Timeline
May 13, 2015:
- 10:59 PM Changeset in webkit [184330] by
-
- 2 edits in trunk/Source/WebKit2
[SOUP] Network Cache: NetworkProcess segfault when file system doesn't support xattrs
https://bugs.webkit.org/show_bug.cgi?id=144953
Reviewed by Martin Robinson.
Return early if we fail to get the birthtime xattr.
- NetworkProcess/cache/NetworkCacheFileSystemPosix.h:
(WebKit::NetworkCache::fileTimes):
- 10:10 PM Changeset in webkit [184329] by
-
- 5 edits in trunk/Source/WebCore
Get the ScriptController from the correct frame for media elements and plug-ins
https://bugs.webkit.org/show_bug.cgi?id=144983
rdar://problem/20692642&19943135
Reviewed by Sam Weinig.
HTMLMediaElement, QuickTimePluginReplacement and HTMLPlugInImageElement were
getting the main frame's ScriptController instead of the one for their frame.
This caused media controls JS to be running in the context of the main frame,
which broke media controls which use getCSSCanvasContext() and -webkit-canvas.
Fix by getting the frame via the element's document.
Also undo r180584 which was working around this bug.
- Modules/mediacontrols/mediaControlsiOS.js:
(ControllerIOS.prototype.drawTimelineBackground):
- Modules/plugins/QuickTimePluginReplacement.mm:
(WebCore::QuickTimePluginReplacement::ensureReplacementScriptInjected):
(WebCore::QuickTimePluginReplacement::installReplacement):
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::updateCaptionContainer):
(WebCore::HTMLMediaElement::ensureMediaControlsInjectedScript):
(WebCore::HTMLMediaElement::didAddUserAgentShadowRoot):
(WebCore::HTMLMediaElement::pageScaleFactorChanged):
- html/HTMLPlugInImageElement.cpp:
(WebCore::HTMLPlugInImageElement::didAddUserAgentShadowRoot):
- 9:19 PM Changeset in webkit [184328] by
-
- 12 edits1 add in trunk/Source/JavaScriptCore
REGRESSION(r180595): same-callee profiling no longer works
https://bugs.webkit.org/show_bug.cgi?id=144787
Reviewed by Filip Pizlo.
This patch introduces a DFG optimization to use NewObject node when the callee of op_create_this is
always the same JSFunction. This condition doesn't hold when the byte code creates multiple
JSFunction objects at runtime as in: function y() { return function () {} }; new y(); new y();
To enable this optimization, LLint and baseline JIT now store the last callee we saw in the newly
added fourth operand of op_create_this. We use this JSFunction's structure in DFG after verifying
our speculation that the callee is the same. To avoid recompiling the same code for different callee
objects in the polymorphic case, the special value of seenMultipleCalleeObjects() is set in
LLint and baseline JIT when multiple callees are observed.
Tests: stress/create-this-with-callee-variants.js
- bytecode/BytecodeList.json: Increased the number of operands to 5.
- bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpBytecode): Dump the newly added callee cache.
(JSC::CodeBlock::finalizeUnconditionally): Clear the callee cache if the callee is no longer alive.
- bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::emitCreateThis): Add the instruction to propertyAccessInstructions so that
we can clear the callee cache in CodeBlock::finalizeUnconditionally. Also initialize the newly added
operand.
- dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::parseBlock): Implement the optimization. Speculate the actual callee to
match the cache. Use the cached callee's structure if the speculation succeeds. Otherwise, OSR exit.
- jit/JITOpcodes.cpp:
(JSC::JIT::emit_op_create_this): Go to the slow path to update the cache unless it's already marked
as seenMultipleCalleeObjects() to indicate the polymorphic behavior and/or we've OSR exited here.
(JSC::JIT::emitSlow_op_create_this):
- jit/JITOpcodes32_64.cpp:
(JSC::JIT::emit_op_create_this): Ditto.
(JSC::JIT::emitSlow_op_create_this):
- llint/LowLevelInterpreter32_64.asm:
(_llint_op_create_this): Ditto.
- llint/LowLevelInterpreter64.asm:
(_llint_op_create_this): Ditto.
- runtime/CommonSlowPaths.cpp:
(slow_path_create_this): Set the callee cache to the actual callee if it's not set. If the cache has
been set to a JSFunction* different from the actual callee, set it to seenMultipleCalleeObjects().
- runtime/JSCell.h:
(JSC::JSCell::seenMultipleCalleeObjects): Added.
- runtime/WriteBarrier.h:
(JSC::WriteBarrierBase::unvalidatedGet): Removed the compile guard around it.
- tests/stress/create-this-with-callee-variants.js: Added.
- 9:07 PM Changeset in webkit [184327] by
-
- 3 edits in trunk/Source/WebCore
Fix trivial typos in ApplyBlockElementCommand
https://bugs.webkit.org/show_bug.cgi?id=144984
Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-13
Reviewed by Alexey Proskuryakov.
No new tests, no behavior change.
- editing/ApplyBlockElementCommand.cpp:
(WebCore::ApplyBlockElementCommand::formatSelection):
(WebCore::ApplyBlockElementCommand::endOfNextParagraphSplittingTextNodesIfNeeded):
(WebCore::ApplyBlockElementCommand::endOfNextParagrahSplittingTextNodesIfNeeded): Deleted.
- editing/ApplyBlockElementCommand.h:
- 8:07 PM Changeset in webkit [184326] by
-
- 13 edits in trunk/Source/WebKit2
Rename ProcessThrottlerClient::{send, cancel}ProcessWillSuspend() to {send, cancel}PrepareToSuspend()
https://bugs.webkit.org/show_bug.cgi?id=144619
<rdar://problem/20812779>
Reviewed by Andy Estes.
The names of the functions ProcessThrottlerClient::{send, cancel}ProcessWillSuspend() are misnomers. These
functions are called when the ProcessThrottler wants to prepare the process that it manages for suspension
and changes its mind, respectively. That is, these functions do not actually correspond to the OS decision
to suspend a process or cancel the suspension of a process, respectively. So, rename these functions and
associated {Network, Web}ProcessProxy message names to better describe their purpose.
- NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::prepareToSuspend):
(WebKit::NetworkProcess::cancelPrepareToSuspend):
(WebKit::NetworkProcess::processWillSuspend): Deleted.
(WebKit::NetworkProcess::cancelProcessWillSuspend): Deleted.
- NetworkProcess/NetworkProcess.h:
- NetworkProcess/NetworkProcess.messages.in:
- UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::sendPrepareToSuspend):
(WebKit::NetworkProcessProxy::sendCancelPrepareToSuspend):
(WebKit::NetworkProcessProxy::sendProcessWillSuspend): Deleted.
(WebKit::NetworkProcessProxy::sendCancelProcessWillSuspend): Deleted.
- UIProcess/Network/NetworkProcessProxy.h:
- UIProcess/ProcessThrottler.cpp:
(WebKit::ProcessThrottler::updateAssertion):
- UIProcess/ProcessThrottlerClient.h:
- UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcess::actualPrepareToSuspend): Formerly named prepareToSuspend.
(WebKit::WebProcessProxy::sendPrepareToSuspend):
(WebKit::WebProcessProxy::sendCancelPrepareToSuspend):
(WebKit::WebProcessProxy::sendProcessWillSuspend): Deleted.
(WebKit::WebProcessProxy::sendCancelProcessWillSuspend): Deleted.
- UIProcess/WebProcessProxy.h:
- WebProcess/WebProcess.cpp:
(WebKit::WebProcess::prepareToSuspend):
(WebKit::WebProcess::cancelPrepareToSuspend):
(WebKit::WebProcess::processWillSuspend): Deleted.
(WebKit::WebProcess::cancelProcessWillSuspend): Deleted.
- WebProcess/WebProcess.h:
- WebProcess/WebProcess.messages.in:
- 6:34 PM Changeset in webkit [184325] by
-
- 7 edits in trunk/Source
Clean up some possible RefPtr to PassRefPtr churn
https://bugs.webkit.org/show_bug.cgi?id=144779
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2015-05-13
Reviewed by Darin Adler.
- runtime/GenericTypedArrayViewInlines.h:
(JSC::GenericTypedArrayView<Adaptor>::create):
(JSC::GenericTypedArrayView<Adaptor>::createUninitialized):
- runtime/JSArrayBufferConstructor.cpp:
(JSC::constructArrayBuffer):
- runtime/Structure.cpp:
(JSC::Structure::toStructureShape):
- runtime/TypedArrayBase.h:
(JSC::TypedArrayBase::create):
(JSC::TypedArrayBase::createUninitialized):
- tools/FunctionOverrides.cpp:
(JSC::initializeOverrideInfo):
Release the last use of a RefPtr as it is passed on.
- 6:32 PM Changeset in webkit [184324] by
-
- 37 edits3 adds in trunk
ES6: Allow duplicate property names
https://bugs.webkit.org/show_bug.cgi?id=142895
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2015-05-13
Reviewed by Geoffrey Garen.
Source/JavaScriptCore:
Introduce new
op_put_getter_by_idandop_put_setter_by_idopcodes
that will define a single getter or setter property on an object.
The existing
op_put_getter_setteropcode is still preferred for
putting both a getter and setter at the same time but cannot be used
for putting an individual getter or setter which is needed in
some cases.
Add a new slow path when generating bytecodes for a property list
with computed properties, as computed properties are the only time
the list of properties cannot be determined statically.
- bytecompiler/NodesCodegen.cpp:
(JSC::PropertyListNode::emitBytecode):
- fast path for all constant properties
- slow but paired getter/setter path if there are no computed properties
- slow path, individual put operation for every property, if there are computed properties
- parser/Nodes.h:
Distinguish a Computed property from a Constant property.
- parser/Parser.cpp:
(JSC::Parser<LexerType>::parseProperty):
(JSC::Parser<LexerType>::parsePropertyMethod):
Distingish Computed and Constant properties.
(JSC::Parser<LexerType>::parseObjectLiteral):
When we drop into strict mode it is because we saw a getter
or setter, so be more explicit.
(JSC::Parser<LexerType>::parseStrictObjectLiteral):
Eliminate duplicate property syntax error exception.
- parser/SyntaxChecker.h:
(JSC::SyntaxChecker::getName):
- parser/ASTBuilder.h:
(JSC::ASTBuilder::getName): Deleted.
No longer used.
- runtime/JSObject.h:
(JSC::JSObject::putDirectInternal):
When updating a property. If the Accessor attribute changed
update the Structure.
- runtime/JSObject.cpp:
(JSC::JSObject::putGetter):
(JSC::JSObject::putSetter):
Called by the opcodes, just perform the same operation that
defineGetter or defineSetter would do.
(JSC::JSObject::putDirectNonIndexAccessor):
This transition is now handled in putDirectInternal.
- runtime/Structure.h:
Add needed export.
- bytecode/BytecodeList.json:
- bytecode/BytecodeUseDef.h:
(JSC::computeUsesForBytecodeOffset):
(JSC::computeDefsForBytecodeOffset):
- bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpBytecode):
- bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::emitPutGetterById):
(JSC::BytecodeGenerator::emitPutSetterById):
- bytecompiler/BytecodeGenerator.h:
- jit/JIT.cpp:
(JSC::JIT::privateCompileMainPass):
- jit/JIT.h:
- jit/JITInlines.h:
(JSC::JIT::callOperation):
- jit/JITOperations.cpp:
- jit/JITOperations.h:
- jit/JITPropertyAccess.cpp:
(JSC::JIT::emit_op_put_getter_by_id):
(JSC::JIT::emit_op_put_setter_by_id):
- jit/JITPropertyAccess32_64.cpp:
(JSC::JIT::emit_op_put_getter_by_id):
(JSC::JIT::emit_op_put_setter_by_id):
- llint/LLIntSlowPaths.cpp:
(JSC::LLInt::LLINT_SLOW_PATH_DECL):
- llint/LLIntSlowPaths.h:
- llint/LowLevelInterpreter.asm:
New bytecodes. Modelled after existing op_put_getter_setter.
LayoutTests:
- js/object-literal-duplicate-properties-expected.txt: Added.
- js/object-literal-duplicate-properties.html: Added.
- js/script-tests/object-literal-duplicate-properties.js: Added.
Include a new test all about testing duplicate property names
and their expected cascading results.
- ietestcenter/Javascript/11.1.5_4-4-b-1-expected.txt:
- ietestcenter/Javascript/11.1.5_4-4-b-2-expected.txt:
- ietestcenter/Javascript/11.1.5_4-4-c-1-expected.txt:
- ietestcenter/Javascript/11.1.5_4-4-c-2-expected.txt:
- ietestcenter/Javascript/11.1.5_4-4-d-1-expected.txt:
- ietestcenter/Javascript/11.1.5_4-4-d-2-expected.txt:
- ietestcenter/Javascript/11.1.5_4-4-d-3-expected.txt:
- ietestcenter/Javascript/11.1.5_4-4-d-4-expected.txt:
ES5 behavior for duplciate properties has changed.
- js/mozilla/strict/11.1.5-expected.txt:
- js/object-literal-syntax-expected.txt:
- js/script-tests/object-literal-syntax.js:
Update other tests and values now that duplicate properties
are allowed, and their cascade order behaves correctly.
- 6:21 PM Changeset in webkit [184323] by
-
- 2 edits in trunk/Source/WebCore
REGRESION (r179958): Crash in WebCore::DocumentLoader::detachFromFrame when -[id<WebPolicyDelegate> decidePolicyForMIMEType:request:frame:decisionListener:] fails to call -[id<WebPolicyDecisionListener> download|ignore|use]
<http://webkit.org/b/144975>
Reviewed by Andy Estes.
This change reverts r179958. It changes RELEASE_ASSERT*()
statements back to Debug-only ASSERT*() statements.
- loader/DocumentLoader.cpp:
(WebCore::DocumentLoader::~DocumentLoader):
(WebCore::DocumentLoader::continueAfterContentPolicy):
(WebCore::DocumentLoader::detachFromFrame):
- 5:39 PM Changeset in webkit [184322] by
-
- 5 edits in branches/safari-600.5.17-branch/Source
Versioning.
- 5:36 PM Changeset in webkit [184321] by
-
- 5 edits in branches/safari-601.1.32-branch/Source
Versioning.
- 5:11 PM Changeset in webkit [184320] by
-
- 2 edits in branches/safari-600.7-branch/LayoutTests
Add a Pass/Failure test expectation on
fast/canvas/webgl/tex-image-and-sub-image-2d-with-potentially-subsampled-image.html.
I don't know why this test expectation was not in the branch given it was added back in r174585.
Perhaps it got lost during some merges.
- platform/mac/TestExpectations:
- 5:08 PM Changeset in webkit [184319] by
-
- 1 copy in tags/Safari-601.1.32.2
New tag.
- 4:57 PM Changeset in webkit [184318] by
-
- 7 edits1 add in trunk/Source/JavaScriptCore
Creating a new blank document in icloud pages causes an AI error: Abstract value (CellBytecodedoubleBoolOther, TOP, TOP) for double node has type outside SpecFullDouble.
https://bugs.webkit.org/show_bug.cgi?id=144856
Reviewed by Benjamin Poulain.
First I made fixTypeForRepresentation() print out better diagnostics when it dies.
Then I fixed the bug: Node::convertToIdentityOn(Node*) needs to make sure that when it
converts to a representation-changing node, it needs to use one of the UseKinds that such
a node expects. For example, DoubleRep(UntypedUse:) doesn't make sense; it needs to be
something like DoubleRep(NumberUse:) since it will speculate that the input is a number.
- dfg/DFGAbstractInterpreter.h:
(JSC::DFG::AbstractInterpreter::setBuiltInConstant):
- dfg/DFGAbstractInterpreterInlines.h:
(JSC::DFG::AbstractInterpreter<AbstractStateType>::executeEffects):
- dfg/DFGAbstractValue.cpp:
(JSC::DFG::AbstractValue::fixTypeForRepresentation):
- dfg/DFGAbstractValue.h:
- dfg/DFGInPlaceAbstractState.cpp:
(JSC::DFG::InPlaceAbstractState::initialize):
- dfg/DFGNode.cpp:
(JSC::DFG::Node::convertToIdentityOn):
- tests/stress/cloned-arguments-get-by-val-double-array.js: Added.
(foo):
- 4:33 PM Changeset in webkit [184317] by
-
- 3 edits in trunk/Source/JavaScriptCore
Unreviewed, rolling out r184313.
https://bugs.webkit.org/show_bug.cgi?id=144974
Introduced an assertion failure in class-syntax-
declaration.js, class-syntax-expression.js, and object-
literal-syntax.js (Requested by rniwa on #webkit).
Reverted changeset:
"Small refactoring before ES6 Arrow function implementation."
https://bugs.webkit.org/show_bug.cgi?id=144954
http://trac.webkit.org/changeset/184313
- 4:18 PM Changeset in webkit [184316] by
-
- 7 edits in trunk/Source
Source/JavaScriptCore:
Ensure that all the smart pointer types in WTF clear their pointer before deref
https://bugs.webkit.org/show_bug.cgi?id=143789
Reviewed by Ryosuke Niwa.
One of the simpler cases of this in JavaScriptCore. There
are other cases where we need to guard the derefs but they
are more complex cases.
- inspector/JSInjectedScriptHost.cpp:
(Inspector::JSInjectedScriptHost::releaseImpl):
- inspector/JSJavaScriptCallFrame.cpp:
(Inspector::JSJavaScriptCallFrame::releaseImpl):
Source/WTF:
Ensure that all the smart pointer types in WTF clear their pointer before deref
https://bugs.webkit.org/show_bug.cgi?id=143789
Reviewed by Ryosuke Niwa.
In order to prevent use after free bugs caused by destructors
that end up trying to access the smart pointer itself, we should
make sure we always clear the m_ptr field before calling deref.
Essentially the UaF path is:
struct Foo : RefCounted<Foo> {
Wibble* m_wibble;
void doSomething();
~Foo() { m_wibble->doSomethingLikeCleanup(); }
};
struct Wibble {
void doSomethingLikeCleanup()
{
if (m_foo) {
/* if this branch is not here we get a null deref */
m_foo->doSomething();
}
}
void replaceFoo(Foo* foo) { m_foo = foo; }
RefPtr<Foo> m_foo;
};
Wibble* someWibble = /* a Wibble with m_foo->m_refCount == 1 */;
/* and m_foo points to someWibble */;
someWibble->replaceFoo(someOtherFoo);
+ someWibble->m_foo->m_ptr->deref();
+ someWibble->m_foo->m_ptr->~Foo()
+ someWibble->m_foo->m_ptr->m_wibble->doSomethingLikeCleanup()
+ someWibble->m_foo->m_ptr->m_wibble /* someWibble */ ->m_foo->m_ptr /*logically dead*/ ->doSomething()
By clearing m_ptr first we either force a null pointer deref or
we force our code down a path that does not use the dead smart
pointer.
- wtf/PassRefPtr.h: (WTF::PassRefPtr::~PassRefPtr):
- wtf/Ref.h: (WTF::Ref::~Ref): (WTF::Ref::operator=):
- wtf/RefPtr.h: (WTF::RefPtr::~RefPtr):
- wtf/RetainPtr.h: (WTF::RetainPtr::~RetainPtr): (WTF::RetainPtr<T>::clear):
- 4:09 PM Changeset in webkit [184315] by
-
- 31 edits5 adds in trunk
Cached CSS image resources don't show up after reloading <http://nightly.webkit.org/start/>
https://bugs.webkit.org/show_bug.cgi?id=144952
Source/WebCore:
rdar://problem/13387307
Reviewed by Oliver Hunt.
This is a symptom of a general problem that we don't revalidate subresources of cached parsed stylesheets.
Fix by tightening the check we perform when choosing to used the cached sheet. If there are expired subresources
we reparse the sheet.
Test: http/tests/cache/stylesheet-sharing.html
- css/CSSCrossfadeValue.cpp:
(WebCore::CSSCrossfadeValue::traverseSubresources):
(WebCore::CSSCrossfadeValue::hasFailedOrCanceledSubresources): Deleted.
Replace hasFailedOrCanceledSubresources with general purpose subresource traversal functions.
- css/CSSCrossfadeValue.h:
- css/CSSFilterImageValue.cpp:
(WebCore::CSSFilterImageValue::traverseSubresources):
(WebCore::CSSFilterImageValue::hasFailedOrCanceledSubresources): Deleted.
- css/CSSFilterImageValue.h:
- css/CSSFontFaceSrcValue.cpp:
(WebCore::CSSFontFaceSrcValue::traverseSubresources):
(WebCore::CSSFontFaceSrcValue::hasFailedOrCanceledSubresources): Deleted.
- css/CSSFontFaceSrcValue.h:
- css/CSSImageSetValue.cpp:
(WebCore::CSSImageSetValue::traverseSubresources):
(WebCore::CSSImageSetValue::hasFailedOrCanceledSubresources): Deleted.
- css/CSSImageSetValue.h:
- css/CSSImageValue.cpp:
(WebCore::CSSImageValue::traverseSubresources):
(WebCore::CSSImageValue::hasFailedOrCanceledSubresources): Deleted.
- css/CSSImageValue.h:
- css/CSSValue.cpp:
(WebCore::CSSValue::traverseSubresources):
(WebCore::CSSValue::hasFailedOrCanceledSubresources): Deleted.
- css/CSSValue.h:
- css/CSSValueList.cpp:
(WebCore::CSSValueList::traverseSubresources):
(WebCore::CSSValueList::hasFailedOrCanceledSubresources): Deleted.
- css/CSSValueList.h:
- css/StyleProperties.cpp:
(WebCore::StyleProperties::traverseSubresources):
(WebCore::StyleProperties::hasFailedOrCanceledSubresources): Deleted.
- css/StyleProperties.h:
- css/StyleSheetContents.cpp:
(WebCore::traverseSubresourcesInRules):
(WebCore::StyleSheetContents::traverseSubresources):
(WebCore::StyleSheetContents::subresourcesAllowReuse):
Disallow reuse if there are expired subresources.
(WebCore::StyleSheetContents::isLoadingSubresources):
Testing support.
(WebCore::childRulesHaveFailedOrCanceledSubresources): Deleted.
(WebCore::StyleSheetContents::hasFailedOrCanceledSubresources): Deleted.
- css/StyleSheetContents.h:
(WebCore::StyleSheetContents::loadCompleted):
- html/HTMLLinkElement.cpp:
(WebCore::HTMLLinkElement::setCSSStyleSheet):
- loader/cache/CachedCSSStyleSheet.cpp:
(WebCore::CachedCSSStyleSheet::restoreParsedStyleSheet):
- loader/cache/CachedCSSStyleSheet.h:
- loader/cache/CachedImage.cpp:
(WebCore::CachedImage::makeRevalidationDecision):
(WebCore::CachedImage::mustRevalidateDueToCacheHeaders): Deleted.
Move the logging code out from this function (it requires frame access this function doesn't otherwise need)
and refactor to return a decision enum.
- loader/cache/CachedImage.h:
- loader/cache/CachedResource.cpp:
(WebCore::CachedResource::makeRevalidationDecision):
(WebCore::logResourceRevalidationReason): Deleted.
(WebCore::CachedResource::mustRevalidateDueToCacheHeaders): Deleted.
- loader/cache/CachedResource.h:
(WebCore::CachedResource::loadFailedOrCanceled):
- loader/cache/CachedResourceLoader.cpp:
(WebCore::logRevalidation):
(WebCore::logResourceRevalidationDecision):
(WebCore::CachedResourceLoader::determineRevalidationPolicy):
Move logging here.
- testing/Internals.cpp:
(WebCore::Internals::isSharingStyleSheetContents):
(WebCore::Internals::isStyleSheetLoadingSubresources):
- testing/Internals.h:
- testing/Internals.idl:
LayoutTests:
Reviewed by Oliver Hunt.
- http/tests/cache/resources/non-shareable.css: Added.
(#foo):
(#bar):
(#test1):
(#test2):
- http/tests/cache/resources/shareable.css: Added.
(#foo):
(#bar):
(#test1):
(#test2):
- http/tests/cache/resources/stylesheet-html.php: Added.
- http/tests/cache/stylesheet-sharing-expected.txt: Added.
- http/tests/cache/stylesheet-sharing.html: Added.
- 3:27 PM Changeset in webkit [184314] by
-
- 3 edits in trunk/Tools
[Content Extensions] Test interactions between multiple extensions and multiple domains.
https://bugs.webkit.org/show_bug.cgi?id=144967
Patch by Alex Christensen <achristensen@webkit.org> on 2015-05-13
Reviewed by Benjamin Poulain.
- DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
Xcode wanted to fix an alphabetization issue.
- TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:
(TestWebKitAPI::TEST_F):
Test interactions that worked but were not explicitly tested before.
- 3:23 PM Changeset in webkit [184313] by
-
- 3 edits in trunk/Source/JavaScriptCore
Small refactoring before ES6 Arrow function implementation.
https://bugs.webkit.org/show_bug.cgi?id=144954
Patch by Alexandr Skachkov <gskachkov@gmail.com> on 2015-05-13
Reviewed by Filip Pizlo.
- parser/Parser.h:
- parser/Parser.cpp:
- 3:16 PM Changeset in webkit [184312] by
-
- 6 edits in trunk/Source/WebCore
[CoordinatedGraphics] Remove scaleFactor from SurfaceUpdateInfo
https://bugs.webkit.org/show_bug.cgi?id=144935
Reviewed by Darin Adler.
The members of SurfaceUpdateInfo are only used to update tile except scaleFactor.
So, this patch removes scaleFactor from SurfaceUpdateInfo.
In addition, removes unnecessary parameters in createTile()
No new tests because there is no behavior change.
- platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.cpp:
(WebCore::CoordinatedGraphicsLayer::createTile):
- platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.h:
- platform/graphics/texmap/coordinated/CoordinatedTile.cpp:
(WebCore::CoordinatedTile::updateBackBuffer):
- platform/graphics/texmap/coordinated/CoordinatedTile.h:
- platform/graphics/texmap/coordinated/SurfaceUpdateInfo.h:
- 3:14 PM Changeset in webkit [184311] by
-
- 8 edits3 adds in trunk/Source/JavaScriptCore
The liveness pruning done by ObjectAllocationSinkingPhase ignores the possibility of an object's bytecode liveness being longer than its DFG liveness
https://bugs.webkit.org/show_bug.cgi?id=144945
Reviewed by Michael Saboff.
We were making the mistake of using DFG liveness for object allocation sinking decisions.
This is wrong. In fact we almost never want to use DFG liveness directly. The only place
where that makes sense is pruning in DFG AI.
So, I created a CombinedLiveness class that combines the DFG liveness with bytecode
liveness.
In the process of doing this, I realized that the DFGForAllKills definition of combined
liveness at block tail was not strictly right; it was using the bytecode liveness at the
block terminal instead of the union of the bytecode live-at-heads of successor blocks. So,
I changed DFGForAllKills to work in terms of CombinedLiveness.
This allows me to unskip the test I added in r184260. I also added a new test that tries to
trigger this bug more directly.
- CMakeLists.txt:
- JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
- JavaScriptCore.xcodeproj/project.pbxproj:
- dfg/DFGArgumentsEliminationPhase.cpp:
- dfg/DFGCombinedLiveness.cpp: Added.
(JSC::DFG::liveNodesAtHead):
(JSC::DFG::CombinedLiveness::CombinedLiveness):
- dfg/DFGCombinedLiveness.h: Added.
(JSC::DFG::CombinedLiveness::CombinedLiveness):
- dfg/DFGForAllKills.h:
(JSC::DFG::forAllKillsInBlock):
(JSC::DFG::forAllLiveNodesAtTail): Deleted.
- dfg/DFGObjectAllocationSinkingPhase.cpp:
(JSC::DFG::ObjectAllocationSinkingPhase::performSinking):
(JSC::DFG::ObjectAllocationSinkingPhase::determineMaterializationPoints):
(JSC::DFG::ObjectAllocationSinkingPhase::placeMaterializationPoints):
(JSC::DFG::ObjectAllocationSinkingPhase::promoteSunkenFields):
- tests/stress/escape-object-in-diamond-then-exit.js: Added.
- tests/stress/sink-object-past-invalid-check-sneaky.js:
- 3:07 PM Changeset in webkit [184310] by
-
- 8 edits in trunk/Source/WebKit2
Don't create a per-pool data store when using the modern API
https://bugs.webkit.org/show_bug.cgi?id=144963
rdar://problem/20331756
Reviewed by Tim Horton.
- UIProcess/API/APIProcessPoolConfiguration.cpp:
(API::ProcessPoolConfiguration::createWithLegacyOptions):
(API::ProcessPoolConfiguration::copy):
- UIProcess/API/APIProcessPoolConfiguration.h:
Keep track of whether the process pool should have a data store.
- UIProcess/API/C/WKContext.cpp:
(WKContextGetWebsiteDataStore):
- UIProcess/WebKeyValueStorageManager.cpp:
(WebKit::WebKeyValueStorageManager::getKeyValueStorageOrigins):
(WebKit::WebKeyValueStorageManager::getStorageDetailsByOrigin):
(WebKit::WebKeyValueStorageManager::deleteEntriesForOrigin):
(WebKit::WebKeyValueStorageManager::deleteAllEntries):
Update now that WebProcessPool::dataStore() no longer returns a reference.
- UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::WebProcessPool):
Only create a data store if the configuration states that we should.
- UIProcess/WebProcessPool.h:
Change dataStore() to return a pointer instead of a reference.
- 3:06 PM Changeset in webkit [184309] by
-
- 3 edits in trunk/Source/WebCore
Modernize ContainerNode::childElementCount
https://bugs.webkit.org/show_bug.cgi?id=144930
Patch by Sam Weinig <sam@webkit.org> on 2015-05-13
Reviewed by Darin Adler.
- dom/ContainerNode.cpp:
(WebCore::ContainerNode::childElementCount):
Use std::distance to compute the number of child elements.
- dom/ElementChildIterator.h:
Add typedefs to make the child element iterators conform STL standards.
- 2:58 PM Changeset in webkit [184308] by
-
- 5 edits2 adds in trunk
REGRESSION(r183770): Crash inside WebEditorClient::shouldApplyStyle when applying underline
https://bugs.webkit.org/show_bug.cgi?id=144949
Source/WebCore:
<rdar://problem/20895753>
Reviewed by Darin Adler.
The crash was caused by the variant of applyStyleToSelection that takes EditingStyle passing
a null pointer to shouldApplyStyle when we're only applying text decoration changes so that
m_mutableStyle in the editing style is null. This didn't reproduce in execCommand since we
wouldn't call shouldApplyStyle in that case. It didn't reproduce in my manual testing because
font panel also sets text shadow, which ends up filling up m_mutableStyle.
Fixed the bug by creating a mutable style properties when one is not provided by EditingStyle.
Also fixed the "FIXME" in the function by converting text decoration changes to a corresponding
text decoration value. The values passed to shouldApplyStyle now matches the old behavior prior
to r183770.
Test: editing/style/underline-by-user.html
- editing/EditingStyle.cpp:
(WebCore::EditingStyle::styleWithResolvedTextDecorations): Added.
- editing/EditingStyle.h:
- editing/Editor.cpp:
(WebCore::Editor::applyStyleToSelection): Use styleWithResolvedTextDecorations to avoid the crash.
LayoutTests:
Reviewed by Darin Adler.
Added a test that emulates underlining of text by the user. Unlike document.execCommand,
testRunner.execCommand simulates a user initiated editing command and therefore invokes
shouldApplyStyle.
- editing/style/underline-by-user-expected.txt: Added.
- editing/style/underline-by-user.html: Added.
- 2:37 PM Changeset in webkit [184307] by
-
- 1 copy in tags/Safari-600.5.17.2
New tag.
- 2:18 PM Changeset in webkit [184306] by
-
- 2 edits in trunk/Source/WebCore
Work around HTMLMediaElement::documentDidResumeFromPageCache being called twice
https://bugs.webkit.org/show_bug.cgi?id=144969
Reviewed by Alexey Proskuryakov.
- dom/Document.cpp:
(WebCore::Document::addPlaybackTargetPickerClient): Replace ASSERT with early
return to work around https://webkit.org/b/144970.
- 2:05 PM Changeset in webkit [184305] by
-
- 3 edits in trunk/Source/JavaScriptCore
I skipped a wrong test in r184270. Fix that.
The failure is tracked by webkit.org/b/144947.
- tests/stress/arith-modulo-node-behaviors.js:
- tests/stress/arith-mul-with-constants.js:
- 1:51 PM Changeset in webkit [184304] by
-
- 7 edits in trunk/Source/WebCore
Going back after resizing causes scroll knob to appear in the middle of the page
https://bugs.webkit.org/show_bug.cgi?id=144968
<rdar://problem/18299827>
Reviewed by Beth Dakin.
- history/CachedPage.cpp:
(WebCore::CachedPage::restore):
(WebCore::CachedPage::clear):
- history/CachedPage.h:
(WebCore::CachedPage::markForContentsSizeChanged):
- history/PageCache.cpp:
(WebCore::PageCache::markPagesForContentsSizeChanged):
- history/PageCache.h:
Add a flag that will cause us to call updateContentsSize() after a page
comes out of the page cache, if necessary.
- page/FrameView.cpp:
(WebCore::FrameView::setContentsSize):
- page/FrameView.h:
Mark all cached pages for this frame as needing updateContentsSize()
when setContentsSize happens. This will ensure that scrollbar layers
are repositioned when coming out of the page cache.
- 1:42 PM Changeset in webkit [184303] by
-
- 2 edits in trunk/Source/WebKit2
[Mac] Sandbox violation reading SubmitDiagInfo.domains
https://bugs.webkit.org/show_bug.cgi?id=144962
rdar://problem/20719330
Reviewed by Darin Adler.
- WebProcess/com.apple.WebProcess.sb.in:
- 1:40 PM Changeset in webkit [184302] by
-
- 3 edits in branches/safari-600.5.17-branch/Source/JavaScriptCore
Merged r184229. rdar://problem/18736465
- 1:10 PM Changeset in webkit [184301] by
-
- 18 edits in branches/safari-601.1.32-branch
Merged r183976.
- 1:08 PM Changeset in webkit [184300] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebCore
Merged r183958.
- 12:59 PM Changeset in webkit [184299] by
-
- 2 edits in trunk/Source/WebKit2
Speculative build fix.
- UIProcess/API/Cocoa/_WKOverlayScrollbarStyle.h:
- 12:56 PM Changeset in webkit [184298] by
-
- 2 edits in trunk/Source/WebKit2
[GTK][CMake] Extra include directory when libnotify is present but disabled
https://bugs.webkit.org/show_bug.cgi?id=144941
Reviewed by Martin Robinson.
Add LIBNOTIFY_INCLUDE_DIRS to WebKit2_INCLUDE_DIRECTORIES only if USE_LIBNOTIFY is true,
rather than checking LIBNOTIFY_FOUND.
- PlatformGTK.cmake:
- 12:28 PM Changeset in webkit [184297] by
-
- 18 edits1 add in trunk/Source
Need SPI to set the overlay scroll bar style
https://bugs.webkit.org/show_bug.cgi?id=144928
-and corresponding-
rdar://problem/20143614
Reviewed by Anders Carlsson.
Source/WebCore:
New ChromeClient function preferredScrollbarOverlayStyle() will fetch the
scrollbar style that was set via the new SPI.
- page/ChromeClient.h:
If the preferredScrollbarOverlayStyle() is anything but None, then use it. None is
used to indicate that the normal heuristic should compute the appropriate color.
- page/FrameView.cpp:
(WebCore::FrameView::recalculateScrollbarOverlayStyle):
- page/FrameView.h:
Source/WebKit2:
Make scrollbarOverlayStyle a part of the creation parameters.
- Shared/WebPageCreationParameters.cpp:
(WebKit::WebPageCreationParameters::encode):
(WebKit::WebPageCreationParameters::decode):
- Shared/WebPageCreationParameters.h:
New SPI.
- UIProcess/API/Cocoa/WKViewPrivate.h:
- UIProcess/API/Cocoa/_WKOverlayScrollbarStyle.h: Added.
- UIProcess/API/mac/WKView.mm:
(-[WKView _setOverlayScrollbarStyle:]):
(-[WKView _overlayScrollbarStyle]):
Store m_scrollbarOverlayStyle on WebPageProxy, and set it to the WebProcess.
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::WebPageProxy):
(WebKit::WebPageProxy::creationParameters):
(WebKit::WebPageProxy::setOverlayScrollbarStyle):
- UIProcess/WebPageProxy.h:
(WebKit::WebPageProxy::overlayScrollbarStyle):
- WebKit2.xcodeproj/project.pbxproj:
Return WebPage’s scrollbarOverlayStyle().
- WebProcess/WebCoreSupport/WebChromeClient.cpp:
(WebKit::WebChromeClient::preferredScrollbarOverlayStyle):
- WebProcess/WebCoreSupport/WebChromeClient.h:
Cache the scrollbarOverlayStyle() here for the WebProcess.
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::WebPage):
(WebKit::WebPage::setScrollbarOverlayStyle):
- WebProcess/WebPage/WebPage.h:
(WebKit::WebPage::scrollbarOverlayStyle):
- WebProcess/WebPage/WebPage.messages.in:
- 12:26 PM Changeset in webkit [184296] by
-
- 3 edits in trunk/Source/WebCore
Scrollbars in overflow regions are not vanishing after scrolling with scroll snap points
https://bugs.webkit.org/show_bug.cgi?id=142521
<rdar://problem/20100706>
Reviewed by Darin Adler.
The scrollbars were not being dismissed because they were not being notified that the wheel
gesture was finished. This was happening because the wheel event 'ended' state has zero
deltaX and deltaY. If the region did not allow stretching, it would exit early, never passing
through the 'handleWheelEventPhase' code that would notify the scrollbar controller that
the gesture had ended.
- platform/ScrollableArea.cpp:
(WebCore::ScrollableArea::mouseExitedContentArea): The wrong ScrollAnimator method was being
called when the mouse exited the content area.
- platform/mac/ScrollAnimatorMac.mm:
(WebCore::ScrollAnimatorMac::handleWheelEvent): Do not early return when the wheel event has
no change in X or Y coordinate.
- 12:24 PM Changeset in webkit [184295] by
-
- 5 edits in trunk/Source/WebKit2
Rename some StorageManager functions to indicate that they work on local storage entries
https://bugs.webkit.org/show_bug.cgi?id=144958
First part of rdar://problem/10690447.
Reviewed by Beth Dakin.
- UIProcess/Storage/StorageManager.cpp:
(WebKit::StorageManager::getLocalStorageOrigins):
(WebKit::StorageManager::getLocalStorageDetailsByOrigin):
(WebKit::StorageManager::deleteLocalStorageEntriesForOrigin):
(WebKit::StorageManager::deleteAllLocalStorageEntries):
(WebKit::StorageManager::deleteLocalStorageEntriesForOrigins):
(WebKit::StorageManager::getOrigins): Deleted.
(WebKit::StorageManager::getStorageDetailsByOrigin): Deleted.
(WebKit::StorageManager::deleteEntriesForOrigin): Deleted.
(WebKit::StorageManager::deleteAllEntries): Deleted.
(WebKit::StorageManager::deleteEntriesForOrigins): Deleted.
- UIProcess/Storage/StorageManager.h:
- UIProcess/WebKeyValueStorageManager.cpp:
(WebKit::WebKeyValueStorageManager::getKeyValueStorageOrigins):
(WebKit::WebKeyValueStorageManager::getStorageDetailsByOrigin):
(WebKit::WebKeyValueStorageManager::deleteEntriesForOrigin):
(WebKit::WebKeyValueStorageManager::deleteAllEntries):
- UIProcess/WebsiteData/WebsiteDataStore.cpp:
(WebKit::WebsiteDataStore::fetchData):
(WebKit::WebsiteDataStore::removeData):
- 12:20 PM Changeset in webkit [184294] by
-
- 2 edits in trunk/Source/WebKit2
Crash under WebKit::WebInspectorProxy::attachAvailabilityChanged sometimes opening new page
https://bugs.webkit.org/show_bug.cgi?id=144957
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2015-05-13
Reviewed by Simon Fraser.
- UIProcess/WebInspectorProxy.cpp:
(WebKit::WebInspectorProxy::attachAvailabilityChanged):
- 11:56 AM Changeset in webkit [184293] by
-
- 2 edits in trunk/Source/WebCore
Don't compute selection painting info when we don't have selection.
https://bugs.webkit.org/show_bug.cgi?id=144920
<rdar://problem/20919920>
Reviewed by Simon Fraser.
- rendering/InlineTextBox.cpp:
(WebCore::InlineTextBox::paint):
Just set the selection paint style to the text paint style when we don't have a selection
at all. Computing the selection style takes time in the case where a ::selection pseudo is
used on the page, so we don't want to waste time computing that info unless it's actually
needed.
- 11:51 AM Changeset in webkit [184292] by
-
- 2 edits in trunk/Source/JavaScriptCore
Avoid always running some debug code in type profiling
https://bugs.webkit.org/show_bug.cgi?id=144775
Reviewed by Daniel Bates.
- runtime/TypeProfilerLog.cpp:
(JSC::TypeProfilerLog::processLogEntries):
- 11:51 AM Changeset in webkit [184291] by
-
- 36 edits in trunk/Source
Pass String as reference in more places
https://bugs.webkit.org/show_bug.cgi?id=144769
Reviewed by Daniel Bates.
Source/JavaScriptCore:
- debugger/Breakpoint.h:
(JSC::Breakpoint::Breakpoint):
- parser/Parser.h:
(JSC::Parser::setErrorMessage):
(JSC::Parser::updateErrorWithNameAndMessage):
- parser/ParserError.h:
(JSC::ParserError::ParserError):
- runtime/RegExp.cpp:
(JSC::RegExpFunctionalTestCollector::outputOneTest):
- runtime/RegExpObject.cpp:
(JSC::regExpObjectSourceInternal):
- runtime/TypeProfiler.cpp:
(JSC::TypeProfiler::typeInformationForExpressionAtOffset):
- runtime/TypeProfilerLog.cpp:
(JSC::TypeProfilerLog::processLogEntries):
- runtime/TypeProfilerLog.h:
- tools/FunctionOverrides.cpp:
(JSC::initializeOverrideInfo):
- inspector/scripts/codegen/generate_objc_conversion_helpers.py:
(ObjCConversionHelpersGenerator._generate_enum_from_protocol_string):
- inspector/scripts/codegen/objc_generator_templates.py:
- inspector/scripts/tests/expected/commands-with-async-attribute.json-result:
- inspector/scripts/tests/expected/commands-with-optional-call-return-parameters.json-result:
- inspector/scripts/tests/expected/domains-with-varying-command-sizes.json-result:
- inspector/scripts/tests/expected/enum-values.json-result:
- inspector/scripts/tests/expected/events-with-optional-parameters.json-result:
- inspector/scripts/tests/expected/generate-domains-with-feature-guards.json-result:
- inspector/scripts/tests/expected/same-type-id-different-domain.json-result:
- inspector/scripts/tests/expected/shadowed-optional-type-setters.json-result:
- inspector/scripts/tests/expected/type-declaration-aliased-primitive-type.json-result:
- inspector/scripts/tests/expected/type-declaration-array-type.json-result:
- inspector/scripts/tests/expected/type-declaration-enum-type.json-result:
- inspector/scripts/tests/expected/type-declaration-object-type.json-result:
- inspector/scripts/tests/expected/type-requiring-runtime-casts.json-result:
Rebaseline tests after updating the generator.
Source/WebCore:
- bindings/js/SerializedScriptValue.cpp:
(WebCore::CloneSerializer::dumpString):
(WebCore::CloneSerializer::dumpStringObject):
- dom/DocumentMarkerController.cpp:
(WebCore::DocumentMarkerController::addMarker):
- dom/DocumentMarkerController.h:
- inspector/InspectorApplicationCacheAgent.cpp:
(WebCore::InspectorApplicationCacheAgent::assertFrameWithDocumentLoader):
- inspector/InspectorApplicationCacheAgent.h:
- inspector/InspectorNodeFinder.cpp:
(WebCore::stripCharacters):
(WebCore::InspectorNodeFinder::InspectorNodeFinder):
- inspector/InspectorNodeFinder.h:
Source/WebKit2:
- WebProcess/WebPage/WebInspectorUI.cpp:
(WebKit::WebInspectorUI::showMainResourceForFrame):
- WebProcess/WebPage/WebInspectorUI.h:
- 11:11 AM Changeset in webkit [184290] by
-
- 11 edits in trunk/Source
View scale changes are temporarily lost after restoring a page from the page cache
https://bugs.webkit.org/show_bug.cgi?id=144934
Reviewed by Brady Eidson.
- history/CachedPage.cpp:
(WebCore::CachedPage::CachedPage):
(WebCore::CachedPage::restore):
(WebCore::CachedPage::clear):
- history/CachedPage.h:
(WebCore::CachedPage::markForDeviceOrPageScaleChanged): Renamed.
- history/PageCache.cpp:
(WebCore::PageCache::markPagesForDeviceOrPageScaleChanged): Renamed.
- history/PageCache.h:
Rename PageCache/CachedPage methods to make it more clear that they
will eventually result in calling deviceOrPageScaleFactorChanged().
Also, use modern initialization for CachedPage members.
- loader/HistoryController.cpp:
(WebCore::HistoryController::saveScrollPositionAndViewStateToItem):
(WebCore::HistoryController::restoreScrollPositionAndViewState):
Store the pageScaleFactor on HistoryItem with the view scale factored out,
because the view scale can change while the page is in the page cache, and
WebCore needs a way - without consulting with WebKit2 - to apply the changed
view scale to the cached page scale.
- page/Page.cpp:
(WebCore::Page::setViewScaleFactor):
(WebCore::Page::setDeviceScaleFactor):
- page/Page.h:
(WebCore::Page::viewScaleFactor):
Keep track of the viewScaleFactor, and mark all pages in the page cache
as needing to call deviceOrPageScaleFactorChanged and do a full style recalc
when they come back from the page cache.
For now, we expect all callers of setPageScaleFactor (including WebKit2 and
HistoryController) to multiply the viewScale in manually, to avoid the
significant amount of change in WebCore that would be required to keep them
totally separately.
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::WebPage):
(WebKit::WebPage::scalePage):
(WebKit::WebPage::scalePageInViewCoordinates):
(WebKit::WebPage::pageScaleFactor):
(WebKit::WebPage::viewScaleFactor):
(WebKit::WebPage::scaleView):
- WebProcess/WebPage/WebPage.h:
(WebKit::WebPage::viewScaleFactor): Deleted.
Get rid of m_viewScaleFactor, instead using Page::viewScaleFactor.
- 10:58 AM Changeset in webkit [184289] by
-
- 2 edits in trunk/Source/JavaScriptCore
com.apple.WebKit.WebContent crashed at JavaScriptCore: JSC::CodeBlock::finalizeUnconditionally
https://bugs.webkit.org/show_bug.cgi?id=144933
Changed the RELEASE_ASSERT_NOT_REACHED into an ASSERT. Added some diagnostic messages to
help determine the cause for any crash.
Reviewed by Geoffrey Garen.
- bytecode/CodeBlock.cpp:
(JSC::CodeBlock::finalizeUnconditionally):
- 10:39 AM Changeset in webkit [184288] by
-
- 7 edits in trunk/Source/JavaScriptCore
REGRESSION(r184260): arguments elimination has stopped working because of Check(UntypedUse:) from SSAConversionPhase
https://bugs.webkit.org/show_bug.cgi?id=144951
Reviewed by Michael Saboff.
There were two issues here:
- In r184260 we expected a small number of possible use kinds in Check nodes, and UntypedUse was not one of them. That seemed like a sensible assumption because we don't create Check nodes unless it's to have a check. But, SSAConversionPhase was creating a Check that could have UntypedUse. I fixed this. It's cleaner for SSAConversionPhase to follow the same idiom as everyone else and not create tautological checks.
- It's clearly not very robust to assume that Checks will not be used tautologically. So, this changes how we validate Checks in the escape analyses. We now use willHaveCheck, which catches cases that AI would have already marked as unnecessary. It then also uses a new helper called alreadyChecked(), which allows us to just ask if the check is unnecessary for objects. That's a good fall-back in case AI hadn't run yet.
- dfg/DFGArgumentsEliminationPhase.cpp:
- dfg/DFGMayExit.cpp:
- dfg/DFGObjectAllocationSinkingPhase.cpp:
(JSC::DFG::ObjectAllocationSinkingPhase::handleNode):
- dfg/DFGSSAConversionPhase.cpp:
(JSC::DFG::SSAConversionPhase::run):
- dfg/DFGUseKind.h:
(JSC::DFG::alreadyChecked):
- dfg/DFGVarargsForwardingPhase.cpp:
- 9:48 AM Changeset in webkit [184287] by
-
- 8 edits2 adds in trunk
[ES6] Implement String.raw
https://bugs.webkit.org/show_bug.cgi?id=144330
Reviewed by Filip Pizlo.
Source/JavaScriptCore:
Implement String.raw. It is intended to be used with tagged-templates syntax.
To implement ToString abstract operation efficiently,
we introduce @toString bytecode intrinsic. It emits op_to_string directly.
- CMakeLists.txt:
- builtins/StringConstructor.js: Added.
(raw):
- bytecompiler/NodesCodegen.cpp:
(JSC::BytecodeIntrinsicNode::emit_intrinsic_toString):
- runtime/CommonIdentifiers.h:
- runtime/StringConstructor.cpp:
- tests/stress/string-raw.js: Added.
(shouldBe):
(.get shouldBe):
(Counter):
LayoutTests:
Add String.raw.
- js/Object-getOwnPropertyNames-expected.txt:
- js/script-tests/Object-getOwnPropertyNames.js:
- 2:32 AM Changeset in webkit [184286] by
-
- 5 edits in branches/safari-601.1.32-branch/Tools
Merged r184018.
- 2:31 AM Changeset in webkit [184285] by
-
- 2 edits in trunk/Source/WebKit2
Minor cleanups to PluginProxy.cpp.
https://bugs.webkit.org/show_bug.cgi?id=144948
Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-13
Reviewed by Gyuyoung Kim.
- Remove unnecessary #include.
- Remove unnecessary return statement from PluginProxy::paint().
No new tests, no behavior change.
- WebProcess/Plugins/PluginProxy.cpp:
(WebKit::PluginProxy::paint):
- 2:23 AM Changeset in webkit [184284] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebCore
Merged r183980. rdar://problem/20769741
- 2:13 AM Changeset in webkit [184283] by
-
- 2 edits in trunk/Source/WebKit2
REGRESSION(r176631): [EFL] Fullscreen feature doesn't work correctly on MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=144906
Patch by Daegyu Lee <daegyu.lee@navercorp.com> on 2015-05-13
Reviewed by Gyuyoung Kim.
- UIProcess/CoordinatedGraphics/PageViewportController.cpp:
(WebKit::PageViewportController::updateMinimumScaleToFit): Recover the r176631 condition to
call applyScaleAfterRenderingContents function to apply correct scale.
- 1:26 AM Changeset in webkit [184282] by
-
- 6 edits in branches/safari-601.1.32-branch/Source/WebKit2
Merged r184028. rdar://problem/20210267
- 12:21 AM Changeset in webkit [184281] by
-
- 2 edits in branches/safari-601.1.32-branch/Tools
Merged r183915.
- 12:10 AM Changeset in webkit [184280] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebKit2
Merged r184241. rdar://problem/20172315
- 12:09 AM Changeset in webkit [184279] by
-
- 8 edits in branches/safari-601.1.32-branch/Source
Merged r184231. rdar://problem/20923031
- 12:04 AM Changeset in webkit [184278] by
-
- 12 edits3 copies in branches/safari-601.1.32-branch/Source/WebKit2
Merged r184215. rdar://problem/19708579
- 12:02 AM Changeset in webkit [184277] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebKit2
Merged r184125. rdar://problem/19708579
- 12:01 AM Changeset in webkit [184276] by
-
- 6 edits in branches/safari-601.1.32-branch/Source/WebKit2
Merged r184061. rdar://problem/20856497
May 12, 2015:
- 11:48 PM Changeset in webkit [184275] by
-
- 4 edits1 copy in branches/safari-601.1.32-branch
Merged r183933. rdar://problem/20681226
- 11:47 PM Changeset in webkit [184274] by
-
- 4 edits in trunk/Source/WebCore
Reduce TransformationMatrix copies in MatrixTransformOperation, Matrix3DTransformOperation
https://bugs.webkit.org/show_bug.cgi?id=144797
Reviewed by Darin Adler.
Using std::swap() on TransformationMatrix objects which don't
provide move constructors will result in copies.
Instead, use a helper function in both MatrixTransformOperation
and Matrix3DTransformOperation that calls TransformationMatrix::blend()
and returns the new Matrix(3D)TransformOperation object, and call it
with fromT and toT arguments switched when blending to identity.
- platform/graphics/transforms/Matrix3DTransformOperation.cpp:
(WebCore::createOperation):
(WebCore::Matrix3DTransformOperation::blend):
- platform/graphics/transforms/Matrix3DTransformOperation.h: No need
to copy the m_matrix member, it won't change when passed to
TransformationMatrix::multiply().
- platform/graphics/transforms/MatrixTransformOperation.cpp:
(WebCore::createOperation):
(WebCore::MatrixTransformOperation::blend):
- 11:18 PM Changeset in webkit [184273] by
-
- 3 edits in trunk/Source/WebCore
[EGL][X11] XPixmap created in GLContextEGL::createPixmapContext() is leaked
https://bugs.webkit.org/show_bug.cgi?id=144909
Reviewed by Sergio Villar Senin and Žan Doberšek.
The pixmap is created and passed to eglCreatePixmapSurface(), but
never released. eglCreatePixmapSurface() doesn't take the
ownership of the pixmap, so we should explicitly free it when the
GLContextEGL is destroyed.
- platform/graphics/egl/GLContextEGL.cpp:
(WebCore::GLContextEGL::createPixmapContext): Use XUniquePixmap
and transfer the ownership to the context by using the new
constructor that receives a XUniquePixmap&&.
(WebCore::GLContextEGL::createContext): createPixmapContext() is
now only defined for X11.
(WebCore::GLContextEGL::GLContextEGL): New constructor that
receives a XUniquePixmap&&.
- platform/graphics/egl/GLContextEGL.h: Add new constructor and
initialize the cairo device when defined to simplify constructors.
- 10:47 PM Changeset in webkit [184272] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebKit/mac
Merged r183968. rdar://problem/20281886
- 10:44 PM Changeset in webkit [184271] by
-
- 2 edits in trunk/Tools
Unreviewed build fix.
Added the missing metric name and wrapped values in an array as done in SunSpider.patch.
- Scripts/webkitpy/benchmark_runner/data/patches/Kraken.patch:
- 10:39 PM Changeset in webkit [184270] by
-
- 2 edits in trunk/Source/JavaScriptCore
Temporarily disable the test on Windows. The failure is tracked in webkit.org/b/144897.
- tests/stress/arith-mul-with-constants.js:
- 10:34 PM Changeset in webkit [184269] by
-
- 12 edits in branches/safari-601.1.32-branch
Fix horked ChangeLogs
- 10:33 PM Changeset in webkit [184268] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r184226. rdar://problem/20707307
- 10:33 PM Changeset in webkit [184267] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r184207. rdar://problem/20707307
- 10:33 PM Changeset in webkit [184266] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r184140. rdar://problem/20907253
- 10:33 PM Changeset in webkit [184265] by
-
- 21 edits in branches/safari-601.1.32-branch/Source
Merge r184139. rdar://problem/20125088
- 10:33 PM Changeset in webkit [184264] by
-
- 54 edits in branches/safari-601.1.32-branch
Merge r184137. rdar://problem/20707307
- 10:33 PM Changeset in webkit [184263] by
-
- 7 edits in branches/safari-601.1.32-branch/Source/WebInspectorUI
Merge r184130. rdar://problem/20829494
- 10:31 PM Changeset in webkit [184262] by
-
- 2 edits in trunk/Source/WebCore
Reindent DIBPixelData.h for consistency.
https://bugs.webkit.org/show_bug.cgi?id=144942
Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-12
Reviewed by Darin Adler.
No new tests, no behavior change.
- platform/graphics/win/DIBPixelData.h:
(WebCore::DIBPixelData::DIBPixelData):
(WebCore::DIBPixelData::buffer):
(WebCore::DIBPixelData::bufferLength):
(WebCore::DIBPixelData::size):
(WebCore::DIBPixelData::bytesPerRow):
(WebCore::DIBPixelData::bitsPerPixel):
- 10:25 PM Changeset in webkit [184261] by
-
- 2 edits in trunk/LayoutTests
fast/text/simple-line-layout-text-stroke-width.html fails on Windows
- platform/win/TestExpectations: Mark it as such. The tets likely needs to be
improved to work cross-platform.
- 10:21 PM Changeset in webkit [184260] by
-
- 11 edits13 adds in trunk
js/dom/stack-trace.html fails with eager compilation
https://bugs.webkit.org/show_bug.cgi?id=144853
Reviewed by Benjamin Poulain.
Source/JavaScriptCore:
All of our escape analyses were mishandling Check(). They were assuming that this is a
non-escaping operation. But, if we do for example a Check(Int32:@x) and @x is an escape
candidate, then we need to do something: if we eliminate or sink @x, then the check no
longer makes any sense since a phantom allocation has no type. This will make us forget
that this operation would have exited. This was causing us to not call a valueOf method in
js/dom/stack-trace.html with eager compilation enabled, because it was doing something like
+o where o had a valueOf method, and o was otherwise sinkable.
This changes our escape analyses to basically pretend that any Check() that isn't obviously
unnecessary is an escape. We don't have to be super careful here. Most checks will be
completely eliminated by constant-folding. If that doesn't run in time, then the most
common check we will see is CellUse. So, we just recognize some very obvious check kinds
that we know would have passed, and for all of the rest we just assume that it's an escape.
This was super tricky to test. The obvious way to test it is to use +o like
stack-trace.html, except that doing so relies on the fact that we still haven't implemented
the optimal behavior for op_to_number. So, I take four approaches in testing this patch:
1) Use +o. These will test what we want it to test for now, but at some point in the future
these tests will just be a good sanity-check that our op_to_number implementation is
right.
2) Do fancy control flow tricks to fool the profiling into thinking that some arithmetic
operation always sees integers even though we eventually feed it an object and that
object is a sink candidate.
3) Introduce a new jsc.cpp intrinsic called isInt32() which returns true if the incoming
value is an int32. This intrinsic is required to be implemented by DFG by
unconditionally speculating that the input is int32. This allows us to write much more
targetted tests of the underlying issue.
4) I made a version of stack-trace.html that runs in run-jsc-stress-tests, so that we can
get regression test coverage of this test in eager mode.
- dfg/DFGArgumentsEliminationPhase.cpp:
- dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::handleIntrinsic):
- dfg/DFGObjectAllocationSinkingPhase.cpp:
(JSC::DFG::ObjectAllocationSinkingPhase::handleNode):
- dfg/DFGVarargsForwardingPhase.cpp:
- ftl/FTLExitValue.cpp:
(JSC::FTL::ExitValue::dumpInContext):
- ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::buildExitArguments):
- ftl/FTLOSRExitCompiler.cpp:
(JSC::FTL::compileFTLOSRExit):
- jsc.cpp:
(GlobalObject::finishCreation):
(functionIsInt32):
- runtime/Intrinsic.h:
- tests/stress/sink-arguments-past-invalid-check-dfg.js: Added.
- tests/stress/sink-arguments-past-invalid-check-int32-dfg.js: Added.
- tests/stress/sink-arguments-past-invalid-check-int32.js: Added.
- tests/stress/sink-arguments-past-invalid-check-sneakier.js: Added.
- tests/stress/sink-arguments-past-invalid-check.js: Added.
- tests/stress/sink-function-past-invalid-check-sneakier.js: Added.
- tests/stress/sink-function-past-invalid-check-sneaky.js: Added.
- tests/stress/sink-object-past-invalid-check-int32.js: Added.
- tests/stress/sink-object-past-invalid-check-sneakier.js: Added.
- tests/stress/sink-object-past-invalid-check-sneaky.js: Added.
- tests/stress/sink-object-past-invalid-check.js: Added.
LayoutTests:
Make a copy of the stack-trace test that only runs in run-jsc-stress-tests. Sadly, we don't
have a good way of having different expectation files for when a test runs in RJST versus
RWT. So, the approach I take is that I make a copy of the test just for RJST and I exclude
the .html file, which makes RWT overlook it. The test has different expectations in the
two harnesses because it does some small DOM things.
- js/script-tests/stack-trace.js: Added.
- js/stack-trace-expected.txt: Added.
- 10:18 PM Changeset in webkit [184259] by
-
- 2 edits in trunk/Source/WebCore
Unreviewed, rolling out r184224.
https://bugs.webkit.org/show_bug.cgi?id=144946
Made inspector/page/main-frame-resource.html assert every time
(Requested by ap on #webkit).
Reverted changeset:
"Web Inspector: REGRESSION (r181625): Timeline recording
started from console.profile is always empty"
https://bugs.webkit.org/show_bug.cgi?id=144882
http://trac.webkit.org/changeset/184224
- 9:38 PM Changeset in webkit [184258] by
-
- 2 edits in trunk/Source/WebKit2
In Safari, Debug > Get Bytecode Profile crashes the Web Content process
https://bugs.webkit.org/show_bug.cgi?id=144944
Reviewed by Darin Adler.
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::getBytecodeProfile): Don’t assert that m_perBytecodeProfiler isn’t null,
because it is when the profiler is disabled, which is the default. In that case, return
after sending back the empty string, rather than continuing to dereference a null pointer.
- 9:25 PM Changeset in webkit [184257] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebKit2
Merge r184121. rdar://problem/20774613
- 9:25 PM Changeset in webkit [184256] by
-
- 27 edits3 adds in branches/safari-601.1.32-branch
Merge r184116. rdar://problem/20774613
- 9:25 PM Changeset in webkit [184255] by
-
- 6 edits in branches/safari-601.1.32-branch/Source/WebInspectorUI
Merge r184108. rdar://problem/20903134
- 9:25 PM Changeset in webkit [184254] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r184104. rdar://problem/20727702
- 9:17 PM Changeset in webkit [184253] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebKit2
Merge r184026. rdar://problem/20757196
- 9:17 PM Changeset in webkit [184252] by
-
- 3 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r184005. rdar://problem/20486538
- 9:17 PM Changeset in webkit [184251] by
-
- 4 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r184001. rdar://problem/20862460
- 9:17 PM Changeset in webkit [184250] by
-
- 3 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r183971. rdar://problem/20065572
- 9:10 PM Changeset in webkit [184249] by
-
- 3 edits in trunk/Source/WebCore
[Mac] Hang in MediaPlayerPrivateAVFoundationObjC::audioSourceProvider() - 'tracks' property not yet loaded
https://bugs.webkit.org/show_bug.cgi?id=144937
Reviewed by Eric Carlson.
Querying for the -[AVAsset tracks] property blocks for network loading; and could
block forever if the asset in question is not reachable. Add a "safe" mechanism for
querying the list of audible tracks, and use that when providing the audioSourceProvider()
with an audible track.
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.h:
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::createAVPlayerItem):
(WebCore::MediaPlayerPrivateAVFoundationObjC::tracksChanged):
(WebCore::MediaPlayerPrivateAVFoundationObjC::audioSourceProvider):
(WebCore::MediaPlayerPrivateAVFoundationObjC::processLegacyClosedCaptionsTracks):
(WebCore::MediaPlayerPrivateAVFoundationObjC::safeAVAssetTracksForAudibleMedia):
- 9:08 PM Changeset in webkit [184248] by
-
- 6 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r183970. rdar://problem/20769741
- 9:08 PM Changeset in webkit [184247] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r183965. rdar://problem/20866590
- 9:08 PM Changeset in webkit [184246] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebKit/mac
Merge r183957. rdar://problem/20811128
- 9:02 PM Changeset in webkit [184245] by
-
- 2 edits in trunk/Source/WTF
If JSC cannot get executable memory, it shouldn't call madvise
https://bugs.webkit.org/show_bug.cgi?id=144931
Reviewed by Mark Lam.
Made calling madvise conditional on really getting mmapped memory.
- wtf/OSAllocatorPosix.cpp:
(WTF::OSAllocator::reserveUncommitted):
- 8:58 PM Changeset in webkit [184244] by
-
- 15 edits in branches/safari-601.1.32-branch
Merge r183954. rdar://problem/20281886
- 8:58 PM Changeset in webkit [184243] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r183953. rdar://problem/19997548
- 8:58 PM Changeset in webkit [184242] by
-
- 7 edits2 adds in branches/safari-601.1.32-branch
Merge r183943. rdar://problem/19913748
- 8:51 PM Changeset in webkit [184241] by
-
- 2 edits in trunk/Source/WebKit2
Crash when using <input type=file>
<https://bugs.webkit.org/show_bug.cgi?id=144939>
<rdar://problem/20172315>
Reviewed by Andy Estes.
- UIProcess/ios/forms/WKFileUploadPanel.mm:
Use a RefPtr to hold the WebOpenPanelResultListenerProxy, because the
WKFileUploadPanel can outlive it.
- 8:45 PM Changeset in webkit [184240] by
-
- 3 edits in branches/safari-600.7-branch/LayoutTests
Added prefixes in the test added in r182985 (merged in r183180) since we don't support unprefixed versions in this branch.
- compositing/animation/animation-backing.html:
- platform/mac-wk2/TestExpectations: Removed merge conflict lines.
- 8:42 PM Changeset in webkit [184239] by
-
- 1 edit1 add in trunk/LayoutTests
AX: [Win] REGRESSION(r184213) breaks aria-menubar-menuitems.html
https://bugs.webkit.org/show_bug.cgi?id=144936
Unreviewed gardening.
Adding the platform-specific expectations for Windows now that the
accessibility tree is being correctly generated for this test.
- platform/win/accessibility/aria-menubar-menuitems-expected.txt: Added.
- 8:22 PM Changeset in webkit [184238] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebKit2
Merge r183937. rdar://problem/20458697
- 8:22 PM Changeset in webkit [184237] by
-
- 3 edits in branches/safari-601.1.32-branch/Source/WebCore
Merge r183927. rdar://problem/20854785
- 8:15 PM Changeset in webkit [184236] by
-
- 2 edits in branches/safari-601.1.32-branch/Source/WebKit2
Merge r183911. rdar://problem/20702435
- 8:15 PM Changeset in webkit [184235] by
-
- 15 edits in branches/safari-601.1.32-branch/Source
Merge r183909. rdar://problem/18894598
- 8:08 PM Changeset in webkit [184234] by
-
- 2 edits in trunk/Source/WebCore
Mac build fix after r184228.
- bindings/scripts/CodeGeneratorJS.pm:
- 8:02 PM Changeset in webkit [184233] by
-
- 4 edits in branches/safari-601.1.32-branch
Merge r183942. rdar://problem/20049088
- 8:02 PM Changeset in webkit [184232] by
-
- 5 edits2 adds in branches/safari-601.1.32-branch
Merge r183894. rdar://problem/20049088
- 7:01 PM Changeset in webkit [184231] by
-
- 8 edits in trunk/Source
Turn antialiased font dilation off by default
https://bugs.webkit.org/show_bug.cgi?id=144940
rdar://problem/20923031
Reviewed by Sam Weinig.
Source/WebCore:
Turn antialised font dilation off by default.
- page/Settings.cpp:
(WebCore::Settings::Settings):
- platform/graphics/FontCascade.cpp:
Source/WebKit/mac:
Turn antialised font dilation off by default.
- WebView/WebPreferences.mm:
(+[WebPreferences initialize]):
Source/WebKit2:
Turn antialised font dilation off by default.
- Shared/WebPreferencesDefinitions.h:
- UIProcess/API/C/WKPreferencesRefPrivate.h:
- 6:52 PM Changeset in webkit [184230] by
-
- 2 edits in trunk/Source/JavaScriptCore
Fix the iteration count of arith-modulo-node-behaviors.js
- tests/stress/arith-modulo-node-behaviors.js:
No need for big numbers for the real testing.
- 6:47 PM Changeset in webkit [184229] by
-
- 2 edits in trunk/Source/JavaScriptCore
Windows: Cannot use HANDLE from GetCurrentThread() to get the CONTEXT of another thread.
https://bugs.webkit.org/show_bug.cgi?id=144924
Reviewed by Alex Christensen.
The present stack scanning code in the Windows port is expecting that the
GetCurrentThread() API will provide a unique HANDLE for each thread. The code
then saves and later uses that HANDLE with GetThreadContext() to get the
runtime state of the target thread from the GC thread. According to
https://msdn.microsoft.com/en-us/library/windows/desktop/ms683182(v=vs.85).aspx,
GetCurrentThread() does not provide this unique HANDLE that we expect:
"The function cannot be used by one thread to create a handle that can
be used by other threads to refer to the first thread. The handle is
always interpreted as referring to the thread that is using it. A
thread can create a "real" handle to itself that can be used by other
threads, or inherited by other processes, by specifying the pseudo
handle as the source handle in a call to the DuplicateHandle function."
As a result of this, GetCurrentThread() always returns the same HANDLE value, and
we end up never scanning the stacks of other threads because we wrongly think that
they are all equal (in identity) to the scanning thread. This, in turn, results
in crashes due to objects that are incorrectly collected.
The fix is to call DuplicateHandle() to create a HANDLE that we can use. The
MachineThreads::Thread class already accurately tracks the period of time when
we need that HANDLE for the VM. Hence, the life-cycle of the HANDLE can be tied
to the life-cycle of the MachineThreads::Thread object for the corresponding thread.
- heap/MachineStackMarker.cpp:
(JSC::getCurrentPlatformThread):
(JSC::MachineThreads::Thread::Thread):
(JSC::MachineThreads::Thread::~Thread):
(JSC::MachineThreads::Thread::suspend):
(JSC::MachineThreads::Thread::resume):
(JSC::MachineThreads::Thread::getRegisters):
- 6:47 PM Changeset in webkit [184228] by
-
- 18 edits in trunk/Source/WebCore
Don't export all JS bindings classes.
https://bugs.webkit.org/show_bug.cgi?id=144932
Reviewed by Dan Bernstein.
This change is good if it compiles and links successfully.
- bindings/scripts/CodeGeneratorJS.pm:
(ExportLabelForClass):
Add a list of the JS bindings classes that need WEBCORE_EXPORT to prevent exporting too many symbols from WebCore.
- bindings/scripts/test/JS/JSTestActiveDOMObject.h:
- bindings/scripts/test/JS/JSTestCustomNamedGetter.h:
- bindings/scripts/test/JS/JSTestEventConstructor.h:
- bindings/scripts/test/JS/JSTestEventTarget.h:
- bindings/scripts/test/JS/JSTestException.h:
- bindings/scripts/test/JS/JSTestGenerateIsReachable.h:
- bindings/scripts/test/JS/JSTestMediaQueryListListener.h:
- bindings/scripts/test/JS/JSTestNamedConstructor.h:
- bindings/scripts/test/JS/JSTestNode.h:
- bindings/scripts/test/JS/JSTestNondeterministic.h:
- bindings/scripts/test/JS/JSTestObj.h:
- bindings/scripts/test/JS/JSTestOverloadedConstructors.h:
- bindings/scripts/test/JS/JSTestSerializedScriptValueInterface.h:
- bindings/scripts/test/JS/JSTestTypedefs.h:
- bindings/scripts/test/JS/JSattribute.h:
- bindings/scripts/test/JS/JSreadonly.h:
Update bindings tests to remove WEBCORE_EXPORT from most of the tests,
but keep one test with WEBCORE_EXPORT and one test with WEBCORE_TESTSUPPORT_EXPORT.
- 6:43 PM Changeset in webkit [184227] by
-
- 2 edits in trunk/LayoutTests
AX: [Win] REGRESSION(r184213) breaks aria-menubar-menuitems.html
https://bugs.webkit.org/show_bug.cgi?id=144936
Reviewed by Chris Fleizach.
Use focusedElement + childAtIndex rather than accessibleElementById
as the latter seems to be failing, and the former worked on Windows
to get the menubar in the original version of this test.
- accessibility/aria-menubar-menuitems.html: Modified.
- 5:57 PM Changeset in webkit [184226] by
-
- 2 edits in trunk/Source/WebCore
[Win] Update DXSDK_DIR settings for build system.
Unreviewed build fix.
- WebCore.vcxproj/WebCore.proj: Add DXSDK_DIR definition for builder.
- 5:30 PM Changeset in webkit [184225] by
-
- 25 edits in trunk/Source/WebCore
Return Ref instead of PassRefPtr in create() factory functions of WebCore/dom
https://bugs.webkit.org/show_bug.cgi?id=144904
Reviewed by Andreas Kling.
We return Ref when we're sure if pointer can't be null. If not, we use RefPtr instead PassRefPtr.
No new tests, no behavior changes.
- dom/AutocompleteErrorEvent.h:
- dom/DOMNamedFlowCollection.h:
(WebCore::DOMNamedFlowCollection::create):
- dom/MessageChannel.h:
(WebCore::MessageChannel::create):
- dom/MessagePort.h:
- dom/MutationObserver.cpp:
(WebCore::MutationObserver::create):
- dom/MutationObserver.h:
- dom/NamedFlowCollection.h:
- dom/NodeFilter.h:
(WebCore::NodeFilter::create):
- dom/NodeIterator.h:
(WebCore::NodeIterator::create):
- dom/ProcessingInstruction.cpp:
(WebCore::ProcessingInstruction::create):
- dom/ProcessingInstruction.h:
- dom/ScriptedAnimationController.h:
(WebCore::ScriptedAnimationController::create):
- dom/SecurityPolicyViolationEvent.h:
- dom/StaticNodeList.h:
- dom/TemplateContentDocumentFragment.h:
- dom/Touch.cpp:
(WebCore::Touch::cloneWithNewTarget):
- dom/Touch.h:
(WebCore::Touch::create):
- dom/TouchEvent.h:
- dom/TouchList.h:
(WebCore::TouchList::create):
- dom/TreeWalker.h:
(WebCore::TreeWalker::create):
- dom/WebKitNamedFlow.cpp:
(WebCore::WebKitNamedFlow::create):
- dom/WebKitNamedFlow.h:
- dom/default/PlatformMessagePortChannel.cpp:
(WebCore::PlatformMessagePortChannel::create):
- dom/default/PlatformMessagePortChannel.h:
(WebCore::PlatformMessagePortChannel::MessagePortQueue::create):
- 4:13 PM Changeset in webkit [184224] by
-
- 2 edits in trunk/Source/WebCore
Web Inspector: REGRESSION (r181625): Timeline recording started from console.profile is always empty
https://bugs.webkit.org/show_bug.cgi?id=144882
Reviewed by Brian Burg.
A timeline recording will always be stopped before the end of the current run loop is observed. Before
dispatching the recording stopped event, the last run loop record must be completed and sent to the frontend
to prevent it from being discarded.
- inspector/InspectorTimelineAgent.cpp:
(WebCore::InspectorTimelineAgent::internalStop):
- 3:06 PM Changeset in webkit [184223] by
-
- 7 edits in trunk/Source/WebKit2
Make {Network, Web}ProcessProxy inherit from ProcessThrottleClient privately
https://bugs.webkit.org/show_bug.cgi?id=144886
Reviewed by Darin Adler.
{Network, Web}ProcessProxy are implemented in terms of ProcessThrottleClient, an
interface that defines the callbacks that ProcessThrottler makes to a client. We
do not want to allow arbitrary callers to make use of the interface ProcessThrottleClient.
So, we should use private inheritance to represent the relationship between
{Network, Web}ProcessProxy and ProcessThrottleClient instead of public inheritance.
Additionally make ProcessThrottler a instance variable of {Network, Web}ProcessProxy instead
of holding a smart pointer to a ProcessThrottler instance because we always want to
have a ProcessThrottler for each instance of {Network, Web}ProcessProxy.
- UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::NetworkProcessProxy): Update code to work now that m_throttler
is an instance of ProcessThrottler instead of a smart pointer to a ProcessThrottler instance.
(WebKit::NetworkProcessProxy::didFinishLaunching): Update code to access functions of
m_throttler by reference instead of by pointer.
(WebKit::NetworkProcessProxy::didCancelProcessSuspension): Ditto.
(WebKit::NetworkProcessProxy::processReadyToSuspend): Ditto.
(WebKit::NetworkProcessProxy::setIsHoldingLockedFiles): Ditto.
- UIProcess/Network/NetworkProcessProxy.h: Inherit from ProcessThrottleClient privately.
(WebKit::NetworkProcessProxy::throttler): Update code to work now that m_throttler is an
instance of ProcessThrottler instead of a smart pointer to a ProcessThrottler instance.
- UIProcess/ProcessThrottler.cpp:
(WebKit::ProcessThrottler::ProcessThrottler): Update code to work now that m_process is
a lvalue reference instead of a pointer.
(WebKit::ProcessThrottler::updateAssertion): Ditto.
(WebKit::ProcessThrottler::assertionWillExpireImminently): Ditto.
- UIProcess/ProcessThrottler.h: Ditto.
- UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::WebProcessProxy): Update code to work now that m_throttler is an
instance of ProcessThrottler instead of a smart pointer to a ProcessThrottler instance.
(WebKit::WebProcessProxy::didFinishLaunching): Update code to access functions of
m_throttler by reference instead of by pointer.
(WebKit::WebProcessProxy::processReadyToSuspend): Ditto.
(WebKit::WebProcessProxy::didCancelProcessSuspension): Ditto.
(WebKit::WebProcessProxy::setIsHoldingLockedFiles): Ditto.
- UIProcess/WebProcessProxy.h: Inherit from ProcessThrottleClient privately.
(WebKit::WebProcessProxy::throttler): Update code to work now that m_throttler is an
instance of ProcessThrottler instead of a smart pointer to a ProcessThrottler instance.
- 2:50 PM Changeset in webkit [184222] by
-
- 2 edits in trunk/Tools
Fix the previous commit, which made both bubbles green.
- BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Styles/StatusLineView.css:
(.status-line.good .bubble):
(.status-line.bad .bubble):
(.status-line.good .message):
- 2:30 PM Changeset in webkit [184221] by
-
- 2 edits in trunk/Tools
People with Deuteranopia can't tell red and green apart on bot watcher's dashboard
https://bugs.webkit.org/show_bug.cgi?id=144929
Reviewed by Eric Carlson.
Tweaked colors as I see fit.
- BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Styles/StatusLineView.css:
(.status-line.good .bubble):
(.status-line.bad .bubble):
- 2:28 PM Changeset in webkit [184220] by
-
- 3 edits1 add in trunk/Source/JavaScriptCore
[JSC] Make the NegZero backward propagated flags of ArithMod stricter
https://bugs.webkit.org/show_bug.cgi?id=144897
Patch by Benjamin Poulain <bpoulain@apple.com> on 2015-05-12
Reviewed by Geoffrey Garen.
The NegZero flags of ArithMod were the same as ArithDiv: both children were
marked as needing to handle NegativeZero.
Lucky for us, ArithMod is quite a bit different than ArithDiv.
First, the sign of the result is completely independent from
the sign of the divisor. A zero on the divisor always produces a NaN.
That's great, we can remove the NodeBytecodeNeedsNegZero
from the flags propagated to child2.
Second, the sign of the result is always the same as the sign of
the dividend. A dividend of zero produces a zero of same sign
unless the divisor is zero (in which case the result is NaN).
This is great too: we can just pass the flags we got into
ArithMod.
With those two out of the way, we can make a faster version of ArithRound
for Kraken's oscillator. Since we no longer care about negative zero,
rounding becomes cast<int32>(value + 0.5). This gives ~3% faster runtime
on the benchmark.
Unfortunatelly, most of the time is spent in FTL and the same optimization
does not apply well just yet: rdar://problem/20904149.
- dfg/DFGBackwardsPropagationPhase.cpp:
(JSC::DFG::BackwardsPropagationPhase::propagate):
Never add NodeBytecodeNeedsNegZero unless needed by the users of this node.
- dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileArithRound):
Faster Math.round() when negative zero is not important.
- tests/stress/arith-modulo-node-behaviors.js: Added.
(moduloWithNegativeZeroDividend):
(moduloWithUnusedNegativeZeroDividend):
(moduloWithNegativeZeroDivisor):
- 2:24 PM Changeset in webkit [184219] by
-
- 3 edits2 adds in trunk
REGRESSION(r175617): Some text doesn't render on internationalculinarycenter.com
https://bugs.webkit.org/show_bug.cgi?id=144917
rdar://problem/20545878
Reviewed by Andreas Kling.
This patch ensures that text stroke width value is taken into account while
calculating visual overflow for simple line layout.
Ceiling the text stroke width value matches the normal text layout behaviour.
Source/WebCore:
Test: fast/text/simple-line-layout-text-stroke-width.html
- rendering/SimpleLineLayoutFunctions.cpp:
(WebCore::SimpleLineLayout::paintFlow):
(WebCore::SimpleLineLayout::collectFlowOverflow):
LayoutTests:
- fast/text/simple-line-layout-text-stroke-width-expected.txt: Added.
- fast/text/simple-line-layout-text-stroke-width.html: Added.
- 2:20 PM Changeset in webkit [184218] by
-
- 2 edits in trunk/Source/JavaScriptCore
Refactor MachineStackMarker.cpp so that it's easier to reason about MachineThreads::Thread.
https://bugs.webkit.org/show_bug.cgi?id=144925
Reviewed by Michael Saboff.
Currently, the code in MachineStackMarker.cpp is written as a bunch of functions that
operate on the platformThread value in the MachineThreads::Thread struct. Instead, we
can apply better OO encapsulation and convert all these functions into methods of the
MachineThreads::Thread struct.
This will also make it easier to reason about the fix for
https://bugs.webkit.org/show_bug.cgi?id=144924 later.
- heap/MachineStackMarker.cpp:
(JSC::getCurrentPlatformThread):
(JSC::MachineThreads::Thread::createForCurrentThread):
(JSC::MachineThreads::Thread::operator!=):
(JSC::MachineThreads::Thread::operator==):
(JSC::MachineThreads::addCurrentThread):
(JSC::MachineThreads::removeThreadIfFound):
(JSC::MachineThreads::Thread::suspend):
(JSC::MachineThreads::Thread::resume):
(JSC::MachineThreads::Thread::getRegisters):
(JSC::MachineThreads::Thread::Registers::stackPointer):
(JSC::MachineThreads::Thread::freeRegisters):
(JSC::MachineThreads::Thread::captureStack):
(JSC::MachineThreads::tryCopyOtherThreadStack):
(JSC::MachineThreads::tryCopyOtherThreadStacks):
(JSC::equalThread): Deleted.
(JSC::suspendThread): Deleted.
(JSC::resumeThread): Deleted.
(JSC::getPlatformThreadRegisters): Deleted.
(JSC::otherThreadStackPointer): Deleted.
(JSC::freePlatformThreadRegisters): Deleted.
(JSC::otherThreadStack): Deleted.
- 2:04 PM Changeset in webkit [184217] by
-
- 4 edits in trunk/Source/JavaScriptCore
Array.slice should have a fast path like Array.splice
https://bugs.webkit.org/show_bug.cgi?id=144901
Reviewed by Geoffrey Garen.
Add a fast memcpy path to Array.prototype.slice as done for Array.prototype.splice.
In Kraken, this appears to be 30% win on stanford-crypto-ccm and 10% win on stanford-crypto-pbkdf2.
- runtime/ArrayPrototype.cpp:
(JSC::arrayProtoFuncSlice):
- runtime/JSArray.cpp:
(JSC::JSArray::fastSlice): Added.
- runtime/JSArray.h:
- 2:02 PM Changeset in webkit [184216] by
-
- 2 edits in trunk/Source/WebCore
Build fix.
- platform/spi/cf/CFNetworkSPI.h:
- 1:46 PM Changeset in webkit [184215] by
-
- 12 edits3 copies in trunk/Source/WebKit2
Production builds should link against the shims directly
https://bugs.webkit.org/show_bug.cgi?id=144922
rdar://problem/19708579
Reviewed by Dan Bernstein.
- Configurations/NetworkService.xcconfig:
Point to an Info.plist that has DYLD_INSERT_LIBRARIES set on 10.10 and earlier.
Link directly against SecItemShim.dylib where possible.
- Configurations/PluginProcessShim.xcconfig:
Set the install name and install path.
- Configurations/PluginService.32.xcconfig:
Point to an Info.plist that has DYLD_INSERT_LIBRARIES set on 10.10 and earlier.
Link directly against PluginProcessShim.dylib where possible.
- Configurations/PluginService.64.xcconfig:
Point to an Info.plist that has DYLD_INSERT_LIBRARIES set on 10.10 and earlier.
Link directly against PluginProcessShim.dylib where possible.
- Configurations/SecItemShim.xcconfig:
Set the install name and install path.
- Configurations/WebContentService.xcconfig:
Point to an Info.plist that has DYLD_INSERT_LIBRARIES set on 10.10 and earlier.
Link directly against WebProcessShim.dylib where possible.
- Configurations/WebProcessShim.xcconfig:
Set the install name and install path.
- NetworkProcess/EntryPoint/mac/XPCService/NetworkService/Info-OSX-10.9-10.10.plist: Copied from Source/WebKit2/NetworkProcess/EntryPoint/mac/XPCService/NetworkService/Info-OSX.plist.
- NetworkProcess/EntryPoint/mac/XPCService/NetworkService/Info-OSX.plist:
Add an Info.plist for 10.9 and 10.10 and remove DYLD_INSERT_LIBRARIES from the non-legacy plist.
- PluginProcess/EntryPoint/mac/XPCService/PluginService.32-64-10.9-10.10.Info.plist: Copied from Source/WebKit2/PluginProcess/EntryPoint/mac/XPCService/PluginService.32-64.Info.plist.
- PluginProcess/EntryPoint/mac/XPCService/PluginService.32-64.Info.plist:
Add an Info.plist for 10.9 and 10.10 and remove DYLD_INSERT_LIBRARIES from the non-legacy plist.
- WebKit2.xcodeproj/project.pbxproj:
Add new files.
- WebProcess/EntryPoint/mac/XPCService/WebContentService/Info-OSX-10.9-10.10.plist: Copied from Source/WebKit2/WebProcess/EntryPoint/mac/XPCService/WebContentService/Info-OSX.plist.
- WebProcess/EntryPoint/mac/XPCService/WebContentService/Info-OSX.plist:
Add an Info.plist for 10.9 and 10.10 and remove DYLD_INSERT_LIBRARIES from the non-legacy plist.
- 12:49 PM Changeset in webkit [184214] by
-
- 2 edits in trunk/Source/WebCore
ROLLOUT [Win] Unreviewed build fix for older DirectX build environments.
- platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
- 12:44 PM Changeset in webkit [184213] by
-
- 9 edits2 adds2 deletes in trunk
AX: [ATK] ARIA menu items should not have anonymous block children
https://bugs.webkit.org/show_bug.cgi?id=144653
Reviewed by Chris Fleizach.
Source/WebCore:
Exclude anonymous block children of menu items from the accessible tree
for ATK. Exclude list markers from the accessibility tree if the parent
is not being exposed as an accessible list item.
No new tests. Instead, modified the existing tests to reflect the correct
behavior.
- accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::computeAccessibilityIsIgnored):
- accessibility/atk/AccessibilityObjectAtk.cpp:
(WebCore::AccessibilityObject::accessibilityPlatformIncludesObject):
LayoutTests:
The existing test was not cross-platform friendly and was rewritten as
part of the fix for this bug. New expectations were generated for Mac,
Efl, and Gtk. The shared expectations and the Windows expectations were
removed so as not to break the Win port. The search-predicate test also
was updated because it included an ARIA treeitem which had a listmarker
child it should not have, and no longer has as a result of the bug fix.
- accessibility/aria-menubar-menuitems-expected.txt: Removed. Platform hierarchies vary vastly.
- accessibility/aria-menubar-menuitems.html: Rewritten to work better as a shared test.
- platform/efl/accessibility/aria-menubar-menuitems-expected.txt: Added.
- platform/gtk/accessibility/aria-menubar-menuitems-expected.txt: Added.
- platform/mac/accessibility/aria-menubar-menuitems-expected.txt: Modified as result of rewrite.
- platform/mac/accessibility/search-predicate.html: Modified to reflect correct treeitem children.
- platform/mac/accessibility/search-predicate-expected.txt: Modified to reflect correct treeitem children.
- platform/win/accessibility/aria-menubar-menuitems-expected.txt: Removed. New file should be generated.
- resources/accessibility-helper.js: Added an argument so dumpAccessibilityTree can print title rather than value.
(dumpAccessibilityTree):
- 12:42 PM Changeset in webkit [184212] by
-
- 3 edits in trunk/Source/WebKit2
Undo a mistaken part of r184211.
- UIProcess/Cocoa/WebProcessPoolCocoa.mm: (WebKit::WebProcessPool::cookieStorageDirectory):
- UIProcess/WebProcessPool.h:
cookieStorageDirectory() is used on iOS, so bring it back.
- 12:37 PM Changeset in webkit [184211] by
-
- 19 edits in trunk
[Mac] Share cookie storage between UI process and secondary processes
https://bugs.webkit.org/show_bug.cgi?id=144820
rdar://problem/20572830
Reviewed by Darin Adler.
Source/WebCore:
- platform/network/mac/CookieJarMac.mm:
- platform/spi/cf/CFNetworkSPI.h:
Moved SPIs into an SPI header, so that WebKit2 could use them too without duplication.
Source/WebKit2:
Pass a cookie storage descriptor that includes file path as well as other data that
CFNetwork needs to use the same cookies storage as in another process (session cookies
are still separate).
While at it, started to clean up code that passes cookie storage path. It was mostly
dead and unimplemented, only used in child processes when USE(SECCOMP_FILTERS) was enabled.
- NetworkProcess/cocoa/NetworkProcessCocoa.mm:
(WebKit::NetworkProcess::platformInitializeNetworkProcessCocoa):
- Shared/Network/NetworkProcessCreationParameters.cpp:
(WebKit::NetworkProcessCreationParameters::encode):
(WebKit::NetworkProcessCreationParameters::decode):
- Shared/Network/NetworkProcessCreationParameters.h:
- Shared/WebProcessCreationParameters.cpp:
(WebKit::WebProcessCreationParameters::encode):
(WebKit::WebProcessCreationParameters::decode):
- Shared/WebProcessCreationParameters.h:
- UIProcess/Cocoa/WebProcessPoolCocoa.mm:
(WebKit::WebProcessPool::platformInitializeWebProcess):
(WebKit::WebProcessPool::platformInitializeNetworkProcess):
(WebKit::WebProcessPool::platformDefaultCookieStorageDirectory): Deleted.
- UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::ensureNetworkProcess):
(WebKit::WebProcessPool::createNewWebProcess):
(WebKit::WebProcessPool::cookieStorageDirectory):
- UIProcess/WebProcessPool.h:
- UIProcess/efl/WebProcessPoolEfl.cpp:
(WebKit::WebProcessPool::platformDefaultCookieStorageDirectory): Deleted.
- UIProcess/gtk/WebProcessPoolGtk.cpp:
(WebKit::WebProcessPool::platformDefaultCookieStorageDirectory): Deleted.
- WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::platformInitializeWebProcess):
Tools:
- WebKitTestRunner/TestController.cpp: (WTR::TestController::initialize):
Don't set cookie storage directory, as it's a no-op, we use an ephemeral session
for testing.
- 12:29 PM Changeset in webkit [184210] by
-
- 5 edits in trunk/Source
[WK2] Log using diagnostic logging how far back in history we navigate
https://bugs.webkit.org/show_bug.cgi?id=144919
<rdar://problem/20920043>
Reviewed by Gavin Barraclough.
Source/WebCore:
Add diagnostic logging keys for logging back history navigation delta.
- page/DiagnosticLoggingKeys.cpp:
(WebCore::DiagnosticLoggingKeys::deltaKey):
(WebCore::DiagnosticLoggingKeys::backNavigationKey):
- page/DiagnosticLoggingKeys.h:
Source/WebKit2:
Log using diagnostic logging how far back in history we navigate.
When a back navigation or a backward indexed navigation occurs, we
log the delta: (HistoryListSize - 1) - NewHistoryItemIndex.
- UIProcess/WebBackForwardList.cpp:
(WebKit::WebBackForwardList::goToItem):
- 12:23 PM Changeset in webkit [184209] by
-
- 9 edits in trunk/Tools
Fix internal build configuration issues
https://bugs.webkit.org/show_bug.cgi?id=144762
Reviewed by David Kilzer.
Modifies DumpRenderTree's xcconfig files to use SDK selectors.
- DumpRenderTree/mac/Configurations/Base.xcconfig:
- DumpRenderTree/mac/Configurations/DebugRelease.xcconfig:
- DumpRenderTree/mac/Configurations/DumpRenderTree.xcconfig:
- DumpRenderTree/mac/Configurations/DumpRenderTreeApp.xcconfig:
- DumpRenderTree/mac/Configurations/DumpRenderTreeLibrary.xcconfig:
- DumpRenderTree/mac/Configurations/ImageDiff.xcconfig:
- DumpRenderTree/mac/Configurations/LayoutTestHelper.xcconfig:
- DumpRenderTree/mac/Configurations/TestNetscapePlugIn.xcconfig:
- 11:40 AM Changeset in webkit [184208] by
-
- 4 edits in trunk/Source/WebCore
Avoid CSSPrimitiveValue allocation when parsing sizes
https://bugs.webkit.org/show_bug.cgi?id=144910
Reviewed by Darin Adler.
Turn non-calc length calculation logic into a static method,
so that it can be called without a CSSPrimitiveValue object.
Then, call that method to calculate the default source-size length.
No new tests, since no functionality change.
- css/CSSPrimitiveValue.cpp:
(WebCore::CSSPrimitiveValue::computeLengthDouble): Split out the logic for everything that's not calc into a static method.
(WebCore::CSSPrimitiveValue::computeNonCalcLengthDouble): Static length computation logic for everything that is not calc.
- css/CSSPrimitiveValue.h:
(WebCore::CSSPrimitiveValue::isFontRelativeLength): Turn into static.
- css/SourceSizeList.cpp:
(WebCore::defaultLength): Use the new static method to compute the length.
(WebCore::computeLength): Use the calc length computation methods directly.
- 11:07 AM Changeset in webkit [184207] by
-
- 2 edits in trunk/Source/WebCore
[Win] Unreviewed build fix for older DirectX build environments.
- platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp: Switch back to our
d3d stub header to avoid build failures on July 2004 DXSDK build environments.
- 11:01 AM Changeset in webkit [184206] by
-
- 5 edits2 adds in trunk/Source/JavaScriptCore
OSR availability analysis would be more scalable (and correct) if it did more liveness pruning
https://bugs.webkit.org/show_bug.cgi?id=143078
Reviewed by Andreas Kling.
In https://bugs.webkit.org/show_bug.cgi?id=144883, we found an example of where liveness
pruning is actually necessary. Well, not quite: we just need to prune out keys from the
heap availability map where the base node doesn't dominate the point where we are asking
for availability. If we don't do this, then eventually the IR gets corrupt because we'll
insert PutHints that reference the base node in places where the base node doesn't
dominate. But if we're going to do any pruning, then it makes sense to prune by bytecode
liveness. This is the strongest possible pruning we can do, and it should be sound. We
shouldn't have a node available for a virtual register if that register is live and the
node doesn't dominate.
Making this work meant reusing the prune-to-liveness algorithm from the FTL backend. So, I
abstracted this a bit better. You can now availabilityMap.pruneByLiveness(graph, origin).
- dfg/DFGAvailabilityMap.cpp:
(JSC::DFG::AvailabilityMap::pruneHeap):
(JSC::DFG::AvailabilityMap::pruneByLiveness):
(JSC::DFG::AvailabilityMap::prune): Deleted.
- dfg/DFGAvailabilityMap.h:
- dfg/DFGOSRAvailabilityAnalysisPhase.cpp:
(JSC::DFG::OSRAvailabilityAnalysisPhase::run):
- ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::buildExitArguments):
- tests/stress/liveness-pruning-needed-for-osr-availability.js: Added. This is a proper regression test.
- tests/stress/liveness-pruning-needed-for-osr-availability-eager.js: Added. This is the original reduced test case, requires eager-no-cjit to fail prior to this changeset.
- 10:30 AM Changeset in webkit [184205] by
-
- 2 edits in trunk/LayoutTests
fast/frames/flattening/iframe-flattening-resize-event-count.html times out on Yosemite WK2
https://bugs.webkit.org/show_bug.cgi?id=144155
- platform/mac-wk2/TestExpectations: Marking as flaky for now.
- 9:41 AM Changeset in webkit [184204] by
-
- 7 edits in trunk/Source/WebCore
[Mac] Refine media playback target client configuration
https://bugs.webkit.org/show_bug.cgi?id=144892
Reviewed by Brent Fulgham.
Client and target picker state changes fequently happen several times in quick succession, so
don't react to immediately so we can batch callbacks to the web process.
- Modules/mediasession/WebMediaSessionManager.cpp:
(WebCore::ClientState::ClientState): Store the client as a reference rather than a pointer
because it can never be NULL.
(WebCore::ClientState::operator == ): New.
(WebCore::WebMediaSessionManager::addPlaybackTargetPickerClient): Schedule the initial client
configuration and a target configuration check.
(WebCore::WebMediaSessionManager::removePlaybackTargetPickerClient): Schedule a target monitoring
update, and a target configuration check.
(WebCore::WebMediaSessionManager::removeAllPlaybackTargetPickerClients): Ditto.
(WebCore::WebMediaSessionManager::showPlaybackTargetPicker): Schedule a target monitoring update.
(WebCore::WebMediaSessionManager::clientStateDidChange): If the client whose state has changed
can play to a target, tell it to start using the target even if it isn't playing as long as
no other client is actively using a target.
(WebCore::WebMediaSessionManager::setPlaybackTarget): Configure clients after a pause.
(WebCore::WebMediaSessionManager::configureNewClients): New, do new client configuration.
(WebCore::WebMediaSessionManager::configurePlaybackTargetClients): New, configure target clients.
(WebCore::WebMediaSessionManager::scheduleDelayedTask): Schedule the timer.
(WebCore::WebMediaSessionManager::taskTimerFired): Execute delayed tasks.
(WebCore::WebMediaSessionManager::find):
- Modules/mediasession/WebMediaSessionManager.h:
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::setMuted): Call updateMediaState.
(WebCore::HTMLMediaElement::setPlaying): Ditto.
(WebCore::HTMLMediaElement::mediaPlayerCurrentPlaybackTargetIsWirelessChanged): Ditto.
(WebCore::HTMLMediaElement::enqueuePlaybackTargetAvailabilityChangedEvent): Expand logging.
(WebCore::HTMLMediaElement::updateMediaState): New, don't broadcast a media state change
unless something actually changed.
- html/HTMLMediaElement.h:
- html/HTMLMediaSession.cpp:
(WebCore::HTMLMediaSession::externalOutputDeviceAvailableDidChange): Update logging.
(WebCore::HTMLMediaSession::setShouldPlayToPlaybackTarget): Ditto.
(WebCore::HTMLMediaSession::mediaEngineUpdated): Cleanup.
- platform/graphics/avfoundation/objc/MediaPlaybackTargetPickerMac.mm:
(WebCore::MediaPlaybackTargetPickerMac::showPlaybackTargetPicker): Remove the call to
deprecated API and the "-Wdeprecated-declarations".
- 9:25 AM Changeset in webkit [184203] by
-
- 6 edits in trunk/Source/WebKit2
ChildProcess should take an os_activity
https://bugs.webkit.org/show_bug.cgi?id=141781
Reviewed by Ryosuke Niwa.
- DatabaseProcess/EntryPoint/mac/XPCService/DatabaseServiceEntryPoint.mm:
(DatabaseServiceInitializer):
- NetworkProcess/EntryPoint/mac/XPCService/NetworkServiceEntryPoint.mm:
(NetworkServiceInitializer):
- PluginProcess/EntryPoint/mac/XPCService/PluginServiceEntryPoint.mm:
(PluginServiceInitializer):
- WebProcess/EntryPoint/mac/XPCService/WebContentServiceEntryPoint.mm:
(WebContentServiceInitializer):
Start an os_activity in the initializer for each XPC service.
- config.h:
Add HAVE_OS_ACTIVITY where available.
- 7:31 AM Changeset in webkit [184202] by
-
- 3 edits in trunk/Tools
Switch to launching iOS Simulator using bundle ID
<rdar://problem/20916096>
This is required to fix the layout tests on internal bots.
- Scripts/webkitdirs.pm:
(relaunchIOSSimulator): Switch to use bundle ID to launch iOS
Simulator. Add FIXME.
(quitIOSSimulator): Add FIXME.
- Scripts/webkitpy/port/ios.py:
(IOSSimulatorPort.setup_test_run): Switch to use bundle ID to
launch iOS Simulator. Add FIXME.
(IOSSimulatorPort.check_sys_deps): Add FIXME.
- 7:22 AM Changeset in webkit [184201] by
-
- 2 edits in trunk/Tools
Fix run-javascriptcore-tests step on the WinCairo bot
https://bugs.webkit.org/show_bug.cgi?id=144866
Reviewed by Csaba Osztrogonác.
Catch exception when failing to create soft link, so we can fall back to copy.
- Scripts/run-jsc-stress-tests:
- 6:28 AM Changeset in webkit [184200] by
-
- 2 edits in trunk/Source/WebKit2
Unreviewed. Fix GTK+ debug build after r184197.
- UIProcess/gtk/RedirectedXCompositeWindow.cpp: Add missing include.
- 5:01 AM Changeset in webkit [184199] by
-
- 10 edits in trunk
AX: [GTK] Defer to WebCore Accessibility for table exposure
https://bugs.webkit.org/show_bug.cgi?id=144896
Reviewed by Chris Fleizach.
Source/WebCore:
Stop unconditionally exposing all HTMLTableElement nodes as AccessibilityTables
for WebKitGtk.
No new tests. Several existing tests already cover table exposure. They have
been updated accordingly.
- accessibility/AccessibilityTable.cpp:
(WebCore::AccessibilityTable::computeIsTableExposableThroughAccessibility):
LayoutTests:
Several platform tests are testing accessibility of tables. In order to ensure
that those tables continue to be exposed as tables, a border was added so that
they are not judged to be layout tables. For the remainder of the tests, the
expectations were updated to reflect the fact that layout tables are no longer
exposed as AccessibilityTables in WebKitGtk.
- platform/gtk/accessibility/caret-offsets.html: Added border to table.
- platform/gtk/accessibility/table-detection-expected.txt: Updated to reflect new behavior.
- platform/gtk/accessibility/table-hierarchy-expected.txt: Updated to reflect new behavior.
- platform/gtk/accessibility/table-one-cell-expected.txt: Updated to reflect new behavior.
- platform/gtk/accessibility/table-with-rules-expected.txt: Updated to reflect new behavior.
- platform/gtk/accessibility/text-for-range-table-cells.html: Added border to table.
- platform/gtk/accessibility/text-for-table.html: Added border to table.
- 4:59 AM Changeset in webkit [184198] by
-
- 6 edits in trunk
[GTK][WK2] rowAtIndex is not implemented in DRT/WKTR
https://bugs.webkit.org/show_bug.cgi?id=116971
Reviewed by Chris Fleizach.
Source/WebCore:
Because ATK lacks API to directly get an accessible row via its index,
the implementation of rowAtIndex gets a cell in the indexed row and
returns the parent row. The failing test continued to fail because
AccessibilityARIAGridCell::parentTable called parentObjectUnignored at
most twice, the second call in place to handle rows which are included
in the tree. However, given a well-formed ARIA grid with a rowgroup that
is interactive, that rowgroup also needs to be in the tree necessitating
parentObjectUnignored be called a third time to get to the grid. Given a
poorly-formed ARIA grid, there may additional objects which pass the test
for inclusion standing in between the cell and grid necessitating more
calls still. Therefore, ascend the hierarchy to find the parent grid.
No new tests. The failing test now passes.
- accessibility/AccessibilityARIAGridCell.cpp:
(WebCore::AccessibilityARIAGridCell::parentTable):
Tools:
Implement rowAtIndex for ATK.
- WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:
(WTR::AccessibilityUIElement::rowAtIndex):
LayoutTests:
- platform/gtk/TestExpectations: Removed previously-failing test.
- 4:11 AM Changeset in webkit [184197] by
-
- 21 edits3 adds in trunk/Source
[X11] Add XUniquePtr and XUniqueResource to automatically free X resources
https://bugs.webkit.org/show_bug.cgi?id=144521
Reviewed by Darin Adler.
Source/WebCore:
Add XUniquePtr as a template alias of std:unique_ptr to handle X
resources using pointers and XUniqueResource as a new class to
handle X resources using a long unsigned identifier. This
simplifies the code and makes it more difficult to leak X resources.
- PlatformEfl.cmake: Add new files to compilation.
- PlatformGTK.cmake: Ditto.
- platform/graphics/cairo/BackingStoreBackendCairoX11.cpp:
(WebCore::BackingStoreBackendCairoX11::BackingStoreBackendCairoX11):
Remove the display member, and use XUnique for Pixmap and GC.
(WebCore::BackingStoreBackendCairoX11::~BackingStoreBackendCairoX11):
Remove code to explicitly free Pixmap and GC.
(WebCore::BackingStoreBackendCairoX11::scroll):
- platform/graphics/cairo/BackingStoreBackendCairoX11.h:
- platform/graphics/glx/GLContextGLX.cpp:
(WebCore::GLContextGLX::createWindowContext): Use XUnique and the
new constructor that receives a XID, since there's no longer
conflict with the one receiving a Pbuffer.
(WebCore::GLContextGLX::createPbufferContext): Use XUnique and the
new constructor that receives a XUniqueGLXPbuffer&&.
(WebCore::GLContextGLX::createPixmapContext):
(WebCore::GLContextGLX::createContext):
(WebCore::GLContextGLX::GLContextGLX):
(WebCore::GLContextGLX::~GLContextGLX): Remove code to explicitly
free X resources.
(WebCore::GLContextGLX::makeContextCurrent):
(WebCore::GLContextGLX::cairoDevice):
(WebCore::GLContextGLX::platformContext):
- platform/graphics/glx/GLContextGLX.h:
- platform/graphics/surfaces/egl/EGLXSurface.cpp:
(WebCore::EGLXTransportSurfaceClient::EGLXTransportSurfaceClient):
(WebCore::EGLXTransportSurfaceClient::destroy):
(WebCore::EGLXTransportSurfaceClient::prepareTexture):
- platform/graphics/surfaces/egl/EGLXSurface.h:
- platform/graphics/surfaces/glx/GLXConfigSelector.h:
(WebCore::GLXConfigSelector::findMatchingConfig): Use XUnique
instead of the custom std::unique X11Deleter.
(WebCore::GLXConfigSelector::findMatchingConfigWithVisualId): Ditto.
- platform/graphics/surfaces/glx/GLXSurface.cpp:
(WebCore::GLXTransportSurface::GLXTransportSurface): Ditto.
(WebCore::GLXOffScreenSurface::initialize):
- platform/graphics/surfaces/glx/X11Helper.cpp:
(WebCore::X11Helper::createOffScreenWindow): Ditto.
(WebCore::X11Helper::createPixmap): Ditto.
- platform/graphics/surfaces/glx/X11Helper.h:
- platform/graphics/x11/XUniquePtr.h: Remove X11Deleter.
(WebCore::XPtrDeleter::operator()):
- platform/graphics/x11/XUniqueResource.cpp: Added.
(WebCore::XUniqueResource<XResource::Colormap>::deleteXResource):
(WebCore::XUniqueResource<XResource::Damage>::deleteXResource):
(WebCore::XUniqueResource<XResource::Pixmap>::deleteXResource):
(WebCore::XUniqueResource<XResource::Window>::deleteXResource):
(WebCore::XUniqueResource<XResource::GLXPbuffer>::deleteXResource):
(WebCore::XUniqueResource<XResource::GLXPixmap>::deleteXResource):
- platform/graphics/x11/XUniqueResource.h: Added.
(WebCore::XUniqueResource::XUniqueResource):
(WebCore::XUniqueResource::operator=):
(WebCore::XUniqueResource::~XUniqueResource):
(WebCore::XUniqueResource::get):
(WebCore::XUniqueResource::release):
(WebCore::XUniqueResource::reset):
(WebCore::XUniqueResource::operator!):
(WebCore::XUniqueResource::operator UnspecifiedBoolType*):
Source/WebKit2:
Use XUniquePtr and XUniqueResource to free X resources.
- PlatformEfl.cmake: Add Source/WebCore/platform/graphics/x11 dir
to the include dir list.
- PlatformGTK.cmake: Ditto.
- UIProcess/cairo/BackingStoreCairo.cpp:
(WebKit::BackingStore::createBackend): Do not pass the display to
the BackingStoreBackendCairoX11 constructor.
- UIProcess/gtk/RedirectedXCompositeWindow.cpp:
(WebKit::RedirectedXCompositeWindow::RedirectedXCompositeWindow):
(WebKit::RedirectedXCompositeWindow::~RedirectedXCompositeWindow):
(WebKit::RedirectedXCompositeWindow::resize):
(WebKit::RedirectedXCompositeWindow::cleanupPixmapAndPixmapSurface):
(WebKit::RedirectedXCompositeWindow::surface):
- UIProcess/gtk/RedirectedXCompositeWindow.h:
(WebKit::RedirectedXCompositeWindow::windowID):
- WebProcess/Plugins/Netscape/NetscapePlugin.h:
- WebProcess/Plugins/Netscape/x11/NetscapePluginX11.cpp:
(WebKit::NetscapePlugin::platformPostInitializeWindowless):
(WebKit::NetscapePlugin::platformDestroy):
(WebKit::NetscapePlugin::platformGeometryDidChange):
(WebKit::NetscapePlugin::platformPaint):
- 3:17 AM Changeset in webkit [184196] by
-
- 1 copy in releases/WebKitGTK/webkit-2.8.2
WebKitGTK+ 2.8.2
- 3:17 AM Changeset in webkit [184195] by
-
- 4 edits in releases/WebKitGTK/webkit-2.8
Unreviewed. Update OptionsGTK.cmake and NEWS for 2.8.2 release.
.:
- Source/cmake/OptionsGTK.cmake: Bump version numbers.
Source/WebKit2:
- gtk/NEWS: Add release notes for 2.8.2.
- 3:16 AM Changeset in webkit [184194] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Unreviewed. Fix the build with RESOURCE_TIMING disabled.
- loader/ThreadableLoader.cpp:
(WebCore::ThreadableLoaderOptions::isolatedCopy):
- 2:53 AM Changeset in webkit [184193] by
-
- 18 edits in trunk/Source/WebCore
Move TransformOperation-based classes off of PassRefPtr
https://bugs.webkit.org/show_bug.cgi?id=144876
Reviewed by Darin Adler.
Remove uses of PassRefPtr from the TransformOperation class and its
derivatives, operating on Ref and RefPtr objects instead.
- platform/graphics/transforms/IdentityTransformOperation.h:
- platform/graphics/transforms/Matrix3DTransformOperation.cpp:
(WebCore::Matrix3DTransformOperation::blend):
- platform/graphics/transforms/Matrix3DTransformOperation.h:
- platform/graphics/transforms/MatrixTransformOperation.cpp:
(WebCore::MatrixTransformOperation::blend):
- platform/graphics/transforms/MatrixTransformOperation.h:
- platform/graphics/transforms/PerspectiveTransformOperation.cpp:
(WebCore::PerspectiveTransformOperation::blend):
- platform/graphics/transforms/PerspectiveTransformOperation.h:
- platform/graphics/transforms/RotateTransformOperation.cpp:
(WebCore::RotateTransformOperation::blend):
- platform/graphics/transforms/RotateTransformOperation.h:
- platform/graphics/transforms/ScaleTransformOperation.cpp:
(WebCore::ScaleTransformOperation::blend):
- platform/graphics/transforms/ScaleTransformOperation.h:
- platform/graphics/transforms/SkewTransformOperation.cpp:
(WebCore::SkewTransformOperation::blend):
- platform/graphics/transforms/SkewTransformOperation.h:
- platform/graphics/transforms/TransformOperation.h:
- platform/graphics/transforms/TransformOperations.cpp:
(WebCore::TransformOperations::blendByMatchingOperations):
- platform/graphics/transforms/TranslateTransformOperation.cpp:
(WebCore::TranslateTransformOperation::blend):
- platform/graphics/transforms/TranslateTransformOperation.h:
- 2:36 AM Changeset in webkit [184192] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/JavaScriptCore
Merge r184170 - Workaround for Cortex-A53 erratum 843419
https://bugs.webkit.org/show_bug.cgi?id=144680
Reviewed by Michael Saboff.
This patch is about to give simple workaround for Cortex-A53 erratum 843419.
It inserts nops after ADRP instruction to avoid wrong address accesses.
- assembler/ARM64Assembler.h:
(JSC::ARM64Assembler::adrp):
(JSC::ARM64Assembler::nopCortexA53Fix843419):
- 2:35 AM Changeset in webkit [184191] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r184154 - Clean up redundant resources in case of failure in GLContextEGL context creation methods
https://bugs.webkit.org/show_bug.cgi?id=144878
Reviewed by Martin Robinson.
GLContextEGL::createWindowContext() and GLContextEGL::createPixmapContext() methods
should clean up the freshly-created resources when prematurely returning due to a
failure.
- platform/graphics/egl/GLContextEGL.cpp:
(WebCore::GLContextEGL::createWindowContext):
(WebCore::GLContextEGL::createPixmapContext):
- 2:25 AM Changeset in webkit [184190] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r184104 - WebContent crash under com.apple.WebCore: WebCore::WebKitCSSResourceValue::isCSSValueNone const + 6
https://bugs.webkit.org/show_bug.cgi?id=144870
rdar://problem/20727702
Reviewed by Simon Fraser.
No repro but we are seeing null pointer crashes like this:
Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0 com.apple.WebCore 0x00007fff92da5706 WebCore::WebKitCSSResourceValue::isCSSValueNone() const + 6
1 com.apple.WebCore 0x00007fff93382b48 WebCore::MaskImageOperation::isCSSValueNone() const + 24
2 com.apple.WebCore 0x00007fff92e0475e WebCore::FillLayer::hasNonEmptyMaskImage() const + 30
- platform/graphics/MaskImageOperation.cpp:
(WebCore::MaskImageOperation::MaskImageOperation):
(WebCore::MaskImageOperation::isCSSValueNone):
This would crash like this if both m_styleImage and m_cssMaskImageValue are null.
There are no obvious guarantees that this doesn't happen. Two of the constructor variants allow it
and there is setImage which may turn m_styleImage null later too.
Fix by making null m_cssMaskImageValue always signify CSSValueNone.
(WebCore::MaskImageOperation::cssValue):
- 2:24 AM Changeset in webkit [184189] by
-
- 3 edits2 adds in releases/WebKitGTK/webkit-2.8
Merge r184097 - AX: Crash at WebCore::AccessibilityMenuList::addChildren()
https://bugs.webkit.org/show_bug.cgi?id=144860
Reviewed by Mario Sanchez Prada.
Source/WebCore:
There were some unsafe pointer accesses in AccessibilityMenuList code that needed to be cleaned up.
Test: accessibility/menu-list-crash2.html
- accessibility/AccessibilityMenuList.cpp:
(WebCore::AccessibilityMenuList::addChildren):
LayoutTests:
- accessibility/menu-list-crash2-expected.txt: Added.
- accessibility/menu-list-crash2.html: Added.
- 2:23 AM Changeset in webkit [184188] by
-
- 3 edits in releases/WebKitGTK/webkit-2.8/Source
Merge r184072 - [GTK] WorkQueue objects are not released
https://bugs.webkit.org/show_bug.cgi?id=144824
Reviewed by Žan Doberšek.
Do not keep a reference of the WorkQueue for the entire life of
its worker thread, since every task scheduled on the WorkQueue
already takes a reference. Instead, take a reference of the main
loop to make sure that when the worker thread starts, the main
loop hasn't been released to avoid runtime warnings (see
webkit.org/b/140998). Also removed the g_main_context_pop_thread_default()
from the thread body, since the thread-specific context queue will
be freed anyway when the thread exits.
If the WorkQueue is released early, before the thread has started,
schedule a main loop quit in the context, to make sure it will
be the first thing run by the main loop and the thread will exit.
- wtf/WorkQueue.h: Remove unused event loop mutex.
- wtf/gtk/WorkQueueGtk.cpp:
(WTF::WorkQueue::platformInitialize):
(WTF::WorkQueue::platformInvalidate):
- 2:17 AM Changeset in webkit [184187] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r184070 - Reduce TransformationMatrix copies in WebKitCSSMatrix operations
https://bugs.webkit.org/show_bug.cgi?id=144795
Reviewed by Darin Adler.
Instead of copying the TransformationMatrix member, performing
the operation on it and then copying it again when creating
the new WebKitCSSMatrix object, copy it just once by first
creating the new WebKitCSSMatrix object and then performing
the operation on that object's TransformationMatrix directly.
- css/WebKitCSSMatrix.cpp:
(WebCore::WebKitCSSMatrix::multiply):
(WebCore::WebKitCSSMatrix::translate):
(WebCore::WebKitCSSMatrix::scale):
(WebCore::WebKitCSSMatrix::rotate):
(WebCore::WebKitCSSMatrix::rotateAxisAngle):
(WebCore::WebKitCSSMatrix::skewX):
(WebCore::WebKitCSSMatrix::skewY):
- 2:17 AM Changeset in webkit [184186] by
-
- 3 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r184005 - Crashes in SocketStreamHandleBase::close
https://bugs.webkit.org/show_bug.cgi?id=144767
rdar://problem/20486538
Reviewed by Brady Eidson.
This is a speculative fix, I could not reproduce the crash.
- Modules/websockets/WebSocketChannel.cpp: (WebCore::WebSocketChannel::processFrame):
Normally, processOutgoingFrameQueue() closes the handle in the end when called in
OutgoingFrameQueueClosing state. But there is no definitive protection against
processing two CLOSE frames, in which case we'd try to close the handle twice.
- platform/network/cf/SocketStreamHandleCFNet.cpp:
(WebCore::SocketStreamHandle::readStreamCallback): Passing empty data to the client
results in the socket being closed, which makes no sense here.
- 2:16 AM Changeset in webkit [184185] by
-
- 6 edits in releases/WebKitGTK/webkit-2.8/Source/bmalloc
Merge r183959 - Release assert in com.apple.WebKit.WebContent under JavaScriptCore: JSC::JSONProtoFuncStringify
https://bugs.webkit.org/show_bug.cgi?id=144758
Reviewed by Andreas Kling.
This was an out-of-memory error when trying to shrink a string builder.
bmalloc was missing the optimization that allowed realloc() to shrink
without copying. So, let's add it.
- bmalloc/Allocator.cpp:
(bmalloc::Allocator::reallocate): Added Large and XLarge cases for
shrinking without copying. This isn't possible for small and medium
objects, and probably not very profitable, either.
- bmalloc/Heap.cpp:
(bmalloc::Heap::findXLarge):
(bmalloc::Heap::deallocateXLarge):
- bmalloc/Heap.h: Refactored this code to return a reference to an
XLarge range. This makes the code reusable, and also makes it easier
for realloc() to update metadata.
- bmalloc/LargeObject.h:
(bmalloc::LargeObject::split): Allow allocated objects to split because
that's what realloc() wants to do, and there's nothing intrinsically
wrong with it.
- 2:02 AM Changeset in webkit [184184] by
-
- 4 edits2 adds in releases/WebKitGTK/webkit-2.8
Merge r183950 - REGRESSION (r164449): Subpixel rendering: http://www.apple.com/iphone-6/ "Faster wireless." image displays vertical black line on 1x displays at specific window width.
https://bugs.webkit.org/show_bug.cgi?id=144723
rdar://problem/18307094
Reviewed by Simon Fraser.
This patch ensures that the backing store's graphics layer is always positioned on a device pixel boundary.
While calculating the backing store's graphics layer location, its ancestor layer's compositing bounds is taken into account.
However the compositing bounds is an unsnapped value, so in order to place the graphics layer properly,
we need to pixel align the ancestor compositing bounds before using it to adjust the child's graphics layer position.
Source/WebCore:
Test: compositing/ancestor-compositing-layer-is-on-subpixel-position.html
- rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::updateGeometry):
LayoutTests:
- compositing/ancestor-compositing-layer-is-on-subpixel-position-expected.html: Added.
- compositing/ancestor-compositing-layer-is-on-subpixel-position.html: Added.
- platform/mac/compositing/layer-creation/overlap-animation-container-expected.txt: progression.
- 2:01 AM Changeset in webkit [184183] by
-
- 3 edits2 adds in releases/WebKitGTK/webkit-2.8
Merge r183948 - Fix sizes crash and add invalid value tests.
https://bugs.webkit.org/show_bug.cgi?id=144739
Reviewed by Darin Adler.
Source/WebCore:
Make sure that only CSS length are allowed when the sizes parser is calling computeLength.
Also make sure that for invalid lengths, the 100vw default is used instead.
Test: fast/dom/HTMLImageElement/sizes/image-sizes-invalids.html
- css/SourceSizeList.cpp:
(WebCore::computeLength):
(WebCore::defaultLength):
(WebCore::parseSizesAttribute):
LayoutTests:
Add tests that make sure that invalid values are properly handled, and a 100vw
source-size length is being used for srcset and for intrinsic dimension calculation.
- fast/dom/HTMLImageElement/sizes/image-sizes-invalids-expected.txt: Added.
- fast/dom/HTMLImageElement/sizes/image-sizes-invalids.html: Added.
- 2:00 AM Changeset in webkit [184182] by
-
- 3 edits in releases/WebKitGTK/webkit-2.8/Source/WebKit2
Merge r183941 - Add PLUGIN_ARCHITECTURE(X11) around m_frameRectInWindowCoordinates in NetscapePlugin.
https://bugs.webkit.org/show_bug.cgi?id=144490
Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-07
Reviewed by Darin Adler.
m_frameRectInWindowCoordinates in NetscapePlugin is currently being used only for
the windowed plugins, and the windowed plugins are only supported on X11. So we can
guard it with PLUGIN_ARCHITECTURE(X11).
No new tests, no behavior change.
- WebProcess/Plugins/Netscape/NetscapePlugin.cpp:
(WebKit::NetscapePlugin::geometryDidChange):
- WebProcess/Plugins/Netscape/NetscapePlugin.h:
- 1:57 AM Changeset in webkit [184181] by
-
- 4 edits1 add in releases/WebKitGTK/webkit-2.8
Merge r183933 - Fix a couple of cases where the backForwardListState's currentIndex is not set correctly in WebBackForwardList::backForwardListState().
https://bugs.webkit.org/show_bug.cgi?id=144666
Reviewed by Darin Adler.
- UIProcess/WebBackForwardList.cpp:
(WebKit::WebBackForwardList::backForwardListState):
If the first item is filtered out and the currentIndex is 0, don't decrement currentIndex.
If all the items are filtered out, set currentIndex to the uninitialized value.
Tools:
Add a test for WKPageCopySessionState() with filtering.
https://bugs.webkit.org/show_bug.cgi?id=144666
Reviewed by Darin Adler.
- TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
- TestWebKitAPI/Tests/WebKit2/WKPageCopySessionStateWithFiltering.cpp: Added.
(TestWebKitAPI::didFinishLoadForFrame):
(TestWebKitAPI::setPageLoaderClient):
(TestWebKitAPI::filterFirstItemCallback):
(TestWebKitAPI::filterAllItemsCallback):
(TestWebKitAPI::createSessionStates):
(TestWebKitAPI::TEST):
- 1:56 AM Changeset in webkit [184180] by
-
- 3 edits in releases/WebKitGTK/webkit-2.8/Source/WTF
Merge r183921 - [GTK] Clean up RunLoop implementation
https://bugs.webkit.org/show_bug.cgi?id=144729
Reviewed by Carlos Garcia Campos.
Clean up the RunLoop implementation for the GTK port,
removing unnecessary methods and using simpler variable names.
Nested GMainLoops in RunLoop::run() are now created for the
RunLoop's GMainContext, and not for the default context (enforced
through the null argument to g_main_loop_new()).
- wtf/RunLoop.h:
- wtf/gtk/RunLoopGtk.cpp:
(WTF::RunLoop::RunLoop):
(WTF::RunLoop::~RunLoop):
(WTF::RunLoop::run):
(WTF::RunLoop::stop):
(WTF::RunLoop::wakeUp):
(WTF::RunLoop::TimerBase::start):
(WTF::RunLoop::innermostLoop): Deleted.
(WTF::RunLoop::pushNestedMainLoop): Deleted.
(WTF::RunLoop::popNestedMainLoop): Deleted.
- 1:53 AM Changeset in webkit [184179] by
-
- 3 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r183891 - Avoid copies in scrollOffsetForFixedPosition() and inline it.
https://bugs.webkit.org/show_bug.cgi?id=144709
Reviewed by Simon Fraser.
- page/FrameView.cpp:
(WebCore::FrameView::frameScaleFactor):
Added so that scrollOffsetForFixedPosition() can be inlined without having to
reference Frame's implementation.
(WebCore::FrameView::scrollOffsetForFixedPosition): Deleted.
Moved this to the header.
- page/FrameView.h:
Inline scrollOffsetForFixedPosition() and get rid of all the copying
it was doing.
- 1:48 AM Changeset in webkit [184178] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r183887 - RenderLayer::currentTransform computes a pixel snapped rect it doesn't use.
https://bugs.webkit.org/show_bug.cgi?id=144708
Reviewed by Simon Fraser.
- rendering/RenderLayer.cpp:
(WebCore::RenderLayer::currentTransform):
Only compute a pixel snapped rect if we actually end up needing it. The common case
is that this rect is not needed, so pushing it inside the two if statements
speeds up the common case.
- 1:47 AM Changeset in webkit [184177] by
-
- 10 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r183885 - Optimize topLeftLocationOffset() addition in updateLayerPosition
https://bugs.webkit.org/show_bug.cgi?id=144704
Reviewed by Dean Jackson.
- page/FrameView.cpp:
(WebCore::FrameView::FrameView):
- page/FrameView.h:
Move the hasFlippedBlocks bit to FrameView instead of RenderView. Works better for inlining
the check in any renderer header, and it also makes more sense conceptually, since the RenderView
itself could be a flipped block.
- rendering/RenderBox.cpp:
(WebCore::RenderBox::layoutOverflowRectForPropagation):
Change over to the FrameView bit.
- rendering/RenderBox.h:
(WebCore::RenderBox::applyTopLeftLocationOffset):
Add a new inlined function that can apply the top left location offset to a point without
multiple LayoutSize creations and copies. It invokes a helper for flipping that is not
inlined only in the case where actual flipped blocks exist in the render tree.
- rendering/RenderBoxModelObject.cpp:
(WebCore::RenderBoxModelObject::updateFromStyle):
Set the bit on the FrameView now instead of the RenderView.
- rendering/RenderLayer.cpp:
(WebCore::RenderLayer::updateLayerPosition):
Call the new applyTopLeftLocationOffset function so that the point can have offsets added
in without any extra copies.
(WebCore::RenderLayer::calculateClipRects):
- rendering/RenderLineBoxList.cpp:
(WebCore::RenderLineBoxList::rangeIntersectsRect):
Switch over to the bit on the FrameView.
- rendering/RenderView.cpp:
(WebCore::RenderView::RenderView):
- rendering/RenderView.h:
Get rid of the bit on the RenderView.
- 1:38 AM Changeset in webkit [184176] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r183879 - Optimize relativePositionOffset() to avoid doing unnecessary work
https://bugs.webkit.org/show_bug.cgi?id=144698
Reviewed by Simon Fraser.
- rendering/RenderBoxModelObject.cpp:
(WebCore::RenderBoxModelObject::relativePositionOffset):
Patch relativePositionOffset to avoid doing unnecessary work in the common case where
all values of top/left/right/bottom are either auto or fixed. We no longer fetch
containingBlock() into a local always, but instead just invoke the function only
when necessary.
Also avoid computing the percentage-relative maximum for the top/right/bottom/left lengths
when they are fixed values, since that maximum won't be examined at all.
- 1:37 AM Changeset in webkit [184175] by
-
- 3 edits in releases/WebKitGTK/webkit-2.8/Source/WebKit2
Merge r183861 - NetworkResourceLoader::cleanup() should clear ResourceHandle client pointer.
https://bugs.webkit.org/show_bug.cgi?id=144641
rdar://problem/20250960
Reviewed by David Kilzer.
- NetworkProcess/NetworkResourceLoader.cpp: (WebKit::NetworkResourceLoader::cleanup):
Clear the client pointer.
- Shared/Authentication/AuthenticationManager.cpp:
(WebKit::AuthenticationManager::useCredentialForChallenge):
(WebKit::AuthenticationManager::continueWithoutCredentialForChallenge):
(WebKit::AuthenticationManager::cancelChallenge):
(WebKit::AuthenticationManager::performDefaultHandling):
(WebKit::AuthenticationManager::rejectProtectionSpaceAndContinue):
Updated comments, which were not accurate, at least on Mac.
- 1:35 AM Changeset in webkit [184174] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r183855 - EventHandler::m_eventHandlerWillResetCapturingMouseEventsElement is incorrectly initialized
https://bugs.webkit.org/show_bug.cgi?id=144583
Reviewed by Daniel Bates.
- page/EventHandler.cpp:
(WebCore::EventHandler::EventHandler): The
m_eventHandlerWillResetCapturingMouseEventsElement is a boolean,
so initialize it to false, instead of to nullptr.
- 1:34 AM Changeset in webkit [184173] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source
Merge r183800 - [GTK] Async operations running in the WorkQueue thread should schedule their sources to the WorkQueue main lopp
https://bugs.webkit.org/show_bug.cgi?id=144541
Reviewed by Žan Doberšek.
Source/WTF:
They are currently sent to the main thread run loop, because we
are not setting the WorkQueue main context as the default one in
the worker thread.
- wtf/gtk/WorkQueueGtk.cpp:
(WTF::WorkQueue::platformInitialize): Call
g_main_context_push_thread_default() to set the WorkQueue main
context as the default of the thread before running the main loop,
and g_main_context_pop_thread_default() when the main loop quits.
- 1:29 AM Changeset in webkit [184172] by
-
- 7 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r183788 - RenderWidget::setWidgetGeometry() can end up destroying *this*.
https://bugs.webkit.org/show_bug.cgi?id=144601
Reviewed by Andreas Kling.
This is a speculative fix to ensure we don't crash on an invalid *this* renderer
while flattening the current iframe.
Calling RenderWidget::setWidgetGeometry() can result in destroying the current renderer.
While it is not a issue in case of normal layout flow as widget positions are updated at post layout,
frame flattening initiates this action in the middle of layout.
This patch re-introduces refcount model for RenderWidgets so that the renderer is protected during layout
when frame flattening is in use.
- rendering/RenderFrameBase.cpp:
(WebCore::RenderFrameBase::layoutWithFlattening): Let's be paranoid about child view.
- rendering/RenderObject.cpp:
(WebCore::RenderObject::destroy):
- rendering/FrameView.cpp:
(WebCore::FrameView::layout):
- rendering/RenderView.h:
- rendering/RenderWidget.cpp:
(WebCore::RenderWidget::~RenderWidget):
- rendering/RenderWidget.h:
(WebCore::RenderWidget::ref):
(WebCore::RenderWidget::deref):
- 1:15 AM Changeset in webkit [184171] by
-
- 6 edits3 adds in releases/WebKitGTK/webkit-2.8
Merge r183781 - Crash at com.apple.WebKit.WebContent at com.apple.WebCore: WebCore::createWindow + 185
https://bugs.webkit.org/show_bug.cgi?id=144597
<rdar://problem/20361579>
Reviewed by Andreas Kling.
Source/WebCore:
Test: fast/dom/Window/window-open-activeWindow-null-frame.html
In our implementation of window.open(), we make sure that the window
which window.open() is called has a frame. However, we did not have the
same check for the activeDOMWindow (i.e. the lexicalGlobalObject) causing
us to crash in WebCore::createWindow() when dereferencing it.
This patch updates WebCore::createWindow() takes a reference to the
openerFrame instead of a pointer to make it clear the implementation
expects it to be non-null. A null check is then added for the frame
at the call site: DOMWindow::createWindow().
- inspector/InspectorFrontendClientLocal.cpp:
(WebCore::InspectorFrontendClientLocal::openInNewTab):
- loader/FrameLoader.cpp:
(WebCore::isDocumentSandboxed):
(WebCore::FrameLoader::submitForm):
(WebCore::createWindow):
Take a reference to openerFrame instead of a pointer as the
implementation expects it to be non-null.
- loader/FrameLoader.h:
- page/DOMWindow.cpp:
(WebCore::DOMWindow::createWindow):
Add null check for activeFrame before passing it to
WebCore::createWindow().
LayoutTests:
Add a layout test to cover the case where window.open() is called on a
window that is different than the activeDOMWindow and where the
activeDOMWindow does not have a frame.
- fast/dom/Window/resources/test-frame.html: Added.
- fast/dom/Window/window-open-activeWindow-null-frame-expected.txt: Added.
- fast/dom/Window/window-open-activeWindow-null-frame.html: Added.
- 1:15 AM Changeset in webkit [184170] by
-
- 2 edits in trunk/Source/JavaScriptCore
Workaround for Cortex-A53 erratum 843419
https://bugs.webkit.org/show_bug.cgi?id=144680
Reviewed by Michael Saboff.
This patch is about to give simple workaround for Cortex-A53 erratum 843419.
It inserts nops after ADRP instruction to avoid wrong address accesses.
- assembler/ARM64Assembler.h:
(JSC::ARM64Assembler::adrp):
(JSC::ARM64Assembler::nopCortexA53Fix843419):
- 1:06 AM Changeset in webkit [184169] by
-
- 6 edits4 adds in releases/WebKitGTK/webkit-2.8
Merge r183765 - REGRESSION (r178156): CSS Parser incorrectly rejects valid calc() in padding-right property
https://bugs.webkit.org/show_bug.cgi?id=144584
<rdar://problem/20796829>
Reviewed by Darin Adler.
Source/WebCore:
The CSS parser was rejecting calculated values at parsing time if it
considered the value was negative and the CSS property did not allow
negative values. However, doing so at this point will not always work
because we don't necessarily know the font-size yet (for e.g. for
calc(0.5em - 2px). Also, rejecting negative calculated values is not
the right behavior as the the specification. The specification says
we should clamp:
http://dev.w3.org/csswg/css-values-3/#calc-range
This patch updates validateCalculationUnit() to stop marking the value
as invalid if it is negative. Instead, let the CSSCalcValue's permitted
range clamp the value as needed.
This bug was causing the bottom graphic on aldentrio.com to not be
rendered properly.
Test: fast/css/negative-calc-values.html
fast/css/padding-calc-value.html
- css/CSSParser.cpp:
(WebCore::CSSParser::validateCalculationUnit):
LayoutTests:
- fast/css/negative-calc-values-expected.txt: Added.
- fast/css/negative-calc-values.html: Added.
Add a layout test that assigns negative calc() values to properties
whose values cannot be negative to verify that values are clamped as
per the specification:
http://dev.w3.org/csswg/css-values-3/#calc-range
- fast/css/padding-calc-value-expected.txt: Added.
- fast/css/padding-calc-value.html: Added.
Add a layout test to test that using calc(.5em - 2px) for padding-right
CSS property works as intended. It used to be resolved as 0px instead
of "2*font-size - 2px".
- fast/css/text-shadow-calc-value-expected.txt:
- fast/css/text-shadow-calc-value.html:
Update test to match what the specification says:
http://dev.w3.org/csswg/css-values-3/#calc-range
"width: calc(5px - 10px);" is equivalent to "width: 0px;" since widths
smaller than 0px are not allowed.
- 1:03 AM Changeset in webkit [184168] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8
Merge r183741 - [cmake] Disable GNU Gold linker on Cortex A53
https://bugs.webkit.org/show_bug.cgi?id=144382
Reviewed by Carlos Garcia Campos.
- Source/cmake/OptionsCommon.cmake:
- 1:00 AM Changeset in webkit [184167] by
-
- 4 edits in releases/WebKitGTK/webkit-2.8/Source
Merge r183740 - [ARM] Don't compare unsigned chars to EOF (-1)
https://bugs.webkit.org/show_bug.cgi?id=144439
Reviewed by Geoffrey Garen.
Source/WebCore:
- platform/linux/MemoryPressureHandlerLinux.cpp:
(WebKit::nextToken): Don't cast return value of fgetc() to char.
Source/WebKit2:
- Shared/linux/WebMemorySamplerLinux.cpp:
(WebKit::nextToken): Don't cast return value of fgetc() to char.
- 12:59 AM Changeset in webkit [184166] by
-
- 3 edits in trunk
Linker fails without -DDEVELOPER_MODE=ON
https://bugs.webkit.org/show_bug.cgi?id=144117
Reviewed by Gyuyoung Kim.
- Source/cmake/OptionsEfl.cmake:
Remove fvisibility flags. linker script will cover the functionality for ewebkit2.so in release mode,
- Source/cmake/eflsymbols.filter: Updated symbol patterns which should be exposed for ewebkit2.
- 12:58 AM Changeset in webkit [184165] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r183729 - [GTK] API tests crashing on debug builds due to extra unref
https://bugs.webkit.org/show_bug.cgi?id=144508
Reviewed by Mario Sanchez Prada.
The problem is that we were assuming that when a new DOMWindow is
created, the DOM object cache was notified about the previous
DOMWindow being destroyed before objects for the new DOMWindow are
added to the cache. However, that's not always the case and we
only create a DOMWindowObserver for the first DOMWindow. We need
to keep a pointer to the DOMWindow being observed to clear() the
cache and create a new DOMWindowObserver when it changes in the
Frame.
Fixes crashes in several unit tests in debug builds.
- bindings/gobject/DOMObjectCache.cpp:
- 12:57 AM Changeset in webkit [184164] by
-
- 3 edits2 adds in releases/WebKitGTK/webkit-2.8
Merge r183706 - Reproducible crash removing name attribute from <img> node
<https://webkit.org/b/144371>
<rdar://problem/17198583>
Reviewed by Darin Adler.
Source/WebCore:
The problem here was with HTMLImageElement::getNameAttribute(), which relies
on Element::hasName() to avoid slow attribute lookups when the attribute
is already known not to be present. Unfortunately hasName() uses an ElementData
flag that wasn't getting updated until after the call to parseAttribute().
This patch fixes the issue by moving the code that updates the hasName() flag
before the parseAttribute() virtual dispatch.
Test: fast/dom/HTMLImageElement/remove-name-id-attribute-from-image.html
- dom/Element.cpp:
(WebCore::Element::attributeChanged):
LayoutTests:
- fast/dom/HTMLImageElement/remove-name-id-attribute-from-image-expected.txt: Added.
- fast/dom/HTMLImageElement/remove-name-id-attribute-from-image.html: Added.
- 12:55 AM Changeset in webkit [184163] by
-
- 12 edits1 add in releases/WebKitGTK/webkit-2.8
Merge r183698 - Back/forward navigation to an error page in Safari breaks the back-forward list
https://bugs.webkit.org/show_bug.cgi?id=144501
Reviewed by Darin Adler.
Test: TestWebKitAPI/Tests/WebKit2Cocoa/LoadAlternateHTMLString.mm
Normally, loading substitute data (such as an error page) creates a new back-forward list
item. FrameLoader has a mechanism that detects when a substitute data load occurs during
handling of a provisional load error and prevents the creation of a new back-forwards list
item in that case if the unreachable URL is the same as the failing provisional URL. This
mechanism was broken in WebKit2, where handling the provisional load error is asynchronous.
The fix is to capture some state (namely, the failing provisional URL) when dispatching the
load error and allow it to be restored when loading the substitute data.
- loader/FrameLoader.cpp:
(WebCore::FrameLoader::FrameLoader): Removed initialization of
m_delegateIsHandlingProvisionalLoadError.
(WebCore::FrameLoader::shouldReloadToHandleUnreachableURL): Instead of checking
m_delegateIsHandlingProvisionalLoadError and if true using the provisional document loader’s
URL, check m_provisionalLoadErrorBeingHandledURL.
(WebCore::FrameLoader::checkLoadCompleteForThisFrame): Instead of checking and setting
m_delegateIsHandlingProvisionalLoadError, use m_provisionalLoadErrorBeingHandledURL.
- loader/FrameLoader.h:
(WebCore::FrameLoader::provisionalLoadErrorBeingHandledURL): Added this getter. The client
can call this from its override of dispatchDidFailProvisionalLoad and store the result.
(WebCore::FrameLoader::setProvisionalLoadErrorBeingHandledURL): Added this setter. The
client can call this prior to loading substitute data if it’s done as part of handling a
previously-dispatched didFailProvisionalLoad.
Source/WebKit2:
WebKit2 part of <rdar://problem/8636045> Back/forward navigation to an error page in Safari breaks the back-forward list
https://bugs.webkit.org/show_bug.cgi?id=144501
Reviewed by Darin Adler.
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::loadAlternateHTMLString): If this is called during
didFailProvisionalLoadForFrame, send back the provisional URL captured at the time of
failure.
(WebKit::WebPageProxy::didFailProvisionalLoadForFrame): Get the provisioinal URL and keep
it in new member variable m_failingProvisionalLoadURL for the duration of the client’s
handling of this message.
- UIProcess/WebPageProxy.h:
- UIProcess/WebPageProxy.messages.in: Added provisionalURL parameter to
DidFailProvisionalLoadForFrame.
- WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::dispatchDidFailProvisionalLoad): Send the URL for this error
to the UI process.
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::loadAlternateHTMLString): Temporarily restore the loader’s state to
reflect the provisional load error being handled.
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/WebPage.messages.in: Added provisionalLoadErrorURL parameter to
LoadAlternateHTMLString.
- 12:40 AM Changeset in webkit [184162] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r183648 - DOM bindings should not be using a reference type to point to a temporary object
https://bugs.webkit.org/show_bug.cgi?id=144474
Reviewed by Beth Dakin.
The DOM bindings will currently try and use a local reference to point
to a temporary object. This currently works as a by product of the compiler's
stack layout. This patch removes the dependency on undefined behaviour
by ensuring that we use a value rather than reference type.
- bindings/scripts/CodeGeneratorJS.pm:
(GenerateParametersCheck):
(GetNativeTypeForCallbacks):
- 12:35 AM Changeset in webkit [184161] by
-
- 5 edits in trunk/LayoutTests
[EFL] Rebaseline some editing tests
Also add some missing baselines.
- platform/efl/editing/execCommand/5142012-1-expected.png:
- platform/efl/editing/execCommand/5142012-1-expected.txt:
- platform/efl/editing/inserting/insert-at-end-02-expected.png:
- platform/efl/editing/inserting/insert-at-end-02-expected.txt:
- 12:16 AM Changeset in webkit [184160] by
-
- 5 edits3 adds in releases/WebKitGTK/webkit-2.8
Merge r183646 - Javascript using WebSQL can create their own WebKit info table.
<rdar://problem/20688792> and https://bugs.webkit.org/show_bug.cgi?id=144466
Reviewed by Alex Christensen.
Source/WebCore:
Test: storage/websql/alter-to-info-table.html
- Modules/webdatabase/DatabaseBackendBase.cpp:
(WebCore::DatabaseBackendBase::databaseInfoTableName): Return the info table name.
(WebCore::fullyQualifiedInfoTableName): Append "main." to the info table name.
(WebCore::DatabaseBackendBase::DatabaseBackendBase): Use the fully qualified name.
(WebCore::DatabaseBackendBase::performOpenAndVerify): Ditto.
(WebCore::DatabaseBackendBase::getVersionFromDatabase): Ditto.
(WebCore::DatabaseBackendBase::setVersionInDatabase): Ditto.
LayoutTests:
- storage/websql/alter-to-info-table-expected.txt: Added.
- storage/websql/alter-to-info-table.html: Added.
- storage/websql/alter-to-info-table.js: Added.
- 12:15 AM Changeset in webkit [184159] by
-
- 8 edits in trunk
[Streams API] ReadableStream reader should not be disposable when having pending promises
https://bugs.webkit.org/show_bug.cgi?id=144869
Reviewed by Darin Adler.
Source/WebCore:
Made error promise callback to take a ref to the reader so that the reader is not disposed as long as the promise callback is not resolved.
Covered by tests added to streams/readable-stream-gc.html.
- Modules/streams/ReadableStreamReader.cpp:
(WebCore::ReadableStreamReader::ReadableStreamReader): Moved initialize() call outside constructor as it can ref/unref.
(WebCore::ReadableStreamReader::releaseStreamAndClean): Added protector.
- Modules/streams/ReadableStreamReader.h:
- bindings/js/JSReadableStreamReaderCustom.cpp:
(WebCore::JSReadableStreamReader::closed): Lambda for error now takes a ref to the reader.
- bindings/js/ReadableJSStream.cpp:
(WebCore::ReadableJSStream::Reader::create): Calling initialize() after adoptRef().
LayoutTests:
- streams/readable-stream-gc.html:
- streams/readable-stream-gc-expected.txt:
- 12:12 AM Changeset in webkit [184158] by
-
- 7 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore
Merge r183636 - Avoid containingBlock() calls when no writing mode flipping is needed.
https://bugs.webkit.org/show_bug.cgi?id=144407
Reviewed by Simon Fraser.
Add a bool to RenderView that indicates whether or not any flipped blocks have been
added to the view. Once tainted, the view just stays dirty forever. If no flipped
blocks are ever seen, we can then optimize away calls to containingBlock().
The motivation for this patch is to improve layer position updating, which makes many
calls to topLeftLocationOffset(), one of the functions that can be optimized by this
change.
- rendering/RenderBox.cpp:
(WebCore::RenderBox::layoutOverflowRectForPropagation):
- rendering/RenderBoxModelObject.cpp:
(WebCore::RenderBoxModelObject::updateFromStyle):
- rendering/RenderLayer.cpp:
(WebCore::RenderLayer::calculateClipRects):
- rendering/RenderLineBoxList.cpp:
(WebCore::RenderLineBoxList::rangeIntersectsRect):
- rendering/RenderView.cpp:
(WebCore::RenderView::RenderView):
- rendering/RenderView.h:
- 12:07 AM Changeset in webkit [184157] by
-
- 2 edits1 add in releases/WebKitGTK/webkit-2.8/Source/JavaScriptCore
Merge r183564 - Safari WebKit crash when loading Google Spreadsheet.
https://bugs.webkit.org/show_bug.cgi?id=144020
Reviewed by Filip Pizlo.
The bug is that the object allocation sinking phase did not account for a case
where a property of a sunken object is only initialized on one path and not
another. As a result, on the path where the property is not initialized, we'll
encounter an Upsilon with a BottomValue (which is not allowed by definition).
The fix is to use a JSConstant(undefined) as the bottom value instead (of
BottomValue). If the property is uninitialized, it should still be accessible
and have the value undefined.
- dfg/DFGObjectAllocationSinkingPhase.cpp:
(JSC::DFG::ObjectAllocationSinkingPhase::promoteSunkenFields):
- tests/stress/object-allocation-sinking-with-uninitialized-property-on-one-path.js: Added.
(foo):
(foo2):
- 12:03 AM Changeset in webkit [184156] by
-
- 3 edits2 adds in releases/WebKitGTK/webkit-2.8
Merge r183538 - Fix crash in WebCore::LogicalSelectionOffsetCaches::ContainingBlockInfo::setBlock().
https://bugs.webkit.org/show_bug.cgi?id=140261
Patch by Hyungwook Lee <hyungwook.lee@navercorp.com> on 2015-04-29
Reviewed by Darin Adler.
Source/WebCore:
We need to check whether RenderObject is valid in RenderView::fooSubtreeSelection functions
because invalid object has caused a crash. This patch adds isValidObjectForNewSelection(), and use it.
- rendering/RenderView.cpp:
(WebCore::isValidObjectForNewSelection):
(WebCore::RenderView::clearSubtreeSelection):
(WebCore::RenderView::applySubtreeSelection):
LayoutTests:
- editing/execCommand/crash-140261-expected.txt: Added.
- editing/execCommand/crash-140261.html: Added.
- 12:01 AM Changeset in webkit [184155] by
-
- 2 edits in releases/WebKitGTK/webkit-2.8/Source/WTF
Merge r183504 - Properly reset deleted count when clearing HashTables.
https://bugs.webkit.org/show_bug.cgi?id=144343
Reviewed by Andreas Kling.
- wtf/HashTable.h:
(WTF::KeyTraits>::clear):
Reset m_deletedCount, which appears to have been forgotten.
