⚠ Archived content — this site is no longer maintained.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

Timeline



Apr 24, 2011:

11:47 PM April 2011 Meeting edited by Adam Roben
(diff)
11:45 PM April 2011 Meeting edited by Philippe Normand
(diff)
11:22 PM Changeset in webkit [84766] by dpranke@chromium.org
  • 2 edits
    2 adds in trunk/LayoutTests

2011-04-24 Dirk Pranke <​dpranke@chromium.org>

Unreviewed, expectations change.

Add Chromium baselines for fast/table/auto-100-percent-width.html.
Mark fast/dom/HTMLMeterElement/meter-element.html and
http/tests/misc/will-send-request-returns-null-on-redirect.html
as failing - need triage.

  • platform/chromium-mac/fast/table/auto-100-percent-width-expected.png: Added.
  • platform/chromium-win/fast/table/auto-100-percent-width-expected.png: Added.
  • platform/chromium/test_expectations.txt:
11:02 PM April 2011 Meeting edited by jocelyn.turcotte@nokia.com
(diff)
10:36 PM Changeset in webkit [84765] by abarth@webkit.org
  • 4 edits in trunk/LayoutTests

2011-04-24 Adam Barth <​abarth@webkit.org>

Update Chromium results for standard-url.html (again).

  • platform/chromium-linux/fast/url/standard-url-expected.txt:
  • platform/chromium-mac/fast/url/standard-url-expected.txt:
  • platform/chromium-win/fast/url/standard-url-expected.txt:
10:12 PM April 2011 Meeting edited by demarchi@webkit.org
adding /me to interesting talkings (diff)
9:56 PM April 2011 Meeting edited by abarth@webkit.org
(diff)
9:55 PM April 2011 Meeting edited by abarth@webkit.org
(diff)
9:22 PM ThreadCommunication edited by levin@chromium.org
(diff)
9:08 PM April 2011 Meeting edited by eric@webkit.org
(diff)
9:06 PM April 2011 Meeting edited by eric@webkit.org
(diff)
8:49 PM April 2011 Meeting edited by dimich@chromium.org
(diff)
8:43 PM April 2011 Meeting edited by jparent@chromium.org
(diff)
7:43 PM April 2011 Meeting edited by ojan@chromium.org
(diff)
5:47 PM Changeset in webkit [84764] by ggaren@apple.com
  • 80 edits in trunk/Source/WebCore

2011-04-24 Geoffrey Garen <​ggaren@apple.com>

Reviewed by Sam Weinig.

Removed a use of markDOMObjectWrapper: WebGL, XHR, workers; plus, more autogeneration
​https://bugs.webkit.org/show_bug.cgi?id=59307

  • WebCore.xcodeproj/project.pbxproj: Added Blob.idl, so it's easier to edit.
  • bindings/js/JSCSSRuleCustom.cpp:
  • bindings/js/JSCSSStyleDeclarationCustom.cpp:
  • bindings/js/JSCanvasRenderingContextCustom.cpp:
  • bindings/js/JSDOMApplicationCacheCustom.cpp: Autogenerate these instead of hand-coding them.
  • bindings/js/JSDOMBinding.h: (WebCore::root): Moved some GC helpers here from JSNodeCustom.h, because they're reasonably generic.
  • bindings/js/JSDOMImplementationCustom.cpp:
  • bindings/js/JSDOMStringMapCustom.cpp:
  • bindings/js/JSDOMTokenListCustom.cpp:
  • bindings/js/JSHTMLCollectionCustom.cpp:
  • bindings/js/JSMediaListCustom.cpp: Autogenerate these instead of hand-coding them.
  • bindings/js/JSMessageChannelCustom.cpp: (WebCore::JSMessageChannel::visitChildren):
  • bindings/js/JSMessagePortCustom.cpp: (WebCore::JSMessagePort::visitChildren): Use addOpaqueRoot instead of markDOMObjectWrapper. This is one of the few cases where a DOM object is considered a stand-alone root. It's not as efficient as the shared root case, but workers and message ports are very rare objects, so it's no big deal.
  • bindings/js/JSNamedNodeMapCustom.cpp: Autogenerate!
  • bindings/js/JSNodeCustom.h: Moved to JSDOMBinding.h.
  • bindings/js/JSSharedWorkerCustom.cpp: (WebCore::JSSharedWorker::visitChildren): Use addOpaqueRoot instead of markDOMObjectWrapper, as above. Once again, a stand-alone root.
  • bindings/js/JSStyleSheetCustom.cpp:
  • bindings/js/JSStyleSheetListCustom.cpp: Autogenerate!
  • bindings/js/JSWebGLRenderingContextCustom.cpp: (WebCore::JSWebGLRenderingContext::visitChildren): Use the opaque roots system for marking WebGL contexts and their associated satellite objects.
  • bindings/js/JSWorkerContextCustom.cpp: (WebCore::JSWorkerContext::visitChildren):
  • bindings/js/JSXMLHttpRequestCustom.cpp: Use addOpaqueRoot instead of markDOMObjectWrapper, as above. Once again, stand-alone roots.

(WebCore::JSXMLHttpRequest::visitChildren):

  • bindings/js/JSXMLHttpRequestUploadCustom.cpp: No need to mark our owner XHR because it is not reachable from us.
  • bindings/scripts/CodeGeneratorJS.pm: Added autogeneration support for a bunch of repetitive cases of isReachableFromOpaqueRoots callbacks.
  • css/CSSRule.idl:
  • css/CSSStyleDeclaration.idl:
  • css/MediaList.idl:
  • css/StyleMedia.idl:
  • css/StyleSheet.idl:
  • css/StyleSheetList.idl:
  • dom/DOMImplementation.idl:
  • dom/DOMStringMap.idl:
  • dom/MessagePort.idl:
  • dom/NamedNodeMap.idl:
  • fileapi/Blob.idl:
  • html/DOMTokenList.idl: Opt in to autogeneration.
  • html/HTMLCanvasElement.cpp: (WebCore::HTMLCanvasElement::getContext): Standardized on PassOwnPtr/OwnPtr usage, to reduce human error and make ownership rules more obvious.
  • html/HTMLCollection.idl:
  • html/canvas/ArrayBuffer.idl: Opt in to autogeneration.
  • html/canvas/CanvasRenderingContext.cpp:
  • html/canvas/CanvasRenderingContext.h: (WebCore::CanvasRenderingContext::ref): (WebCore::CanvasRenderingContext::deref):
  • html/canvas/CanvasRenderingContext2D.h: (WebCore::CanvasRenderingContext2D::create): (WebCore::CanvasRenderingContext2D::state): Standardized on PassOwnPtr/OwnPtr usage, to reduce human error and make ownership rules more obvious.


Inlined some functions to match our standard idiom and to make the code
in the header document itself.

  • html/canvas/CanvasRenderingContext.idl: Opt in to autogeneration.
  • html/canvas/OESStandardDerivatives.cpp: (WebCore::OESStandardDerivatives::OESStandardDerivatives): (WebCore::OESStandardDerivatives::create):
  • html/canvas/OESStandardDerivatives.h:
  • html/canvas/OESStandardDerivatives.idl:
  • html/canvas/OESTextureFloat.cpp: (WebCore::OESTextureFloat::OESTextureFloat): (WebCore::OESTextureFloat::create):
  • html/canvas/OESTextureFloat.h:
  • html/canvas/OESTextureFloat.idl:
  • html/canvas/OESVertexArrayObject.cpp: (WebCore::OESVertexArrayObject::OESVertexArrayObject): (WebCore::OESVertexArrayObject::create):
  • html/canvas/OESVertexArrayObject.h:
  • html/canvas/OESVertexArrayObject.idl:
  • html/canvas/WebGLExtension.cpp: (WebCore::WebGLExtension::WebGLExtension):
  • html/canvas/WebGLExtension.h: (WebCore::WebGLExtension::ref): (WebCore::WebGLExtension::deref): (WebCore::WebGLExtension::context):
  • html/canvas/WebGLRenderingContext.cpp: (WebCore::WebGLRenderingContext::~WebGLRenderingContext): (WebCore::WebGLRenderingContext::getExtension):
  • html/canvas/WebGLRenderingContext.h:
  • html/canvas/WebKitLoseContext.cpp: (WebCore::WebKitLoseContext::WebKitLoseContext): (WebCore::WebKitLoseContext::create): (WebCore::WebKitLoseContext::loseContext):
  • html/canvas/WebKitLoseContext.h:
  • html/canvas/WebKitLoseContext.idl: Updated the ownership model for WebGL canavs rendering contexts to match the model for 2D canvas rendering contexts. This makes garbage collection a lot more straightforward, but it also makes the behavior of these objects more reliable. (Previously, satellite objects would magically stop working when the last reference to their owner object was dropped. Now, satellite objects keep their owner alive through reference counting.)
  • loader/appcache/DOMApplicationCache.idl:
  • page/BarInfo.idl:
  • page/Console.idl:
  • page/DOMSelection.idl:
  • page/Geolocation.idl:
  • page/History.idl:
  • page/Location.idl:
  • page/Navigator.idl:
  • page/Screen.idl:
  • page/WorkerNavigator.idl:
  • plugins/DOMMimeTypeArray.h: (WebCore::DOMMimeTypeArray::frame):
  • plugins/DOMMimeTypeArray.idl:
  • plugins/DOMPluginArray.h: (WebCore::DOMPluginArray::frame):
  • plugins/DOMPluginArray.idl:
  • storage/Storage.idl:
  • workers/WorkerLocation.idl: Opt in to autogeneration.
  • xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::~XMLHttpRequest):
  • xml/XMLHttpRequest.h:
  • xml/XMLHttpRequestUpload.cpp: (WebCore::XMLHttpRequestUpload::scriptExecutionContext):
  • xml/XMLHttpRequestUpload.h: (WebCore::XMLHttpRequestUpload::create): (WebCore::XMLHttpRequestUpload::ref): (WebCore::XMLHttpRequestUpload::deref): (WebCore::XMLHttpRequestUpload::xmlHttpRequest): (WebCore::XMLHttpRequestUpload::toXMLHttpRequestUpload):
  • xml/XMLHttpRequestUpload.idl: Updated the ownership model for XHR and its associated upload object to match the model for canvas. This makes garbage collection a lot more straightforward, but it also makes the behavior of these objects more reliable.
5:38 PM April 2011 Meeting edited by rniwa@webkit.org
Make the title of RenderObject guard more descriptive (diff)
4:39 PM April 2011 Meeting edited by alex
(diff)
4:13 PM April 2011 Meeting edited by aestes@apple.com
(diff)
4:09 PM Changeset in webkit [84763] by commit-queue@webkit.org
  • 3 edits
    2 deletes in trunk

2011-04-24 Sheriff Bot <​webkit.review.bot@gmail.com>

Unreviewed, rolling out r84759.
​http://trac.webkit.org/changeset/84759
​https://bugs.webkit.org/show_bug.cgi?id=59306

Caused gc-shadow.html to start crashing in V8. (Requested by
dglazkov|away on #webkit).

  • fast/dom/shadow/append-child-text-expected.txt: Removed.
  • fast/dom/shadow/append-child-text.html: Removed.

2011-04-24 Sheriff Bot <​webkit.review.bot@gmail.com>

Unreviewed, rolling out r84759.
​http://trac.webkit.org/changeset/84759
​https://bugs.webkit.org/show_bug.cgi?id=59306

Caused gc-shadow.html to start crashing in V8. (Requested by
dglazkov|away on #webkit).

  • dom/Node.cpp: (WebCore::NodeRendererFactory::createRendererAndStyle): (WebCore::Node::styleForRenderer):
3:53 PM Changeset in webkit [84762] by mjs@apple.com
  • 9 edits in trunk

2011-04-24 Maciej Stachowiak <​mjs@apple.com>

Reviewed by George Staikos.

Handling of URLs like http:/example.com/ is incorrect
​https://bugs.webkit.org/show_bug.cgi?id=59300
<rdar://problem/9231956>

URLs like http:/example.com/ or http:example.com/ are now correctly
canonicalized as ​http://example.com/


The code still doesn't quite match other browsers - at least some
other browsers seem to base parsing behavior on whether they
recongize a scheme from a fixed list, and ignore whether is
present in the URL or not.


  • platform/KURL.cpp: (WebCore::isNonFileHierarchicalScheme): New helper function. (WebCore::KURL::parse): For a particular list of whitelisted schemes, assume they are hierarchical and need an authority even if there is no after the :/

2011-04-24 Maciej Stachowiak <​mjs@apple.com>

Reviewed by George Staikos.

Handling of URLs like http:/example.com/ is incorrect
​https://bugs.webkit.org/show_bug.cgi?id=59300
<rdar://problem/9231956>


Added some new test cases. Also updated results for the
change. The changes are progressions.

  • fast/url/relative-expected.txt:
  • fast/url/script-tests/segments.js:
  • fast/url/script-tests/standard-url.js:
  • fast/url/segments-expected.txt:
  • fast/url/segments-from-data-url-expected.txt:
  • fast/url/standard-url-expected.txt:
2:48 PM Changeset in webkit [84761] by abarth@webkit.org
  • 3 edits
    1 add in trunk/LayoutTests

2011-04-24 Adam Barth <​abarth@webkit.org>

Update Chromium results for standard-url.html

  • platform/chromium-linux/fast/url/standard-url-expected.txt: Added.
  • platform/chromium-mac/fast/url/standard-url-expected.txt:
  • platform/chromium-win/fast/url/standard-url-expected.txt:
2:37 PM Changeset in webkit [84760] by mitz@apple.com
  • 2 edits in trunk/Source/WebCore

LLVM Compiler build fix.

  • page/ContentSecurityPolicy.cpp: Removed an unused function.
2:20 PM Changeset in webkit [84759] by commit-queue@webkit.org
  • 3 edits
    2 adds in trunk

2011-04-24 Dominic Cooney <​dominicc@chromium.org>

Reviewed by Dimitri Glazkov.

Crash when adding a text node to a shadow root
​https://bugs.webkit.org/show_bug.cgi?id=59304

  • fast/dom/shadow/append-child-text-expected.txt: Added.
  • fast/dom/shadow/append-child-text.html: Added.

2011-04-24 Dominic Cooney <​dominicc@chromium.org>

Reviewed by Dimitri Glazkov.

Crash when adding a text node to a shadow root
​https://bugs.webkit.org/show_bug.cgi?id=59304

Text nodes need to consult their host element for style.

Test: fast/dom/shadow/append-child-text.html

  • dom/Node.cpp: (WebCore::NodeRendererFactory::styleForRenderer): (WebCore::NodeRendererFactory::createRendererAndStyle): (WebCore::Node::styleForRenderer): forward to NodeRendererFactory
11:38 AM Component Model edited by Dimitri Glazkov
(diff)
11:38 AM Component Model edited by Dimitri Glazkov
(diff)
11:37 AM Component Model edited by Dimitri Glazkov
(diff)
11:36 AM Changeset in webkit [84758] by abarth@webkit.org
  • 15 edits in trunk

2011-04-24 Adam Barth <​abarth@webkit.org>

Reviewed by Eric Seidel.

Update Content-Security-Policy syntax to match new version of spec
​https://bugs.webkit.org/show_bug.cgi?id=59291

Update tests to use the new syntax.

  • http/tests/security/contentSecurityPolicy/eval-scripts-setInterval-allowed.html:
  • http/tests/security/contentSecurityPolicy/eval-scripts-setInterval-blocked.html:
  • http/tests/security/contentSecurityPolicy/eval-scripts-setTimeout-allowed.html:
  • http/tests/security/contentSecurityPolicy/eval-scripts-setTimeout-blocked.html:
  • http/tests/security/contentSecurityPolicy/image-allowed.html:
  • http/tests/security/contentSecurityPolicy/image-blocked.html:
  • http/tests/security/contentSecurityPolicy/inline-script-allowed.html:
  • http/tests/security/contentSecurityPolicy/style-allowed.html:
  • http/tests/security/contentSecurityPolicy/style-blocked.html:
  • http/tests/security/contentSecurityPolicy/xsl-allowed.php:
  • http/tests/security/contentSecurityPolicy/xsl-blocked.php:

2011-04-24 Adam Barth <​abarth@webkit.org>

Reviewed by Eric Seidel.

Update Content-Security-Policy syntax to match new version of spec
​https://bugs.webkit.org/show_bug.cgi?id=59291

Brandon removed disable-xss-protection in favor of unsafe-inline and
allow-eval in favor of unsafe-eval. This change in syntax also means
the options directive no longer exists.

  • page/ContentSecurityPolicy.cpp: (WebCore::CSPSourceList::allowInline): (WebCore::CSPSourceList::allowEval): (WebCore::CSPSourceList::CSPSourceList): (WebCore::CSPSourceList::parseSource): (WebCore::CSPSourceList::addSourceUnsafeInline): (WebCore::CSPSourceList::addSourceUnsafeEval): (WebCore::CSPDirective::allowInline): (WebCore::CSPDirective::allowEval): (WebCore::ContentSecurityPolicy::allowJavaScriptURLs): (WebCore::ContentSecurityPolicy::allowInlineEventHandlers): (WebCore::ContentSecurityPolicy::allowInlineScript): (WebCore::ContentSecurityPolicy::allowEval): (WebCore::ContentSecurityPolicy::addDirective):
  • page/ContentSecurityPolicy.h:
11:29 AM Changeset in webkit [84757] by mitz@apple.com
  • 3 edits
    1 add in trunk/Source

../WebCore: Manual test for <rdar://problem/9329741> Reproducible crash in WebChromeClient::invalidateContentsAndWindow() on simonscat.com
​https://bugs.webkit.org/show_bug.cgi?id=59299

Reviewed by Maciej Stachowiak.

  • manual-tests/back-from-document-with-scrollbar.html: Added.

../WebKit2: <rdar://problem/9329741> Reproducible crash in WebChromeClient::invalidateContentsAndWindow() on simonscat.com
​https://bugs.webkit.org/show_bug.cgi?id=59299

Reviewed by Maciej Stachowiak.

  • WebProcess/WebCoreSupport/WebChromeClient.cpp:

(WebKit::WebChromeClient::invalidateContentsAndWindow): Null-check the document.
(WebKit::WebChromeClient::invalidateContentsForSlowScroll): Ditto.

10:50 AM April 2011 Meeting edited by vestbo@webkit.org
(diff)
9:25 AM Component Model edited by Dimitri Glazkov
(diff)
9:25 AM Component Model edited by Dimitri Glazkov
(diff)
9:23 AM Component Model edited by Dimitri Glazkov
(diff)
9:02 AM Component Model created by Dimitri Glazkov
2:41 AM April 2011 Meeting edited by mjs@apple.com
(diff)
2:41 AM April 2011 Meeting edited by mjs@apple.com
(diff)
2:40 AM April 2011 Meeting edited by mjs@apple.com
(diff)
2:23 AM April 2011 Meeting edited by andreas.kling@nokia.com
(diff)
2:14 AM Changeset in webkit [84756] by abarth@webkit.org
  • 5 edits in trunk/LayoutTests

2011-04-24 Adam Barth <​abarth@webkit.org>

Reviewed by Maciej Stachowiak.

Test canonicalization of scheme:/host
​https://bugs.webkit.org/show_bug.cgi?id=59296

  • fast/url/script-tests/segments.js:
  • fast/url/script-tests/standard-url.js:
  • fast/url/segments-expected.txt:
  • fast/url/standard-url-expected.txt:
1:17 AM Changeset in webkit [84755] by commit-queue@webkit.org
  • 3 edits
    2 adds in trunk

2011-04-24 Rik Cabanier <​cabanier@adobe.com>

Reviewed by Simon Fraser.

Tables are not rendered correctly
​https://bugs.webkit.org/show_bug.cgi?id=59138

  • fast/table/auto-100-percent-width.html: Added.
  • platform/mac/fast/table/auto-100-percent-width-expected.txt: Added.

2011-04-24 Rik Cabanier <​cabanier@adobe.com>

Reviewed by Simon Fraser.

Tables are not rendered correctly
​https://bugs.webkit.org/show_bug.cgi?id=59138

Test: fast/table/auto-100-percent-width.html

  • rendering/AutoTableLayout.cpp: (WebCore::AutoTableLayout::computePreferredLogicalWidths):

Apr 23, 2011:

11:22 PM Changeset in webkit [84754] by bfulgham@webkit.org
  • 2 edits in trunk/WebKitLibraries

[WinCairo] Unreviewed build change.

Activate Geolocation API, as it will work with WinCairo assuming
an appropriate Geolocation Provider is registered.

  • win/tools/vsprops/FeatureDefinesCairo.vsprops: Activate the Geolocation features.
11:01 PM Changeset in webkit [84753] by Simon Fraser
  • 2 edits in trunk/Source/WebCore

2011-04-23 Simon Fraser <Simon Fraser>

Fix Windows build, which complains about unreachable code.

  • platform/ScrollableArea.cpp: (WebCore::ScrollableArea::hasLayerForHorizontalScrollbar): (WebCore::ScrollableArea::hasLayerForVerticalScrollbar): (WebCore::ScrollableArea::hasLayerForScrollCorner):
10:53 PM Changeset in webkit [84752] by Simon Fraser
  • 5 edits in trunk/Source/WebCore

2011-04-23 Simon Fraser <Simon Fraser>

Reviewed by Dan Bernstein.

Area under composited scrollbars not repainted in WebKit2 on scrolling
​https://bugs.webkit.org/show_bug.cgi?id=59294
<rdar://problem/9299062>

When computing the rect that can be copied on scrolling, we normally
exclude the areas occupied by overlay scrollbars to avoid smeared scrollbars;
the assumption is that other code will repaint these areas. However, when
scrollbars are in their own compositing layers, we can, and should
copy the entire area.

Not testable in DRT because it does not allow tests to use overlay scrollbars.

  • platform/ScrollView.h:
  • platform/ScrollView.cpp: (WebCore::ScrollView::rectToCopyOnScroll): New method, with code moved from scrollContents and fixed to look for scrollbars in layers. (WebCore::ScrollView::scrollContents): Call rectToCopyOnScroll().
  • platform/ScrollableArea.h:
  • platform/ScrollableArea.cpp: (WebCore::ScrollableArea::hasLayerForHorizontalScrollbar): (WebCore::ScrollableArea::hasLayerForVerticalScrollbar): (WebCore::ScrollableArea::hasLayerForScrollCorner): Methods we can call outside of the ACCELERATED_COMPOSITING #ifdef.
8:59 PM Changeset in webkit [84751] by commit-queue@webkit.org
  • 13 edits in trunk/Source

2011-04-23 Sheriff Bot <​webkit.review.bot@gmail.com>

Unreviewed, rolling out r84740.
​http://trac.webkit.org/changeset/84740
​https://bugs.webkit.org/show_bug.cgi?id=59290

change breaks apple internal builds and is incorrect
(Requested by smfr on #webkit).

  • WebCore.exp.in:
  • page/FrameView.cpp:
  • page/FrameView.h:
  • platform/ScrollTypes.h:
  • platform/ScrollableArea.h:
  • platform/mac/ScrollbarThemeMac.mm: (WebCore::ScrollbarThemeMac::paint):
  • platform/mac/WebCoreSystemInterface.h:
  • platform/mac/WebCoreSystemInterface.mm:

2011-04-23 Sheriff Bot <​webkit.review.bot@gmail.com>

Unreviewed, rolling out r84740.
​http://trac.webkit.org/changeset/84740
​https://bugs.webkit.org/show_bug.cgi?id=59290

change breaks apple internal builds and is incorrect
(Requested by smfr on #webkit).

  • WebCoreSupport/WebSystemInterface.mm: (InitWebCoreSystemInterface):

2011-04-23 Sheriff Bot <​webkit.review.bot@gmail.com>

Unreviewed, rolling out r84740.
​http://trac.webkit.org/changeset/84740
​https://bugs.webkit.org/show_bug.cgi?id=59290

change breaks apple internal builds and is incorrect
(Requested by smfr on #webkit).

  • WebProcess/WebCoreSupport/mac/WebSystemInterface.mm: (InitWebCoreSystemInterface):
8:56 PM Changeset in webkit [84750] by mitz@apple.com
  • 5 edits in trunk

<rdar://problem/8970549> WebFindOptionsAtWordStarts fails with Japanese words
​https://bugs.webkit.org/show_bug.cgi?id=59288

Reviewed by Simon Fraser.

Source/WebCore:

  • platform/text/mac/TextBoundaries.mm:

(WebCore::wordStringTokenizer): Added this helper function.
(WebCore::findNextWordFromIndex): Changed to use a CFStringTokenizer with kCFStringTokenizerUnitWord,
whose behavior is not locale-dependent.

LayoutTests:

  • editing/text-iterator/findString-expected.txt:
  • editing/text-iterator/findString.html:
7:24 PM April 2011 Meeting edited by noam.rosenthal@nokia.com
(diff)
7:12 PM April 2011 Meeting edited by yael.aharon@nokia.com
(diff)
7:10 PM April 2011 Meeting edited by sam@webkit.org
(diff)
7:08 PM April 2011 Meeting edited by alex@milowski.com
(diff)
7:07 PM April 2011 Meeting edited by alex@milowski.com
(diff)
6:31 PM Changeset in webkit [84749] by morrita@google.com
  • 5 edits
    2 adds in trunk

2011-04-23 Dominic Cooney <​dominicc@chromium.org>

Reviewed by Dimitri Glazkov.

[V8] Nodes in shadow DOM should not be GCed while their hosts are alive
​https://bugs.webkit.org/show_bug.cgi?id=59284

  • fast/dom/shadow/gc-shadow-expected.txt: Added.
  • fast/dom/shadow/gc-shadow.html: Added.

2011-04-23 Dominic Cooney <​dominicc@chromium.org>

Reviewed by Dimitri Glazkov.

[V8] Nodes in shadow DOM should not be GCed while their hosts are alive
​https://bugs.webkit.org/show_bug.cgi?id=59284

Test: fast/dom/shadow/gc-shadow.html

  • bindings/v8/V8GCController.cpp: (WebCore::calculateGroupId): group shadow nodes with their hosts

2011-04-23 Dominic Cooney <​dominicc@chromium.org>

Reviewed by Dimitri Glazkov.

[V8] Nodes in shadow DOM should not be GCed while their hosts are alive
​https://bugs.webkit.org/show_bug.cgi?id=59284

Chromium DRT over-refcounted the NPObject wrappers it used to
marshal shadowRoot nodes back to tests (WebBindings::makeNode
births NPObjects with referenceCount == 1). This masked the GC bug
59284 in tests by making shadowRoot nodes live forever.

  • DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::shadowRoot): (LayoutTestController::ensureShadowRoot):
6:05 PM Changeset in webkit [84748] by barraclough@apple.com
  • 3 edits in trunk/Source/JavaScriptCore

Bug 59287 - DFG JIT - Handle temporaries as vars, allowing support for ?:

Reviewed by Oliver Hunt.

SetLocals to temporaries will only be generated if they are used within other
blocks, due to the SSA based DCE.

  • dfg/DFGByteCodeParser.cpp:

(JSC::DFG::ByteCodeParser::ByteCodeParser):
(JSC::DFG::ByteCodeParser::get):
(JSC::DFG::ByteCodeParser::set):
(JSC::DFG::ByteCodeParser::getLocal):
(JSC::DFG::ByteCodeParser::setLocal):
(JSC::DFG::ByteCodeParser::parseBlock):
(JSC::DFG::ByteCodeParser::processPhiStack):
(JSC::DFG::ByteCodeParser::allocateVirtualRegisters):
(JSC::DFG::ByteCodeParser::parse):

  • dfg/DFGGraph.h:

(JSC::DFG::BasicBlock::BasicBlock):

5:51 PM Changeset in webkit [84747] by morrita@google.com
  • 4 edits in trunk/Source/WebCore

2011-04-23 MORITA Hajime <​morrita@google.com>

Reviewed by Dimitri Glazkov.

[Refactoring] NodeRenderParentDetector should be NodeRenderFactory
​https://bugs.webkit.org/show_bug.cgi?id=59280

  • Rename NodeRenderParentDetector to NodeRendererFactory
  • move createRendererAndStyle() and createRendererIfNeeded to NodeRendererFactory

No new test, no behavioral change.

  • dom/Node.cpp: (WebCore::NodeRendererFactory::NodeRendererFactory): (WebCore::NodeRendererFactory::document): (WebCore::NodeRendererFactory::nextRenderer): (WebCore::NodeRendererFactory::findVisualParent): (WebCore::NodeRendererFactory::shouldCreateRenderer): (WebCore::NodeRendererFactory::createRendererAndStyle): (WebCore::NodeRendererFactory::createRendererIfNeeded): (WebCore::Node::parentNodeForRenderingAndStyle): (WebCore::Node::createRendererIfNeeded):
  • dom/Node.h:
  • html/HTMLDetailsElement.cpp: (WebCore::DetailsSummaryElement::detailsElement): Added const_cast due to the chage on parentNodeForRenderingAndStyle()
5:17 PM Changeset in webkit [84746] by morrita@google.com
  • 2 edits in trunk/LayoutTests

2011-04-23 MORITA Hajime <​morrita@google.com>

Unreviewed test_expectations.txt update.
r84745 didn't fix meter-element.html break on chromium.

  • platform/chromium/test_expectations.txt:
4:16 PM Changeset in webkit [84745] by morrita@google.com
  • 4 edits
    6 adds in trunk

2011-04-23 MORITA Hajime <​morrita@google.com>

Reviewed by Kent Tamura.

REGRESSION: (non-Mac) <meter>'s bar part isn't rendered for with -webkit-writing-mode: vertical-lr from r82899
​https://bugs.webkit.org/show_bug.cgi?id=59281

  • fast/dom/HTMLMeterElement/meter-writing-mode.html: Added.
  • fast/dom/HTMLProgressElement/progress-writing-mode.html: Added.
  • platform/chromium/test_expectations.txt:
  • platform/mac/fast/dom/HTMLMeterElement/meter-writing-mode-expected.png: Added.
  • platform/mac/fast/dom/HTMLMeterElement/meter-writing-mode-expected.txt: Added.
  • platform/mac/fast/dom/HTMLProgressElement/progress-writing-mode-expected.png: Added.
  • platform/mac/fast/dom/HTMLProgressElement/progress-writing-mode-expected.txt: Added.

2011-04-23 MORITA Hajime <​morrita@google.com>

Reviewed by Kent Tamura.

REGRESSION: (non-Mac) <meter>'s bar part isn't rendered for with -webkit-writing-mode: vertical-lr from r82899
​https://bugs.webkit.org/show_bug.cgi?id=59281

Styles for -webkit-meter-bar and -webkit-progress-bar missed width property, which caused unexpected layout results
for vertical writing modes. This fix added a width property for each of them.

Tests: fast/dom/HTMLMeterElement/meter-writing-mode.html

fast/dom/HTMLProgressElement/progress-writing-mode.html

  • css/html.css: (meter::-webkit-meter-bar): (progress::-webkit-progress-bar):
1:16 PM Changeset in webkit [84744] by barraclough@apple.com
  • 8 edits in trunk/Source/JavaScriptCore

Bug 59266 - DFG JIT - Add SSA style DCE

Reviewed by Sam Weinig & Geoff Garen.

This works by making GetLocal nodes reference SetLocal nodes from prior blocks,
via intermediate Phi nodes. Whenever we add a GetLocal to the graph, also add a
matching child Phi, and add the Phi to a work queue to add references to prior
definitions once we have the full CFG & can determine predecessors. This process
is iterative, inserting new phis into predecessors as necessary.

  • dfg/DFGByteCodeParser.cpp:

(JSC::DFG::ByteCodeParser::getVariable):
(JSC::DFG::ByteCodeParser::setVariable):
(JSC::DFG::ByteCodeParser::getArgument):
(JSC::DFG::ByteCodeParser::setArgument):
(JSC::DFG::ByteCodeParser::parseBlock):
(JSC::DFG::ByteCodeParser::processWorkQueue):
(JSC::DFG::ByteCodeParser::allocateVirtualRegisters):
(JSC::DFG::ByteCodeParser::parse):

  • dfg/DFGGraph.cpp:

(JSC::DFG::Graph::dump):
(JSC::DFG::Graph::refChildren):

  • dfg/DFGGraph.h:

(JSC::DFG::Graph::ref):

  • dfg/DFGNode.h:

(JSC::DFG::Node::ref):

  • dfg/DFGNonSpeculativeJIT.cpp:

(JSC::DFG::NonSpeculativeJIT::compile):

  • dfg/DFGScoreBoard.h:

(JSC::DFG::ScoreBoard::~ScoreBoard):
(JSC::DFG::ScoreBoard::dump):

  • dfg/DFGSpeculativeJIT.cpp:

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

1:05 PM Changeset in webkit [84743] by kevino@webkit.org
  • 7 edits in trunk

[wx] Unreviewed build fixes for recent trunk changes.

12:14 PM Changeset in webkit [84742] by ap@apple.com
  • 10 edits in trunk

2011-04-23 Alexey Proskuryakov <​ap@apple.com>

Unreviewed.

REGRESSION (r80812): window.print();window.close() doesn't work in WebKit2
​https://bugs.webkit.org/show_bug.cgi?id=59241
<rdar://problem/9150861>

  • http/tests/misc/will-send-request-returns-null-on-redirect-expected.txt:
  • http/tests/security/XFrameOptions/x-frame-options-deny-expected.txt:
  • http/tests/security/XFrameOptions/x-frame-options-parent-same-origin-deny-expected.txt: Resource load delegates for cancelled loads are now correctly sent before frame load delegate's didFinishLoading, so cancellation errors get logged by DumpRenderTree.

2011-04-23 Alexey Proskuryakov <​ap@apple.com>

Reviewed by Maciej Stachowiak.

REGRESSION (r80812): window.print();window.close() doesn't work in WebKit2
​https://bugs.webkit.org/show_bug.cgi?id=59241
<rdar://problem/9150861>

Covered by manual-tests/print-after-window-close.html.

Re-fixing <​https://bugs.webkit.org/show_bug.cgi?id=51357> in a way that doesn't cause the
regression.

  • loader/MainResourceLoader.cpp: (WebCore::MainResourceLoader::didCancel): We shouldn't be doing complicated work while the loader is half-canceled.
  • manual-tests/print-after-window-close.html: Extended to actually print a non-empty page, thus being more practical.
  • page/Chrome.cpp: (WebCore::Chrome::print): Added a FIXME about a need for PageGroupLoadDeferrer. It's too scary for me to try now.
12:11 AM Changeset in webkit [84741] by vitalyr@chromium.org
  • 6 edits in trunk/Source/WebCore

2011-04-19 Vitaly Repeshko <​vitalyr@chromium.org>

Reviewed by Adam Barth.

[V8] Use implicit references for V8 listeners on DOM nodes.
​https://bugs.webkit.org/show_bug.cgi?id=58953

Instead of allocating an auxiliary V8 array referencing V8
listener objects associated with a DOM node and using an extra
pointer in every DOM node wrapper, we can register implicit
references between nodes and their listeners during GC. This also
makes V8 bindings more aligned with JSC bindings.

No new tests because this is a refactoring.

  • bindings/scripts/CodeGeneratorV8.pm: Stopped generating the listener cache internal field for DOM nodes.
  • bindings/v8/V8AbstractEventListener.h: Exposed the lister handle to the GC. (WebCore::V8AbstractEventListener::existingListenerObjectPeristentHandle):
  • bindings/v8/V8GCController.cpp: Started using implicit references. (WebCore::GrouperVisitor::visitDOMWrapper):
  • dom/EventTarget.h: Implemented an iterator over all listeners. (WebCore::EventListenerIterator):
Note: See TracTimeline for information about the timeline view.