WebKit-https.git
4 years ago[iOS] WebSQL operations are not performed after device is locked
dbates@webkit.org [Fri, 8 May 2015 18:59:31 +0000 (18:59 +0000)]
[iOS] WebSQL operations are not performed after device is locked
https://bugs.webkit.org/show_bug.cgi?id=137503
<rdar://problem/20844952>

Rubber-stamped by Alexey Proskuryakov.

Add a manual test to help verify that we do not regress this issue.

* ManualTests/ios/execute-sql-transaction-callback-after-locking-unlocking-device-in-earlier-transaction.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184002 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Mac] Playback target clients do not unregister on page reload
eric.carlson@apple.com [Fri, 8 May 2015 18:45:40 +0000 (18:45 +0000)]
[Mac] Playback target clients do not unregister on page reload
https://bugs.webkit.org/show_bug.cgi?id=144761

Reviewed by Brady Eidson.

* dom/Document.cpp:
(WebCore::Document::prepareForDestruction): Unregister all target picker clients.

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::registerWithDocument): Register for page cache callback.
(WebCore::HTMLMediaElement::unregisterWithDocument): Unregister for page cache callback.
(WebCore::HTMLMediaElement::documentWillSuspendForPageCache): New.
(WebCore::HTMLMediaElement::documentDidResumeFromPageCache): New.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184001 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Styles sidebar editing with incomplete property looks poor in UI
commit-queue@webkit.org [Fri, 8 May 2015 18:37:49 +0000 (18:37 +0000)]
Web Inspector: Styles sidebar editing with incomplete property looks poor in UI
https://bugs.webkit.org/show_bug.cgi?id=141692

Patch by Tobias Reiss <tobi+webkit@basecode.de> on 2015-05-08
Reviewed by Timothy Hatcher.

Add "css-rule" Formatter that breaks CSS declarations into multiple lines,
keeps comments and invalid styles and adds whitespace.

* Tools/PrettyPrinting/css-rule-tests/*.css: Added.
Add test cases.

* Tools/PrettyPrinting/index.html:
Enable Test setup to be able to run "css-rule" Formatter tests.

* UserInterface/Controllers/Formatter.js:
(Formatter.prototype._handleToken):
* UserInterface/Controllers/FormatterContentBuilder.js:
(FormatterContentBuilder.prototype.removeLastNewline):
(FormatterContentBuilder.prototype.removeLastWhitespace):
(FormatterContentBuilder.prototype._popFormattedContent):
(FormatterContentBuilder.prototype._popNewLine): Deleted.
* UserInterface/Views/CSSStyleDeclarationTextEditor.js:
(WebInspector.CSSStyleDeclarationTextEditor.prototype._formattedContentFromEditor):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._resetContent.update.set this):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._resetContent.update.get this):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._resetContent.update):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._resetContent):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._resetContent.update.countNewLineCharacters): Deleted.
(WebInspector.CSSStyleDeclarationTextEditor.prototype._resetContent.update.else): Deleted.
* UserInterface/Views/CodeMirrorFormatters.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184000 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix performance tests after r183954
bdakin@apple.com [Fri, 8 May 2015 18:20:59 +0000 (18:20 +0000)]
Fix performance tests after r183954
https://bugs.webkit.org/show_bug.cgi?id=144805

Reviewed by Alexey Proskuryakov.

Web* is traditionally a WK1 prefix, so use the WK2 prefix instead in WK2 so that
we don’t have two classes with the same name in the two projects.
* UIProcess/mac/WKImmediateActionController.mm:
(-[WKImmediateActionController _defaultAnimationController]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183999 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoThrottle RequestAnimationFrame in subframes that are outside the viewport
cdumez@apple.com [Fri, 8 May 2015 18:15:57 +0000 (18:15 +0000)]
Throttle RequestAnimationFrame in subframes that are outside the viewport
https://bugs.webkit.org/show_bug.cgi?id=144718
<rdar://problem/20688782>

Reviewed by Simon Fraser.

Source/WebCore:

Throttle RequestAnimationFrame in subframes that are outside the
viewport or have "display: none" for performance and power.

Tests: fast/animation/request-animation-frame-throttle-subframe-display-none.html
       fast/animation/request-animation-frame-throttle-subframe.html

* dom/Document.h:
(WebCore::Document::scriptedAnimationController):
* dom/ScriptedAnimationController.cpp:
(WebCore::ScriptedAnimationController::setThrottled):
(WebCore::ScriptedAnimationController::isThrottled):
* dom/ScriptedAnimationController.h:
* loader/FrameLoader.cpp:
(WebCore::FrameLoader::open):
Call FrameView::setFrameRect() only *after* the view has been
set on the Frame. Otherwise, setFrameRect() ends up calling
viewportContentsChanged() and we hit the
ASSERT(frame().view() == this) assertion in windowClipRect()
because the Frame still has its old FrameView. This is covered
by loader/go-back-to-different-window-size.html layout test.

* page/FrameView.cpp:
(WebCore::FrameView::viewportContentsChanged):
(WebCore::FrameView::applyRecursivelyWithVisibleRect):
(WebCore::FrameView::resumeVisibleImageAnimations):
(WebCore::FrameView::updateScriptedAnimationsThrottlingState):
(WebCore::FrameView::resumeVisibleImageAnimationsIncludingSubframes):
(WebCore::FrameView::updateThrottledDOMTimersState):
(WebCore::FrameView::scrollPositionChanged): Deleted.
(WebCore::FrameView::sendResizeEventIfNeeded): Deleted.
* page/FrameView.h:
* testing/Internals.cpp:
(WebCore::Internals::isRequestAnimationFrameThrottled):
(WebCore::Internals::isTimerThrottled): Deleted.
* testing/Internals.h:
* testing/Internals.idl:

LayoutTests:

* fast/animation/request-animation-frame-throttle-subframe-display-none-expected.txt: Added.
* fast/animation/request-animation-frame-throttle-subframe-display-none.html: Added.
Add layout test to test that RequestAnimationFrame is properly throttled
in "display: none" subframes.

* fast/animation/request-animation-frame-throttle-subframe-expected.txt: Added.
* fast/animation/request-animation-frame-throttle-subframe.html: Added.
Add layout test to test that RequestAnimationFrame is properly throttled
in frames that are outside the viewport.

* fast/animation/resources/requestAnimationFrame-frame-2.html: Added.
* fast/animation/resources/requestAnimationFrame-frame.html: Added.

* platform/win/TestExpectations:
Skip the 2 new tests on Windows as requestAnimationFrame throttling is
only supported on Cocoa.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183998 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix the iOS Simulator external SDK build following <http://trac.webkit.org/changeset...
dbates@webkit.org [Fri, 8 May 2015 17:56:25 +0000 (17:56 +0000)]
Fix the iOS Simulator external SDK build following <trac.webkit.org/changeset/181918>

Remove unused iOS Simulator-variant of CARenderServerRenderLayerWithTransform(). Also guard
IOSurfaceRef-variant of CARenderServerRenderLayerWithTransform() with USE(IOSURFACE)
since it is only applicable on platforms that use IOSurface.

* platform/spi/cocoa/QuartzCoreSPI.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183997 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove convenience constructors for TextRun
mmaxfield@apple.com [Fri, 8 May 2015 17:18:00 +0000 (17:18 +0000)]
Remove convenience constructors for TextRun
https://bugs.webkit.org/show_bug.cgi?id=144752

Source/WebCore:

These convenience constructors are unnecessary. Moving the code that makes the StringView
back to the call site will also help us make things more elegant in future refactoring.

Reviewed by Darin Adler.

No new tests because there is no behavior change.

* css/CSSPrimitiveValue.cpp:
(WebCore::CSSPrimitiveValue::formatNumberForCustomCSSText): Remove ambiguous call.
* platform/graphics/StringTruncator.cpp:
(WebCore::stringWidth):
* platform/graphics/TextRun.h:
(WebCore::TextRun::TextRun):
* platform/mac/DragImageMac.mm:
(WebCore::widthWithFont):
(WebCore::drawAtPoint):
* rendering/SimpleLineLayout.cpp:
(WebCore::SimpleLineLayout::canUseFor):
* rendering/SimpleLineLayoutTextFragmentIterator.cpp:
(WebCore::SimpleLineLayout::TextFragmentIterator::Style::Style):
(WebCore::SimpleLineLayout::TextFragmentIterator::runWidth):
* rendering/TextPainter.cpp:
(WebCore::TextPainter::paintText):

Source/WebKit/mac:

These convenience constructors are unnecessary. Moving the code that makes the StringView
back to the call site will also help us make things more elegant in future refactoring.

Reviewed by Darin Adler.

No new tests because there is no behavior change.

* Misc/WebKitNSStringExtras.mm:
(-[NSString _web_drawAtPoint:font:textColor:allowingFontSmoothing:]):
(-[NSString _web_widthWithFont:]):

Source/WTF:

Reviewed by Anders Carlsson.

No reason why StringView shouldn't have a StringImpl* constructor.

Test: StringView8Bit in TestWebKitAPI

* wtf/text/StringView.h: Add the constructor.

Tools:

Reviewed by Anders Carlsson.

Test the StringView which takes a StringImpl*.

* TestWebKitAPI/Tests/WTF/StringView.cpp:
(StringView8Bit): Testing is8Bit() on StringView

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183996 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] WTR doesn't correctly handle the Escape key
carlosgc@webkit.org [Fri, 8 May 2015 16:09:42 +0000 (16:09 +0000)]
[GTK] WTR doesn't correctly handle the Escape key
https://bugs.webkit.org/show_bug.cgi?id=144788

Reviewed by Martin Robinson.

Tools:

Correctly handle 0x001B character code in GTK+ event sender to be
recognized as Escape key, because gdk_unicode_to_keyval() doesn't
handle it. This fixes layout tests using keyDown(String.fromCharCode(0x001B), null);
to dismiss popup menus.

Fixes editing/selection/5354455-1.html.

* WebKitTestRunner/gtk/EventSenderProxyGtk.cpp:
(WTR::getGDKKeySymForKeyRef):

LayoutTests:

Unskip editing/selection/5354455-1.html.

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183995 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win] Unskip passing test case after r183981
commit-queue@webkit.org [Fri, 8 May 2015 15:12:50 +0000 (15:12 +0000)]
[Win] Unskip passing test case after r183981
https://bugs.webkit.org/show_bug.cgi?id=140517

Patch by Hyungwook Lee <hyungwook.lee@navercorp.com> on 2015-05-08
Reviewed by Brent Fulgham.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183994 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed EFL gardening on 8th May.
gyuyoung.kim@webkit.org [Fri, 8 May 2015 15:00:55 +0000 (15:00 +0000)]
Unreviewed EFL gardening on 8th May.

Mark flaky tests and timeout tests after using xorgdriver on EFL port.

* platform/efl/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183993 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Expose allowFileAccessFromFileURLs setting to WebKit2 GTK+ API
carlosgc@webkit.org [Fri, 8 May 2015 11:53:56 +0000 (11:53 +0000)]
[GTK] Expose allowFileAccessFromFileURLs setting to WebKit2 GTK+ API
https://bugs.webkit.org/show_bug.cgi?id=144748

Reviewed by Sergio Villar Senin.

Source/WebKit2:

This is needed by local applications loaded as a file URI that do XMLHttpRequests.

* UIProcess/API/gtk/WebKitSettings.cpp:
(webKitSettingsSetProperty):
(webKitSettingsGetProperty):
(webkit_settings_class_init):
(webkit_settings_get_allow_file_access_from_file_urls):
(webkit_settings_set_allow_file_access_from_file_urls):
* UIProcess/API/gtk/WebKitSettings.h:
* UIProcess/API/gtk/docs/webkit2gtk-4.0-sections.txt:
* UIProcess/API/gtk/docs/webkit2gtk-docs.sgml:

Tools:

* TestWebKitAPI/Tests/WebKit2Gtk/TestWebKitSettings.cpp:
(testWebKitSettings): Check the new setting is correctly
initialized and updated.
* TestWebKitAPI/Tests/WebKit2Gtk/TestWebKitWebContext.cpp:
(testWebContextSecurityFileXHR): Check XHR to local files is
allowed from file URLs after changing the setting.
(beforeAll): Add new test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183992 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Streams API] ReadableStream constructor start function should be able to error the...
youenn.fablet@crf.canon.fr [Fri, 8 May 2015 10:44:06 +0000 (10:44 +0000)]
[Streams API] ReadableStream constructor start function should be able to error the stream
https://bugs.webkit.org/show_bug.cgi?id=141162

Reviewed by Darin Adler.

Source/WebCore:

This patch implements the functionality of the ReadableStreamController error function.
It basically changes the state of the stream to errored, resolves the ready promise and rejects the closed promise.
Adding support to reject promise with any JSValue.

Support for storing the error is added to both reader and stream.

Test: streams/readable-stream-controller-error.html and rebased tests

* Modules/streams/ReadableStream.cpp:
(WebCore::ReadableStream::changeStateToErrored):
* Modules/streams/ReadableStream.h:
* Modules/streams/ReadableStreamReader.cpp:
(WebCore::ReadableStreamReader::initialize):
(WebCore::ReadableStreamReader::closed):
(WebCore::ReadableStreamReader::changeStateToClosed):
(WebCore::ReadableStreamReader::changeStateToErrored):
* Modules/streams/ReadableStreamReader.h:
* bindings/js/JSDOMPromise.h:
(WebCore::DeferredWrapper::reject):
* bindings/js/JSReadableStreamControllerCustom.cpp:
(WebCore::JSReadableStreamController::error):
* bindings/js/JSReadableStreamReaderCustom.cpp:
(WebCore::JSReadableStreamReader::closed):
* bindings/js/ReadableJSStream.cpp:
(WebCore::ReadableJSStream::createReader):
(WebCore::ReadableJSStream::storeError):
(WebCore::ReadableJSStream::Reader::storeError):
(WebCore::ReadableJSStream::jsController): Deleted.
* bindings/js/ReadableJSStream.h:

LayoutTests:

* streams/readable-stream-controller-error-expected.txt: Added.
* streams/readable-stream-controller-error.html: Added.
* streams/reference-implementation/bad-underlying-sources-expected.txt:
* streams/reference-implementation/readable-stream-templated-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183991 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r183985.
commit-queue@webkit.org [Fri, 8 May 2015 09:44:28 +0000 (09:44 +0000)]
Unreviewed, rolling out r183985.
https://bugs.webkit.org/show_bug.cgi?id=144796

broke loader/go-back-to-different-window-size.html (Requested
by kling on #webkit).

Reverted changeset:

"Throttle RequestAnimationFrame in subframes that are outside
the viewport"
https://bugs.webkit.org/show_bug.cgi?id=144718
http://trac.webkit.org/changeset/183985

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183990 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION(r183072): dfg-put-by-id-prototype-check.js.layout-dfg-eager-no-cjit fails...
ossy@webkit.org [Fri, 8 May 2015 09:37:56 +0000 (09:37 +0000)]
REGRESSION(r183072): dfg-put-by-id-prototype-check.js.layout-dfg-eager-no-cjit fails on AArch64 Linux
https://bugs.webkit.org/show_bug.cgi?id=144256

Unreviewed gardening, skip the failing test on AArch64 Linux.

* js/script-tests/dfg-put-by-val-direct-to-generic-array.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183989 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoOptimize serialization of quoted JSON strings.
akling@apple.com [Fri, 8 May 2015 08:44:23 +0000 (08:44 +0000)]
Optimize serialization of quoted JSON strings.
<https://webkit.org/b/144754>

Reviewed by Darin Adler.

Source/JavaScriptCore:

Optimized the serialization of quoted strings into JSON by moving the logic into
StringBuilder so it can make smarter decisions about buffering.

12% progression on Kraken/json-stringify-tinderbox (on my Mac Pro.)

* bytecompiler/NodesCodegen.cpp:
(JSC::ObjectPatternNode::toString): Use the new StringBuilder API.

* runtime/JSONObject.h:
* runtime/JSONObject.cpp:
(JSC::Stringifier::Holder::appendNextProperty):
(JSC::appendStringToStringBuilder): Deleted.
(JSC::appendQuotedJSONStringToBuilder): Deleted.
(JSC::Stringifier::appendQuotedString): Deleted.
(JSC::Stringifier::appendStringifiedValue): Moved the bulk of this logic
to StringBuilder and call that from here.

Source/WebKit2:

* NetworkProcess/cache/NetworkCacheEntry.cpp:
(WebKit::NetworkCache::Entry::asJSON): Use the new StringBuilder API.

Source/WTF:

Add a StringBuilder API for appending a quoted JSON string. This is used by
JSON.stringify() to implement efficient appending of strings while escaping
quotes, control characters and \uNNNN-style characters.

The main benefit comes from only doing a single buffer expansion up front,
instead of doing it every time we append something. The fudge factor is pretty
large, since the maximum number of output characters per input character is 6.

The first landing of this patch had two bugs in it:

- Made \uNNNN escapes uppercase hexadecimal instead of lowercase.
- Didn't preallocate enough space for 8-bit input strings.

Both were caught by existing tests on our bots, and both were due to last-minute
changes before landing. :/

* wtf/text/StringBuilder.cpp:
(WTF::appendQuotedJSONStringInternal):
(WTF::StringBuilder::appendQuotedJSONString):
* wtf/text/StringBuilder.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183988 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSharedBuffer::m_size should be initialized to 0 when defining it
youenn.fablet@crf.canon.fr [Fri, 8 May 2015 08:04:07 +0000 (08:04 +0000)]
SharedBuffer::m_size should be initialized to 0 when defining it
https://bugs.webkit.org/show_bug.cgi?id=144740

Reviewed by Darin Adler.

Setting m_size to 0 when declaring it and updating constructor definitions.

No change in behaviour.

* platform/SharedBuffer.cpp:
(WebCore::SharedBuffer::SharedBuffer):
* platform/SharedBuffer.h:
* platform/cf/SharedBufferCF.cpp:
(WebCore::SharedBuffer::SharedBuffer):
* platform/soup/SharedBufferSoup.cpp:
(WebCore::SharedBuffer::SharedBuffer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183987 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r183945.
commit-queue@webkit.org [Fri, 8 May 2015 07:56:51 +0000 (07:56 +0000)]
Unreviewed, rolling out r183945.
https://bugs.webkit.org/show_bug.cgi?id=144789

"It broke all the GTK+ tests" (Requested by KaL on #webkit).

Reverted changeset:

"[GTK] Checks for DEVELOPMENT_BUILD are all wrong"
https://bugs.webkit.org/show_bug.cgi?id=144746
http://trac.webkit.org/changeset/183945

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183986 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoThrottle RequestAnimationFrame in subframes that are outside the viewport
cdumez@apple.com [Fri, 8 May 2015 06:51:55 +0000 (06:51 +0000)]
Throttle RequestAnimationFrame in subframes that are outside the viewport
https://bugs.webkit.org/show_bug.cgi?id=144718
<rdar://problem/20688782>

Reviewed by Simon Fraser.

Source/WebCore:

Throttle RequestAnimationFrame in subframes that are outside the
viewport or have "display: none" for performance and power.

Tests: fast/animation/request-animation-frame-throttle-subframe-display-none.html
       fast/animation/request-animation-frame-throttle-subframe.html

* dom/Document.h:
(WebCore::Document::scriptedAnimationController):
* dom/ScriptedAnimationController.cpp:
(WebCore::ScriptedAnimationController::setThrottled):
(WebCore::ScriptedAnimationController::isThrottled):
* dom/ScriptedAnimationController.h:
* page/FrameView.cpp:
(WebCore::FrameView::viewportContentsChanged):
(WebCore::FrameView::applyRecursivelyWithVisibleRect):
(WebCore::FrameView::resumeVisibleImageAnimations):
(WebCore::FrameView::updateScriptedAnimationsThrottlingState):
(WebCore::FrameView::resumeVisibleImageAnimationsIncludingSubframes):
(WebCore::FrameView::updateThrottledDOMTimersState):
(WebCore::FrameView::scrollPositionChanged): Deleted.
(WebCore::FrameView::sendResizeEventIfNeeded): Deleted.
* page/FrameView.h:
* testing/Internals.cpp:
(WebCore::Internals::isRequestAnimationFrameThrottled):
(WebCore::Internals::isTimerThrottled): Deleted.
* testing/Internals.h:
* testing/Internals.idl:

LayoutTests:

* fast/animation/request-animation-frame-throttle-subframe-display-none-expected.txt: Added.
* fast/animation/request-animation-frame-throttle-subframe-display-none.html: Added.
Add layout test to test that RequestAnimationFrame is properly throttled
in "display: none" subframes.

* fast/animation/request-animation-frame-throttle-subframe-expected.txt: Added.
* fast/animation/request-animation-frame-throttle-subframe.html: Added.
Add layout test to test that RequestAnimationFrame is properly throttled
in frames that are outside the viewport.

* fast/animation/resources/requestAnimationFrame-frame-2.html: Added.
* fast/animation/resources/requestAnimationFrame-frame.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183985 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoCalculate source-size length as a float
yoav@yoav.ws [Fri, 8 May 2015 06:50:57 +0000 (06:50 +0000)]
Calculate source-size length as a float
https://bugs.webkit.org/show_bug.cgi?id=144766

Reviewed by Dean Jackson.

Source/WebCore:

Make sure that the source-size length is calculated as a float,
to align with the spec.

Test: fast/dom/HTMLImageElement/sizes/image-sizes-1x.html
      fast/dom/HTMLImageElement/sizes/image-sizes-2x.html

* css/SourceSizeList.cpp:
(WebCore::parseSizesAttribute):
(WebCore::defaultLength):
(WebCore::computeLength):
* css/SourceSizeList.h:
* html/HTMLImageElement.cpp:
(WebCore::HTMLImageElement::parseAttribute):
* html/parser/HTMLPreloadScanner.cpp:
(WebCore::TokenPreloadScanner::StartTagScanner::processAttributes):
* html/parser/HTMLSrcsetParser.cpp:
(WebCore::pickBestImageCandidate):
(WebCore::bestFitSourceForImageAttributes):
* html/parser/HTMLSrcsetParser.h:

LayoutTests:

Add tests that make sure fractional source-size lengths are supported
and are calculated as floats.

* fast/dom/HTMLImageElement/sizes/image-sizes-1x-expected.txt:
* fast/dom/HTMLImageElement/sizes/image-sizes-1x.html:
* fast/dom/HTMLImageElement/sizes/image-sizes-2x-expected.txt:
* fast/dom/HTMLImageElement/sizes/image-sizes-2x.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183984 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r183895.
commit-queue@webkit.org [Fri, 8 May 2015 06:32:26 +0000 (06:32 +0000)]
Unreviewed, rolling out r183895.
https://bugs.webkit.org/show_bug.cgi?id=144786

Many tests became flaky (Requested by ap on #webkit).

Reverted changeset:

"Expose more font weights for -apple-system"
https://bugs.webkit.org/show_bug.cgi?id=144707
http://trac.webkit.org/changeset/183895

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183983 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove touchadjustment tests
mrobinson@webkit.org [Fri, 8 May 2015 06:30:23 +0000 (06:30 +0000)]
Remove touchadjustment tests
https://bugs.webkit.org/show_bug.cgi?id=144780

Reviewed by Daniel Bates.

These tests test a feature that has been removed from WebKit entirely and all
ports skip the tests. This code was removed in http://trac.webkit.org/changeset/162231.

* platform/efl/TestExpectations: Remove expectations.
* platform/gtk/TestExpectations: Ditto.
* platform/ios-simulator/TestExpectations: Ditto.
* platform/mac/TestExpectations: Ditto.
* platform/win/TestExpectations: Ditto.
* platform/wincairo/TestExpectations:
* touchadjustment: Removed all files in this directory.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183982 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd first-letter assert exception in RenderButton::setupInnerStyle().
commit-queue@webkit.org [Fri, 8 May 2015 06:13:02 +0000 (06:13 +0000)]
Add first-letter assert exception in RenderButton::setupInnerStyle().
https://bugs.webkit.org/show_bug.cgi?id=140517OB

Patch by Hyungwook Lee <hyungwook.lee@navercorp.com> on 2015-05-07
Reviewed by Darin Adler.

StyleResolver references inner block's render style temporary in
RenderBlock::styleDidChange() when we have first-letter css pseudo
element.

* rendering/RenderButton.cpp:
(WebCore::RenderButton::setupInnerStyle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183981 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, fix build after r183970.
cdumez@apple.com [Fri, 8 May 2015 06:01:17 +0000 (06:01 +0000)]
Unreviewed, fix build after r183970.

WEBCORE_EXPORT should not be used on inlined functions.

* platform/graphics/ca/GraphicsLayerCA.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183980 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoNew results after r183970. Fixed elements are now always considered to intersect
simon.fraser@apple.com [Fri, 8 May 2015 04:57:25 +0000 (04:57 +0000)]
New results after r183970. Fixed elements are now always considered to intersect
the coverage rect.

* platform/mac-wk2/tiled-drawing/scrolling/fixed/four-bars-zoomed-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183979 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDon't register for Lookup notifications until needed
conrad_shultz@apple.com [Fri, 8 May 2015 04:23:41 +0000 (04:23 +0000)]
Don't register for Lookup notifications until needed
https://bugs.webkit.org/show_bug.cgi?id=144783

Reviewed by Tim Horton.

* UIProcess/API/mac/WKView.mm:
Add a WKViewData ivar to track whether we've already registered for popover notifications.
(-[WKView _prepareForDictionaryLookup]):
Added; moved notification registration from -[WKView initWithFrame:processPool:configuration:webView:.
(-[WKView initWithFrame:processPool:configuration:webView:]):
Move notification registration to -_prepareForDictionaryLookup.

* UIProcess/API/mac/WKViewInternal.h:
Expose -_prepareForDictionaryLookup.

* UIProcess/mac/PageClientImpl.mm:
(WebKit::PageClientImpl::didPerformDictionaryLookup):
Send -[WKView _prepareForDictionaryLookup].

* UIProcess/mac/WKImmediateActionController.mm:
(-[WKImmediateActionController _animationControllerForText]):
Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183978 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r183961.
commit-queue@webkit.org [Fri, 8 May 2015 03:27:11 +0000 (03:27 +0000)]
Unreviewed, rolling out r183961.
https://bugs.webkit.org/show_bug.cgi?id=144784

Broke js/dom/JSON-stringify.html (Requested by kling on
#webkit).

Reverted changeset:

"Optimize serialization of quoted JSON strings."
https://bugs.webkit.org/show_bug.cgi?id=144754
http://trac.webkit.org/changeset/183961

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183977 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fixes.
andersca@apple.com [Fri, 8 May 2015 02:59:37 +0000 (02:59 +0000)]
Build fixes.

Source/WebCore:

* accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(-[WebAccessibilityObjectWrapper doAXRTFForRange:]):
* editing/cocoa/HTMLConverter.mm:
(HTMLConverter::_addMarkersToList):
* platform/mac/PasteboardMac.mm:
(WebCore::writeFileWrapperAsRTFDAttachment):
* platform/mac/PlatformPasteboardMac.mm:
(WebCore::PlatformPasteboard::setTypes):
Pass empty arrays and dictionaries instead of nil.

* rendering/RenderThemeMac.mm:
(WebCore::RenderThemeMac::paintSliderThumb):
Pass the document view to the AppKit paint method.

Source/WebKit/mac:

* Misc/WebNSPasteboardExtras.mm:
(-[NSPasteboard _web_writeFileWrapperAsRTFDAttachment:]):
* WebView/WebHTMLView.mm:
(-[WebHTMLView _writeSelectionWithPasteboardTypes:toPasteboard:cachedAttributedString:]):
(-[WebHTMLView _selectionStartFontAttributesAsRTF]):
* WebView/WebPDFView.mm:
(-[WebPDFView writeSelectionWithPasteboardTypes:toPasteboard:]):
Pass empty arrays and dictionaries instead of nil.

* WebView/WebView.mm:
(-[WebView _commonInitializationWithFrameName:groupName:]):
Use the regular init method.

Source/WebKit2:

* UIProcess/API/mac/WKView.mm:
(-[WKView initWithFrame:processPool:configuration:webView:]):
Use the regular init method.

* UIProcess/mac/WKSharingServicePickerDelegate.mm:
(-[WKSharingServicePickerDelegate sharingService:didShareItems:]):
* WebProcess/WebPage/mac/WebPageMac.mm:
(WebKit::WebPage::handleSelectionServiceClick):
Pass empty dictionaries instead of nil.

Tools:

* TestWebKitAPI/Tests/mac/HTMLCollectionNamedItem.mm:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/mac/HTMLFormCollectionNamedItem.mm:
(TestWebKitAPI::TEST):
Add casts.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183976 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoPurge PassRefPtr in WebCore/html - 3
commit-queue@webkit.org [Fri, 8 May 2015 02:18:32 +0000 (02:18 +0000)]
Purge PassRefPtr in WebCore/html - 3
https://bugs.webkit.org/show_bug.cgi?id=144686

Patch by Gyuyoung Kim <gyuyoung.kim@samsung.com> on 2015-05-07
Reviewed by Andreas Kling.

Replace PassRefPtr with Ref in create() factory functions, because
it doesn't have to change to be null.

No new tests, no behavior changes.

* html/FTPDirectoryDocument.h:
* html/HTMLDetailsElement.cpp:
(WebCore::DetailsContentElement::create):
* html/HTMLKeygenElement.cpp:
* html/HTMLSummaryElement.cpp:
(WebCore::SummaryContentElement::create):
* html/TimeRanges.cpp:
(WebCore::TimeRanges::create):
* html/shadow/SliderThumbElement.cpp:
(WebCore::SliderContainerElement::create):
* html/shadow/SliderThumbElement.h:
(WebCore::SliderThumbElement::create):
* html/shadow/SpinButtonElement.cpp:
(WebCore::SpinButtonElement::create):
* html/shadow/SpinButtonElement.h:
* html/shadow/TextControlInnerElements.cpp:
(WebCore::TextControlInnerContainer::create):
(WebCore::TextControlInnerElement::create):
(WebCore::TextControlInnerTextElement::create):
(WebCore::SearchFieldResultsButtonElement::create):
(WebCore::SearchFieldCancelButtonElement::create):
* html/shadow/TextControlInnerElements.h:
* html/shadow/YouTubeEmbedShadowElement.cpp:
(WebCore::YouTubeEmbedShadowElement::create):
* html/track/InbandDataTextTrack.cpp:
(WebCore::InbandDataTextTrack::create):
* html/track/InbandDataTextTrack.h:
* html/track/InbandGenericTextTrack.cpp:
(WebCore::InbandGenericTextTrack::create):
* html/track/InbandGenericTextTrack.h:
* html/track/InbandWebVTTTextTrack.cpp:
(WebCore::InbandWebVTTTextTrack::create):
* html/track/InbandWebVTTTextTrack.h:
* html/track/LoadableTextTrack.h:
* html/track/TextTrack.h:
(WebCore::TextTrack::create):
* html/track/TextTrackCueGeneric.cpp:
* html/track/TextTrackCueGeneric.h:
* html/track/TextTrackList.h:
* html/track/VTTRegionList.h:
(WebCore::VTTRegionList::create):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183975 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoGC has trouble with pathologically large array allocations
fpizlo@apple.com [Fri, 8 May 2015 02:12:35 +0000 (02:12 +0000)]
GC has trouble with pathologically large array allocations
https://bugs.webkit.org/show_bug.cgi?id=144609

Reviewed by Geoffrey Garen.
Source/JavaScriptCore:

The bug was that SlotVisitor::copyLater() would return early for oversize blocks (right
after pinning them), and would skip the accounting. The GC calculates the size of the heap
in tandem with the scan to save time, and that accounting was part of how the GC would
know how big the heap was. The GC would then think that oversize copied blocks use no
memory, and would then mess up its scheduling of the next GC.

Fixing this bug is harder than it seems. When running an eden GC, we figure out the heap
size by summing the size from the last collection and the size by walking the eden heap.
But this breaks when we eagerly delete objects that the last collection touched. We can do
that in one corner case: copied block reallocation. The old block will be deleted from old
space during the realloc and a new block will be allocated in new space. In order for the
GC to know that the size of old space actually shrank, we need a field to tell us how much
such shrinkage could occur. Since this is a very dirty corner case and it only works for
very particular reasons arising from the special properties of copied space (single owner,
and the realloc is used in places where the compiler already knows that it cannot register
allocate a pointer to the old block), I opted for an equally dirty shrinkage counter
devoted just to this case. It's called bytesRemovedFromOldSpaceDueToReallocation.

To test this, I needed to add an Option to force a particular RAM size in the GC. This
allows us to write tests that assert that the GC heap size is some value X, without
worrying about machine-to-machine variations due to GC heuristics changing based on RAM
size.

* heap/CopiedSpace.cpp:
(JSC::CopiedSpace::CopiedSpace): Initialize the dirty shrinkage counter.
(JSC::CopiedSpace::tryReallocateOversize): Bump the dirty shrinkage counter.
* heap/CopiedSpace.h:
(JSC::CopiedSpace::takeBytesRemovedFromOldSpaceDueToReallocation): Swap out the counter. Used by the GC when it does its accounting.
* heap/Heap.cpp:
(JSC::Heap::Heap): Allow the user to force the RAM size.
(JSC::Heap::updateObjectCounts): Use the dirty shrinkage counter to good effect. Also, make this code less confusing.
* heap/SlotVisitorInlines.h:
(JSC::SlotVisitor::copyLater): The early return for isOversize() was the bug. We still need to report these bytes as live. Otherwise the GC doesn't know that it owns this memory.
* jsc.cpp: Add size measuring hooks to write the largeish test.
(GlobalObject::finishCreation):
(functionGCAndSweep):
(functionFullGC):
(functionEdenGC):
(functionHeapSize):
* runtime/Options.h:
* tests/stress/new-array-storage-array-with-size.js: Fix this so that it actually allocates ArrayStorage arrays and tests the thing it was supposed to test.
* tests/stress/new-largeish-contiguous-array-with-size.js: Added. This tests what the other test accidentally started testing, but does so without running your system out of memory.
(foo):
(test):

Tools:

Add a --filter option that restricts the set of tests we run. I needed it to fix this bug
and it's a frequently requested feature.

Also add the ability to run a test pretending that your system has a particular RAM size.
This is useful for GC tests, and the new GC test that I added uses this.

* Scripts/run-javascriptcore-tests:
(runJSCStressTests):
* Scripts/run-jsc-stress-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183974 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[EFL] Bump EFL version to 1.14.0
ossy@webkit.org [Fri, 8 May 2015 02:05:15 +0000 (02:05 +0000)]
[EFL] Bump EFL version to 1.14.0
https://bugs.webkit.org/show_bug.cgi?id=144743

Reviewed by Darin Adler.

Bump EFL and elementary version to 1.14.0 and removed
md5sum, because it isn't used at all if hash is present.

* efl/jhbuild.modules:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183973 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoGlobal functions should be initialized as JSFunctions in byte code
saambarati1@gmail.com [Fri, 8 May 2015 01:59:14 +0000 (01:59 +0000)]
Global functions should be initialized as JSFunctions in byte code
https://bugs.webkit.org/show_bug.cgi?id=144178

Reviewed by Geoffrey Garen.

This patch makes the initialization of global functions more explicit by
moving initialization into bytecode. It also prepares JSC for having ES6
style lexical scoping because initializing global functions in bytecode
easily allows global functions to be initialized with the proper scope that
will have access to global lexical variables. Global lexical variables
should be visible to global functions but don't live on the global object.

* bytecode/UnlinkedCodeBlock.cpp:
(JSC::UnlinkedProgramCodeBlock::visitChildren):
* bytecode/UnlinkedCodeBlock.h:
* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::generate):
(JSC::BytecodeGenerator::BytecodeGenerator):
* bytecompiler/BytecodeGenerator.h:
* runtime/Executable.cpp:
(JSC::ProgramExecutable::initializeGlobalProperties):
* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::addGlobalVar):
(JSC::JSGlobalObject::addFunction):
* runtime/JSGlobalObject.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183972 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] MediaControls: disappear while scrubbing
dino@apple.com [Fri, 8 May 2015 01:55:08 +0000 (01:55 +0000)]
[iOS] MediaControls: disappear while scrubbing
https://bugs.webkit.org/show_bug.cgi?id=144777
<rdar://problem/20065572>

Reviewed by Eric Carlson.

If we are scrubbing we shouldn't hide the controls.

* Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.hideControls): Return early if we are scrubbing.
* Modules/mediacontrols/mediaControlsiOS.js:
(ControllerIOS.prototype): Add initial value for _potentiallyScrubbing and
rename from non-underscored value throughout the file.
(ControllerIOS.prototype.handleTimelineTouchEnd): When we finish scrubbing, reset
the timer to hide the controls.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183971 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION (r183300): Fixed elements flash when scrolling
simon.fraser@apple.com [Fri, 8 May 2015 01:46:59 +0000 (01:46 +0000)]
REGRESSION (r183300): Fixed elements flash when scrolling
https://bugs.webkit.org/show_bug.cgi?id=144778
rdar://problem/20769741

Reviewed by Dean Jackson.

After r183300 we can detached layer backing store when outside the coverage region.
However, position:fixed layers are moved around by the ScrollingCoordinator behind
GraphicsLayer's back, so we can do layer flushes with stale information about layer
geometry.

To avoid dropping backing store for layers in this situation, prevent backing
store detachment on layers registered with the ScrollingCoordinator as viewport-constrained
layers. Preventing detachment on a layer also prevents detachment on all descendant
layers.

* platform/graphics/GraphicsLayer.h:
(WebCore::GraphicsLayer::setAllowsBackingStoreDetachment):
(WebCore::GraphicsLayer::allowsBackingStoreDetachment):
* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::GraphicsLayerCA):
(WebCore::GraphicsLayerCA::setVisibleAndCoverageRects): Set m_intersectsCoverageRect to true
if backing store detachment is prevented.
(WebCore::GraphicsLayerCA::recursiveCommitChanges): Set a bit in the CommitState to
communicate to descendants that detachment is prevented.
* platform/graphics/ca/GraphicsLayerCA.h:
(WebCore::GraphicsLayerCA::CommitState::CommitState): Deleted.
* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::setIsScrollCoordinatedWithViewportConstrainedRole):
* rendering/RenderLayerBacking.h:
(WebCore::RenderLayerBacking::setScrollingNodeIDForRole): If registering with a non-zero
nodeID for the ViewportConstrained role, turn off backing store detachment.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183970 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoASSERTION when pasting text into the WebInspector console
cdumez@apple.com [Fri, 8 May 2015 01:23:49 +0000 (01:23 +0000)]
ASSERTION when pasting text into the WebInspector console
https://bugs.webkit.org/show_bug.cgi?id=144774

Reviewed by Ryosuke Niwa.

Fix assertion in didChangeSelection() meant to check that calling
editorState() does not cause a synchronous layout. The assertion
was not correct as it was relying on FrameView::needsLayout() and
we would hit it if calling editorState() would schedule a layout.
Instead, the new assertion relies on FrameView::layoutCount(),
which is more accurate.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183969 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix iPhone 6 iOS simulator Build.
said@apple.com [Fri, 8 May 2015 01:08:54 +0000 (01:08 +0000)]
Fix iPhone 6 iOS simulator Build.
following http://trac.webkit.org/changeset/183954.

Reviewed by Daniel Bates.

* WebView/WebHTMLView.mm:
(-[WebHTMLView pressureChangeWithEvent:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183968 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoConsider implementing Document.scrollingElement
weinig@apple.com [Fri, 8 May 2015 00:59:01 +0000 (00:59 +0000)]
Consider implementing Document.scrollingElement
<rdar://problem/20845213>
https://bugs.webkit.org/show_bug.cgi?id=143609

Reviewed by Simon Fraser.

Source/WebCore:

Add the non-compliant implementation of Document.scrollingElement that the spec,
found at http://dev.w3.org/csswg/cssom-view/#dom-document-scrollingelement,
says we should have while we have non-compliant implementations of Element.scrollTop
and Element.scrollLeft (fixing that is tracked by https://bugs.webkit.org/show_bug.cgi?id=106133).

Tests: fast/dom/Document/scrollingElement-quirks-mode.html
       fast/dom/Document/scrollingElement-standards-mode.html

* dom/Document.cpp:
(WebCore::Document::scrollingElement):
* dom/Document.h:
* dom/Document.idl:

LayoutTests:

* fast/dom/Document/scrollingElement-quirks-mode-expected.txt: Added.
* fast/dom/Document/scrollingElement-quirks-mode.html: Added.
* fast/dom/Document/scrollingElement-standards-mode-expected.txt: Added.
* fast/dom/Document/scrollingElement-standards-mode.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183967 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix the x86 32bits build
benjamin@webkit.org [Fri, 8 May 2015 00:54:54 +0000 (00:54 +0000)]
Fix the x86 32bits build

Patch by Benjamin Poulain <bpoulain@apple.com> on 2015-05-07

* assembler/X86Assembler.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183966 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] While scrubbing and holding down, video continues to play
dino@apple.com [Fri, 8 May 2015 00:45:00 +0000 (00:45 +0000)]
[iOS] While scrubbing and holding down, video continues to play
https://bugs.webkit.org/show_bug.cgi?id=144776
<rdar://problem/20863757>

Reviewed by Simon Fraser.

When we are scrubbing a video, we should pause playback. As we
let go of the scrubber playback can resume (but only if it was
playing originally).

* Modules/mediacontrols/mediaControlsiOS.js:
(ControllerIOS.prototype.createControls): Listen for touchstart on the scrubber.
(ControllerIOS.prototype.handleTimelineInput): Call the prototype, but pause if necessary.
(ControllerIOS.prototype.handleTimelineChange): Just moved this to be with the other timeline functions.
(ControllerIOS.prototype.handleTimelineTouchStart): Add the listeners for end and cancel. Remember that we are
potentially about to scrub.
(ControllerIOS.prototype.handleTimelineTouchEnd): Remove the listeners.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183965 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[JSC] Add basic DFG/FTL support for Math.round
benjamin@webkit.org [Fri, 8 May 2015 00:23:32 +0000 (00:23 +0000)]
[JSC] Add basic DFG/FTL support for Math.round
https://bugs.webkit.org/show_bug.cgi?id=144725

Patch by Benjamin Poulain <bpoulain@apple.com> on 2015-05-07
Reviewed by Filip Pizlo.

This patch adds two optimizations targeting Math.round():
-Add a DFGNode ArithRound corresponding to the intrinsic RoundIntrinsic.
-Change the MacroAssembler to be stricter on how we fail to convert a double
 to ingeter. Previously, any number valued zero would fail, now we only
 fail for -0.

Since ArithRound speculate it produces int32, the MacroAssembler assembler
part became necessary because zero is a pretty common output of Math.round()
and we would OSR exit a lot (and eventually recompile for doubles).

The implementation itself of the inline Math.round() is exactly the same
as the C function that exists for Math.round(). We can very likely do better
but it is a good start known to be valid and inlining alone alread provides
significant speedups.

* assembler/X86Assembler.h:
(JSC::X86Assembler::movmskpd_rr):
* assembler/MacroAssemblerX86Common.h:
(JSC::MacroAssemblerX86Common::branchConvertDoubleToInt32):
When we have a zero, get the sign bit out of the double and check if is one.

I'll look into doing the same improvement for ARM.

* bytecode/SpeculatedType.cpp:
(JSC::typeOfDoubleRounding):
(JSC::typeOfDoubleFRound): Deleted.
* bytecode/SpeculatedType.h:
* dfg/DFGAbstractInterpreterInlines.h:
(JSC::DFG::AbstractInterpreter<AbstractStateType>::executeEffects):
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::handleIntrinsic):
* dfg/DFGClobberize.h:
(JSC::DFG::clobberize):
* dfg/DFGDoesGC.cpp:
(JSC::DFG::doesGC):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
* dfg/DFGGraph.h:
(JSC::DFG::Graph::roundShouldSpeculateInt32):
(JSC::DFG::Graph::negateShouldSpeculateMachineInt): Deleted.
* dfg/DFGNode.h:
(JSC::DFG::Node::arithNodeFlags):
(JSC::DFG::Node::hasHeapPrediction):
(JSC::DFG::Node::hasArithMode):
* dfg/DFGNodeType.h:
* dfg/DFGPredictionPropagationPhase.cpp:
(JSC::DFG::PredictionPropagationPhase::propagate):
* dfg/DFGSafeToExecute.h:
(JSC::DFG::safeToExecute):
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileArithRound):
* dfg/DFGSpeculativeJIT.h:
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* ftl/FTLCapabilities.cpp:
(JSC::FTL::canCompile):
* ftl/FTLIntrinsicRepository.h:
* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::compileNode):
(JSC::FTL::LowerDFGToLLVM::convertDoubleToInt32):
(JSC::FTL::LowerDFGToLLVM::compileDoubleAsInt32):
(JSC::FTL::LowerDFGToLLVM::compileArithRound):
* ftl/FTLOutput.h:
(JSC::FTL::Output::ceil64):
* jit/ThunkGenerators.cpp:
* runtime/MathCommon.cpp:
* runtime/MathCommon.h:
* runtime/MathObject.cpp:
(JSC::mathProtoFuncRound):
* tests/stress/math-round-basics.js: Added.
(mathRoundOnIntegers):
(mathRoundOnDoubles):
(mathRoundOnBooleans):
(uselessMathRound):
(mathRoundWithOverflow):
(mathRoundConsumedAsDouble):
(mathRoundDoesNotCareAboutMinusZero):
(mathRoundNoArguments):
(mathRoundTooManyArguments):
(testMathRoundOnConstants):
(mathRoundStructTransition):
(Math.round):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183963 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoexceptionFuzz tests should explicitly initialize the exceptionFuzz boolean in JavaScr...
saambarati1@gmail.com [Fri, 8 May 2015 00:20:58 +0000 (00:20 +0000)]
exceptionFuzz tests should explicitly initialize the exceptionFuzz boolean in JavaScript code through a function in jsc.cpp
https://bugs.webkit.org/show_bug.cgi?id=144753

Reviewed by Mark Lam.

Source/JavaScriptCore:

This allows the BytecodeGenerator to freely emit startup code that "may"
throw exceptions without worrying that this startup code will trigger
the exceptionFuzz exception. The exceptionFuzz counter will only begin
ticking when the 'enableExceptionFuzz' function is explicitly called in
the exceptionFuzz tests.

* jsc.cpp:
(GlobalObject::finishCreation):
(functionEnableExceptionFuzz):
* tests/exceptionFuzz/3d-cube.js:
* tests/exceptionFuzz/date-format-xparb.js:
* tests/exceptionFuzz/earley-boyer.js:

Tools:

* Scripts/jsc-stress-test-helpers/js-exception-fuzz:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183962 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoOptimize serialization of quoted JSON strings.
akling@apple.com [Fri, 8 May 2015 00:17:48 +0000 (00:17 +0000)]
Optimize serialization of quoted JSON strings.
<https://webkit.org/b/144754>

Reviewed by Darin Adler.

Source/JavaScriptCore:

Optimized the serialization of quoted strings into JSON by moving the logic into
StringBuilder so it can make smarter decisions about buffering.

12% progression on Kraken/json-stringify-tinderbox (on my Mac Pro.)

* bytecompiler/NodesCodegen.cpp:
(JSC::ObjectPatternNode::toString): Use the new StringBuilder API.

* runtime/JSONObject.h:
* runtime/JSONObject.cpp:
(JSC::Stringifier::Holder::appendNextProperty):
(JSC::appendStringToStringBuilder): Deleted.
(JSC::appendQuotedJSONStringToBuilder): Deleted.
(JSC::Stringifier::appendQuotedString): Deleted.
(JSC::Stringifier::appendStringifiedValue): Moved the bulk of this logic
to StringBuilder and call that from here.

Source/WebKit2:

* NetworkProcess/cache/NetworkCacheEntry.cpp:
(WebKit::NetworkCache::Entry::asJSON): Use the new StringBuilder API.

Source/WTF:

Add a StringBuilder API for appending a quoted JSON string. This is used by
JSON.stringify() to implement efficient appending of strings while escaping
quotes, control characters and \uNNNN-style characters.

The main benefit comes from only doing a single buffer expansion up front,
instead of doing it every time we append something. The fudge factor is pretty
large for 16-bit strings, since the maximum number of output characters per
input character is 6.

* wtf/text/StringBuilder.cpp:
(WTF::appendQuotedJSONStringInternal):
(WTF::StringBuilder::appendQuotedJSONString):
* wtf/text/StringBuilder.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183961 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoApplying a filter on an SVG element, which is larger than 4096 pixels, causes this...
said@apple.com [Thu, 7 May 2015 23:32:45 +0000 (23:32 +0000)]
Applying a filter on an SVG element, which is larger than 4096 pixels, causes this element to be rendered shifted to the left
https://bugs.webkit.org/show_bug.cgi?id=144335

Reviewed by Daniel Bates.

Address comments raised by Darin Adler in review.

* platform/graphics/ImageBuffer.cpp:
(WebCore::ImageBuffer::sizeNeedsClamping):
(WebCore::ImageBuffer::clampedSize):
(WebCore::ImageBuffer::isSizeClamped): Deleted.
* platform/graphics/ImageBuffer.h:
The condition for the negative width/height in isSizeClamped() was wrong. Use
FloatSize::isEmpty() instead and rename the function to sizeNeedsClamping().
The new function should return the opposite of the old function return value.

* platform/graphics/filters/FilterEffect.cpp:
(WebCore::FilterEffect::apply):
(WebCore::FilterEffect::asUnmultipliedImage):
(WebCore::FilterEffect::asPremultipliedImage):
(WebCore::FilterEffect::copyUnmultipliedImage):
(WebCore::FilterEffect::copyPremultipliedImage):
(WebCore::FilterEffect::createUnmultipliedImageResult):
(WebCore::FilterEffect::createPremultipliedImageResult):
* rendering/FilterEffectRenderer.cpp:
(WebCore::FilterEffectRenderer::updateBackingStoreRect):
(WebCore::FilterEffectRendererHelper::beginFilterEffect):
* rendering/svg/RenderSVGResourceFilter.cpp:
(WebCore::RenderSVGResourceFilter::applyResource):
Call the new function and negate the condition for the return value.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183960 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRelease assert in com.apple.WebKit.WebContent under JavaScriptCore: JSC::JSONProtoFun...
ggaren@apple.com [Thu, 7 May 2015 23:29:15 +0000 (23:29 +0000)]
Release assert in com.apple.WebKit.WebContent under JavaScriptCore: JSC::JSONProtoFuncStringify
https://bugs.webkit.org/show_bug.cgi?id=144758

Reviewed by Andreas Kling.

This was an out-of-memory error when trying to shrink a string builder.
bmalloc was missing the optimization that allowed realloc() to shrink
without copying. So, let's add it.

* bmalloc/Allocator.cpp:
(bmalloc::Allocator::reallocate): Added Large and XLarge cases for
shrinking without copying. This isn't possible for small and medium
objects, and probably not very profitable, either.

* bmalloc/Heap.cpp:
(bmalloc::Heap::findXLarge):
(bmalloc::Heap::deallocateXLarge):
* bmalloc/Heap.h: Refactored this code to return a reference to an
XLarge range. This makes the code reusable, and also makes it easier
for realloc() to update metadata.

* bmalloc/LargeObject.h:
(bmalloc::LargeObject::split): Allow allocated objects to split because
that's what realloc() wants to do, and there's nothing intrinsically
wrong with it.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183959 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix.
andersca@apple.com [Thu, 7 May 2015 23:08:53 +0000 (23:08 +0000)]
Build fix.

* editing/mac/EditorMac.mm:
(WebCore::dataInRTFDFormat):
(WebCore::dataInRTFFormat):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183958 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoOccasional null deref in WebImmediateActionController
timothy_horton@apple.com [Thu, 7 May 2015 23:08:01 +0000 (23:08 +0000)]
Occasional null deref in WebImmediateActionController
https://bugs.webkit.org/show_bug.cgi?id=144772
<rdar://problem/20811128>

Reviewed by Beth Dakin.

* WebView/WebImmediateActionController.mm:
(-[WebImmediateActionController _defaultAnimationController]):
(-[WebImmediateActionController _animationControllerForDataDetectedText]):
(-[WebImmediateActionController _animationControllerForDataDetectedLink]):
Null-check TextIndicators before dereferencing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183957 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoApplying a filter on an SVG element, which is larger than 4096 pixels, causes this...
commit-queue@webkit.org [Thu, 7 May 2015 22:46:43 +0000 (22:46 +0000)]
Applying a filter on an SVG element, which is larger than 4096 pixels, causes this element to be rendered shifted to the left
https://bugs.webkit.org/show_bug.cgi?id=144335

Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2015-05-07
Reviewed by Darin Adler.

Source/WebCore:

When the SVG element is larger than 4096x4096 pixels and it has a filter,
a clipper or a masker, the ImageBuffer which is created for drawing the
sourceGraphicBuffer has to be clamped to this size. The reason for this
clamping is the FilterEffect simply ignores processing any ImageBuffer
larger than this size.

The bug was happening because we did not propagate the clamping logic to
the FilterEffect. The FilterEffect result ImageBuffer was not clamped as
what we do for drawing the sourceGraphicBuffer. If only the destination
point is specified, the GraphicsContext::drawImageBuffer() assumes the
source and the destination images have the same size which is not true
with the clamping.

The fix is to add the clamping logic to the FilterEffect. Two places we
need to change. (1) FilterEffect::createImageBufferResult() has to apply
the same clamping we do in SVGRenderingContext::createImageBuffer(). (2)
FilterEffect::drawingRegionOfInputImage() has to consider the clamping
when mapping from absolute coordinates to the ImageBuffer coordinates.

Tests: svg/filters/big-sized-off-viewport-filter.svg

* platform/graphics/ImageBuffer.cpp:
(WebCore::ImageBuffer::isSizeClamped):
(WebCore::ImageBuffer::clampedSize):
(WebCore::ImageBuffer::clampedRect):
* platform/graphics/ImageBuffer.h:
(WebCore::ImageBuffer::setSpaceSize): Move all the clamping helper methods
from SVGRenderingContext and RenderSVGResourceFilter to the ImageBuffer.

* platform/graphics/filters/FEColorMatrix.cpp:
(WebCore::FEColorMatrix::platformApplySoftware): Use logicalSize() of the
result image instead of using absolutePaintRect(). The later returns the
rectangle of the element without clamping.

* platform/graphics/filters/FETile.cpp:
(WebCore::FETile::platformApplySoftware):
* rendering/svg/RenderSVGResourceClipper.cpp:
(WebCore::RenderSVGResourceClipper::applyClippingToContext):
* rendering/svg/RenderSVGResourceFilter.cpp:
(WebCore::RenderSVGResourceFilter::applyResource):
(WebCore::RenderSVGResourceFilter::fitsInMaximumImageSize): Deleted.
* rendering/svg/RenderSVGResourceFilter.h:
* rendering/svg/RenderSVGResourceGradient.cpp:
(WebCore::createMaskAndSwapContextForTextGradient):
(WebCore::clipToTextMask):
* rendering/svg/RenderSVGResourceMasker.cpp:
(WebCore::RenderSVGResourceMasker::applySVGMask):
* rendering/svg/RenderSVGResourcePattern.cpp:
(WebCore::RenderSVGResourcePattern::buildPattern):
(WebCore::RenderSVGResourcePattern::createTileImage): Make the modifications
which are needed because of moving the clamping helper methods to ImageBuffer
and because of changing the prototype of SVGRenderingContext methods.

* platform/graphics/filters/FilterEffect.cpp:
(WebCore::FilterEffect::drawingRegionOfInputImage): Consider the clamping
when mapping from absolute coordinates to the ImageBuffer coordinates.

(WebCore::FilterEffect::createImageBufferResult): Apply the same clamping
we do in SVGRenderingContext::createImageBuffer() when creating the
FilterEffect result ImageBuffer.

(WebCore::FilterEffect::apply):
(WebCore::FilterEffect::asUnmultipliedImage):
(WebCore::FilterEffect::asPremultipliedImage):
(WebCore::FilterEffect::copyUnmultipliedImage):
(WebCore::FilterEffect::copyPremultipliedImage):
(WebCore::FilterEffect::createUnmultipliedImageResult):
(WebCore::FilterEffect::createPremultipliedImageResult):
(WebCore::FilterEffect::maxFilterArea): Deleted.
(WebCore::FilterEffect::isFilterSizeValid): Deleted.
* platform/graphics/filters/FilterEffect.h: Use the new ImageBuffer clamping
helper methods and delete the local ones.

* platform/graphics/transforms/AffineTransform.cpp:
(WebCore::AffineTransform::scaleNonUniform):
(WebCore::AffineTransform::scale):
(WebCore::AffineTransform::translate):
* platform/graphics/transforms/AffineTransform.h: Add new scale and translate
overloads to AffineTransform.

* rendering/FilterEffectRenderer.cpp:
(WebCore::FilterEffectRenderer::updateBackingStoreRect):
(WebCore::FilterEffectRendererHelper::beginFilterEffect): Code clean up.

* rendering/svg/SVGRenderingContext.cpp:
(WebCore::SVGRenderingContext::calculateScreenFontSizeScalingFactor):
(WebCore::SVGRenderingContext::calculateTransformationToOutermostCoordinateSystem):
Return the AffineTransform instead of passing it through the reference of
an argument.

(WebCore::SVGRenderingContext::createImageBuffer):
(WebCore::SVGRenderingContext::createImageBufferForPattern): Deleted.
Code clean up and method rename.

(WebCore::SVGRenderingContext::clampedAbsoluteTargetRect): Deleted.
(WebCore::SVGRenderingContext::clampedAbsoluteSize): Deleted.
* rendering/svg/SVGRenderingContext.h: Move the clamping helper methods to
the ImageBuffer class.

LayoutTests:

* svg/filters/big-sized-off-viewport-filter-expected.svg: Added.
* svg/filters/big-sized-off-viewport-filter.svg: Added.
Ensure when applying a filter on a large SVG element, the viewport of the
SVG includes the correct drawing part of this element.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183956 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFunctionCallBracketNode should store the base value to the temporary when subscript...
utatane.tea@gmail.com [Thu, 7 May 2015 22:35:56 +0000 (22:35 +0000)]
FunctionCallBracketNode should store the base value to the temporary when subscript has assignment
https://bugs.webkit.org/show_bug.cgi?id=144678

Reviewed by Geoffrey Garen.

Currently, FunctionCallBracketNode directly use the RegisterID returned by emitNode.
But if the base part is the local register and the subscript part has assignment to it, the base result is accidentally rewritten.

function t() { var ok = {null: function () { } }; ok[ok = null](); }
t();  // Should not throw error.

This patch takes care about `subscriptHasAssignment`.
By using `emitNodeForLeftHandSide`, when there's assignment to local variables in RHS,
it correctly moves the LHS value to a temporary register.

* bytecompiler/NodesCodegen.cpp:
(JSC::FunctionCallBracketNode::emitBytecode):
* parser/ASTBuilder.h:
(JSC::ASTBuilder::makeFunctionCallNode):
* parser/NodeConstructors.h:
(JSC::FunctionCallBracketNode::FunctionCallBracketNode):
* parser/Nodes.h:
* tests/stress/assignment-in-function-call-bracket-node.js: Added.
(shouldBe):
(shouldBe.):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183955 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoNew force-related DOM events should fire in WK1 views
bdakin@apple.com [Thu, 7 May 2015 22:32:07 +0000 (22:32 +0000)]
New force-related DOM events should fire in WK1 views
https://bugs.webkit.org/show_bug.cgi?id=144663
-and corresponding-
rdar://problem/20281886

Reviewed by Sam Weinig.

Source/WebCore:

All of the WK1 mouse events need to take the correspondingPressureEvent.
* page/EventHandler.h:

Make correspondingPressureEvent a part of CurrentEventScope. This is needed to
have accurate pressure information for all of the mouse events in subframes.
* page/mac/EventHandlerMac.mm:
(WebCore::correspondingPressureEventSlot):
(WebCore::EventHandler::correspondingPressureEvent):
(WebCore::CurrentEventScope::CurrentEventScope):
(WebCore::CurrentEventScope::~CurrentEventScope):

These events don’t have an associated pressure, so send nil for the
correspondingPressureEvent.
(WebCore::EventHandler::wheelEvent):
(WebCore::EventHandler::keyEvent):

Pipe through correspondingPressureEvent.
(WebCore::EventHandler::mouseDown):
(WebCore::EventHandler::mouseDragged):
(WebCore::EventHandler::mouseUp):
(WebCore::EventHandler::mouseMoved):

New function to handle pressure change events.
(WebCore::EventHandler::pressureChange):

Pipe through correspondingPressureEvent.
(WebCore::EventHandler::passMouseMovedEventToScrollbars):
(WebCore::EventHandler::currentPlatformMouseEvent):

Take the correspondingPressureEvent in order to build a PlatformMouseEvent with
the correct pressure information.
* platform/mac/PlatformEventFactoryMac.h:
* platform/mac/PlatformEventFactoryMac.mm:
(WebCore::globalPointForEvent):
(WebCore::pointForEvent):
(WebCore::mouseButtonForEvent):
(WebCore::PlatformMouseEventBuilder::PlatformMouseEventBuilder):
(WebCore::PlatformEventFactory::createPlatformMouseEvent):

Source/WebKit:

Pass the lastPressureEvent to WebCore.
* WebView/WebHTMLView.mm:
(-[WebHTMLView _updateMouseoverWithEvent:]):
(-[WebHTMLView rightMouseUp:]):
(-[WebHTMLView menuForEvent:]):
(-[WebHTMLView acceptsFirstMouse:]):
(-[WebHTMLView shouldDelayWindowOrderingForEvent:]):
(-[WebHTMLView mouseDown:mouseDown:]):
(-[WebHTMLView mouseDragged:]):
(-[WebHTMLView mouseUp:mouseUp:]):

New NSRespnder method for pressure changes.
(-[WebHTMLView pressureChangeWithEvent:]):

New BOOL _contentPreventsDefault tracks whether the HitTestResult prevented the
default action. Get rid of willHandleMouseDown; now that the gesture recognizer
sets delaysPrimaryMouseButtonEvents to NO, we don’t need this.
* WebView/WebImmediateActionController.h:
* WebView/WebImmediateActionController.mm:
(-[WebImmediateActionController _clearImmediateActionState]):

Set all of the immediateActionStages on EventHandler. This is critical to keep
link navigation happening at the right time now that
delaysPrimaryMouseButtonEvents is set to NO.
(-[WebImmediateActionController performHitTestAtPoint:]):
(-[WebImmediateActionController immediateActionRecognizerDidUpdateAnimation:]):
(-[WebImmediateActionController immediateActionRecognizerDidCancelAnimation:]):
(-[WebImmediateActionController immediateActionRecognizerDidCompleteAnimation:]):

Use a dummy animation controller if the content prevents default.
(-[WebImmediateActionController _defaultAnimationController]):
(-[WebImmediateActionController _updateImmediateActionItem]):
(-[WebImmediateActionController webView:willHandleMouseDown:]): Deleted.

Set delaysPrimaryMouseButtonEvents to NO so that we get existing mouse events when
we expect to.
* WebView/WebView.mm:
(-[WebView _commonInitializationWithFrameName:groupName:]):

Cache the most recent pressure event so that we can send it to WebCore for all of
the mouse events.
(-[WebView _pressureEvent]):
(-[WebView _setPressureEvent:]):
* WebView/WebViewData.h:
* WebView/WebViewData.mm:
(-[WebViewPrivate dealloc]):
* WebView/WebViewInternal.h:

Source/WebKit/mac:

Pass the lastPressureEvent to WebCore.
* WebView/WebHTMLView.mm:
(-[WebHTMLView _updateMouseoverWithEvent:]):
(-[WebHTMLView rightMouseUp:]):
(-[WebHTMLView menuForEvent:]):
(-[WebHTMLView acceptsFirstMouse:]):
(-[WebHTMLView shouldDelayWindowOrderingForEvent:]):
(-[WebHTMLView mouseDown:mouseDown:]):
(-[WebHTMLView mouseDragged:]):
(-[WebHTMLView mouseUp:mouseUp:]):

New NSRespnder method for pressure changes.
(-[WebHTMLView pressureChangeWithEvent:]):

New BOOL _contentPreventsDefault tracks whether the HitTestResult prevented the
default action. Get rid of willHandleMouseDown; now that the gesture recognizer
sets delaysPrimaryMouseButtonEvents to NO, we don’t need this.
* WebView/WebImmediateActionController.h:
* WebView/WebImmediateActionController.mm:
(-[WebImmediateActionController _clearImmediateActionState]):

Set all of the immediateActionStages on EventHandler. This is critical to keep
link navigation happening at the right time now that
delaysPrimaryMouseButtonEvents is set to NO.
(-[WebImmediateActionController performHitTestAtPoint:]):
(-[WebImmediateActionController immediateActionRecognizerDidUpdateAnimation:]):
(-[WebImmediateActionController immediateActionRecognizerDidCancelAnimation:]):
(-[WebImmediateActionController immediateActionRecognizerDidCompleteAnimation:]):

Use a dummy animation controller if the content prevents default.
(-[WebImmediateActionController _defaultAnimationController]):
(-[WebImmediateActionController _updateImmediateActionItem]):
(-[WebImmediateActionController webView:willHandleMouseDown:]): Deleted.

Set delaysPrimaryMouseButtonEvents to NO so that we get existing mouse events when
we expect to.
* WebView/WebView.mm:
(-[WebView _commonInitializationWithFrameName:groupName:]):

Cache the most recent pressure event so that we can send it to WebCore for all of
the mouse events.
(-[WebView _pressureEvent]):
(-[WebView _setPressureEvent:]):
* WebView/WebViewData.h:
* WebView/WebViewData.mm:
(-[WebViewPrivate dealloc]):
* WebView/WebViewInternal.h:

Tools:

PlatformEventFactory::createPlatformMouseEvent() takes the last pressure event
now. Just send nil.
* TestWebKitAPI/Tests/mac/MenuTypesForMouseEvents.mm:
(TestWebKitAPI::buildAndPerformTest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183954 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMedia Controls: Scrubber should be independent of actual video time, causes scrubber...
roger_fong@apple.com [Thu, 7 May 2015 22:25:36 +0000 (22:25 +0000)]
Media Controls: Scrubber should be independent of actual video time, causes scrubber to be jumpy.
https://bugs.webkit.org/show_bug.cgi?id=144700.
<rdar://problem/19997548>

Reviewed by Jer Noble.

Update time and timeline during the timeline input event instead of the wrapper's mousemove.
(Controller.prototype.handleWrapperMouseMove):
(Controller.prototype.handleTimelineMouseMove):
(Controller.prototype.drawTimelineBackground):

(Controller.prototype.updateControlsWhileScrubbing):
Updates time and scrubber to reflect timeline user input.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183953 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoCleanup after r183940
mmaxfield@apple.com [Thu, 7 May 2015 22:23:05 +0000 (22:23 +0000)]
Cleanup after r183940
https://bugs.webkit.org/show_bug.cgi?id=144768

Unreviewed.

Looks like the empty directories were not deleted.

* Source/PAL: Removed.
* Source/PAL/Configurations: Removed.
* Source/PAL/PAL.xcodeproj: Removed.
* Source/PAL/graphics: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183952 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoHandle meta viewport in HTMLPreloadScanner
yoav@yoav.ws [Thu, 7 May 2015 22:16:27 +0000 (22:16 +0000)]
Handle meta viewport in HTMLPreloadScanner
https://bugs.webkit.org/show_bug.cgi?id=144640

Reviewed by Dean Jackson.

Source/WebCore:

Add support for the <meta> tag to the HTMLPreloadScanner
and when the meta tag's name is viewport, get the document
to process the tag's attribute value, and set its ViewportArgument.

Tests: fast/dom/HTMLMetaElement/meta-preloader.html

* html/parser/HTMLPreloadScanner.cpp:
(WebCore::TokenPreloadScanner::tagIdFor): Added meta.
(WebCore::TokenPreloadScanner::initiatorFor): Added meta.
(WebCore::TokenPreloadScanner::StartTagScanner::StartTagScanner): Call document.processViewport is meta viewport encountered.
(WebCore::TokenPreloadScanner::StartTagScanner::processAttributes):
(WebCore::TokenPreloadScanner::StartTagScanner::processAttribute): Keep track of meta viewport tags.
(WebCore::testPreloadScannerViewportSupport): Testing function that makes sure this functionality is working.
* html/parser/HTMLPreloadScanner.h:
* testing/Internals.cpp:
(WebCore::Internals::testPreloaderSettingViewport): Calls the testing function.
* testing/Internals.h:
* testing/Internals.idl:

LayoutTests:

Added a layout test that is simply calling an internals method
that tests the HTMLPreloadScanner.

* fast/dom/HTMLMetaElement/meta-preloader-expected.txt: Added.
* fast/dom/HTMLMetaElement/meta-preloader.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183951 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION (r164449): Subpixel rendering: http://www.apple.com/iphone-6/ "Faster...
zalan@apple.com [Thu, 7 May 2015 22:03:12 +0000 (22:03 +0000)]
REGRESSION (r164449): Subpixel rendering: apple.com/iphone-6/ "Faster wireless." image displays vertical black line on 1x displays at specific window width.
https://bugs.webkit.org/show_bug.cgi?id=144723
rdar://problem/18307094

Reviewed by Simon Fraser.

This patch ensures that the backing store's graphics layer is always positioned on a device pixel boundary.

While calculating the backing store's graphics layer location, its ancestor layer's compositing bounds is taken into account.
However the compositing bounds is an unsnapped value, so in order to place the graphics layer properly,
we need to pixel align the ancestor compositing bounds before using it to adjust the child's graphics layer position.

Source/WebCore:

Test: compositing/ancestor-compositing-layer-is-on-subpixel-position.html

* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::updateGeometry):

LayoutTests:

* compositing/ancestor-compositing-layer-is-on-subpixel-position-expected.html: Added.
* compositing/ancestor-compositing-layer-is-on-subpixel-position.html: Added.
* platform/mac/compositing/layer-creation/overlap-animation-container-expected.txt: progression.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183950 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFollowup fixes to r183936
mrobinson@webkit.org [Thu, 7 May 2015 21:45:30 +0000 (21:45 +0000)]
Followup fixes to r183936

I neglected to include a few review comments when I landed the patch.

* UIProcess/gtk/TextCheckerGtk.cpp:
(WebKit::checkerState): Remove double semicolons.
(WebKit::updateStateForAllProcessPools): Use range-based-for loop here and
give the function a better name.
(WebKit::TextChecker::setContinuousSpellCheckingEnabled): Use new function name
(WebKit::TextChecker::setGrammarCheckingEnabled):Use new function name
(WebKit::TextChecker::ignoreWord): Fix unused parameter warning.
(WebKit::updateStateForAllContexts): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183949 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix sizes crash and add invalid value tests.
yoav@yoav.ws [Thu, 7 May 2015 21:22:35 +0000 (21:22 +0000)]
Fix sizes crash and add invalid value tests.
https://bugs.webkit.org/show_bug.cgi?id=144739

Reviewed by Darin Adler.

Source/WebCore:

Make sure that only CSS length are allowed when the sizes parser is calling computeLength.
Also make sure that for invalid lengths, the 100vw default is used instead.

Test: fast/dom/HTMLImageElement/sizes/image-sizes-invalids.html

* css/SourceSizeList.cpp:
(WebCore::computeLength):
(WebCore::defaultLength):
(WebCore::parseSizesAttribute):

LayoutTests:

Add tests that make sure that invalid values are properly handled, and a 100vw
source-size length is being used for srcset and for intrinsic dimension calculation.

* fast/dom/HTMLImageElement/sizes/image-sizes-invalids-expected.txt: Added.
* fast/dom/HTMLImageElement/sizes/image-sizes-invalids.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183948 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Expanding Object with only __proto__ looks poor should have a label
commit-queue@webkit.org [Thu, 7 May 2015 21:16:59 +0000 (21:16 +0000)]
Web Inspector: Expanding Object with only __proto__ looks poor should have a label
https://bugs.webkit.org/show_bug.cgi?id=144755

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-05-07
Reviewed by Timothy Hatcher.

Better handle cases where expanding an object has only a __proto__
and no other properties.

* UserInterface/Views/ObjectTreePropertyTreeElement.js:
(WebInspector.ObjectTreePropertyTreeElement.mode.prototype.this.children.length):
(WebInspector.ObjectTreePropertyTreeElement):
* UserInterface/Views/ObjectTreeView.css:
(.object-tree-outline li .empty-message):
* UserInterface/Views/ObjectTreeView.js:
(WebInspector.ObjectTreeView.createEmptyMessageElement):
(WebInspector.ObjectTreeView.comparePropertyDescriptors):
(WebInspector.ObjectTreeView.prototype._updateProperties):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183947 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Fix querySelector in ResourceContentView.js, caught by assertion
commit-queue@webkit.org [Thu, 7 May 2015 21:15:21 +0000 (21:15 +0000)]
Web Inspector: Fix querySelector in ResourceContentView.js, caught by assertion
https://bugs.webkit.org/show_bug.cgi?id=144756

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-05-07
Reviewed by Timothy Hatcher.

* UserInterface/Views/ResourceContentView.js:
(WebInspector.ResourceContentView.prototype._hasContent):
Ensure querySelector call is using the class name correctly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183946 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Checks for DEVELOPMENT_BUILD are all wrong
mcatanzaro@igalia.com [Thu, 7 May 2015 21:09:02 +0000 (21:09 +0000)]
[GTK] Checks for DEVELOPMENT_BUILD are all wrong
https://bugs.webkit.org/show_bug.cgi?id=144746

Reviewed by Martin Robinson.

.:

Don't set DEVELOPMENT_BUILD. We check whether it is defined for conditional compilation, not
what it is defined to, so defining it to 0 effectively turned it on always.

* Source/cmake/OptionsGTK.cmake:

Source/WebCore:

Use DEVELOPER_MODE rather than DEVELOPMENT_BUILD.

* platform/gtk/GtkUtilities.cpp:
* platform/gtk/GtkUtilities.h:
* platform/text/gtk/HyphenationLibHyphen.cpp:
(WebCore::availableLocales):

Source/WebKit2:

Use DEVELOPER_MODE rather than DEVELOPMENT_BUILD.

* Shared/gtk/ProcessExecutablePathGtk.cpp:
(WebKit::findWebKitProcess):
* UIProcess/API/gtk/WebKitWebContext.cpp:
(injectedBundleDirectory):
* UIProcess/gtk/TextCheckerGtk.cpp:
(WebKit::enchantTextChecker):

Tools:

Use DEVELOPER_MODE rather than DEVELOPMENT_BUILD.

* MiniBrowser/gtk/main.c:
(main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183945 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, add missing braces on a single-line if that got expanded in r183939
basile_clement@apple.com [Thu, 7 May 2015 20:59:44 +0000 (20:59 +0000)]
Unreviewed, add missing braces on a single-line if that got expanded in r183939

* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::buildExitArguments):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183944 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove the WK1-only code path for independently composited iframes
simon.fraser@apple.com [Thu, 7 May 2015 20:45:42 +0000 (20:45 +0000)]
Remove the WK1-only code path for independently composited iframes
https://bugs.webkit.org/show_bug.cgi?id=144722

Reviewed by Dean Jackson.
Source/WebCore:

In WebKit1 on Mac, we allowed iframes to be composited independently of their
parent document, relying on the fact that the frame's platform view can host
a layer-backed view. However, this ran into bugs (rdar://problem/18862298),
and triggers the assertion at the end of FrameView::updateLayoutAndStyleIfNeededRecursive(),
because the compositing update after a layout can dirty style in notifyIFramesOfCompositingChange().

Removing the WK1-only code path solves these problems. It also eliminates the need
to do compositing-specific frame overlap testing.

* page/FrameView.cpp:
(WebCore::FrameView::setIsOverlapped): No need to do compositing-related things here.
Any iframe that gets composited will participate in the normal compositing overlap
testing in its parent frame.
(WebCore::FrameView::hasCompositedContentIncludingDescendants): Deleted.
(WebCore::FrameView::hasCompositingAncestor): Deleted.
* page/FrameView.h:
* rendering/RenderLayerCompositor.cpp: Replace ownerElement() checks in this file
with an isMainFrameCompositor() for readability. Some 0->nullptr.
(WebCore::RenderLayerCompositor::cacheAcceleratedCompositingFlags):
(WebCore::RenderLayerCompositor::chromeClient):
(WebCore::RenderLayerCompositor::enclosingCompositorFlushingLayers):
(WebCore::RenderLayerCompositor::updateCompositingLayers):
(WebCore::RenderLayerCompositor::appendDocumentOverlayLayers):
(WebCore::RenderLayerCompositor::updateBacking):
(WebCore::RenderLayerCompositor::layerTreeAsText):
(WebCore::RenderLayerCompositor::frameContentsCompositor):
(WebCore::RenderLayerCompositor::setIsInWindow):
(WebCore::RenderLayerCompositor::requiresCompositingForScrollableFrame):
(WebCore::RenderLayerCompositor::requiresCompositingForFrame): frameRenderer.requiresAcceleratedCompositing()
already bails on no content RenderView, so the shouldPropagateCompositingToEnclosingFrame() check does
nothing and is removed.
(WebCore::RenderLayerCompositor::isAsyncScrollableStickyLayer):
(WebCore::RenderLayerCompositor::requiresScrollLayer):
(WebCore::RenderLayerCompositor::documentUsesTiledBacking):
(WebCore::RenderLayerCompositor::isMainFrameCompositor):
(WebCore::RenderLayerCompositor::shouldCompositeOverflowControls):
(WebCore::RenderLayerCompositor::requiresOverhangAreasLayer):
(WebCore::RenderLayerCompositor::requiresContentShadowLayer):
(WebCore::RenderLayerCompositor::updateLayerForTopOverhangArea):
(WebCore::RenderLayerCompositor::updateLayerForBottomOverhangArea):
(WebCore::RenderLayerCompositor::updateLayerForHeader):
(WebCore::RenderLayerCompositor::updateLayerForFooter):
(WebCore::RenderLayerCompositor::ensureRootLayer): Main frame attaches via ChromeClient,
all other frames attach via parent frame.
(WebCore::RenderLayerCompositor::notifyIFramesOfCompositingChange): This call to
scheduleSetNeedsStyleRecalc(SyntheticStyleChange) was the source of dirtying style after
layout, but is no longer needed so remove it.
(WebCore::RenderLayerCompositor::registerAllViewportConstrainedLayers):
(WebCore::RenderLayerCompositor::unregisterAllViewportConstrainedLayers):
(WebCore::RenderLayerCompositor::scrollingCoordinator):
(WebCore::RenderLayerCompositor::graphicsLayerFactory):
(WebCore::RenderLayerCompositor::allowsIndependentlyCompositedFrames): Deleted.
(WebCore::RenderLayerCompositor::shouldPropagateCompositingToEnclosingFrame): Deleted.
(WebCore::RenderLayerCompositor::mainFrameBackingIsTiled): Deleted. This was mis-named;
it really asks whether the document uses tiled backing, but does not check for main frame.
* rendering/RenderLayerCompositor.h:
* rendering/RenderWidget.cpp:
(WebCore::RenderWidget::paintContents): No need to do frame overlap testing for
compositing now.

LayoutTests:

Results different from WK2, because WK1 does not make layers for scrollbars.

* platform/mac-wk1/compositing/visible-rect/iframe-no-layers-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183943 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoHandle backdrop views that have to tile
dino@apple.com [Thu, 7 May 2015 20:13:47 +0000 (20:13 +0000)]
Handle backdrop views that have to tile
https://bugs.webkit.org/show_bug.cgi?id=142317
<rdar://problem/20049088>

Reviewed by Simon Fraser.

Take 2 - this was rolled out because Mavericks was crashing.

Source/WebCore:

Make sure backdrop layers don't tile. If they are big
enough, we'll leave it to the platform compositor to handle.

This also fixes a bug where if a layer changed from a backdrop
type to a tiled type, it would still retain its custom appearance
and we'd try to add children to the wrong layer.

Test: compositing/media-controls-bar-appearance-big.html

* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::commitLayerChangesBeforeSublayers): Check if
a layer needs a backdrop before checking if it needs to tile.

Source/WebKit2:

Add some better logging for custom appearance.

* Shared/mac/RemoteLayerTreeTransaction.mm:
(WebKit::RemoteLayerTreeTextStream::operator<<):

LayoutTests:

A test that creates some backdrop views, then makes them
big enough that it would trigger tiling (which we don't want
to happen).

* compositing/media-controls-bar-appearance-big-expected.txt: Added.
* compositing/media-controls-bar-appearance-big.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183942 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd PLUGIN_ARCHITECTURE(X11) around m_frameRectInWindowCoordinates in NetscapePlugin.
commit-queue@webkit.org [Thu, 7 May 2015 20:06:13 +0000 (20:06 +0000)]
Add PLUGIN_ARCHITECTURE(X11) around m_frameRectInWindowCoordinates in NetscapePlugin.
https://bugs.webkit.org/show_bug.cgi?id=144490

Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-07
Reviewed by Darin Adler.

m_frameRectInWindowCoordinates in NetscapePlugin is currently being used only for
the windowed plugins, and the windowed plugins are only supported on X11. So we can
guard it with PLUGIN_ARCHITECTURE(X11).

No new tests, no behavior change.

* WebProcess/Plugins/Netscape/NetscapePlugin.cpp:
(WebKit::NetscapePlugin::geometryDidChange):
* WebProcess/Plugins/Netscape/NetscapePlugin.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183941 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRevert "Introducing the Platform Abstraction Layer (PAL)"
mmaxfield@apple.com [Thu, 7 May 2015 18:52:59 +0000 (18:52 +0000)]
Revert "Introducing the Platform Abstraction Layer (PAL)"
https://bugs.webkit.org/show_bug.cgi?id=144751

Unreviewed.

PAL should be a new target inside WebCore, rather than a top-level folder.

.:

* WebKit.xcworkspace/contents.xcworkspacedata: Add it to the workspace.

Source/JavaScriptCore:

* Configurations/FeatureDefines.xcconfig: Updated

Source/WebCore:

* WebCore.xcodeproj/project.pbxproj:
* Configurations/FeatureDefines.xcconfig:

Source/WebKit/mac:

* Configurations/FeatureDefines.xcconfig: Updated

Source/WebKit2:

* Configurations/FeatureDefines.xcconfig: Updated

Tools:

* Scripts/build-webkit: Add it to build-webkit.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183940 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDumping OSR ExitValue should expand materializations only once
basile_clement@apple.com [Thu, 7 May 2015 18:39:46 +0000 (18:39 +0000)]
Dumping OSR ExitValue should expand materializations only once
https://bugs.webkit.org/show_bug.cgi?id=144694

Reviewed by Filip Pizlo.

Currently, dumping OSR exit values will print the full materialization
information each time it is encountered. We change it to print only a
brief description (only the materialization's address), and print the
whole set of materializations later on.

This makes the dump less confusing (less likely to think that two
instances of the same materialization are different), and will be a
necessary change if/when we support materialization cycles.

* ftl/FTLCompile.cpp:
(JSC::FTL::mmAllocateDataSection):
* ftl/FTLExitValue.cpp:
(JSC::FTL::ExitValue::dumpInContext):
* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::buildExitArguments):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183939 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWorker threads leak WeakBlocks (as seen on leaks bot)
akling@apple.com [Thu, 7 May 2015 18:32:29 +0000 (18:32 +0000)]
Worker threads leak WeakBlocks (as seen on leaks bot)
<https://webkit.org/b/144721>
<rdar://problem/20848288>

Reviewed by Darin Adler.

Nuke any remaining empty WeakBlocks when the Heap is being torn down.
Trying to peek into these blocks after the VM is dead would be a bug anyway.

This fixes a ~750 KB leak seen on the leaks bot.

* heap/Heap.cpp:
(JSC::Heap::~Heap):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183938 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2][Fullscreen] Fullscreen video does not enter low-power mode.
jer.noble@apple.com [Thu, 7 May 2015 18:26:14 +0000 (18:26 +0000)]
[WK2][Fullscreen] Fullscreen video does not enter low-power mode.
https://bugs.webkit.org/show_bug.cgi?id=144744

Reviewed by Darin Adler.

One of the requirements of entering low-power compositing mode is that no masking layers
are present in any of the ancestors of the fullscreen video layer. So once our fullscreen
transition animation completes, remove the mask layer entirely from our clipping layer.
This means it needs to be re-created and added when entering fullscreen, rather than just
at initialization time.

* UIProcess/mac/WKFullScreenWindowController.mm:
(-[WKFullScreenWindowController initWithWindow:webView:]):
(-[WKFullScreenWindowController finishedEnterFullScreenAnimation:]):
(-[WKFullScreenWindowController _startEnterFullScreenAnimationWithDuration:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183937 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] All spell checking layout tests fail
mrobinson@webkit.org [Thu, 7 May 2015 18:24:31 +0000 (18:24 +0000)]
[GTK] All spell checking layout tests fail
https://bugs.webkit.org/show_bug.cgi?id=144648

Reviewed by Carlos Garcia Campos.

Source/WebKit2:

Simplify and complete the GTK+ spell-checking implementation. Instead of using
the EFL/GTK+ only WKTextChecker interface, directly implement Enchant spell-checking
in TextCheckerGtk. The per-context spell checking API is an illusion, since all
WKTextChecker implementations will adjust the same singleton. This can be a first step
toward removing it entirely.

Also complete the implementation of the TextChecker interface by duplicating the EFL code,
which reuses other methods in the same class. Eventually this duplication will collapse,
once GTK+ and EFL can share an Enchant implementation.

* PlatformGTK.cmake:
* UIProcess/API/gtk/WebKitPrivate.h:
* UIProcess/API/gtk/WebKitTextChecker.cpp: Removed.
* UIProcess/API/gtk/WebKitTextChecker.h: Removed.
* UIProcess/API/gtk/WebKitWebContext.cpp: Remove WebTextChecker usage and
replace it with calls directly to the TextChecker singleton via static methods.
* UIProcess/TextChecker.h:
* UIProcess/gtk/TextCheckerGtk.cpp: Move the WebTextChecker enchant implementation
here and reuse the EFL version of WebKit::TextChecker::requestCheckingOfString.
Also initialize languages properly when building for developer builds. This
allows WebKitTestRunner to work even when it doesn't explicitly set the languages.

Source/WTF:

* wtf/Platform.h: Activate UNIFIED_TEXT_CHECKING for GTK+.

LayoutTests:

* platform/gtk/TestExpectations: Unskip and reclassify previously failing tests.
* platform/gtk/editing/spelling/spellcheck-paste-expected.txt: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183936 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDon't branch when accessing the callee
ggaren@apple.com [Thu, 7 May 2015 18:01:46 +0000 (18:01 +0000)]
Don't branch when accessing the callee
https://bugs.webkit.org/show_bug.cgi?id=144645

Reviewed by Michael Saboff.

The branch was added in <http://trac.webkit.org/changeset/81040> without
explanation.

kling found it to be a performance problem. See <https://webkit.org/b/144586>.

Our theory of access to Registers is that it's up to the client to access
them in the right way. So, let's do that.

* interpreter/CallFrame.h:
(JSC::ExecState::callee):
(JSC::ExecState::setCallee): Call the field object instead of function
because nothing guarantees that it's a function.
* interpreter/ProtoCallFrame.h:
(JSC::ProtoCallFrame::callee):
(JSC::ProtoCallFrame::setCallee):
* interpreter/Register.h:
* runtime/JSObject.h:
(JSC::Register::object): Just do a cast like our other accessors do.
(JSC::Register::operator=):
(JSC::Register::function): Deleted.
(JSC::Register::withCallee): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183935 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoTried to fix the build.
mitz@apple.com [Thu, 7 May 2015 17:59:23 +0000 (17:59 +0000)]
Tried to fix the build.

* Configurations/WebCoreTestSupport.xcconfig:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183934 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSource/WebKit2:
adachan@apple.com [Thu, 7 May 2015 17:45:51 +0000 (17:45 +0000)]
Source/WebKit2:
Fix a couple of cases where the backForwardListState's currentIndex is not set correctly in WebBackForwardList::backForwardListState().
https://bugs.webkit.org/show_bug.cgi?id=144666

Reviewed by Darin Adler.

* UIProcess/WebBackForwardList.cpp:
(WebKit::WebBackForwardList::backForwardListState):
If the first item is filtered out and the currentIndex is 0, don't decrement currentIndex.
If all the items are filtered out, set currentIndex to the uninitialized value.

Tools:
Add a test for WKPageCopySessionState() with filtering.
https://bugs.webkit.org/show_bug.cgi?id=144666

Reviewed by Darin Adler.

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/WKPageCopySessionStateWithFiltering.cpp: Added.
(TestWebKitAPI::didFinishLoadForFrame):
(TestWebKitAPI::setPageLoaderClient):
(TestWebKitAPI::filterFirstItemCallback):
(TestWebKitAPI::filterAllItemsCallback):
(TestWebKitAPI::createSessionStates):
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183933 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAX: [ATK] The pre element should be mapped to ATK_ROLE_SECTION; not ATK_ROLE_PANEL
jdiggs@igalia.com [Thu, 7 May 2015 17:30:00 +0000 (17:30 +0000)]
AX: [ATK] The pre element should be mapped to ATK_ROLE_SECTION; not ATK_ROLE_PANEL
https://bugs.webkit.org/show_bug.cgi?id=144702

Reviewed by Chris Fleizach.

Source/WebCore:

Add a new PreRole AccessibilityRole so that per-platform mapping is possible.
Map PreRole to ATK_ROLE_SECTION for Efl and Gtk; to NSAccessibilityGroupRole
for the Mac. Add PreRole to the list of "text type" ATK roles.

No new tests. Existing tests were updated to reflect the corrected mapping.

* accessibility/AccessibilityObject.h:
* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::determineAccessibilityRole):
* accessibility/atk/AccessibilityObjectAtk.cpp:
(WebCore::AccessibilityObject::allowsTextRanges):
* accessibility/atk/WebKitAccessibleWrapperAtk.cpp:
(atkRole):
(roleIsTextType):
* accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(createAccessibilityRoleMap):

LayoutTests:

In addition to updating the platform expectations for Efl and Gtk, the computed
roles test needed to be updated because there is not a one-to-one mapping between
the HTML pre element and any ARIA role type.

* accessibility/roles-computedRoleString-expected.txt: Updated.
* accessibility/roles-computedRoleString.html: Updated.
* platform/efl/accessibility/roles-exposed-expected.txt: Updated.
* platform/gtk/accessibility/roles-exposed-expected.txt: Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183932 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] Remove Windows cruft
ossy@webkit.org [Thu, 7 May 2015 17:18:13 +0000 (17:18 +0000)]
[WK2] Remove Windows cruft
https://bugs.webkit.org/show_bug.cgi?id=144745

Reviewed by Darin Adler.

* NetworkProcess/soup/NetworkProcessSoup.cpp:
(WebKit::getMemorySize):
* Shared/CoordinatedGraphics/CoordinatedGraphicsArgumentCoders.cpp:
(IPC::ArgumentCoder<WebCore::GraphicsSurfaceToken>::encode):
(IPC::ArgumentCoder<WebCore::GraphicsSurfaceToken>::decode):
* WebProcess/soup/WebProcessSoup.cpp:
(WebKit::getMemorySize):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183931 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION(r182032): [GTK] Production build is broken after r182032
carlosgc@webkit.org [Thu, 7 May 2015 16:46:05 +0000 (16:46 +0000)]
REGRESSION(r182032): [GTK] Production build is broken after r182032
https://bugs.webkit.org/show_bug.cgi?id=144736

Reviewed by Martin Robinson.

The problem is that r182032 moved the NetworkProcess sources from
WebKit2 sources to NetworkProcess sources. That's not a problem
for developer builds where all internal symbols are exported, but
doesn't work for production builds because all those internal
symbols are private.

* CMakeLists.txt: Use a different name for common network process
sources variable.
* PlatformMac.cmake: Append common network process sources to the
network process source list for Mac.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183930 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago<rdar://problem/19317140> [Xcode] Remove usage of AspenFamily.xcconfig in Source/
mitz@apple.com [Thu, 7 May 2015 16:26:59 +0000 (16:26 +0000)]
<rdar://problem/19317140> [Xcode] Remove usage of AspenFamily.xcconfig in Source/
https://bugs.webkit.org/show_bug.cgi?id=144727

Reviewed by Darin Adler.

Source/bmalloc:

* Configurations/Base.xcconfig: Dont’s include AspenFamily.xcconfig, and define
INSTALL_PATH_PREFIX for the iOS 8.x Simulator.

Source/JavaScriptCore:

* Configurations/Base.xcconfig: Don’t include AspenFamily.xcconfig, and define
INSTALL_PATH_PREFIX and LD_DYLIB_INSTALL_NAME for the iOS 8.x Simulator.

Source/PAL:

* Configurations/Base.xcconfig: Don’t include AspenFamily.xcconfig, and define
INSTALL_PATH_PREFIX for the iOS 8.x Simulator.

Source/ThirdParty:

* gtest/xcode/Config/General.xcconfig: Don’t include AspenFamily.xcconfig.

Source/ThirdParty/ANGLE:

* Configurations/Base.xcconfig: Don’t include AspenFamily.xcconfig, and define
INSTALL_PATH_PREFIX and LD_DYLIB_INSTALL_NAME for the iOS 8.x Simulator.

Source/WebCore:

* Configurations/Base.xcconfig: Don’t include AspenFamily.xcconfig.
* Configurations/WebCore.xcconfig: For iOS, define INSTALL_PATH in terms of
INSTALL_PATH_PREFIX, and set INSTALL_PATH_PREFIX for the iOS 8.x Simulator.
* Configurations/WebCoreTestSupport.xcconfig: Define INSTALL_PATH in terms of
INSTALL_PATH_PREFIX and INSTALL_PATH_ACTUAL, and define INSTALL_PATH_PREFIX for the iOS 8.x
Simulator.

Source/WebKit/mac:

* Configurations/Base.xcconfig: Don’t include AspenFamily.xcconfig.
* Configurations/WebKitLegacy.xcconfig: Define INSTALL_PATH in terms of
INSTALL_PATH_PREFIX and INSTALL_PATH_ACTUAL, and define INSTALL_PATH_PREFIX for the iOS 8.x
Simulator.

Source/WebKit2:

* Configurations/Base.xcconfig: Don’t include AspenFamily.xcconfig, and define
INSTALL_PATH_PREFIX and LD_DYLIB_INSTALL_NAME for the iOS 8.x Simulator.
* Configurations/WebKit2.xcconfig: Define INSTALL_PATH_ACTUAL and let INSTALL_PATH be
derived from it and INSTALL_PATH_PREFIX. Don’t override DYLIB_INSTALL_NAME_BASE for iOS,
because we define LD_DYLIB_INSTALL_NAME directly where needed.

Source/WTF:

* Configurations/Base.xcconfig: Don’t include AspenFamily.xcconfig, and define
INSTALL_PATH_PREFIX and LD_DYLIB_INSTALL_NAME for the iOS 8.x Simulator.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183929 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSpecial-case Int32 values in JSON.stringify().
akling@apple.com [Thu, 7 May 2015 15:56:40 +0000 (15:56 +0000)]
Special-case Int32 values in JSON.stringify().
<https://webkit.org/b/144731>

Reviewed by Michael Saboff.

Add a fast path for serializing Int32 values to JSON. This is far faster than dragging
simple integers through the full-blown dtoa() machinery.

~50% speedup on Kraken/json-stringify-tinderbox.

* runtime/JSONObject.cpp:
(JSC::Stringifier::appendStringifiedValue):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183928 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Mac] Playback target isn't set on new element
eric.carlson@apple.com [Thu, 7 May 2015 14:12:17 +0000 (14:12 +0000)]
[Mac] Playback target isn't set on new element
https://bugs.webkit.org/show_bug.cgi?id=144724

Reviewed by Jer Noble.

* platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.h:
* platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::createAVPlayer): Call setShouldPlayToPlaybackTarget
if necessary.
(WebCore::MediaPlayerPrivateAVFoundationObjC::setShouldPlayToPlaybackTarget): Remember
the setting in case we don't have an AVPlayer yet.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183927 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[EFL] Unreviewed gardening.
ossy@webkit.org [Thu, 7 May 2015 14:03:43 +0000 (14:03 +0000)]
[EFL] Unreviewed gardening.

Removed stale expected files, because these tests are reftests now. (as RWT suggested)
Commented old expectations after r183925 to avoid having ambiguous expectations.

* platform/efl/TestExpectations:
* platform/efl/animations/cross-fade-background-image-expected.png: Removed.
* platform/efl/animations/cross-fade-background-image-expected.txt: Removed.
* platform/efl/css3/masking/clip-path-circle-relative-overflow-expected.png: Removed.
* platform/efl/css3/masking/clip-path-polygon-evenodd-expected.png: Removed.
* platform/efl/css3/masking/clip-path-polygon-expected.png: Removed.
* platform/efl/css3/masking/clip-path-polygon-nonzero-expected.png: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183926 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[EFL] Mouseevent/hover tests fail with XorgDriver
ossy@webkit.org [Thu, 7 May 2015 13:48:10 +0000 (13:48 +0000)]
[EFL] Mouseevent/hover tests fail with XorgDriver
https://bugs.webkit.org/show_bug.cgi?id=144741

Unreviewed gardening, mark failing tests, skip timeouting tests.

* platform/efl/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183925 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[EFL] REGRESSION(r183907): It made all inspector-protocol test crash
ossy@webkit.org [Thu, 7 May 2015 10:14:02 +0000 (10:14 +0000)]
[EFL] REGRESSION(r183907): It made all inspector-protocol test crash
https://bugs.webkit.org/show_bug.cgi?id=144737

Unreviewed gardening, skip crashing tests.

* platform/efl/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183923 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed. Update OptionsGTK.cmake and NEWS for 2.9.1 release.
carlosgc@webkit.org [Thu, 7 May 2015 10:05:16 +0000 (10:05 +0000)]
Unreviewed. Update OptionsGTK.cmake and NEWS for 2.9.1 release.

.:

* Source/cmake/OptionsGTK.cmake: Bump version numbers.

Source/WebKit2:

* gtk/NEWS: Add release notes for 2.9.1.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183922 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Clean up RunLoop implementation
zandobersek@gmail.com [Thu, 7 May 2015 08:26:56 +0000 (08:26 +0000)]
[GTK] Clean up RunLoop implementation
https://bugs.webkit.org/show_bug.cgi?id=144729

Reviewed by Carlos Garcia Campos.

Clean up the RunLoop implementation for the GTK port,
removing unnecessary methods and using simpler variable names.

Nested GMainLoops in RunLoop::run() are now created for the
RunLoop's GMainContext, and not for the default context (enforced
through the null argument to g_main_loop_new()).

* wtf/RunLoop.h:
* wtf/gtk/RunLoopGtk.cpp:
(WTF::RunLoop::RunLoop):
(WTF::RunLoop::~RunLoop):
(WTF::RunLoop::run):
(WTF::RunLoop::stop):
(WTF::RunLoop::wakeUp):
(WTF::RunLoop::TimerBase::start):
(WTF::RunLoop::innermostLoop): Deleted.
(WTF::RunLoop::pushNestedMainLoop): Deleted.
(WTF::RunLoop::popNestedMainLoop): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183921 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix <script srcset> being loaded by the preloader.
yoav@yoav.ws [Thu, 7 May 2015 08:09:03 +0000 (08:09 +0000)]
Fix <script srcset> being loaded by the preloader.
https://bugs.webkit.org/show_bug.cgi?id=144675

Reviewed by Darin Adler.

Source/WebCore:

Fix HTMLPreloadScanner to make sure that srcset is only loaded
when it is present on an <img>.

Tests: http/tests/loading/preload-script-src.html
       http/tests/loading/preload-script-srcset-ignored.html

* html/parser/HTMLPreloadScanner.cpp:
(WebCore::TokenPreloadScanner::StartTagScanner::processAttributes):

LayoutTests:

Add preloader tests to make sure that <script src> is preloaded,
and <script srcset> is not.

* http/tests/loading/preload-script-src-expected.txt: Added.
* http/tests/loading/preload-script-src.html: Added.
* http/tests/loading/preload-script-srcset-ignored-expected.txt: Added.
* http/tests/loading/preload-script-srcset-ignored.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183920 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Remove dock button from Web Inspector window
carlosgc@webkit.org [Thu, 7 May 2015 08:02:24 +0000 (08:02 +0000)]
[GTK] Remove dock button from Web Inspector window
https://bugs.webkit.org/show_bug.cgi?id=144728

Reviewed by Sergio Villar Senin.

We added the dock buttons to the headerbar, because they were
removed from the inspector toolbar, but with the new tabbed
design, the dock buttons are back in the toolbar and look redundant.

* UIProcess/WebInspectorProxy.h:
* UIProcess/gtk/WebInspectorProxyGtk.cpp:
(WebKit::WebInspectorProxy::createInspectorWindow):
(WebKit::WebInspectorProxy::platformAttachAvailabilityChanged):
(WebKit::WebInspectorProxy::dockButtonClicked): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183919 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd missing type traits specialization for PlatformDisplayWayland
zandobersek@gmail.com [Thu, 7 May 2015 07:09:06 +0000 (07:09 +0000)]
Add missing type traits specialization for PlatformDisplayWayland
https://bugs.webkit.org/show_bug.cgi?id=144671

Reviewed by Carlos Garcia Campos.

* platform/graphics/wayland/PlatformDisplayWayland.h:
Add missing type traits specialization for the PlatformDisplayWayland
class via the SPECIALIZE_TYPE_TRAITS_PLATFORM_DISPLAY macro.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183918 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoVersioning.
bshafiei@apple.com [Thu, 7 May 2015 06:53:41 +0000 (06:53 +0000)]
Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183917 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoiOS build fix attempt
ap@apple.com [Thu, 7 May 2015 06:20:20 +0000 (06:20 +0000)]
iOS build fix attempt

Link WebKitTestRunnerApp to WebKit, now that we don't have WebKit2.

* WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183915 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd srcset preloader tests
yoav@yoav.ws [Thu, 7 May 2015 06:13:06 +0000 (06:13 +0000)]
Add srcset preloader tests
https://bugs.webkit.org/show_bug.cgi?id=144674

Reviewed by Darin Adler.

Add several preloader tests making sure that the when srcset is being preloaded,
the right resource is fetched, and the other resource is not.

* http/tests/loading/preload-img-src-expected.txt: Added.
* http/tests/loading/preload-img-src.html: Added.
* http/tests/loading/preload-img-srcset-expected.txt: Added.
* http/tests/loading/preload-img-srcset-sizes-expected.txt: Added.
* http/tests/loading/preload-img-srcset-sizes.html: Added.
* http/tests/loading/preload-img-srcset.html: Added.
* http/tests/resources/checkPreload.js: Added.
(checkForPreload):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183914 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[CSS Grid Layout] Mark grid shorthands as layout dependent
rego@igalia.com [Thu, 7 May 2015 06:07:25 +0000 (06:07 +0000)]
[CSS Grid Layout] Mark grid shorthands as layout dependent
https://bugs.webkit.org/show_bug.cgi?id=144687

Reviewed by Darin Adler.

Source/WebCore:

grid-template and grid shorthands were not marked us layout dependent,
so you can end up calling them before the grid has been laid out and
you'll get a crash at valueForGridTrackList() because of trackPositions
won't be initialized yet.

Tests: fast/css-grid-layout/grid-shorthand-computed-style-crash.html
       fast/css-grid-layout/grid-template-shorthand-computed-style-crash.html

* css/CSSComputedStyleDeclaration.cpp:
(WebCore::isLayoutDependent): Add the grid shorthands only if the object
is a RenderGrid as otherwise it's not needed (using the same condition
for grid-template-columns and grid-template-rows properties too).

LayoutTests:

* fast/css-grid-layout/grid-shorthand-computed-style-crash-expected.txt: Added.
* fast/css-grid-layout/grid-shorthand-computed-style-crash.html: Added.
* fast/css-grid-layout/grid-template-shorthand-computed-style-crash-expected.txt: Added.
* fast/css-grid-layout/grid-template-shorthand-computed-style-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183913 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoToT WebKit crashes while loading ES6 compatibility table
rniwa@webkit.org [Thu, 7 May 2015 05:15:56 +0000 (05:15 +0000)]
ToT WebKit crashes while loading ES6 compatibility table
https://bugs.webkit.org/show_bug.cgi?id=144726

Reviewed by Filip Pizlo.

Source/JavaScriptCore:

The bug was caused by parseClass superfluously avoiding to build up the string after seeing {.

Always build the identifier here as it could be a method name.

* parser/Parser.cpp:
(JSC::Parser<LexerType>::parseClass):

LayoutTests:

Added new test cases.

* js/class-syntax-string-and-numeric-names-expected.txt:
* js/script-tests/class-syntax-string-and-numeric-names.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183912 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2][Fullscreen] Elements whose children extend beyond their bounds are clipped...
jer.noble@apple.com [Thu, 7 May 2015 04:56:11 +0000 (04:56 +0000)]
[WK2][Fullscreen] Elements whose children extend beyond their bounds are clipped in fullscreen mode.
https://bugs.webkit.org/show_bug.cgi?id=144716

Reviewed by Darin Adler.

We create a mask animation for the transition between windowed and fullscreen modes, on the
assumption that the element being taken into fullscreen mode does not have visible children
who extend beyond that elements bounds. This assumption breaks down in the case where div
with absolutely positioned children is taken fullscreen. While we can't necessarily make the
transition look correct in this case, we can remove the mask after the transition completes.

* UIProcess/mac/WKFullScreenWindowController.mm:
(-[WKFullScreenWindowController finishedEnterFullScreenAnimation:]):
* WebProcess/MediaCache/WebMediaKeyStorageManager.cpp:
(WebKit::removeAllMediaKeyStorageForOriginPath):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183911 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix a build break on EFL debug build
commit-queue@webkit.org [Thu, 7 May 2015 04:46:07 +0000 (04:46 +0000)]
Fix a build break on EFL debug build
https://bugs.webkit.org/show_bug.cgi?id=144654

Patch by Jinwoo Jeong <jw00.jeong@samsung.com> on 2015-05-06
Reviewed by Gyuyoung Kim.

No new tests, just build fix.

* Shared/CoordinatedGraphics/CoordinatedGraphicsScene.cpp:
(WebKit::CoordinatedGraphicsScene::paintToCurrentGLContext):
(WebKit::CoordinatedGraphicsScene::paintToGraphicsContext):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183910 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS][WK2] Pause/resume database thread when UIProcess enters/leaves the background
dbates@webkit.org [Thu, 7 May 2015 03:23:57 +0000 (03:23 +0000)]
[iOS][WK2] Pause/resume database thread when UIProcess enters/leaves the background
https://bugs.webkit.org/show_bug.cgi?id=144657
<rdar://problem/18894598>

Reviewed by Andy Estes.

Source/WebCore:

Export WebCore functionality to pause and resume the database thread so that we can
make use of this functionality from WebKit2.

* Modules/webdatabase/AbstractDatabaseServer.h:
* Modules/webdatabase/DatabaseManager.cpp:
(WebCore::DatabaseManager::setPauseAllDatabases): Added; turns around and calls DatabaseServer::setPauseAllDatabases().
* Modules/webdatabase/DatabaseManager.h:
* Modules/webdatabase/DatabaseServer.cpp:
(WebCore::DatabaseServer::setPauseAllDatabases): Added; turns around and calls
DatabaseTracker::tracker().setDatabasesPaused() to pause or resume the database thread.
For now, we guard this call with PLATFORM(IOS). We'll look to remove this guard once
we fix <https://bugs.webkit.org/show_bug.cgi?id=144660>.
* Modules/webdatabase/DatabaseServer.h:

Source/WebKit2:

Pause and resume the database thread when the UIProcess enters and leaves the background,
respectively, so that we avoid WebProcess termination due to holding a locked SQLite
database file when the WebProcess is suspended. This behavior matches the analagous
behavior in Legacy WebKit.

* UIProcess/WebPageProxy.h:
* UIProcess/ios/WKContentView.mm:
(-[WKContentView _applicationDidEnterBackground:]): Call WebPageProxy::applicationDidEnterBackground()
when the UIProcess enters the background.
* UIProcess/ios/WebPageProxyIOS.mm:
(WebKit::WebPageProxy::applicationDidEnterBackground): Added; notify the WebProcess to pause the database thread.
We temporarily take out background assertion on the WebProcess before sending this notification to ensure that the
WebProcess is running to receive it. We'll release this assertion when the WebProcess replies that it received the
notification.
* WebProcess/WebCoreSupport/WebDatabaseManager.cpp:
(WebKit::WebDatabaseManager::setPauseAllDatabases): Added; turns around and calls DatabaseManager::setPauseAllDatabases().
* WebProcess/WebCoreSupport/WebDatabaseManager.h:
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in: Add message ApplicationDidEnterBackground(). Also,
add empty lines to help demarcate this message and the other UIKit application lifecycle-related
messages from the rest of the list of messages.
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::applicationWillEnterForeground): Resume the database thread.
(WebKit::WebPage::applicationDidEnterBackground): Pause the database thread.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183909 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[EFL] ewk_view_page_contents_get() API test is flaky
ryuan.choi@navercorp.com [Thu, 7 May 2015 03:19:42 +0000 (03:19 +0000)]
[EFL] ewk_view_page_contents_get() API test is flaky
https://bugs.webkit.org/show_bug.cgi?id=144683

Reviewed by Gyuyoung Kim.

Callback of WKPageGetContentsAsMHTMLData contains WKData which does not contain null terminated character.

* UIProcess/API/efl/ewk_view.cpp: Pass the data of string instead of WKData directly.
(ewkViewPageContentsAsMHTMLCallback):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183908 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Docking Toolbar Buttons do not respect Docking Availability
commit-queue@webkit.org [Thu, 7 May 2015 03:04:32 +0000 (03:04 +0000)]
Web Inspector: Docking Toolbar Buttons do not respect Docking Availability
https://bugs.webkit.org/show_bug.cgi?id=144714

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-05-06
Reviewed by Timothy Hatcher.

Source/WebInspectorUI:

* UserInterface/Base/Main.js:
(WebInspector.contentLoaded):
(WebInspector.updateDockingAvailability):
(WebInspector._updateDockNavigationItems):
Update the Toolbar UI based on docking availability.
Assume docking is unavailable by default.

* UserInterface/Protocol/InspectorFrontendAPI.js:
(InspectorFrontendAPI.setDockingUnavailable):
Add back this API as the backend may be calling it when
docking availability changes. This just calls out to
global WebInspector functions to update the Toolbar UI.

* UserInterface/Base/Test.js:
(WebInspector.updateDockedState):
(WebInspector.updateDockingAvailability):
Do nothing in tests.

Source/WebKit2:

* UIProcess/WebInspectorProxy.cpp:
(WebKit::WebInspectorProxy::createInspectorPage):
(WebKit::WebInspectorProxy::attachAvailabilityChanged):
* WebProcess/WebPage/WebInspectorUI.cpp:
(WebKit::WebInspectorUI::setDockingUnavailable):
* WebProcess/WebPage/WebInspectorUI.h:
* WebProcess/WebPage/WebInspectorUI.messages.in:
When docking availability changes, inform the inspector page
so it can respond by updating its UI.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183907 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoScroll-snap points do not handle margins and padding propertly
bfulgham@apple.com [Thu, 7 May 2015 02:53:00 +0000 (02:53 +0000)]
Scroll-snap points do not handle margins and padding propertly
https://bugs.webkit.org/show_bug.cgi?id=144647
<rdar://problem/20829473>

Reviewed by Simon Fraser.

Source/WebCore:

The calculation of scroll snap points was incorrect because it did not account for margins or padding.
This was fixed by using the "paddingBoxRect" to represent the overall size of the view area, and
subtracting the relevant padding when computing the size of the scroll snap offsets.

Extend testing internals with accessor methods to retrieve string representations of the scroll snap
offsets and scroll snap coordinates computed during layout. These values are used in the new
'css3/scroll-snap/scroll-snap-offsets-and-coordinates.html' test.

New tests:
* css3/scroll-snap/scroll-snap-offsets-and-coordinates.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-borders.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-padding.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-rotated.html: Added.

* css/LengthFunctions.h: Expose the 'valueForLength' method so we can use it in WebCoreTestSupport.
* dom/Node.h: Expose the 'renderBox' method so we can use it in WebCoreTestSupport.
* page/scrolling/AxisScrollSnapOffsets.cpp:
(WebCore::appendChildSnapOffsets): Use 'contentBoxRect' for overall size, so that we don't use padding
as part of our offset calculations.
(WebCore::updateSnapOffsetsForScrollableArea): Ditto.
* rendering/RenderBox.h: Expose 'canBeScrolledAndHasScrollableArea' for use in WebCoreTestSupport.
* rendering/style/RenderStyle.h: Expose 'scrollSnapCoordinates' for use in WebCoreTestSupport.
* testing/Internals.cpp:
(WebCore::appendOffsets): Helper function.
(WebCore::Internals::scrollSnapOffsets): Added.
* testing/Internals.h:
* testing/Internals.idl:

LayoutTests:

Add new scroll-snap tests that cover elements with borders, padding, and rotation. Also correct the
cause of the 'scroll-snap-mandatory-overflow.html' test flakiness by making sure the mouse is always
inside the element when starting new wheel gestures.

Also add a new 'generic' test that confirms that we have expected results for scroll-snap coordinates
and calculated scroll-snap-offset values.

* css3/scroll-snap/scroll-snap-offsets.html: Added.
* platform/mac-wk2/TestExpectations: Remove flakiness annotation for scroll-snap-mandatory-overflow.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-iframe.html:
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-borders-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-borders.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-overflow.html:
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-padding-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-padding.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-rotated-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-rotated.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183906 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: The text in the left pane overlaps the "Filter Time Events" field...
mattbaker@apple.com [Thu, 7 May 2015 02:32:57 +0000 (02:32 +0000)]
Web Inspector: The text in the left pane overlaps the "Filter Time Events" field in the Timeline after the Web Inspector is resized
https://bugs.webkit.org/show_bug.cgi?id=144720

Reviewed by Timothy Hatcher.

* UserInterface/Views/FilterBar.css:
(.filter-bar):
(.filter-bar > .navigation-bar):
(.filter-bar > input[type="search"]):
(.filter-bar > .navigation-bar > .item): Deleted.
Filter bar now has opaque background and height updated to the ubiquitous 29px.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183905 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoClean up TextRun constructors
mmaxfield@apple.com [Thu, 7 May 2015 02:09:26 +0000 (02:09 +0000)]
Clean up TextRun constructors
https://bugs.webkit.org/show_bug.cgi?id=144712

Reviewed by Zalan Bujtas.

This patch uses constructor forwarding to make TextRun's constructors much simpler. It then
updates the implementations of RenderBlock::constructTextRun() to be more consistent with
TextRun (via using StringViews).

No new tests because there is no behavior change.

* mathml/MathMLMencloseElement.cpp:
(WebCore::MathMLMencloseElement::longDivLeftPadding):
* platform/graphics/TextRun.h:
(WebCore::TextRun::TextRun):
* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::constructTextRun):
(WebCore::constructTextRunInternal): Deleted.
* rendering/RenderBlock.h:
* rendering/RenderEmbeddedObject.cpp:
(WebCore::RenderEmbeddedObject::paintReplaced):
(WebCore::RenderEmbeddedObject::unavailablePluginIndicatorBounds):
(WebCore::RenderEmbeddedObject::isInUnavailablePluginIndicator):
* rendering/svg/SVGTextMetricsBuilder.cpp:
(WebCore::SVGTextMetricsBuilder::SVGTextMetricsBuilder):
* platform/win/DragImageWin.cpp:
(WebCore::createDragImageForLink):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183904 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] Simplify AsyncRequestImpl::ArgumentType parameters
commit-queue@webkit.org [Thu, 7 May 2015 01:23:12 +0000 (01:23 +0000)]
[WK2] Simplify AsyncRequestImpl::ArgumentType parameters
https://bugs.webkit.org/show_bug.cgi?id=144689

Patch by Joonghun Park <jh718.park@samsung.com> on 2015-05-06
Reviewed by Darin Adler.

Remove duplicated typename and ::type from AsyncRequestImpl related
function parameters.

* Shared/AsyncRequest.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183903 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, updated my list of email addresses.
mattbaker@apple.com [Thu, 7 May 2015 00:59:05 +0000 (00:59 +0000)]
Unreviewed, updated my list of email addresses.

* Scripts/webkitpy/common/config/contributors.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183902 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRename URL::copy() to URL::isolatedCopy() to match String.
commit-queue@webkit.org [Thu, 7 May 2015 00:28:22 +0000 (00:28 +0000)]
Rename URL::copy() to URL::isolatedCopy() to match String.
https://bugs.webkit.org/show_bug.cgi?id=144684

Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-06
Reviewed by Darin Adler.

No new tests, no behavior change.

* Modules/websockets/WebSocketHandshake.cpp:
(WebCore::WebSocketHandshake::setURL):
(WebCore::WebSocketHandshake::httpURLForAuthenticationAndCookies):
* fileapi/ThreadableBlobRegistry.cpp:
(WebCore::BlobRegistryContext::BlobRegistryContext):
* html/parser/HTMLPreloadScanner.cpp:
(WebCore::TokenPreloadScanner::updatePredictedBaseURL):
* html/parser/HTMLResourcePreloader.h:
(WebCore::PreloadRequest::PreloadRequest):
* html/parser/XSSAuditor.cpp:
(WebCore::XSSAuditor::init):
* platform/CrossThreadCopier.cpp:
(WebCore::URL>::copy):
* platform/URL.cpp:
(WebCore::URL::isolatedCopy):
(WebCore::URL::copy): Deleted.
* platform/URL.h:
(WebCore::URLCapture::URLCapture):
* platform/network/BlobPart.h:
(WebCore::BlobPart::detachFromCurrentThread):
* platform/network/ResourceRequestBase.cpp:
(WebCore::ResourceRequestBase::copyData):
* platform/network/ResourceResponseBase.cpp:
(WebCore::ResourceResponseBase::copyData):
* workers/WorkerThread.cpp:
(WebCore::WorkerThreadStartupData::WorkerThreadStartupData):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183901 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r183894.
commit-queue@webkit.org [Thu, 7 May 2015 00:26:48 +0000 (00:26 +0000)]
Unreviewed, rolling out r183894.
https://bugs.webkit.org/show_bug.cgi?id=144719

crashes on mavericks (Requested by dino on #webkit).

Reverted changeset:

"Handle backdrop views that have to tile"
https://bugs.webkit.org/show_bug.cgi?id=142317
http://trac.webkit.org/changeset/183894

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183900 268f45cc-cd09-0410-ab3c-d52691b4dbfc