Timeline



Jun 3, 2012:

11:20 PM Changeset in webkit [119372] by rniwa@webkit.org
  • 2 edits in trunk/Source/WebKit/chromium

Roll Chromium DEPS from r140222 to r140260.

  • DEPS:
10:38 PM Changeset in webkit [119371] by kinuko@chromium.org
  • 9 edits in trunk

[chromium] DataTransferItem.webkitGetAsEntry() should synchronously return entry
​https://bugs.webkit.org/show_bug.cgi?id=87969

Reviewed by Tony Chang.

Source/WebCore:

Per discussion on whatwg we should make webkitGetAsEntry synchronous.
​http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Apr/0079.html

Tests: editing/pasteboard/data-transfer-items-drag-drop-entry.html

fast/filesystem/cross-filesystem-op.html

  • Modules/filesystem/DataTransferItemFileSystem.h:

(DataTransferItemFileSystem):

  • Modules/filesystem/DataTransferItemFileSystem.idl:
  • Modules/filesystem/chromium/DataTransferItemFileSystemChromium.cpp:

(WebCore::DataTransferItemFileSystem::webkitGetAsEntry): Changed it to synchronously return entry.

LayoutTests:

  • editing/pasteboard/data-transfer-items-drag-drop-entry-expected.txt:
  • editing/pasteboard/data-transfer-items-drag-drop-entry.html:
  • fast/filesystem/cross-filesystem-op.html:
  • platform/chromium/test_expectations.txt:
10:24 PM Changeset in webkit [119370] by shawnsingh@chromium.org
  • 2 edits in trunk/LayoutTests

Unreviewed fix of chromium TestExpectations

Some extra junk got added by accident when I tried to change
TestExpectations. This patch removes it.

  • platform/chromium/TestExpectations:
10:10 PM Changeset in webkit [119369] by shawnsingh@chromium.org
  • 17 edits
    43 adds in trunk/LayoutTests

Unreviewed rebasline for transforms/3d test cases.

9 tests re-baselined.
The remainin 6 tests are explicitly enumerated in the TestExpectations file now.

  • platform/chromium-linux/transforms/3d/general/perspective-units-expected.png: Added.
  • platform/chromium-linux/transforms/3d/hit-testing/backface-hit-test-expected.png: Added.
  • platform/chromium-linux/transforms/3d/hit-testing/backface-no-transform-hit-test-expected.png: Added.
  • platform/chromium-linux/transforms/3d/point-mapping/3d-point-mapping-3-expected.png: Added.
  • platform/chromium-linux/transforms/3d/point-mapping/3d-point-mapping-expected.png: Added.
  • platform/chromium-linux/transforms/3d/point-mapping/3d-point-mapping-origins-expected.png: Added.
  • platform/chromium-linux/transforms/3d/point-mapping/3d-point-mapping-overlapping-expected.png: Added.
  • platform/chromium-linux/transforms/3d/point-mapping/3d-point-mapping-preserve-3d-expected.png: Added.
  • platform/chromium-mac-leopard/transforms/3d/general/perspective-units-expected.png:
  • platform/chromium-mac-leopard/transforms/3d/hit-testing/backface-hit-test-expected.png: Added.
  • platform/chromium-mac-leopard/transforms/3d/hit-testing/backface-no-transform-hit-test-expected.png: Added.
  • platform/chromium-mac-leopard/transforms/3d/point-mapping/3d-point-mapping-3-expected.png: Added.
  • platform/chromium-mac-leopard/transforms/3d/point-mapping/3d-point-mapping-expected.png:
  • platform/chromium-mac-leopard/transforms/3d/point-mapping/3d-point-mapping-origins-expected.png:
  • platform/chromium-mac-leopard/transforms/3d/point-mapping/3d-point-mapping-overlapping-expected.png:
  • platform/chromium-mac-leopard/transforms/3d/point-mapping/3d-point-mapping-preserve-3d-expected.png:
  • platform/chromium-mac-snowleopard/transforms/3d/general/perspective-units-expected.png:
  • platform/chromium-mac-snowleopard/transforms/3d/hit-testing/backface-hit-test-expected.png: Added.
  • platform/chromium-mac-snowleopard/transforms/3d/hit-testing/backface-no-transform-hit-test-expected.png: Added.
  • platform/chromium-mac-snowleopard/transforms/3d/point-mapping/3d-point-mapping-3-expected.png: Added.
  • platform/chromium-mac-snowleopard/transforms/3d/point-mapping/3d-point-mapping-expected.png:
  • platform/chromium-mac-snowleopard/transforms/3d/point-mapping/3d-point-mapping-origins-expected.png:
  • platform/chromium-mac-snowleopard/transforms/3d/point-mapping/3d-point-mapping-overlapping-expected.png: Added.
  • platform/chromium-mac-snowleopard/transforms/3d/point-mapping/3d-point-mapping-preserve-3d-expected.png:
  • platform/chromium-mac/transforms/3d/general/matrix-with-zoom-3d-expected.png: Added.
  • platform/chromium-mac/transforms/3d/general/perspective-units-expected.png: Added.
  • platform/chromium-mac/transforms/3d/hit-testing/backface-hit-test-expected.png: Added.
  • platform/chromium-mac/transforms/3d/hit-testing/backface-no-transform-hit-test-expected.png: Added.
  • platform/chromium-mac/transforms/3d/point-mapping/3d-point-mapping-3-expected.png: Added.
  • platform/chromium-mac/transforms/3d/point-mapping/3d-point-mapping-expected.png: Added.
  • platform/chromium-mac/transforms/3d/point-mapping/3d-point-mapping-origins-expected.png: Added.
  • platform/chromium-mac/transforms/3d/point-mapping/3d-point-mapping-overlapping-expected.png: Added.
  • platform/chromium-mac/transforms/3d/point-mapping/3d-point-mapping-preserve-3d-expected.png: Added.
  • platform/chromium-win/transforms/3d/general/matrix-with-zoom-3d-expected.png: Added.
  • platform/chromium-win/transforms/3d/general/perspective-units-expected.png: Added.
  • platform/chromium-win/transforms/3d/general/perspective-units-expected.txt: Added.
  • platform/chromium-win/transforms/3d/hit-testing/backface-hit-test-expected.png: Added.
  • platform/chromium-win/transforms/3d/hit-testing/backface-hit-test-expected.txt: Added.
  • platform/chromium-win/transforms/3d/hit-testing/backface-no-transform-hit-test-expected.png: Added.
  • platform/chromium-win/transforms/3d/hit-testing/backface-no-transform-hit-test-expected.txt: Added.
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-3-expected.png: Added.
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-3-expected.txt: Added.
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-expected.png:
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-expected.txt:
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-origins-expected.png:
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-origins-expected.txt:
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-overlapping-expected.png: Added.
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-overlapping-expected.txt: Added.
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-preserve-3d-expected.png:
  • platform/chromium-win/transforms/3d/point-mapping/3d-point-mapping-preserve-3d-expected.txt:
  • platform/chromium/TestExpectations:
9:19 PM TestExpectations edited by rniwa@webkit.org
Update per TestExpectations rename (diff)
9:18 PM NewRunWebKitTests edited by rniwa@webkit.org
Update per TestExpectations rename (diff)
9:18 PM Rebaseline edited by rniwa@webkit.org
Update per TestExpectations rename (diff)
9:16 PM April 2012 Keeping the bots green edited by rniwa@webkit.org
Update per TestExpectations rename (diff)
9:14 PM WikiStart edited by rniwa@webkit.org
Rename test_expectations.txt to TestExpectations (diff)
8:14 PM Changeset in webkit [119368] by haraken@chromium.org
  • 3 edits
    1 copy in trunk/PerformanceTests

[perf-test] Add a benchmark of querySelector() for an element that appears in the depths in the document
​https://bugs.webkit.org/show_bug.cgi?id=88202

Reviewed by Ryosuke Niwa.

query-selector-first.html tests querySelector() for an element
that appears at the head of the document. query-selector-last.html
tests querySelector() for an element that appears at the tail of
the document. In addition, we want a benchmark of querySelector()
for an element that appears in the depths in the document.

Test results in my Linux desktop:

RESULT Parser: query-selector-deep= 460.776980611 runs/s
median= 460.947503201 runs/s, stdev= 1.72274941942 runs/s, min= 455.696202532 runs/s, max= 463.32046332 runs/s

RESULT Parser: query-selector-deep= 458.925050915 runs/s
median= 459.183673469 runs/s, stdev= 1.47930124139 runs/s, min= 454.545454545 runs/s, max= 460.947503201 runs/s

RESULT Parser: query-selector-deep= 461.866981491 runs/s
median= 461.538461538 runs/s, stdev= 1.14613123359 runs/s, min= 459.183673469 runs/s, max= 463.917525773 runs/s

RESULT Parser: query-selector-deep= 453.922010788 runs/s
median= 454.258855898 runs/s, stdev= 1.77279123317 runs/s, min= 447.204968944 runs/s, max= 455.696202532 runs/s

RESULT Parser: query-selector-deep= 454.666321221 runs/s
median= 454.545454545 runs/s, stdev= 1.65775667417 runs/s, min= 451.127819549 runs/s, max= 456.852791878 runs/s

  • Parser/query-selector-deep.html: Added.
8:06 PM Changeset in webkit [119367] by dominicc@chromium.org
  • 15 edits
    1 move
    1 delete in trunk/Source/WebCore

EventTarget.idl should generate JS bindings
​https://bugs.webkit.org/show_bug.cgi?id=88120

This is one step in a larger change to expose a constructor for
EventTarget and put EventTarget on the prototype chain of DOM
objects that are event targets, in line with DOM Core. See
<​https://bugs.webkit.org/show_bug.cgi?id=67312>

Reviewed by Adam Barth.

No new tests. This change does not change any functionality.

  • CMakeLists.txt: Generate JSEventTarget.
  • DerivedSources.cpp:
  • DerivedSources.make:
  • DerivedSources.pri:
  • GNUmakefile.list.am:
  • Target.pri:
  • UseJSC.cmake:
  • WebCore.gyp/WebCore.gyp:
  • WebCore.gypi:
  • WebCore.vcproj/WebCore.vcproj:
  • WebCore.xcodeproj/project.pbxproj:
  • bindings/js/JSBindingsAllInOne.cpp: JSEventTarget.cpp was renamed.
  • bindings/js/JSEventTarget.h: Removed. Generated now.
  • bindings/js/JSEventTargetCustom.cpp: Renamed from Source/WebCore/bindings/js/JSEventTarget.cpp.
  • bindings/scripts/CodeGeneratorV8.pm:

(GetInternalFields): EventTarget is an EventTarget.

  • dom/EventTarget.idl: Defer to toJS, toEventTarget in JSEventTargetCustom.cpp.
7:24 PM Changeset in webkit [119366] by haraken@chromium.org
  • 2 edits in trunk/PerformanceTests

[perf-test] Make query-selector-last.html more realistic
​https://bugs.webkit.org/show_bug.cgi?id=88203

Reviewed by Ryosuke Niwa.

query-selector-last.html tests querySelector() for an element
that appears after 1000 div elements, all of which have the same
id and class. To test the performance of querySelector() under
a lot of ids and classes, we should make the ids and classes of
the 1000 elements different.

This patch degrades the performance of query-selector-last.html
by 3.6%.

[Before]
RESULT Parser: query-selector-last= 476.382274152 runs/s
median= 476.821192053 runs/s, stdev= 1.14578875814 runs/s, min= 474.308300395 runs/s, max= 478.087649402 runs/s

RESULT Parser: query-selector-last= 478.423061861 runs/s
median= 478.723404255 runs/s, stdev= 1.54876867255 runs/s, min= 473.684210526 runs/s, max= 481.347773767 runs/s

RESULT Parser: query-selector-last= 479.849287174 runs/s
median= 480.0 runs/s, stdev= 1.69849752498 runs/s, min= 475.561426684 runs/s, max= 481.927710843 runs/s

RESULT Parser: query-selector-last= 476.765438846 runs/s
median= 477.453580902 runs/s, stdev= 1.8724488664 runs/s, min= 471.821756225 runs/s, max= 478.723404255 runs/s

RESULT Parser: query-selector-last= 481.715340644 runs/s
median= 481.637742305 runs/s, stdev= 1.81446361145 runs/s, min= 475.561426684 runs/s, max= 483.675937122 runs/s

[After]
RESULT Parser: query-selector-last= 458.866623582 runs/s
median= 459.183673469 runs/s, stdev= 1.48881464737 runs/s, min= 453.972257251 runs/s, max= 460.947503201 runs/s

RESULT Parser: query-selector-last= 460.363532897 runs/s
median= 460.358056266 runs/s, stdev= 1.58512056995 runs/s, min= 456.273764259 runs/s, max= 462.130937099 runs/s

RESULT Parser: query-selector-last= 457.727448451 runs/s
median= 458.015267176 runs/s, stdev= 1.23681915516 runs/s, min= 453.972257251 runs/s, max= 459.183673469 runs/s

RESULT Parser: query-selector-last= 465.57336853 runs/s
median= 465.717981889 runs/s, stdev= 1.6616925374 runs/s, min= 460.947503201 runs/s, max= 467.532467532 runs/s

RESULT Parser: query-selector-last= 459.836252019 runs/s
median= 459.770114943 runs/s, stdev= 1.84485630505 runs/s, min= 455.696202532 runs/s, max= 462.724935733 runs/s

  • Parser/query-selector-last.html:
5:35 PM WebKit Team edited by Joone Hur
(diff)
3:35 PM Changeset in webkit [119365] by hayato@chromium.org
  • 5 edits in trunk

Supports traversing nodes in an orphaned shadow subtree in ComposedShadowTreeParentWalker.
​https://bugs.webkit.org/show_bug.cgi?id=87924

Reviewed by Dimitri Glazkov.

Source/WebCore:

Some events, such as a 'click' event, can be dispatched on nodes in an orphaned shadow subtree.
When resolving a parent node of a shadow root which is not assigned to any shadow insertion point,
we have to return its shadow host as a parent node.

The related shadow DOM spec section is:
​https://dvcs.w3.org/hg/webcomponents/raw-file/tip/spec/shadow/index.html#events

Test: fast/dom/shadow/shadow-dom-event-dispatching.html

  • dom/ComposedShadowTreeWalker.cpp:

(WebCore::ComposedShadowTreeParentWalker::traverseParentIncludingInsertionPointAndShadowRoot):

LayoutTests:

  • fast/dom/shadow/shadow-dom-event-dispatching-expected.txt:
  • fast/dom/shadow/shadow-dom-event-dispatching.html:
2:16 PM Changeset in webkit [119364] by ggaren@apple.com
  • 9 edits in trunk/Source/JavaScriptCore

Weak pointer finalization should be lazy
​https://bugs.webkit.org/show_bug.cgi?id=87599

Reviewed by Sam Weinig.

This time for sure!

  • heap/Heap.cpp:

(JSC::Heap::collect): Don't sweep eagerly -- we'll sweep lazily instead.

  • heap/MarkedBlock.cpp:

(JSC::MarkedBlock::sweep): Sweep our weak set before we sweep our other
destructors -- this is our last chance to run weak set finalizers before
we recycle our memory.

  • heap/MarkedBlock.h:

(JSC::MarkedBlock::resetAllocator):

  • heap/MarkedSpace.cpp:

(JSC::MarkedSpace::resetAllocators):

  • heap/MarkedSpace.h:

(JSC::MarkedSpace::resetAllocators): Don't force allocator reset anymore.
It will happen automatically when a weak set is swept. It's simpler to
have only one canonical way for this to happen, and it wasn't buying
us anything to do it eagerly.

  • heap/WeakBlock.cpp:

(JSC::WeakBlock::sweep): Don't short-circuit a sweep unless we know
the sweep would be a no-op. If even one finalizer is pending, we need to
run it, since we won't get another chance.

  • heap/WeakSet.cpp:

(JSC::WeakSet::sweep): This loop can be simpler now that
WeakBlock::sweep() does what we mean.

Reset our allocator after a sweep because this is the optimal time to
start trying to recycle old weak pointers.

(JSC::WeakSet::tryFindAllocator): Don't sweep when searching for an
allocator because we've swept already, and forcing a new sweep would be
wasteful.

  • heap/WeakSet.h:

(JSC::WeakSet::shrink): Be sure to reset our allocator after a shrink
because the shrink may have removed the block the allocator was going to
allocate out of.

12:13 PM Changeset in webkit [119363] by commit-queue@webkit.org
  • 2 edits in trunk/LayoutTests

[EFL] http/tests/media/video-buffered.html is failing after r119268 and need to be skipped
​https://bugs.webkit.org/show_bug.cgi?id=88194

Unreviewed EFL gardening. Skipping http/tests/media/video-buffered.html
which is failing after r119268.

Patch by Christophe Dumez <Christophe Dumez> on 2012-06-03

  • platform/efl/TestExpectations:
11:49 AM Changeset in webkit [119362] by commit-queue@webkit.org
  • 3 edits in trunk/LayoutTests

[EFL] legacy-viewport tests added in r119256 need to be skipped
​https://bugs.webkit.org/show_bug.cgi?id=88192

Unreviewed EFL gardening. Skip new tests added in r119256 since the
LEGACY_VIEWPORT_ADAPTION flag is disabled on EFL port.

Patch by Christophe Dumez <Christophe Dumez> on 2012-06-03

  • platform/efl/Skipped:
  • platform/efl/TestExpectations:
11:30 AM Changeset in webkit [119361] by commit-queue@webkit.org
  • 1 edit
    4 adds in trunk/LayoutTests

[EFL] New tests need a baseline
​https://bugs.webkit.org/show_bug.cgi?id=88190

Unreviewed EFL gardening. Add EFL baseline for new test cases.

Patch by Christophe Dumez <Christophe Dumez> on 2012-06-03

  • platform/efl/fast/reflections/reflection-with-zoom-expected.png: Added.
  • platform/efl/fast/reflections/reflection-with-zoom-expected.txt: Added.
  • platform/efl/svg/repaint/text-mask-update-expected.png: Added.
  • platform/efl/svg/repaint/text-mask-update-expected.txt: Added.
8:51 AM Changeset in webkit [119360] by commit-queue@webkit.org
  • 10 edits in trunk

Unreviewed, rolling out r119359.
​http://trac.webkit.org/changeset/119359
​https://bugs.webkit.org/show_bug.cgi?id=88189

This broke platform/mac/fast/dom/wrapper-classes-objc.html.
(Requested by dominicc on #webkit).

Patch by Sheriff Bot <​webkit.review.bot@gmail.com> on 2012-06-03

Source/WebCore:

  • bindings/objc/PublicDOMInterfaces.h:
  • bindings/scripts/CodeGeneratorObjC.pm:

(GetBaseClass):

  • dom/MouseEvent.h:

(MouseEvent):

  • dom/WheelEvent.cpp:

(WebCore::WheelEvent::WheelEvent):

  • dom/WheelEvent.h:

(WheelEvent):

  • dom/WheelEvent.idl:

LayoutTests:

  • fast/events/event-creation-expected.txt:
  • fast/events/event-creation.html:
7:33 AM Changeset in webkit [119359] by dominicc@chromium.org
  • 10 edits in trunk

WheelEvent should inherit from MouseEvent
​https://bugs.webkit.org/show_bug.cgi?id=76104

Reviewed by Kentaro Hara.

Source/WebCore:

The spec for WheelEvent is
<​http://www.w3.org/TR/DOM-Level-3-Events/#webidl-events-WheelEvent>

Tests: fast/events/event-creation.html

​http://samples.msdn.microsoft.com/ietestcenter/dominheritance/showdominheritancetest.htm?Prototype_WheelEvent

  • bindings/objc/PublicDOMInterfaces.h: Remove redundant MouseEvent API from WheelEvent
  • bindings/scripts/CodeGeneratorObjC.pm: MouseEvents are Events
  • dom/MouseEvent.h: Expose no-arg constructor to WheelEvent
  • dom/WheelEvent.cpp:

(WebCore::WheelEvent::WheelEvent): Call MouseEvent superconstructor

  • dom/WheelEvent.h: Extend MouseEvent
  • dom/WheelEvent.idl: "

LayoutTests:

  • fast/events/event-creation-expected.txt:
  • fast/events/event-creation.html: Also check WheelEvent instanceof MouseEvent as well as UIEvent, Event
2:35 AM Changeset in webkit [119358] by gyuyoung.kim@samsung.com
  • 4 edits
    1 copy in trunk/Source/WebKit/efl

Add ewk_custom_handler_private.h for ewk_custom_handler.cpp
​https://bugs.webkit.org/show_bug.cgi?id=88163

Reviewed by Ryosuke Niwa.

EFL WK1 decided to divide up ewk_private.h into each file's XXX_private.h file
in order to maintain internal functions more easily. Bug 73176 is adding new
internal functions to ewk_custom_hanlder.cpp. So, ewk_custom_handler.cpp needs
to have _private.h file as well.

  • WebCoreSupport/ChromeClientEfl.cpp:
  • ewk/ewk_custom_handler.cpp:
  • ewk/ewk_custom_handler_private.h: Added.
  • ewk/ewk_view_private.h:
1:21 AM Changeset in webkit [119357] by rniwa@webkit.org
  • 2 edits in trunk/Source/WebKit/chromium

Roll Chromium DEPS from r140000 to r140222.

  • DEPS:
12:26 AM Changeset in webkit [119356] by commit-queue@webkit.org
  • 12 edits
    2 adds in trunk

[chromium] Add new gesture type (two finger tap) that triggers context menu.
​https://bugs.webkit.org/show_bug.cgi?id=88173

Patch by Varun Jain <​varunjain@google.com> on 2012-06-03
Reviewed by Adam Barth.

Source/WebCore:

Test: fast/events/touch/gesture/context-menu-on-two-finger-tap.html

  • platform/PlatformEvent.h:

Source/WebKit/chromium:

  • public/WebInputEvent.h:
  • src/PageWidgetDelegate.cpp:

(WebKit::PageWidgetDelegate::handleInputEvent):

  • src/WebInputEventConversion.cpp:

(WebKit::PlatformGestureEventBuilder::PlatformGestureEventBuilder):

  • src/WebPopupMenuImpl.cpp:

(WebKit::WebPopupMenuImpl::handleInputEvent):

  • src/WebViewImpl.cpp:

(WebKit::WebViewImpl::handleGestureEvent):

Tools:

  • DumpRenderTree/chromium/EventSender.cpp:

(EventSender::EventSender):
(EventSender::gestureTwoFingerTap):
(EventSender::gestureEvent):

  • DumpRenderTree/chromium/EventSender.h:

(EventSender):

LayoutTests:

  • fast/events/touch/gesture/context-menu-on-two-finger-tap.html: Added.
  • platform/chromium/fast/events/touch/gesture/context-menu-on-two-finger-tap-expected.txt: Added.
12:06 AM Changeset in webkit [119355] by rjkroege@chromium.org
  • 2 edits in trunk/Source/WebKit/chromium

[chromium] replace isScrollGestureEventType with isGestureEventType
​https://bugs.webkit.org/show_bug.cgi?id=88097

This change replaces the unused and incorrect isScrollGestureEventType
method with a correct and useful isGestureEventType method.

Reviewed by Adam Barth.

  • public/WebInputEvent.h:

(WebInputEvent):
(WebKit::WebInputEvent::isGestureEventType):

Jun 2, 2012:

11:29 PM Changeset in webkit [119354] by aestes@apple.com
  • 2 edits in trunk/Source/WebCore

ZNK7WebCore4Node27traverseNextAncestorSiblingEv should be exported even if ENABLE(FULLSCREEN_API) is false
​https://bugs.webkit.org/show_bug.cgi?id=88184

Reviewed by Dan Bernstein.

Node::traverseNextAncestorSibling() is called in several inline
functions defined in ContainerNode.h (a private header), so its symbol
needs to be exported regardless of whether full-screen API is enabled.

  • WebCore.exp.in:
10:06 PM Changeset in webkit [119353] by wangxianzhu@chromium.org
  • 15 edits
    2 adds in trunk

SVGImageCache leaks image data
​https://bugs.webkit.org/show_bug.cgi?id=87792

Source/WebCore:

There are two functions to remove a client from a CachedImage:

  • CachedResource::removeClient()
  • CachedImage::removeClientForRenderer().

It's easy to make error to call the former which will leak the cached
image buffers in SVGImageCache.

This change combined the two by adding the virtual
CachedResource::didRemoveClient(). CachedImage will do SVGImageCache
cleanup in the function.

Reviewed by Nikolas Zimmermann.

Test: svg/as-image/svg-image-leak-cached-data.html

  • loader/cache/CachedFont.h:

(WebCore::CachedFontClient::resourceClientType): Added 'const'.

  • loader/cache/CachedImage.cpp:

(WebCore):
(WebCore::CachedImage::didRemoveClient): Removes the client from SVGImageCache.
(WebCore::CachedImage::lookupOrCreateImageForRenderer):

  • loader/cache/CachedImage.h:

(CachedImage):
(WebCore::CachedImageClient::resourceClientType): Added 'const'.

  • loader/cache/CachedRawResource.h:

(WebCore::CachedRawResourceClient::resourceClientType): Added 'const'.

  • loader/cache/CachedResource.cpp:

(WebCore::CachedResource::removeClient): Added invocation of didRemoveClient().

  • loader/cache/CachedResource.h:

(WebCore::CachedResource::didRemoveClient): Added for subclasses to do additional works.

  • loader/cache/CachedResourceClient.h:

(WebCore::CachedResourceClient::resourceClientType): Added 'const'.

  • loader/cache/CachedSVGDocument.h:

(WebCore::CachedSVGDocumentClient::resourceClientType): Added 'const'.

  • loader/cache/CachedStyleSheetClient.h:

(WebCore::CachedStyleSheetClient::resourceClientType): Added 'const'.

  • rendering/style/StyleCachedImage.cpp:

(WebCore::StyleCachedImage::removeClient):

  • rendering/style/StyleCachedImageSet.cpp:

(WebCore::StyleCachedImageSet::removeClient):

  • svg/graphics/SVGImageCache.cpp:

(WebCore::SVGImageCache::~SVGImageCache): Added checking for leaks.
(WebCore::SVGImageCache::removeClientFromCache):
(WebCore::SVGImageCache::setRequestedSizeAndScales):
(WebCore::SVGImageCache::requestedSizeAndScales):
(WebCore::SVGImageCache::lookupOrCreateBitmapImageForClient):

  • svg/graphics/SVGImageCache.h:

(WebCore):
(SVGImageCache):

LayoutTests:

Reviewed by Nikolas Zimmermann.

New test case.

  • svg/as-image/svg-image-leak-cached-data-expected.txt: Added.
  • svg/as-image/svg-image-leak-cached-data.html: Added.
9:10 PM Changeset in webkit [119352] by rniwa@webkit.org
  • 2 edits in trunk/LayoutTests

Remove the failing expectation on fast/forms/textarea-scrollbar-height.html since it has been passing.

  • platform/chromium/TestExpectations:
8:45 PM Changeset in webkit [119351] by ggaren@apple.com
  • 3 edits in trunk/Source/WebKit2

2012-06-02 Geoffrey Garen <​ggaren@apple.com>

Try to fix the Windows build.

  • win/WebKit2.def:
  • win/WebKit2CFLite.def:
8:35 PM Changeset in webkit [119350] by rniwa@webkit.org
  • 1 edit
    3 adds in trunk/LayoutTests

Add back files inadvertently removed in r119323.

7:46 PM Changeset in webkit [119349] by aestes@apple.com
  • 20 edits in trunk/Source/WebKit2

[WebKit2] Use USE(), ENABLE(), and HAVE() macros in more places
​https://bugs.webkit.org/show_bug.cgi?id=88182

Reviewed by Sam Weinig.

Use USE(APPKIT):

  • Shared/mac/NativeWebKeyboardEventMac.mm:
  • Shared/mac/NativeWebMouseEventMac.mm:
  • Shared/mac/NativeWebWheelEventMac.mm:
  • Shared/mac/WebEventFactory.h:
  • UIProcess/API/mac/WKTextInputWindowController.h:
  • UIProcess/API/mac/WKTextInputWindowController.mm:
  • UIProcess/PageClient.h:
  • UIProcess/WebPageProxy.h:

Use ENABLE(NETSCAPE_PLUGIN_API):

  • UIProcess/API/C/WKPluginSiteDataManager.cpp:

(WKPluginSiteDataManagerClearSiteData):
(WKPluginSiteDataManagerClearAllSiteData):

Use USE(AUTOCORRECTION_PANEL):

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

(WebKit::PageClientImpl::showCorrectionPanel):
(WebKit::PageClientImpl::dismissCorrectionPanel):
(WebKit::PageClientImpl::dismissCorrectionPanelSoon):

  • UIProcess/mac/CorrectionPanel.mm:

Use HAVE(HOSTED_CORE_ANIMATION):

  • UIProcess/mac/WebContextMac.mm:

Put ENABLE(FULLSCREEN_API) in the right place:

  • UIProcess/mac/WebFullScreenManagerProxyMac.mm:

Remove unnecessary use of USE(APPKIT):

  • UIProcess/mac/WebPageProxyMac.mm:

Use USE(SECURITY_FRAMEWORK):

  • WebProcess/Authentication/mac/AuthenticationManager.mac.mm:

Use ENABLE(CONTEXT_MENUS):

  • WebProcess/WebCoreSupport/mac/WebContextMenuClientMac.mm:

Use USE(SECURITY_FRAMEWORK):

  • WebProcess/WebProcess.h:

Remove unnecessary use of USE(SECURITY_FRAMEWORK):

  • WebProcess/mac/WebProcessMac.mm:
7:38 PM Changeset in webkit [119348] by rniwa@webkit.org
  • 3 edits in trunk/Source/WebCore

Windows build fix after r119330. Unnest the class for now.

  • dom/SelectorQuery.cpp:

(WebCore::SelectorQueryCacheEntry::SelectorQueryCacheEntry):
(WebCore::SelectorQueryCache::add):

  • dom/SelectorQuery.h:

(SelectorQueryCacheEntry):
(WebCore::SelectorQueryCacheEntry::selectorQuery):
(WebCore):
(SelectorQueryCache):

7:33 PM Changeset in webkit [119347] by aestes@apple.com
  • 2 edits in trunk/Source/WebKit2

Remove a duplicate INIT() from WebSystemInterface.mm
​https://bugs.webkit.org/show_bug.cgi?id=88181

Reviewed by Anders Carlsson.

  • WebProcess/WebCoreSupport/mac/WebSystemInterface.mm:

(InitWebCoreSystemInterface):

7:23 PM Changeset in webkit [119346] by aestes@apple.com
  • 2 edits in trunk/Source/WebKit2

Remove unnecessary import of <Cocoa/Cocoa.h>
​https://bugs.webkit.org/show_bug.cgi?id=88180

Reviewed by Sam Weinig.

  • UIProcess/mac/WKFullKeyboardAccessWatcher.h:
5:53 PM Changeset in webkit [119345] by aestes@apple.com
  • 2 edits in trunk/Source/WebKit2

Ensure cpu_type_t is defined by including <mach/machine.h>
​https://bugs.webkit.org/show_bug.cgi?id=88178

Reviewed by Anders Carlsson.

  • Shared/Plugins/PluginModuleInfo.h:
5:47 PM Changeset in webkit [119344] by weinig@apple.com
  • 2 edits in trunk/Source/WebKit2

InjectedBundles' sandbox extensions don't do what we think they do
​https://bugs.webkit.org/show_bug.cgi?id=88177

Reviewed by Anders Carlsson.

This cannot be tested with our current infrastructure (or any modifications of the
infrastructure I can think of) since it relies on the bundle being put outside
the build directory where WebKit is put.

  • WebProcess/InjectedBundle/mac/InjectedBundleMac.cpp:

(WebKit::InjectedBundle::load):
Use SandboxExtension::consumePermanently() rather than consume(), otherwise,
nulling out the SandboxExtension a few lines later invalidates the extension
we just consumed.

5:41 PM Changeset in webkit [119343] by fpizlo@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

If the DFG bytecode parser detects that op_method_check has gone polymorphic, it
shouldn't revert all the way to GetById/GetByIdFlush
​https://bugs.webkit.org/show_bug.cgi?id=88176

Reviewed by Geoffrey Garen.

Refactored the code so that the op_method_check case of the parser gracefully falls
through to all of the goodness of the normal op_get_by_id case.

  • dfg/DFGByteCodeParser.cpp:

(ByteCodeParser):
(JSC::DFG::ByteCodeParser::handleGetById):
(DFG):
(JSC::DFG::ByteCodeParser::parseBlock):

3:58 PM Changeset in webkit [119342] by fpizlo@apple.com
  • 18 edits in trunk/Source/JavaScriptCore

DFG CSE should be able to eliminate unnecessary flushes of arguments and captured variables
​https://bugs.webkit.org/show_bug.cgi?id=87929

Reviewed by Geoffrey Garen.

Slight speed-up on V8. Big win (up to 50%) on programs that inline very small functions.

This required a bunch of changes:

  • The obvious change is making CSE essentially ignore whether or not the set of operations between the Flush and the SetLocal can exit, and instead focus on whether or not that set of operations can clobber the world or access local variables. This code is now refactored to return a set of flags indicating any of these events, and the CSE decides what to do based on those flags. If the set of operations is non-clobbering and non-accessing, then the Flush is turned into a Phantom on the child of the SetLocal. This expands the liveness of the relevant variable but virtually guarantees that it will be register allocated and not flushed to the stack. So, yeah, this patch is a lot of work to save a few stores to the stack.


  • Previously, CheckArgumentsNotCreated was optimized "lazily" in that you only knew if it was a no-op if you were holding onto a CFA abstract state. But this would make the CSE act pessimistically, since it doesn't use the CFA. Hence, this patch changes the constant folding phase into something more broad; it now fixes up CheckArgumentsNotCreated nodes by turning them into phantoms if it knows that they are no-ops.


  • Arguments simplification was previously relying on this very strange PhantomArguments node, which had two different meanings: for normal execution it meant the empty value but for OSR exit it meant that the arguments should be reified. This produces problems when set SetLocals to the captured arguments registers are CSE'd away, since we'd be triggering reification of arguments without having initialized the arguments registers to empty. The cleanest solution was to fix PhantomArguments to have one meaning: namely, arguments reification on OSR exit. Hence, this patch changes arguments simplification to change SetLocal of CreateArguments on the arguments registers to be a SetLocal of Empty.


  • Argument value recoveries were previously derived from the value source of the arguments at the InlineStart. But that relies on all SetLocals to arguments having been flushed. It's possible that we could have elided the SetLocal to the arguments at the callsite because there were subsequent SetLocals to the arguments inside of the callee, in which case the InlineStart would get the wrong information. Hence, this patch changes argument value recovery computation to operate over the ArgumentPositions directly.


  • But that doesn't actually work, because previously, there was no way to link an InlineStart back to the corresponding ArgumentPositions, at least not without some ugliness. So this patch instates the rule that the m_argumentPositions vector consists of disjoint subsequences such that each subsequence corresponds to an inline callsite and can be identified by its first index, and within each subsequence are the ArgumentPositions of all of the arguments ordered by argument index. This required flipping the order in which ArgumentPositions are added to the vector, and giving InlineStart an operand that indicates the start of that inline callsite's ArgumentPosition subsequence.


  • This patch also revealed a nasty bug in the reification of arguments in inline call frames on OSR exit. Since the reification was happening after the values of virtual registers were recovered, the value recoveries of the inline arguments were wrong. Hence using operationCreateInlinedArguments is wrong. For example a value recovery might say that you have to box a double, but if we had already boxed it then boxing it a second time will result in garbage. The specific case of this bug was this patch uncovered was that now it is possible for an inline call frame to not have any valid value recoveries for any inline arguments, if the optimization elides all argument flushes, while at the same time optimizing away arguments creation. Then OSR exit would try to recover the arguments using the inline call frame, which had bogus information, and humorous crashes would ensue. This patch fixes this issue by moving arguments reification to after call frame reification, so that arguments reification can always use operationCreateArguments instead of operationCreateInlinedArguments.


  • This patch may turn a Flush into a Phantom. That's kind of the whole point. But that broke forward speculation checks, which knew to look for a Flush prior to a SetLocal but didn't know that there could alternatively be a Phantom in place of the Flush. This patch fixes that by augmenting the forward speculation check logic.


  • Finally, in the process of having fun with all of the above, I realized that my DFG validation was not actually running on every phase like I had originally designed it to. In fact it was only running just after bytecode parsing. I initially tried to make it run in every phase but found that this causes some tests to timeout (specifically the evil fuzzing ones), so I decided on a compromise where: (i) in release mode validation never runs, (ii) in debug mode validation will run just after parsing and just before the backend, and (iii) it's possible with a simple switch to enable validation to run on every phase.


Luckily all of the above issues were already covered by the 77 or so DFG-specific
layout tests. Hence, this patch does not introduce any new tests despite being so
meaty.

  • dfg/DFGAbstractState.cpp:

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

  • dfg/DFGArgumentPosition.h:

(JSC::DFG::ArgumentPosition::prediction):
(JSC::DFG::ArgumentPosition::doubleFormatState):
(JSC::DFG::ArgumentPosition::shouldUseDoubleFormat):
(ArgumentPosition):

  • dfg/DFGArgumentsSimplificationPhase.cpp:

(JSC::DFG::ArgumentsSimplificationPhase::run):

  • dfg/DFGByteCodeParser.cpp:

(JSC::DFG::ByteCodeParser::handleInlining):
(JSC::DFG::ByteCodeParser::InlineStackEntry::InlineStackEntry):

  • dfg/DFGCSEPhase.cpp:

(JSC::DFG::CSEPhase::SetLocalStoreEliminationResult::SetLocalStoreEliminationResult):
(SetLocalStoreEliminationResult):
(JSC::DFG::CSEPhase::setLocalStoreElimination):
(JSC::DFG::CSEPhase::performNodeCSE):

  • dfg/DFGCommon.h:
  • dfg/DFGConstantFoldingPhase.cpp:

(JSC::DFG::ConstantFoldingPhase::run):

  • dfg/DFGDriver.cpp:

(JSC::DFG::compile):

  • dfg/DFGNode.h:

(Node):
(JSC::DFG::Node::hasArgumentPositionStart):
(JSC::DFG::Node::argumentPositionStart):

  • dfg/DFGOSRExitCompiler32_64.cpp:

(JSC::DFG::OSRExitCompiler::compileExit):

  • dfg/DFGOSRExitCompiler64.cpp:

(JSC::DFG::OSRExitCompiler::compileExit):

  • dfg/DFGPhase.cpp:

(DFG):

  • dfg/DFGPhase.h:

(Phase):

  • dfg/DFGSpeculativeJIT.cpp:

(JSC::DFG::SpeculativeJIT::compile):

  • dfg/DFGSpeculativeJIT.h:

(JSC::DFG::SpeculativeJIT::forwardSpeculationCheck):

  • dfg/DFGSpeculativeJIT32_64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

  • dfg/DFGSpeculativeJIT64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

3:49 PM Changeset in webkit [119341] by ggaren@apple.com
  • 8 edits in trunk/Source

DOM string cache should hash pointers, not characters
​https://bugs.webkit.org/show_bug.cgi?id=88175

Reviewed by Phil Pizlo and Sam Weinig.

../JavaScriptCore:

  • heap/Weak.h:

(JSC::weakAdd):
(JSC::weakRemove): Made these function templates slightly more generic
to accommodate new client types.

../WebCore:

Dromaeo DOM Core reports no change.

​http://trac.webkit.org/changeset/84934 accidentally changed from hashing
pointers to hashing characters, due to template defaults. Let's change back.

Hashing characters is not so good because:

(1) It's not memory-safe with HashMap::set(). HashMap::set() replaces
the value but not the key. Since our values own our keys, we need to
ensure object identity between key and value, or the key can be freed
prematurely. (This is impossible to demonstrate with our current
eager sweep behavior, but it shows up as crashes in layout tests if you
change to lazy sweep.)

(2) It's slower.

  • bindings/js/DOMWrapperWorld.h:

(WebCore): Override the default hash, which hashes based on characters.

2:52 PM Changeset in webkit [119340] by efidler@rim.com
  • 2 edits in trunk/Source/WebCore

Don't crash if we ask for fonts that don't exist.
​https://bugs.webkit.org/show_bug.cgi?id=88106

Reviewed by Dan Bernstein.

RIM PR 161219

If the site doesn't ask for a reasonable font and the system doesn't
provide a good fallback, FontFallBackList::primaryFontData can be 0,
which can cause a crash.

  • platform/graphics/skia/FontCacheSkia.cpp:

(WebCore::FontCache::getLastResortFallbackFont):

2:30 PM Changeset in webkit [119339] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/mac

loadRequest should check for empty URLs.
​https://bugs.webkit.org/show_bug.cgi?id=88154

In [WebFrame loadRequest], if the url is invalid and not nil, we transform it to file: url.
However, WebKit client could send down an empty URL that makes us transform it to file: url
as well. We should skip that as we did for nil URLs.

Patch by Yongjun Zhang <​yongjun_zhang@apple.com> on 2012-06-02
Reviewed by Brady Eidson.

  • WebView/WebFrame.mm:

(-[WebFrame loadRequest:]):

1:34 PM WebKitGTK/1.8.x edited by kov@webkit.org
(diff)
1:32 PM Changeset in webkit [119338] by kov@webkit.org
  • 5 edits in releases/WebKitGTK/webkit-1.8

Merge 115763 - Ensure HTMLElementStack fails gracefully if it has a non-Element.
​https://bugs.webkit.org/show_bug.cgi?id=85167

Reviewed by Adam Barth.

Source/WebCore:

Test: Added to html5lib/resources/webkit02.dat

  • html/parser/HTMLElementStack.cpp:

(WebCore::HTMLElementStack::oneBelowTop):

  • html/parser/HTMLTreeBuilder.cpp:

(WebCore::HTMLTreeBuilder::processEndTag):

LayoutTests:

  • html5lib/resources/webkit02.dat:
1:31 PM Changeset in webkit [119337] by kov@webkit.org
  • 4 edits in releases/WebKitGTK/webkit-1.8

Merge 108311 - Invalid cast in WebCore::toElement / WebCore::HTMLElementStack::ElementRecord::element
​https://bugs.webkit.org/show_bug.cgi?id=78975

Reviewed by Eric Seidel.

Source/WebCore:

We're supposed to set the action attribute on the form element we just
created. Previously, we assumed the newly created form element would
be on the top of the stack of open elements, but if we're in the table
body insertion mode, the form element gets treated as self closing and
is therefore popped off the stack of open elements.

Fortunately, we already cache a pointer to the most recently inserted
form element on the HTMLConstructionSite, so we can just grab the
element from there.

Test: html5lib/runner.html

  • html/parser/HTMLTreeBuilder.cpp:

(WebCore::HTMLTreeBuilder::processIsindexStartTagForInBody):
(WebCore):

LayoutTests:

  • html5lib/resourcesl/webkit-02.dat:
1:31 PM Changeset in webkit [119336] by kov@webkit.org
  • 1 add in releases/WebKitGTK/webkit-1.8/LayoutTests/platform/gtk/fast/dom/htmlcollection-non-html-expected.txt

Rebseline for 115398.

1:31 PM Changeset in webkit [119335] by kov@webkit.org
  • 3 edits
    2 adds in releases/WebKitGTK/webkit-1.8

Merge 115646 - loadOrRedirectSubframe should return the owner element's frame
​https://bugs.webkit.org/show_bug.cgi?id=84780

Reviewed by Nate Chapin.

Source/WebCore:

Test: fast/loader/javascript-url-iframe-remove-on-navigate.html

  • loader/SubframeLoader.cpp:

(WebCore::SubframeLoader::loadOrRedirectSubframe):

LayoutTests:

  • fast/loader/javascript-url-iframe-remove-on-navigate-expected.txt: Added.
  • fast/loader/javascript-url-iframe-remove-on-navigate.html: Added.
1:30 PM Changeset in webkit [119334] by kov@webkit.org
  • 3 edits
    2 adds in releases/WebKitGTK/webkit-1.8

Merge 115398 - Invalid cast in WebCore::HTMLCollection::isAcceptableElement
​https://bugs.webkit.org/show_bug.cgi?id=84626

Reviewed by Darin Adler.

Source/WebCore:

Check if the object is an HTMLElement before casting.

Test: fast/dom/htmlcollection-non-html.html

  • html/HTMLCollection.cpp:

(WebCore::HTMLCollection::isAcceptableElement):

LayoutTests:

Add tests to make sure only HTML elements are present in most HTMLCollection objects.

  • fast/dom/htmlcollection-non-html-option-expected.txt: Added.
  • fast/dom/htmlcollection-non-html.html: Added.
1:30 PM Changeset in webkit [119333] by kov@webkit.org
  • 3 edits
    2 adds in releases/WebKitGTK/webkit-1.8

Source/WebCore: Crash in CachedRawResource::didAddClient() due to missing protector.
​https://bugs.webkit.org/show_bug.cgi?id=83632

Reviewed by Eric Seidel.

Test: http/tests/xmlhttprequest/access-control-repeated-failed-preflight-crash.html

  • loader/cache/CachedRawResource.cpp:

(WebCore::CachedRawResource::didAddClient):

LayoutTests: Test for ​https://bugs.webkit.org/show_bug.cgi?id=83632.

Reviewed by Eric Seidel.

  • http/tests/xmlhttprequest/access-control-repeated-failed-preflight-crash-expected.txt: Added.
  • http/tests/xmlhttprequest/access-control-repeated-failed-preflight-crash.html: Added.
1:30 PM Changeset in webkit [119332] by kov@webkit.org
  • 3 edits
    3 adds in releases/WebKitGTK/webkit-1.8

2012-03-09 Nate Chapin <Nate Chapin>

CachedRawResource breaks when trying to load
a resource with an empty response body from cache.

Reviewed by Alexey Proskuryakov.

Test: http/tests/cache/zero-length-xhr.html

  • loader/cache/CachedRawResource.cpp: (WebCore::CachedRawResource::didAddClient): Don't exit early

if m_data is empty, we may still need to notifyFinished().

1:29 PM Changeset in webkit [119331] by kov@webkit.org
  • 7 edits in releases/WebKitGTK/webkit-1.8

Source/WebCore: Prevent CachedRawResource from sending the same data
chunk multiple times.
​https://bugs.webkit.org/show_bug.cgi?id=78810

Reviewed by Adam Barth.

If a CachedRawResource receives data while a CachedRawResourceCallback
timer is active, the incremental data will be sent to the client, followed
but all data received so far, resulting in invalid data. Resolving this adds
complexity to CachedRawResource and requires making a few more CachedResource
functions virtual, so push the callback logic into CachedResource where it can
be implemented more cleanly.

Test: inspector/debugger/script-formatter-console.html

should no longer be flaky.

  • loader/cache/CachedRawResource.cpp: CachedRawResourceCallback renamed and moved to CachedResource.

(WebCore::CachedRawResource::didAddClient): More or less the same as sendCallbacks() was.

  • loader/cache/CachedRawResource.h:
  • loader/cache/CachedResource.cpp:

(WebCore::CachedResource::addClient): Check the return value of addClientToSet() to determine whether

or not to call didAddClient.

(WebCore::CachedResource::didAddClient): May be called during addClient(), or may be called on a timer.

If called on a timer, move the client between sets.

(WebCore::CachedResource::addClientToSet): Determine whether didAddClient() can be called synchronously and

return true if it can.

(WebCore::CachedResource::removeClient): Ensure we cancel pending callbacks if necessary.
(WebCore::CachedResource::CachedResourceCallback::CachedResourceCallback): Renamed and moved from CachedRawResource.

  • loader/cache/CachedResource.h:

(WebCore::CachedResource::hasClients): Check m_clientsAwaitingCallback as well as m_clients.
(WebCore::CachedResource::CachedResourceCallback::schedule):
(WebCore::CachedResource::hasClient): Helper for calling contains() on both m_clientsAwaitingCallback and m_clients.

LayoutTests: inspector/debugger/script-formatter-console.html should
no longer be flaky.
​https://bugs.webkit.org/show_bug.cgi?id=78810

Reviewed by Adam Barth.

  • platform/chromium/test_expectations.txt:
12:23 PM Changeset in webkit [119330] by haraken@chromium.org
  • 7 edits in trunk/Source/WebCore

[Performance] Optimize querySelector() by caching SelectorQuery objects
​https://bugs.webkit.org/show_bug.cgi?id=87942

Reviewed by Antti Koivisto.

This patch improves performance of Node::querySelector() by 7.1x in Safari/Mac
and by 8.5x in Chromium/Linux.

Performance test: Parser/query-selector-first.html, Parser/query-selector-last.html

[query-selector-first.html]
Safari/Mac 264.97 runs/s => 1872.78 runs/s (7.06x speed-up)
Chromium/Linux 244.84 runs/s => 2071.60 runs/s (8.46x speed-up)

[query-selector-last.html]
Safari/Mac 393.73 runs/s => 466.05 runs/s (1.18x speed-up)
Chromium/Linux 401.15 runs/s => 484.45 runs/s (1.20x speed-up)

Previously Node::querySelector() and Node::querySelectorAll() had been
parsing CSS queries every time. This patch optimizes the performance by caching
parsed results onto a Document.

The cache is invalidated when any of CSS related variables is updated.
As per the current implementation of CSSParserContext::operator==(), the CSS related
variables are as follows:

  • baseURI
  • charset
  • mode
  • isHTMLDocument
  • isCSSCustomFilterEnabled
  • isCSSRegionsEnabled
  • needsSiteSpecificQuirks
  • enforcesCSSMIMETypeInNoQuirksMode

Actually, we do not need to watch all of these variables:

  • The current implementation does not watch the change of charset.

charset is always set to a null String by CSSParserContext::CSSParserContext().

  • isHTMLDocument never changes.
  • isCSSCustomFilterEnabled, isCSSRegionsEnabled, needsSiteSpecificQuirks and

enforcesCSSMIMETypeInNoQuirksMode are not flipped in a user scenario.
If someone changes them, it would be reasonable to expect them
to take the effect only on subsequent document loads.
Thus we do not need to invalidate the cache when these variables are updated.

Consequently, the condition under which we have to invalidate the cache is
that any of the following variables is updated:

  • baseURI
  • mode

Tests: fast/dom/SelectorAPI/*. No change in test results.

  • dom/SelectorQuery.h: SelectorQueryCache is a cache from CSS selectors to parsed results.

SelectorQueryCache::Entry is an entry of the cache.
SelectorQueryCache::Entry holds a SelectorQuery object and a CSSSelectorList object.
The reason why SelectorQueryCache::Entry needs to hold the CSSSelectorList object
is that the CSSSelectorList object keeps the lifetime of CSSSelector objects
in the SelectorQuery object. Since the SelectorQuery object just holds pointers
to CSSSelector objects, the CSSSelectorList object must not be destructed
before the SelectorQuery object is destructed.
(WebCore):
(SelectorDataList):
(WebCore::SelectorQuery::SelectorQuery):
(SelectorQuery):
(SelectorQueryCache):
(WebCore::SelectorQueryCache::SelectorQueryCache):
(Entry):
(WebCore::SelectorQueryCache::Entry::selectorQuery):

  • dom/SelectorQuery.cpp:

(WebCore::SelectorQuery::initialize):
(WebCore::SelectorQueryCache::Entry::Entry):
(WebCore::SelectorQueryCache::add): Returns a cached SelectorQuery object if any.
Otherwise, parses a given CSS selector, creates a SelectorQuery object,
adds the SelectorQuery object to a new entry in the cache, returns the SelectorQuery
object.
(WebCore::SelectorQueryCache::invalidate): Clears the cache.

  • dom/Document.h:

(WebCore):
(Document):

  • dom/Document.cpp:

(WebCore::Document::selectorQueryCache):
(WebCore):
(WebCore::Document::setCompatibilityMode): Invalidates the cache
when m_compatibilityMode is updated.
(WebCore::Document::updateBaseURL): Invalidates the cache
when m_baseURL is updated.

  • dom/Node.h: Changed String to AtomicString, since the key of the cache

should be AtomicString.
(Node):

  • dom/Node.cpp: Optimized the code by using the cache.

(WebCore::Node::querySelector):
(WebCore::Node::querySelectorAll):

12:05 PM Changeset in webkit [119329] by mitz@apple.com
  • 2 edits in trunk/Source/WebCore

Reverted the last change.

  • WebCore.xcodeproj/project.pbxproj:
12:02 PM Changeset in webkit [119328] by mitz@apple.com
  • 2 edits in trunk/Source/WebKit2

Reverted the last change.

  • WebKit2.xcodeproj/project.pbxproj:
11:59 AM Changeset in webkit [119327] by mitz@apple.com
  • 2 edits in trunk/Source/WebCore

Stop treating Perl code as private headers.

  • WebCore.xcodeproj/project.pbxproj: Removed CodeGenerator.pm, generate-bindings.pl,

IDLParser.pm, IDLStructure.pm, and preprocessor.pm from the Copy Headers build phase.

11:57 AM Changeset in webkit [119326] by fischman@chromium.org
  • 2 edits in trunk/LayoutTests

[chromium] SKIP http/tests/media/video-buffered.html because it requires overly large data
​https://bugs.webkit.org/show_bug.cgi?id=88172

Unreviewed gardening.

  • platform/chromium/TestExpectations:
11:54 AM Changeset in webkit [119325] by kevino@webkit.org
  • 2 edits in trunk

[wx] Unreviewed build fix. Temporarily disable DerivedSources cleanup on Windows.

11:43 AM Changeset in webkit [119324] by mitz@apple.com
  • 2 edits in trunk/Source/WebKit2

Stop treating two Python scripts as private headers.

  • WebKit2.xcodeproj/project.pbxproj: Removed model.py and parser.py from the Copy Headers

build phase.

9:58 AM Changeset in webkit [119323] by Csaba Osztrogonác
  • 6 edits
    37 moves
    17 adds
    5 deletes in trunk/LayoutTests

[Qt] Replace QtXmlPatterns usage with libxslt dependency
​https://bugs.webkit.org/show_bug.cgi?id=76820

Unreviewed gardening.

  • fast/xsl/xslt-mismatched-tags-in-xslt-expected.txt: Renamed from LayoutTests/platform/efl/fast/xsl/xslt-mismatched-tags-in-xslt-expected.txt.
  • platform/efl/fast/xsl/xslt-mismatched-tags-in-xslt-expected.png: Removed.
  • platform/gtk/fast/xsl/xslt-mismatched-tags-in-xslt-expected.png: Removed.
  • platform/gtk/fast/xsl/xslt-mismatched-tags-in-xslt-expected.txt: Removed.
  • platform/mac/fast/xsl/xslt-mismatched-tags-in-xslt-expected.png: Removed.
  • platform/mac/fast/xsl/xslt-mismatched-tags-in-xslt-expected.txt: Removed.
  • platform/qt-4.8/Skipped:
  • platform/qt-4.8/fast/css/dumpAsText/xml-stylesheet-pi-not-in-prolog-expected.txt: Renamed from LayoutTests/platform/qt/fast/css/dumpAsText/xml-stylesheet-pi-not-in-prolog-expected.txt.
  • platform/qt-4.8/fast/dom/Range/surround-contents-font-face-crash-expected.txt: Renamed from LayoutTests/platform/qt/fast/dom/Range/surround-contents-font-face-crash-expected.txt.
  • platform/qt-4.8/fast/dom/xml-parser-error-message-crash-expected.txt: Renamed from LayoutTests/platform/qt/fast/dom/xml-parser-error-message-crash-expected.txt.
  • platform/qt-4.8/fast/encoding/dumpAsText/utf-16-no-bom-expected.txt: Renamed from LayoutTests/platform/qt/fast/encoding/dumpAsText/utf-16-no-bom-expected.txt.
  • platform/qt-4.8/fast/events/xsl-onload-expected.txt: Renamed from LayoutTests/platform/qt/fast/events/xsl-onload-expected.txt.
  • platform/qt-4.8/fast/parser/xml-colon-entity-expected.txt: Renamed from LayoutTests/platform/qt/fast/parser/xml-colon-entity-expected.txt.
  • platform/qt-4.8/fast/parser/xml-declaration-missing-ending-mark-expected.txt: Renamed from LayoutTests/platform/qt/fast/parser/xml-declaration-missing-ending-mark-expected.txt.
  • platform/qt-4.8/fast/parser/xslt-with-html-expected.txt: Renamed from LayoutTests/platform/qt/fast/parser/xslt-with-html-expected.txt.
  • platform/qt-4.8/fast/xsl/subframe-location-expected.txt: Renamed from LayoutTests/platform/qt/fast/xsl/subframe-location-expected.txt.
  • platform/qt-4.8/fast/xsl/transform-to-html-expected.txt: Renamed from LayoutTests/platform/qt/fast/xsl/transform-to-html-expected.txt.
  • platform/qt-4.8/fast/xsl/utf8-chunks-expected.txt: Renamed from LayoutTests/platform/qt/fast/xsl/utf8-chunks-expected.txt.
  • platform/qt-4.8/fast/xsl/xslt-bad-import-uri-expected.txt: Renamed from LayoutTests/platform/qt/fast/xsl/xslt-bad-import-uri-expected.txt.
  • platform/qt-4.8/fast/xsl/xslt-doc-noenc-expected.txt: Renamed from LayoutTests/platform/qt/fast/xsl/xslt-doc-noenc-expected.txt.
  • platform/qt-4.8/fast/xsl/xslt-entity-enc-expected.txt: Renamed from LayoutTests/platform/qt/fast/xsl/xslt-entity-enc-expected.txt.
  • platform/qt-4.8/fast/xsl/xslt-entity-expected.png: Renamed from LayoutTests/platform/qt/fast/xsl/xslt-entity-expected.png.
  • platform/qt-4.8/fast/xsl/xslt-mismatched-tags-in-xslt-expected.txt: Renamed from LayoutTests/platform/qt/fast/xsl/xslt-mismatched-tags-in-xslt-expected.txt.
  • platform/qt-4.8/fast/xsl/xslt-recursion-expected.txt: Renamed from LayoutTests/platform/qt/fast/xsl/xslt-recursion-expected.txt.
  • platform/qt-4.8/http/tests/misc/location-test-xsl-style-sheet-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/misc/location-test-xsl-style-sheet-expected.txt.
  • platform/qt-4.8/http/tests/security/contentSecurityPolicy/xsl-allowed-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/security/contentSecurityPolicy/xsl-allowed-expected.txt.
  • platform/qt-4.8/http/tests/security/contentSecurityPolicy/xsl-img-blocked-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/security/contentSecurityPolicy/xsl-img-blocked-expected.txt.
  • platform/qt-4.8/http/tests/security/contentSecurityPolicy/xsl-unaffected-by-style-src-2-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/security/contentSecurityPolicy/xsl-unaffected-by-style-src-2-expected.txt.
  • platform/qt-4.8/http/tests/security/cookies/first-party-cookie-allow-xslt-expected.png: Renamed from LayoutTests/platform/qt/http/tests/security/cookies/first-party-cookie-allow-xslt-expected.png.
  • platform/qt-4.8/http/tests/security/cookies/first-party-cookie-allow-xslt-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/security/cookies/first-party-cookie-allow-xslt-expected.txt.
  • platform/qt-4.8/http/tests/security/cookies/third-party-cookie-blocking-xslt-expected.png: Renamed from LayoutTests/platform/qt/http/tests/security/cookies/third-party-cookie-blocking-xslt-expected.png.
  • platform/qt-4.8/http/tests/security/cookies/third-party-cookie-blocking-xslt-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/security/cookies/third-party-cookie-blocking-xslt-expected.txt.
  • platform/qt-4.8/http/tests/security/xss-DENIED-xml-external-entity-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/security/xss-DENIED-xml-external-entity-expected.txt.
  • platform/qt-4.8/http/tests/security/xss-DENIED-xsl-document-redirect-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/security/xss-DENIED-xsl-document-redirect-expected.txt.
  • platform/qt-4.8/http/tests/security/xss-DENIED-xsl-external-entity-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/security/xss-DENIED-xsl-external-entity-expected.txt.
  • platform/qt-4.8/http/tests/security/xss-DENIED-xsl-external-entity-redirect-expected.txt: Renamed from LayoutTests/platform/qt/http/tests/security/xss-DENIED-xsl-external-entity-redirect-expected.txt.
  • platform/qt-4.8/svg/custom/bug45331-expected.png: Renamed from LayoutTests/platform/qt/svg/custom/bug45331-expected.png.
  • platform/qt-4.8/svg/custom/bug45331-expected.txt: Renamed from LayoutTests/platform/qt/svg/custom/bug45331-expected.txt.
  • platform/qt-4.8/svg/custom/bug78807-expected.txt: Renamed from LayoutTests/platform/qt/svg/custom/bug78807-expected.txt.
  • platform/qt-4.8/svg/custom/use-invalid-html-expected.txt: Renamed from LayoutTests/platform/qt/svg/custom/use-invalid-html-expected.txt.
  • platform/qt-4.8/svg/custom/use-invalid-style-expected.txt: Renamed from LayoutTests/platform/qt/svg/custom/use-invalid-style-expected.txt.
  • platform/qt-4.8/svg/hixie/error/dumpAsText/004-expected.txt: Renamed from LayoutTests/platform/qt/svg/hixie/error/dumpAsText/004-expected.txt.
  • platform/qt-4.8/svg/hixie/error/dumpAsText/005-expected.txt: Renamed from LayoutTests/platform/qt/svg/hixie/error/dumpAsText/005-expected.txt.
  • platform/qt-5.0-wk1/Skipped:
  • platform/qt-5.0/Skipped:
  • platform/qt-5.0/http/tests/security/xss-DENIED-xsl-document-redirect-expected.txt: Added.
  • platform/qt-5.0/http/tests/security/xss-DENIED-xsl-external-entity-redirect-expected.txt: Added.
  • platform/qt-arm/Skipped:
  • platform/qt/Skipped:
7:53 AM Changeset in webkit [119322] by tonikitoo@webkit.org
  • 3 edits in trunk/Source/WebKit/blackberry

[BlackBerry] browser video player fullscreen mode (portrait) - out of screen/focus - cannot navigate or use the buttons on the screen (PART III)
​https://bugs.webkit.org/show_bug.cgi?id=88019

Reviewed by George Staikos.
Patch by Antonio Gomes <​agomes@rim.com>

Enter 'pure-with-mouse-conversion' mode when going fullscreen, so
that it prevents user from scrolling the WebPage, pinch zooming,
touch-and-hold, enter selection mode, etc ...

Internally reviewed by Gen Mak.

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::WebPagePrivate):
(BlackBerry::WebKit::WebPagePrivate::enterFullScreenForElement):
(BlackBerry::WebKit::WebPagePrivate::exitFullScreenForElement):

  • Api/WebPage_p.h:

(WebPagePrivate):

7:52 AM Changeset in webkit [119321] by tonikitoo@webkit.org
  • 3 edits in trunk/Source/WebKit/blackberry

[BlackBerry] browser video player fullscreen mode (portrait) - out of screen/focus - cannot navigate or use the buttons on the screen (PART II)
​https://bugs.webkit.org/show_bug.cgi?id=88019

Reviewed by George Staikos.
Patch by Antonio Gomes <​agomes@rim.com>

When an element goes fullscreen, its wrapper/container obeys all
BlackBerry specific fixed position customizations: we fixed
against Y, but not X. Then, in order to have the wrapper element
properly positioned when entering fullscreen mode, we
temporarily scroll x to 0.

The original x scroll position is restored when we leave
fullscreen.

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::WebPagePrivate):
(BlackBerry::WebKit::WebPagePrivate::enterFullScreenForElement):
(BlackBerry::WebKit::WebPagePrivate::exitFullScreenForElement):

  • Api/WebPage_p.h:

(WebPagePrivate):

7:52 AM Changeset in webkit [119320] by tonikitoo@webkit.org
  • 3 edits in trunk/Source/WebKit/blackberry

[BlackBerry] browser video player fullscreen mode (portrait) - out of screen/focus - cannot navigate or use the buttons on the screen (PART I)
​https://bugs.webkit.org/show_bug.cgi?id=88019
PR #158266

Reviewed by George Staikos.
Patch by Antonio Gomes <​agomes@rim.com>

The way elements go fullscreen with the new FULLSCREEN_API
is that they get cloned and added to an out-of-DOM wrapper
element. The wrapper is a normal fixed position element and
then zoom in/out accordingly to how other layers do: following
WebPage's scale.

When going fullscreen, we have to take the current WebPage scale
into account in order to properly fit the element to the screen,
regardless the web page scale.

  • WebCoreSupport/ChromeClientBlackBerry.cpp:

(WebCore):
(WebCore::ChromeClientBlackBerry::fullScreenRendererChanged):

  • WebCoreSupport/ChromeClientBlackBerry.h:

(ChromeClientBlackBerry):

7:48 AM Changeset in webkit [119319] by Csaba Osztrogonác
  • 2 edits in trunk/LayoutTests

[Qt] Unreviewed fix after r119255.

  • platform/qt/svg/css/getComputedStyle-basic-expected.txt:
7:01 AM Changeset in webkit [119318] by zeno.albisser@nokia.com
  • 20 edits in trunk

Fix and enable WebGL for WebKit2 on Qt.
​https://bugs.webkit.org/show_bug.cgi?id=86214

Source/WebCore:

Make GraphicsContext3DPrivate use GraphicsSurfaces
for WK2. The GraphicsContext3D then uses the existing
RenderBuffer for multisample rendering.
When WebGraphicsLayer::syncCompositingState is being
executed, the canvas is being synced as well. This means
that the RenderBuffer contents are being blit onto
a GraphicsSurface, and the GraphicsSurface token is
being sent to the UIProcess.
The WebLayerTreeRenderer then creates a
TextureMapperSurfaceBackingStore for the canvas and
passes the GraphicsSurface token as an argument.
The token can then be used to identify the GraphicsSurface
from the UIProcess side.

Reviewed by Noam Rosenthal.

  • platform/graphics/GraphicsContext3D.h:

Added createGraphicsSurfaces function. This is currently only
being used by the Qt port.

  • platform/graphics/opengl/GraphicsContext3DOpenGLCommon.cpp:

(WebCore::GraphicsContext3D::reshape):
Calling the createGraphicsSurfaces function when the GraphicsContext3D
is reshaped. This is currently only relevant for the Qt port.

  • platform/graphics/qt/GraphicsContext3DQt.cpp:

(GraphicsContext3DPrivate):
Added m_frontBufferGraphicsSurface, m_backBufferGraphicsSurface
and m_surfaceFlags members.
(WebCore::GraphicsContext3DPrivate::GraphicsContext3DPrivate):
In case of WK2, create a QOpenGLContext and two GraphicsSurface
for sharing the WebGL content with the UIProcess. One GraphicsSurface
is being used as the front, the other one as the backbuffer.
Creating a QOpenGLContext currently requires showing a QWindow.
For the moment we therefore create a minimal QWindow and place
it offscreen.
(WebCore::GraphicsContext3DPrivate::copyToGraphicsSurface):
This new function is called from the WebGraphicsLayer,
to blit the multisample framebuffer and copy its contents
onto the GraphicsSurface.
(WebCore::GraphicsContext3DPrivate::createGraphicsSurfaces):
Whenever the GraphicsContext3D is being reshaped,
new GraphicsSurfaces must be created with the updated dimensions.
(WebCore::GraphicsContext3D::createGraphicsSurfaces):

  • platform/graphics/texmap/TextureMapperPlatformLayer.h:

Added a new virtual function copyToGraphicsSurface.
(TextureMapperPlatformLayer):
(WebCore::TextureMapperPlatformLayer::copyToGraphicsSurface):

Source/WebKit2:

Added glue code to make use of GraphicsSurface
as a backend for the webgl-canvas in case of WK2.

Reviewed by Noam Rosenthal.

  • UIProcess/API/qt/qwebpreferences.cpp:

(QWebPreferencesPrivate::testAttribute):
(QWebPreferencesPrivate::setAttribute):
(QWebPreferences::webGLEnabled):
(QWebPreferences::setWebGLEnabled):

  • UIProcess/API/qt/qwebpreferences_p.h:
  • UIProcess/API/qt/qwebpreferences_p_p.h:

Added WebGLEnabled enum to allow enabling/disabling of
WebGL using QWebPreferences.

  • UIProcess/LayerTreeHostProxy.cpp:

(WebKit::LayerTreeHostProxy::syncCanvas):
Dispatch syncCanvas calls to the apropriate
WebLayerTreeRenderer.

  • UIProcess/LayerTreeHostProxy.h:

(LayerTreeHostProxy):

  • UIProcess/LayerTreeHostProxy.messages.in:
  • UIProcess/WebLayerTreeRenderer.h:

(WebLayerTreeRenderer):

  • UIProcess/WebLayerTreeRenderer.cpp:

(WebKit::WebLayerTreeRenderer::syncCanvas):
(WebKit::WebLayerTreeRenderer::deleteLayer):
(WebKit::WebLayerTreeRenderer::purgeGLResources):
Create a TextureMapperSurfaceBackingStore for the canvas
if necessary and pass or update the graphicsSurfaceToken
for to be used with the backing store.

  • WebProcess/WebCoreSupport/WebGraphicsLayer.cpp:

(WebCore::WebGraphicsLayer::WebGraphicsLayer):
(WebCore):
(WebCore::WebGraphicsLayer::setContentsToCanvas):
(WebCore::WebGraphicsLayer::syncCanvas):
Copy the multisample framebuffer contents onto the GraphicsSurface.
Notify the UIProcess of the availability of a new texture.
(WebCore::WebGraphicsLayer::syncCompositingStateForThisLayerOnly):
Sync the canvas as well.

  • WebProcess/WebCoreSupport/WebGraphicsLayer.h:

Added a pure virtual function syncCanvas.
This is guarded by PLATFORM(QT).
(WebGraphicsLayerClient):
(WebGraphicsLayer):

  • WebProcess/WebPage/qt/LayerTreeHostQt.cpp:

(WebKit::LayerTreeHostQt::syncLayerChildren):
(WebKit):
(WebKit::LayerTreeHostQt::syncCanvas):

  • WebProcess/WebPage/qt/LayerTreeHostQt.h:

(LayerTreeHostQt):

Tools:

Enable WebGL by default for Qt MiniBrowser.

Reviewed by Noam Rosenthal.

  • MiniBrowser/qt/qml/BrowserWindow.qml:
2:46 AM Changeset in webkit [119317] by rniwa@webkit.org
  • 3 edits
    6 moves in trunk

Tools: Teach svn-apply how to apply changes in test_expectations.txt to TestExpectations
​https://bugs.webkit.org/show_bug.cgi?id=88164

Reviewed by Kentaro Hara.

Just like r74856, adjust git/svn headers from test_expectations.txt to TestExpectations.
Also replace all occurrences of /test_expectations.txt: by /TestExpectations: in change log entries.

  • Scripts/VCSUtils.pm:

(adjustPathForRecentRenamings):
(fixChangeLogPatch):

LayoutTests: Rename test_expectations.txt to TestExpectations.

  • platform/chromium/TestExpectations: Copied from LayoutTests/platform/chromium/test_expectations.txt.
  • platform/chromium/test_expectations.txt: Removed.
  • platform/efl/TestExpectations: Copied from LayoutTests/platform/efl/test_expectations.txt.
  • platform/efl/test_expectations.txt: Removed.
  • platform/gtk/TestExpectations: Copied from LayoutTests/platform/gtk/test_expectations.txt.
  • platform/gtk/test_expectations.txt: Removed.
  • platform/mac/TestExpectations: Copied from LayoutTests/platform/mac/test_expectations.txt.
  • platform/mac/test_expectations.txt: Removed.
  • platform/qt/TestExpectations: Copied from LayoutTests/platform/qt/test_expectations.txt.
  • platform/qt/test_expectations.txt: Removed.
  • platform/win/TestExpectations: Copied from LayoutTests/platform/win/test_expectations.txt.
  • platform/win/test_expectations.txt: Removed.
1:15 AM Changeset in webkit [119316] by tony@chromium.org
  • 55 edits in trunk

Rename the flexbox CSS propery values from start to flex-start and end to flex-end
​https://bugs.webkit.org/show_bug.cgi?id=88152

Reviewed by Ojan Vafai.

Source/WebCore:

The spec changed. This is to differentiate from start/end which will
depend on writing mode direction, rather than flex direction.

No new tests, just updated the existing values.

  • css/CSSParser.cpp:

(WebCore::isValidKeywordPropertyAndValue):

  • css/CSSPrimitiveValueMappings.h:

(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
(WebCore::CSSPrimitiveValue::operator EAlignItems):
(WebCore::CSSPrimitiveValue::operator EJustifyContent):
(WebCore::CSSPrimitiveValue::operator EAlignContent):

  • css/CSSValueKeywords.in:
  • rendering/RenderFlexibleBox.cpp:

(WebCore::initialJustifyContentOffset):
(WebCore::alignmentForChild):
(WebCore::initialAlignContentOffset):
(WebCore::RenderFlexibleBox::alignFlexLines):
(WebCore::RenderFlexibleBox::alignChildren):

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

LayoutTests:

Rename start to flex-start and end to flex-end.

  • css3/flexbox/css-properties-expected.txt:
  • css3/flexbox/css-properties.html:
  • css3/flexbox/flex-align-column.html:
  • css3/flexbox/flex-align-end.html:
  • css3/flexbox/flex-align-vertical-writing-mode.html:
  • css3/flexbox/flex-align.html:
  • css3/flexbox/flex-pack.html:
  • css3/flexbox/line-wrapping.html:
  • css3/flexbox/multiline-align.html:
  • css3/flexbox/multiline-column-auto.html:
  • css3/flexbox/multiline-line-pack-horizontal-column-expected.txt:
  • css3/flexbox/multiline-line-pack-horizontal-column.html:
  • css3/flexbox/multiline-line-pack.html:
  • css3/flexbox/multiline-pack-expected.txt:
  • css3/flexbox/multiline-pack.html:
  • css3/flexbox/multiline-reverse-wrap-baseline.html:
  • css3/flexbox/multiline-reverse-wrap-overflow.html:
  • css3/flexbox/multiline-shrink-to-fit.html:
  • css3/flexbox/multiline.html:
  • css3/flexbox/nested-stretch.html:
  • css3/flexbox/orthogonal-flex-directions.html:
  • platform/chromium-linux-x86/svg/css/getComputedStyle-basic-expected.txt:
  • platform/chromium-linux/svg/css/getComputedStyle-basic-expected.txt:
  • platform/chromium-mac-leopard/svg/css/getComputedStyle-basic-expected.txt:
  • platform/chromium-mac-snowleopard/svg/css/getComputedStyle-basic-expected.txt:
  • platform/chromium-mac/svg/css/getComputedStyle-basic-expected.txt:
  • platform/chromium-win-vista/svg/css/getComputedStyle-basic-expected.txt:
  • platform/chromium-win-xp/svg/css/getComputedStyle-basic-expected.txt:
  • platform/chromium-win/svg/css/getComputedStyle-basic-expected.txt:
  • platform/gtk/svg/css/getComputedStyle-basic-expected.txt:
  • platform/mac/svg/css/getComputedStyle-basic-expected.txt:
  • platform/qt/svg/css/getComputedStyle-basic-expected.txt:
  • svg/css/getComputedStyle-basic-expected.txt:
12:12 AM Changeset in webkit [119315] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/qt

[Qt] Use 'Ok' and 'Cancel' buttons in JavaScript confirm box.
​https://bugs.webkit.org/show_bug.cgi?id=76190
Based on Johannes Obermayrs (​johannesobermayr@gmx.de) patch.
See: ​http://www.javascripter.net/faq/confirm.htm
Fixes: ​https://bugs.kde.org/show_bug.cgi?id=287629

Patch by Zoltan Arvai <​zarvai@inf.u-szeged.hu> on 2012-06-02
Reviewed by Simon Hausmann.

  • Api/qwebpage.cpp:

(QWebPage::javaScriptConfirm):

Note: See TracTimeline for information about the timeline view.