changed functions to take FontFace& instead of RefPtr<FontFace>. Removed unneeded
WebCore namespace prefixes. Use final instead of override for virtual functions.

* css/FontFaceSet.idl: Removed UsePointersEvenForNonNullableObjectArguments, which
was preserving incorrect behavior for null as demonstrated by the test cases.

LayoutTests:

* fast/text/font-face-set-javascript-expected.txt: Added expected results for new tests.
* fast/text/font-face-set-javascript.html: Added tests for handling of null, also added tests for
the has function.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199216 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agosync-buildbot.js doesn't mark disappeared builds as failed
rniwa@webkit.org [Fri, 8 Apr 2016 04:21:32 +0000 (04:21 +0000)]
sync-buildbot.js doesn't mark disappeared builds as failed
https://bugs.webkit.org/show_bug.cgi?id=156386

Reviewed by Chris Dumez.

Fix a bug that new syncing script doesn't mark builds that it scheduled but doesn't appear when queried
by buildbot's JSON API. These are builds that got canceled by humans (e.g. buildbot was restarted, data
loss, pending build was canceled, etc...)

* server-tests/tools-buildbot-triggerable-tests.js: Added a test case.
* tools/js/buildbot-triggerable.js:
(BuildbotTriggerable.prototype._pullBuildbotOnAllSyncers): Added a set of build requests we've matched
against BuildbotBuildEntry's. Mark build requests that didn't have any entry but supposed to be in either
'scheduled' or 'running' status as failed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199215 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove ENABLE(ENABLE_ES6_CLASS_SYNTAX) guards
commit-queue@webkit.org [Fri, 8 Apr 2016 04:17:23 +0000 (04:17 +0000)]
Remove ENABLE(ENABLE_ES6_CLASS_SYNTAX) guards
https://bugs.webkit.org/show_bug.cgi?id=156384

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-04-07
Reviewed by Ryosuke Niwa.

.:

* Source/cmake/WebKitFeatures.cmake:

Source/JavaScriptCore:

* Configurations/FeatureDefines.xcconfig:
* features.json: Mark as Done.
* parser/Parser.cpp:
(JSC::Parser<LexerType>::parseExportDeclaration):
(JSC::Parser<LexerType>::parseStatementListItem):
(JSC::Parser<LexerType>::parsePrimaryExpression):
(JSC::Parser<LexerType>::parseMemberExpression):

Source/WebCore:

* Configurations/FeatureDefines.xcconfig:

Source/WebKit/mac:

* Configurations/FeatureDefines.xcconfig:

Source/WebKit2:

* Configurations/FeatureDefines.xcconfig:

Source/WTF:

* wtf/FeatureDefines.h:

Tools:

* Scripts/webkitperl/FeatureList.pm:
* TestWebKitAPI/Configurations/FeatureDefines.xcconfig:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199214 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] Media playback button should use appearance
dino@apple.com [Fri, 8 Apr 2016 03:59:45 +0000 (03:59 +0000)]
[iOS] Media playback button should use appearance
https://bugs.webkit.org/show_bug.cgi?id=156388

Reviewed by Simon Fraser.

With the recent change in backdrop appearance, we can
now use the system style directly for the play button.

While I was here I also updated the artwork to the
latest style (slightly rounded corners on the triangle).

Covered by the test in ManualTests/ios/start-playback-button-appearance.html.

* Modules/mediacontrols/mediaControlsiOS.css: Move the clip onto the backdrop
element. Use an appearance insted.
* Modules/mediacontrols/mediaControlsiOS.js: Remove the tint element, and
set the highlight on the glyph instead.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199213 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] Play button on video is too dark
dino@apple.com [Fri, 8 Apr 2016 03:59:21 +0000 (03:59 +0000)]
[iOS] Play button on video is too dark
https://bugs.webkit.org/show_bug.cgi?id=156383
<rdar://problem/23540816>

Reviewed by Simon Fraser.

.:

Add a manual test for iOS that shows the expected appearance
of a video element. Unfortunately, due to the way we take
snapshots on iOS within our test runner, we don't get the
platform blurring effect, which means an automated test
won't work.

* ManualTests/ios/start-playback-button-appearance-expected.html: Added.
* ManualTests/ios/start-playback-button-appearance.html: Added.

Source/WebKit2:

Elements that are backed by a layer with either LightBackdropAppearance
or DarkBackdropAppearance are actually a combination of a
few layers (inside a special view). If we apply a mask to one
of those layers, it needs to be attached to the correct
child layer.

* Shared/mac/RemoteLayerTreePropertyApplier.mm:
(WebKit::RemoteLayerTreePropertyApplier::applyProperties): If we have
one of the special appearance flags, apply the mask layer to
a particular child, rather than the layer itself.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199212 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWatchlist modifications:
dino@apple.com [Fri, 8 Apr 2016 03:58:46 +0000 (03:58 +0000)]
Watchlist modifications:
- remove roger_fong
- add myself to a few areas
- remove graouts from WebInspectorAPI

* Scripts/webkitpy/common/config/watchlist:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199211 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoA/B testing bots should prioritize user created test groups
rniwa@webkit.org [Fri, 8 Apr 2016 03:19:22 +0000 (03:19 +0000)]
A/B testing bots should prioritize user created test groups
https://bugs.webkit.org/show_bug.cgi?id=156375

Reviewed by Chris Dumez.

Order build requests preferring user created ones over ones automatically created by detect-changes.js.

Also fixed a bug in BuildbotSyncer.scheduleFirstRequestInGroupIfAvailable that it was scheduling a new
build request on a builder/slave even when we had previously scheduled another build request.

* public/include/build-requests-fetcher.php:
(BuildRequestsFetcher::fetch_incomplete_requests_for_triggerable): Order build requested based on
author_order which is 0 when it's created by an user and 1 when it's created by detect-changes.js.
Since we're using ascending order, this would put user created test groups first.
* server-tests/api-build-requests-tests.js: Updated an existing test case and added a new test case
for testing that build requests for an user created test group shows up first.
* server-tests/resources/mock-data.js:
(MockData.addAnotherMockTestGroup): Takes an extra argument to specify the author name.
* server-tests/tools-buildbot-triggerable-tests.js: Added a test case for testing that build requests
for an user created test group shows up first.
* tools/js/buildbot-syncer.js:
(BuildbotSyncer): Added _slavesWithNewRequests to keep track of build slaves on which we have already
scheduled new build requests. Don't schedule more requests on these slaves.
(BuildbotSyncer.prototype.scheduleRequest):
(BuildbotSyncer.prototype.scheduleFirstRequestInGroupIfAvailable): Add the specified slave name (or null
when slaveList is not specified) to _slavesWithNewRequests.
(BuildbotSyncer.prototype.pullBuildbot): Clear the set after pulling buildbot since any build request
we have previously scheduled should be included in one of the entires now.
* unit-tests/buildbot-syncer-tests.js: Added test cases for the aforementioned bug.
(sampleiOSConfig): Added a second slave for new test cases.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199210 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoImplementing caching transition puts that need to reallocate with indexing storage
fpizlo@apple.com [Fri, 8 Apr 2016 02:11:48 +0000 (02:11 +0000)]
Implementing caching transition puts that need to reallocate with indexing storage
https://bugs.webkit.org/show_bug.cgi?id=130914

Reviewed by Saam Barati.

Source/JavaScriptCore:

This enables the IC's put_by_id path to handle reallocating the out-of-line storage even if
the butterfly has indexing storage. Like the DFG, we do this by calling operations that
reallocate the butterfly. Those use JSObject API and do all of the nasty work for us, like
triggering a barrier.

This does a bunch of refactoring to how PolymorphicAccess makes calls. It's a lot easier to
do it now because the hard work is hidden under AccessGenerationState methods. This means
that custom accessors now share logic with put_by_id transitions.

* bytecode/PolymorphicAccess.cpp:
(JSC::AccessGenerationState::succeed):
(JSC::AccessGenerationState::calculateLiveRegistersForCallAndExceptionHandling):
(JSC::AccessGenerationState::preserveLiveRegistersToStackForCall):
(JSC::AccessGenerationState::originalCallSiteIndex):
(JSC::AccessGenerationState::emitExplicitExceptionHandler):
(JSC::AccessCase::AccessCase):
(JSC::AccessCase::transition):
(JSC::AccessCase::generate):
(JSC::PolymorphicAccess::regenerate):
* bytecode/PolymorphicAccess.h:
(JSC::AccessGenerationState::needsToRestoreRegistersIfException):
(JSC::AccessGenerationState::liveRegistersToPreserveAtExceptionHandlingCallSite):
* dfg/DFGOperations.cpp:
* dfg/DFGOperations.h:
* jit/JITOperations.cpp:
* jit/JITOperations.h:

LayoutTests:

* js/regress/put-by-id-transition-with-indexing-header-expected.txt: Added.
* js/regress/put-by-id-transition-with-indexing-header.html: Added.
* js/regress/script-tests/put-by-id-transition-with-indexing-header.js: Added.
(allocate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199209 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemote Inspector: When disallowing remote inspection on a debuggable, a listing is...
commit-queue@webkit.org [Fri, 8 Apr 2016 01:27:42 +0000 (01:27 +0000)]
Remote Inspector: When disallowing remote inspection on a debuggable, a listing is still sent to debuggers
https://bugs.webkit.org/show_bug.cgi?id=156380
<rdar://problem/25323727>

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-04-07
Reviewed by Timothy Hatcher.

* inspector/remote/RemoteInspector.mm:
(Inspector::RemoteInspector::updateTarget):
(Inspector::RemoteInspector::updateAutomaticInspectionCandidate):
When a target has been updated and it no longer generates a listing,
we should remove the old listing as that is now stale and should
not be sent. Not generating a listing means this target is no
longer allowed to be debugged.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199208 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: OpenResourceDialog should keep its resources list up-to-date
mattbaker@apple.com [Fri, 8 Apr 2016 01:24:47 +0000 (01:24 +0000)]
Web Inspector: OpenResourceDialog should keep its resources list up-to-date
https://bugs.webkit.org/show_bug.cgi?id=155321
<rdar://problem/25093890>

Reviewed by Timothy Hatcher.

The Quick Open dialog should listen for resource change events, refreshing
the resource list and current query results as needed.

* UserInterface/Views/OpenResourceDialog.js:
(WebInspector.OpenResourceDialog):
(WebInspector.OpenResourceDialog.prototype.didDismissDialog):
Unregister resource event handlers.

(WebInspector.OpenResourceDialog.prototype.didPresentDialog):
Register resource event handlers and add main frame resources.

(WebInspector.OpenResourceDialog.prototype._addResource):
Add resource to the query controller, if valid. Optionally suppress
the potentially expensive filter update, which is useful when adding
multiple resources at once.

(WebInspector.OpenResourceDialog.prototype._addResourcesForFrame):
Add the entire frame resource tree and update dialog filter.

(WebInspector.OpenResourceDialog.prototype._mainResourceDidChange):
(WebInspector.OpenResourceDialog.prototype._resourceWasAdded):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199207 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRoll out the css change in mediaControlsApple.css that has been causing assertions...
adachan@apple.com [Fri, 8 Apr 2016 01:22:11 +0000 (01:22 +0000)]
Roll out the css change in mediaControlsApple.css that has been causing assertions in layout for multiple tests
https://bugs.webkit.org/show_bug.cgi?id=156381

Rubber-stamped by Alexey Proskuryakov.

Source/WebCore:

* Modules/mediacontrols/mediaControlsApple.css:
(::-webkit-media-controls):
Remove overflow: hidden.

LayoutTests:

* platform/mac/TestExpectations:
* platform/mac/media/media-document-audio-repaint-expected.txt:
* platform/mac/media/video-zoom-controls-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199206 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Not necessary to validate webinspectord connection on iOS
commit-queue@webkit.org [Fri, 8 Apr 2016 00:40:21 +0000 (00:40 +0000)]
Web Inspector: Not necessary to validate webinspectord connection on iOS
https://bugs.webkit.org/show_bug.cgi?id=156377
<rdar://problem/25612460>

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-04-07
Reviewed by Simon Fraser.

* inspector/remote/RemoteInspectorXPCConnection.h:
* inspector/remote/RemoteInspectorXPCConnection.mm:
(Inspector::RemoteInspectorXPCConnection::handleEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199205 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoClearing the application cache doesn't work.
commit-queue@webkit.org [Fri, 8 Apr 2016 00:19:45 +0000 (00:19 +0000)]
Clearing the application cache doesn't work.
https://bugs.webkit.org/show_bug.cgi?id=156354
rdar://problem/22369239

Patch by Jeremy Jones <jeremyj@apple.com> on 2016-04-07
Reviewed by Brady Eidson.

Use the correct "ApplicationCache" directory.
Delete the caches, not just the entries.

* UIProcess/WebsiteData/WebsiteDataStore.cpp:
(WebKit::WebsiteDataStore::fetchData):
(WebKit::WebsiteDataStore::removeData):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199204 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRename ArrayMode::supportsLength to supportsSelfLength
keith_miller@apple.com [Fri, 8 Apr 2016 00:10:20 +0000 (00:10 +0000)]
Rename ArrayMode::supportsLength to supportsSelfLength
https://bugs.webkit.org/show_bug.cgi?id=156374

Reviewed by Filip Pizlo.

The name supportsLength is confusing because TypedArray have a
length function however it is on the prototype and not on the
instance. supportsSelfLength makes more sense since we use the
function during fixup to tell if we can intrinsic the length
property lookup on self accesses.

* dfg/DFGArrayMode.h:
(JSC::DFG::ArrayMode::supportsSelfLength):
(JSC::DFG::ArrayMode::supportsLength): Deleted.
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::attemptToMakeGetArrayLength):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199203 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r199199.
jiewen_tan@apple.com [Thu, 7 Apr 2016 23:38:32 +0000 (23:38 +0000)]
Unreviewed, rolling out r199199.

Revision breaks layout tests

Reverted changeset:

"fast/loader/opaque-base-url.html crashing during mac and ios
debug tests"
https://bugs.webkit.org/show_bug.cgi?id=156179
http://trac.webkit.org/changeset/199199

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199202 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: ProfileView source links are off by 1 line, worse in pretty printed...
commit-queue@webkit.org [Thu, 7 Apr 2016 23:28:08 +0000 (23:28 +0000)]
Web Inspector: ProfileView source links are off by 1 line, worse in pretty printed code
https://bugs.webkit.org/show_bug.cgi?id=156371

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-04-07
Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

* inspector/protocol/ScriptProfiler.json:
Clarify that these locations are 1-based.

Source/WebInspectorUI:

* UserInterface/Views/ProfileDataGridNode.js:
(WebInspector.ProfileDataGridNode.prototype._displayContent):
Switch the 1-based locations in the CCT data structure to 0-based for SourceCodeLocation.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199201 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMake it possible to test effect of view exposed rect on tiled backing
simon.fraser@apple.com [Thu, 7 Apr 2016 23:14:25 +0000 (23:14 +0000)]
Make it possible to test effect of view exposed rect on tiled backing
https://bugs.webkit.org/show_bug.cgi?id=156365

Reviewed by Tim Horton.

Source/WebCore:

Implement Internals::setViewExposedRect().

When the viewExposedRect is non-null, assume that we're scrollable on both axes
to avoid creation of huge tiles in this scenario.

We also need to call adjustTiledBackingScrollability() when setViewExposedRect()
has been called.

Tests: tiled-drawing/tile-coverage-view-exposed-rect.html
       tiled-drawing/tile-size-view-exposed-rect.html

* page/FrameView.cpp:
(WebCore::FrameView::adjustTiledBackingScrollability):
(WebCore::FrameView::setViewExposedRect):
* testing/Internals.cpp:
(WebCore::Internals::setViewExposedRect):
* testing/Internals.h:
* testing/Internals.idl:

Tools:

Expose testRunner.setViewSize() and internals.setViewExposedRect() to enable
testing of tile coverage when setViewExposedRect() is passed a non-null rectangle.

testRunner.setViewSize() is used instead of using window.resizeTo(), since we
can't easily resize a window to larger than the screen being tested on.

* DumpRenderTree/TestRunner.cpp:
(setViewSizeCallback):
(TestRunner::staticFunctions):
* DumpRenderTree/TestRunner.h:
* DumpRenderTree/mac/TestRunnerMac.mm:
(TestRunner::setViewSize):
* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::postSetViewSize):
* WebKitTestRunner/InjectedBundle/InjectedBundle.h:
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::setViewSize):
* WebKitTestRunner/InjectedBundle/TestRunner.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):

LayoutTests:

Tests for tile size and tile coverage when we have an exposed view rect.

* tiled-drawing/tile-coverage-view-exposed-rect-expected.txt: Added.
* tiled-drawing/tile-coverage-view-exposed-rect.html: Added.
* tiled-drawing/tile-size-view-exposed-rect-expected.txt: Added.
* tiled-drawing/tile-size-view-exposed-rect.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199200 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agofast/loader/opaque-base-url.html crashing during mac and ios debug tests
jiewen_tan@apple.com [Thu, 7 Apr 2016 22:44:37 +0000 (22:44 +0000)]
fast/loader/opaque-base-url.html crashing during mac and ios debug tests
https://bugs.webkit.org/show_bug.cgi?id=156179
<rdar://problem/25507719>

Reviewed by Andy Estes.

Source/WebCore:

A relative URL other than "#" with a non-hierarchical base is invalid, but prior to this
change the URL's string would still contain the invalid relative URL. To avoid mistakes
where we might later treat this URL string as a parsed URL string, set the string to
"about:blank" instead.

Test: fast/url/data-uri-based-urls.html

* platform/URL.cpp:
(WebCore::URL::init):

LayoutTests:

* TestExpectations:
* fast/url/data-uri-based-urls-expected.txt: Added.
* fast/url/data-uri-based-urls.html: Added.
* fast/url/relative-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199199 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSkip a couple more tests that are asserting in FrameView::scheduleRelayoutOfSubtree().
adachan@apple.com [Thu, 7 Apr 2016 22:35:30 +0000 (22:35 +0000)]
Skip a couple more tests that are asserting in FrameView::scheduleRelayoutOfSubtree().

Unreviewed test gardening.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199198 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Automation: implement Automation.addSingleCookie
bburg@apple.com [Thu, 7 Apr 2016 22:34:58 +0000 (22:34 +0000)]
Web Automation: implement Automation.addSingleCookie
https://bugs.webkit.org/show_bug.cgi?id=156319
<rdar://problem/25589605>

Reviewed by Timothy Hatcher.

Source/WebCore:

* platform/Cookie.h: Document the units used by the 'expires' field.

Source/WebKit2:

Implement this command by converting the protocol cookie to
WebCore::Cookie, then sending the cookie to NetworkProcess to
be added to the storage session using CookieJar::addCookie.

* UIProcess/Automation/Automation.json:
Clarify the units used in the 'expires' field and how the default value
for the domain field should be computed.

* UIProcess/Automation/WebAutomationSession.cpp:
(WebKit::WebAutomationSession::addSingleCookie):
Parse the cookie and send it out via WebCookieManagerProxy.

* UIProcess/WebCookieManagerProxy.cpp:
(WebKit::WebCookieManagerProxy::addCookie): Added.
* UIProcess/WebCookieManagerProxy.h:
* WebProcess/Cookies/WebCookieManager.cpp:
(WebKit::WebCookieManager::addCookie):
* WebProcess/Cookies/WebCookieManager.h:
* WebProcess/Cookies/WebCookieManager.messages.in:
Forward the message to WebCore::addCookie.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199197 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd ImageBitmap as under consideration on Feature Status page
jond@apple.com [Thu, 7 Apr 2016 22:31:56 +0000 (22:31 +0000)]
Add ImageBitmap as under consideration on Feature Status page
https://bugs.webkit.org/show_bug.cgi?id=156362

Reviewed by Timothy Hatcher.

* features.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199196 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoInclude Conical Gradients on the Feature Status page.
jond@apple.com [Thu, 7 Apr 2016 22:31:38 +0000 (22:31 +0000)]
Include Conical Gradients on the Feature Status page.
https://bugs.webkit.org/show_bug.cgi?id=156363

Reviewed by Timothy Hatcher.

* features.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199195 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix.
bdakin@apple.com [Thu, 7 Apr 2016 22:28:56 +0000 (22:28 +0000)]
Build fix.

* platform/mac/WebVideoFullscreenInterfaceMac.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199194 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoIn WK1 WebVideoFullscreen interface may be accessed from WK1 main thread instead...
commit-queue@webkit.org [Thu, 7 Apr 2016 22:26:27 +0000 (22:26 +0000)]
In WK1 WebVideoFullscreen interface may be accessed from WK1 main thread instead of UI thread.
https://bugs.webkit.org/show_bug.cgi?id=154252
rdar://problem/22460539

Patch by Jeremy Jones <jeremyj@apple.com> on 2016-04-07
Reviewed by Eric Carlson.

In WebKit1, Javascript can cause enter fullscreen to happen on the main thead, which is not
necessarily the UI thread. This can cause autolayout errors. Move this code to the UI thread.

* platform/ios/WebVideoFullscreenControllerAVKit.mm:
(WebVideoFullscreenControllerContext::setUpFullscreen): Move setup to the UI thread.
* platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
(-[WebAVPlayerLayer layoutSublayers]): Move call to resolveBounds to the UI thread.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199193 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMark http/tests/cache/disk-cache/disk-cache-validation-back-navigation-policy.html...
jiewen_tan@apple.com [Thu, 7 Apr 2016 22:25:51 +0000 (22:25 +0000)]
Mark http/tests/cache/disk-cache/disk-cache-validation-back-navigation-policy.html as timeout in ios-simulator-wk2
https://bugs.webkit.org/show_bug.cgi?id=149087

Unreviewed, test gardening.

* platform/ios-simulator-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199192 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMigrate legacy perf dashboard tests to mocha.js based tests
rniwa@webkit.org [Thu, 7 Apr 2016 22:10:50 +0000 (22:10 +0000)]
Migrate legacy perf dashboard tests to mocha.js based tests
https://bugs.webkit.org/show_bug.cgi?id=156335

Reviewed by Chris Dumez.

Migrated all legacy run-tests.js tests to mocha.js based tests. Since the new harness uses Promise
for most of asynchronous operations, refactored the tests to use Promises as well, and added more
assertions where appropriate.

Also consolidated common helper functions into server-tests/resources/common-operations.js.
Unfortunately there were multiple inconsistent implementations of addBuilder/addSlave. Some were
taking an array of reports while others were taking a single report. New shared implementation in
common-operations.js now takes a single report.

Also decreased the timeout in most tests from 10s to 1s so that tests fail early when they timeout.
Most of tests are passing under 100ms on my computer so 1s should be plenty still.

* run-tests.js: Removed.
* server-tests/admin-platforms-tests.js: Moved from tests/admin-platforms.js.
(reportsForDifferentPlatforms):
* server-tests/admin-reprocess-report-tests.js: Moved from tests/admin-reprocess-report.js.
(.addBuilder): Moved to common-operations.js.
* server-tests/api-build-requests-tests.js:
* server-tests/api-manifest.js: Use MockData.resetV3Models() instead of manually clearing maps.
* server-tests/api-measurement-set-tests.js: Moved from tests/api-measurement-set.js.
(.queryPlatformAndMetric):
(.format):
* server-tests/api-report-commits-tests.js: Moved from tests/api-report-commits.js.
* server-tests/api-report-tests.js: Moved from tests/api-report.js.
(.emptyReport):
(.emptySlaveReport):
(.reportWithSameSubtestName):
* server-tests/resources/common-operations.js: Added.
(addBuilderForReport): Extracted from tests.
(addSlaveForReport): Ditto.
(connectToDatabaseInEveryTest): Added.
(submitReport): Extracted from admin-platforms-tests.js.
* server-tests/resources/test-server.js:
(TestServer): Make TestServer a singleton since it doesn't make any sense for each module to start
its own Apache instance (that would certainly will fail).
* server-tests/tools-buildbot-triggerable-tests.js:
* tests: Removed.
* tools/js/database.js:
(Database.prototype.selectAll): Added.
(Database.prototype.selectFirstRow): Added.
(Database.prototype.selectRows): Added. Dynamically construct a query string based on arguments.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199191 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix.
bdakin@apple.com [Thu, 7 Apr 2016 22:02:30 +0000 (22:02 +0000)]
Build fix.

* platform/mac/WebVideoFullscreenInterfaceMac.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199190 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd Web Animations API to Feature Status Page
jond@apple.com [Thu, 7 Apr 2016 22:01:34 +0000 (22:01 +0000)]
Add Web Animations API to Feature Status Page
https://bugs.webkit.org/show_bug.cgi?id=156360

Reviewed by Timothy Hatcher.

* features.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@199189 268f45cc-cd09-0410-ab3c-d52691b4dbfc