WebKit-https.git
4 years agoProperly reset deleted count when clearing HashTables.
achristensen@apple.com [Tue, 28 Apr 2015 20:23:52 +0000 (20:23 +0000)]
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.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183504 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd support for lookup on iOS.
enrica@apple.com [Tue, 28 Apr 2015 20:18:36 +0000 (20:18 +0000)]
Add support for lookup on iOS.
https://bugs.webkit.org/show_bug.cgi?id=144342
rdar://problem/19994090

Reviewed by Tim Horton.

* UIProcess/WebPageProxy.h:
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _lookup:]):
(-[WKContentView canPerformAction:withSender:]):
* UIProcess/ios/WebPageProxyIOS.mm:
(WebKit::WebPageProxy::getLookupContextAtPoint):
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::getLookupContextAtPoint):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183503 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUpdate the features.json for WeakSet, WeakMap, Template literals, Tagged templates
utatane.tea@gmail.com [Tue, 28 Apr 2015 19:58:42 +0000 (19:58 +0000)]
Update the features.json for WeakSet, WeakMap, Template literals, Tagged templates
https://bugs.webkit.org/show_bug.cgi?id=144328

Reviewed by Andreas Kling.

Update the status of ES6 features.

* features.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183500 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Content Extensions] Use less memory for CombinedURLFilters.
commit-queue@webkit.org [Tue, 28 Apr 2015 19:40:19 +0000 (19:40 +0000)]
[Content Extensions] Use less memory for CombinedURLFilters.
https://bugs.webkit.org/show_bug.cgi?id=144290

Patch by Alex Christensen <achristensen@webkit.org> on 2015-04-28
Reviewed by Andreas Kling.

Source/WebCore:

* contentextensions/CombinedURLFilters.cpp:
(WebCore::ContentExtensions::recursiveMemoryUsed):
(WebCore::ContentExtensions::CombinedURLFilters::addPattern):
(WebCore::ContentExtensions::generateNFAForSubtree):
(WebCore::ContentExtensions::CombinedURLFilters::createNFAs):
* contentextensions/NFA.cpp:
(WebCore::ContentExtensions::NFA::memoryUsed):
(WebCore::ContentExtensions::NFA::setActions):
* contentextensions/NFA.h:
* contentextensions/NFANode.h:
* contentextensions/Term.h:
(WebCore::ContentExtensions::Term::Term::generateGraph):
(WebCore::ContentExtensions::Term::generateSubgraphForAtom):
Use Vectors instead of HashTables in PrefixTreeVertex because the sets stay small and need to be more memory efficient.

Source/WTF:

* wtf/Forward.h:
* wtf/Vector.h:
Added a minCapacity template parameter to allow changing the minimum size of an
allocated buffer. The default minCapacity is kept at 16 unless otherwise specified
to have no change on existing code, but this could be changed later. A smaller
default minCapacity would use less memory with small Vectors but spend more time
copying when expanding to large Vectors.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183499 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoConsolidate most "frame load" arguments into FrameLoadRequest.
beidson@apple.com [Tue, 28 Apr 2015 19:38:13 +0000 (19:38 +0000)]
Consolidate most "frame load" arguments into FrameLoadRequest.
https://bugs.webkit.org/show_bug.cgi?id=144276

Reviewed by Alexey Proskuryakov.

Source/WebCore:

No new tests (No change in behavior).

This patch starts the long overdue process of wrangling the insane load methods in FrameLoader.

Humble beginnings of this long process:
- Put most of the various "frame load" arguments that are passed around within FrameLoader on FrameLoadRequest.
- Get rid of the "easy" constructors of FrameLoadRequest, forcing users to actually think about what they're doing.
- Change a private FrameLoader load method (urlSelected) to take a FrameLoadRequest instead of argument gobbledygook.
- Change a public FrameLoader load methods (changeLocation) to take a FrameLoadRequest instead of gobbledygook.

These cover the straightforward changes where it was easy to figure out what the values of the FrameLoadRequest
arguments were from these various call sites.

Further refactoring can be done in much smaller patches, handling fewer cases at a time. Little by little we will be
able to reduce the number of "load" methods on FrameLoader and make sure that they mainly only take a FrameLoadRequest
as an argument.

* inspector/InspectorFrontendClientLocal.cpp:
(WebCore::InspectorFrontendClientLocal::openInNewTab):

* inspector/InspectorPageAgent.cpp:
(WebCore::InspectorPageAgent::navigate):

* loader/FrameLoadRequest.cpp:
(WebCore::FrameLoadRequest::FrameLoadRequest):
* loader/FrameLoadRequest.h:
(WebCore::FrameLoadRequest::FrameLoadRequest):
(WebCore::FrameLoadRequest::lockHistory):
(WebCore::FrameLoadRequest::lockBackForwardList):
(WebCore::FrameLoadRequest::shouldSendReferrer):
(WebCore::FrameLoadRequest::allowNavigationToInvalidURL):
(WebCore::FrameLoadRequest::newFrameOpenerPolicy):
(WebCore::FrameLoadRequest::shouldReplaceDocumentIfJavaScriptURL):

* loader/FrameLoader.cpp:
(WebCore::FrameLoader::changeLocation):
(WebCore::FrameLoader::urlSelected):
(WebCore::FrameLoader::loadFrameRequest):
* loader/FrameLoader.h:

* loader/NavigationScheduler.cpp:
(WebCore::NavigationScheduler::scheduleLocationChange):

* page/ContextMenuController.cpp:
(WebCore::openNewWindow):
(WebCore::ContextMenuController::contextMenuItemSelected):

* page/DOMWindow.cpp:
(WebCore::DOMWindow::createWindow):

Source/WebKit/ios:

* WebView/WebPDFViewPlaceholder.mm:
(-[WebPDFViewPlaceholder simulateClickOnLinkToURL:]):

Source/WebKit/mac:

* WebView/WebPDFView.mm:
(-[WebPDFView PDFViewWillClickOnLink:withURL:]):

Source/WebKit/win:

* Plugins/PluginView.cpp:
(WebCore::PluginView::start):
(WebCore::PluginView::getURLNotify):
(WebCore::PluginView::getURL):
(WebCore::PluginView::handlePost):

Source/WebKit2:

* WebProcess/Plugins/PluginView.cpp:
(WebKit::PluginView::loadURL):

* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::dispatchCreatePage):

* WebProcess/WebPage/WebInspector.cpp:
(WebKit::WebInspector::openInNewTab):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183498 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDFG should not use or preserve Phantoms during transformations
fpizlo@apple.com [Tue, 28 Apr 2015 19:27:23 +0000 (19:27 +0000)]
DFG should not use or preserve Phantoms during transformations
https://bugs.webkit.org/show_bug.cgi?id=143736

Reviewed by Geoffrey Garen.

Since http://trac.webkit.org/changeset/183207 and http://trac.webkit.org/changeset/183406, it is
no longer necessary to preserve Phantoms during transformations. They are still useful just
before FixupPhase to support backwards propagation analyses. They are still inserted late in the
game in the DFG backend. But transformations don't need to worry about them. Inside a basic
block, we can be sure that so long as the IR pinpoints the place where the value becomes
available in a bytecode register (using MovHint) and so long as there is a SetLocal anytime some
other block would need the value (either for OSR or for DFG execution), then we don't need any
liveness markers.

So, this removes any places where we inserted Phantoms just for liveness during transformation
and it replaces convertToPhantom() with remove(), which just converts the node to a Check. A
Check node only keeps its children so long as those children have checks.

The fact that we no longer convertToPhantom() means that we have to be more careful when
constant-folding GetLocal. Previously we would convertToPhantom() and use the fact that
Phantom(Phi) was a valid construct. It's not valid anymore. So, when constant folding encounters
a GetLocal it needs to insert a PhantomLocal directly. This allows us to simplify
Graph::convertToConstant() a bit. Luckily, none of the other users of this method would see
GetLocals.

The only Phantom-like cruft left over after this patch is:

- Phantoms before FixupPhase. I kind of like these. It means that before FixupPhase, we can do
  backwards analyses and rely on the fact that the users of a node in DFG IR are a superset of
  the users of the original local's live range in bytecode. This is essential for supporting our
  BackwardsPropagationPhase, which is an important optimization for things like asm.js.

- PhantomLocals and GetLocals being NodeMustGenerate. See discussion in
  https://bugs.webkit.org/show_bug.cgi?id=144086. It appears that this is not as evil as the
  alternatives. The best long-term plan is to simply ditch the ThreadedCPS IR entirely and have
  the DFG use SSA. For now, so long as any new DFG optimizations we add are block-local and
  treat GetLocal/SetLocal conservatively, this should all be sound.

This change should be perf-neutral although it does reduce the total work that the compiler
does.

* CMakeLists.txt:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* dfg/DFGAdjacencyList.h:
(JSC::DFG::AdjacencyList::justChecks):
* dfg/DFGArgumentsEliminationPhase.cpp:
* dfg/DFGBasicBlock.cpp:
(JSC::DFG::BasicBlock::replaceTerminal):
* dfg/DFGBasicBlock.h:
(JSC::DFG::BasicBlock::findTerminal):
* dfg/DFGCFGSimplificationPhase.cpp:
(JSC::DFG::CFGSimplificationPhase::keepOperandAlive):
(JSC::DFG::CFGSimplificationPhase::mergeBlocks):
* dfg/DFGCPSRethreadingPhase.cpp:
(JSC::DFG::CPSRethreadingPhase::CPSRethreadingPhase):
(JSC::DFG::CPSRethreadingPhase::clearVariables):
(JSC::DFG::CPSRethreadingPhase::canonicalizeFlushOrPhantomLocalFor):
(JSC::DFG::CPSRethreadingPhase::canonicalizeLocalsInBlock):
* dfg/DFGCSEPhase.cpp:
* dfg/DFGCleanUpPhase.cpp: Copied from Source/JavaScriptCore/dfg/DFGPhantomRemovalPhase.cpp.
(JSC::DFG::CleanUpPhase::CleanUpPhase):
(JSC::DFG::CleanUpPhase::run):
(JSC::DFG::performCleanUp):
(JSC::DFG::PhantomRemovalPhase::PhantomRemovalPhase): Deleted.
(JSC::DFG::PhantomRemovalPhase::run): Deleted.
(JSC::DFG::performPhantomRemoval): Deleted.
* dfg/DFGCleanUpPhase.h: Copied from Source/JavaScriptCore/dfg/DFGPhantomRemovalPhase.h.
* dfg/DFGConstantFoldingPhase.cpp:
(JSC::DFG::ConstantFoldingPhase::foldConstants):
(JSC::DFG::ConstantFoldingPhase::addBaseCheck):
(JSC::DFG::ConstantFoldingPhase::fixUpsilons):
* dfg/DFGDCEPhase.cpp:
(JSC::DFG::DCEPhase::run):
(JSC::DFG::DCEPhase::fixupBlock):
(JSC::DFG::DCEPhase::cleanVariables):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupBlock):
(JSC::DFG::FixupPhase::fixupNode):
(JSC::DFG::FixupPhase::convertStringAddUse):
(JSC::DFG::FixupPhase::attemptToMakeFastStringAdd):
(JSC::DFG::FixupPhase::checkArray):
(JSC::DFG::FixupPhase::fixIntConvertingEdge):
(JSC::DFG::FixupPhase::fixIntOrBooleanEdge):
(JSC::DFG::FixupPhase::fixDoubleOrBooleanEdge):
(JSC::DFG::FixupPhase::injectTypeConversionsInBlock):
(JSC::DFG::FixupPhase::tryToRelaxRepresentation):
(JSC::DFG::FixupPhase::injectTypeConversionsForEdge):
(JSC::DFG::FixupPhase::addRequiredPhantom): Deleted.
(JSC::DFG::FixupPhase::addPhantomsIfNecessary): Deleted.
* dfg/DFGGraph.cpp:
(JSC::DFG::Graph::convertToConstant):
(JSC::DFG::Graph::mergeRelevantToOSR): Deleted.
* dfg/DFGGraph.h:
* dfg/DFGInsertionSet.h:
(JSC::DFG::InsertionSet::insertCheck):
* dfg/DFGIntegerCheckCombiningPhase.cpp:
(JSC::DFG::IntegerCheckCombiningPhase::handleBlock):
* dfg/DFGLICMPhase.cpp:
(JSC::DFG::LICMPhase::attemptHoist):
* dfg/DFGNode.cpp:
(JSC::DFG::Node::remove):
* dfg/DFGNode.h:
(JSC::DFG::Node::replaceWith):
(JSC::DFG::Node::convertToPhantom): Deleted.
(JSC::DFG::Node::convertToCheck): Deleted.
(JSC::DFG::Node::willHaveCodeGenOrOSR): Deleted.
* dfg/DFGNodeFlags.h:
* dfg/DFGNodeType.h:
* dfg/DFGObjectAllocationSinkingPhase.cpp:
(JSC::DFG::ObjectAllocationSinkingPhase::lowerNonReadingOperationsOnPhantomAllocations):
(JSC::DFG::ObjectAllocationSinkingPhase::handleNode):
* dfg/DFGPhantomCanonicalizationPhase.cpp: Removed.
* dfg/DFGPhantomCanonicalizationPhase.h: Removed.
* dfg/DFGPhantomRemovalPhase.cpp: Removed.
* dfg/DFGPhantomRemovalPhase.h: Removed.
* dfg/DFGPlan.cpp:
(JSC::DFG::Plan::compileInThreadImpl):
* dfg/DFGPutStackSinkingPhase.cpp:
* dfg/DFGResurrectionForValidationPhase.cpp: Removed.
* dfg/DFGResurrectionForValidationPhase.h: Removed.
* dfg/DFGSSAConversionPhase.cpp:
(JSC::DFG::SSAConversionPhase::run):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::fillSpeculateDouble):
* dfg/DFGStoreBarrierElisionPhase.cpp:
(JSC::DFG::StoreBarrierElisionPhase::elideBarrier):
* dfg/DFGStrengthReductionPhase.cpp:
(JSC::DFG::StrengthReductionPhase::handleNode):
(JSC::DFG::StrengthReductionPhase::convertToIdentityOverChild):
* dfg/DFGValidate.cpp:
(JSC::DFG::Validate::validate):
(JSC::DFG::Validate::validateCPS):
(JSC::DFG::Validate::validateSSA):
* dfg/DFGVarargsForwardingPhase.cpp:
* ftl/FTLLink.cpp:
(JSC::FTL::link):
* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::compileNode):
(JSC::FTL::LowerDFGToLLVM::compileNoOp):
(JSC::FTL::LowerDFGToLLVM::compilePhantom): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183497 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK][jhbuild] Mesa build failure with llvm 3.6
jdiggs@igalia.com [Tue, 28 Apr 2015 18:56:51 +0000 (18:56 +0000)]
[GTK][jhbuild] Mesa build failure with llvm 3.6
https://bugs.webkit.org/show_bug.cgi?id=144250

Reviewed by Žan Doberšek.

Bump the version of Mesa to 10.5.4

* gtk/jhbuild.modules:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183496 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSource/JavaScriptCore:
akling@apple.com [Tue, 28 Apr 2015 18:54:12 +0000 (18:54 +0000)]
Source/JavaScriptCore:
DFG+FTL should generate efficient code for branching on a string's boolean value.
<https://webkit.org/b/144317>

Reviewed by Geoff Garen & Filip Pizlo

Teach Branch nodes about StringUse and have them generate an efficient zero-length string check
instead of dropping out to C++ whenever we branch on a string.

The FTL JIT already handled Branch nodes with StringUse through its use of boolify(), so only
the DFG JIT gets some new codegen logic in this patch.

Test: js/regress/branch-on-string-as-boolean.js (~4.5x speedup)

* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::emitStringBranch):
* dfg/DFGSpeculativeJIT.h:
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::emitBranch):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::emitBranch):

LayoutTests:
DFG should generate efficient code for branching on a string's boolean value.
<https://webkit.org/b/144317>

Reviewed by Geoff Garen & Filip Pizlo.

* js/regress/branch-on-string-as-boolean-expected.txt: Added.
* js/regress/branch-on-string-as-boolean.html: Added.
* js/regress/script-tests/branch-on-string-as-boolean.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183495 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSource/WebCore:
mmaxfield@apple.com [Tue, 28 Apr 2015 18:36:59 +0000 (18:36 +0000)]
Source/WebCore:
[Mac] [iOS] Implement font-synthesis CSS property
https://bugs.webkit.org/show_bug.cgi?id=144305

Reviewed by Andreas Kling.

This patch is fairly straightforward, though there are some particular pieces to the patch listed below.

The implementation of this font-synthesis CSS property lives inside FontCacheIOS and FontCacheMac. We already
determine if we should synthesize bold and italics, this CSS property simply guards that decision.

Because FontCache interacts with FontDescriptions and not RenderStyles, this patch moves the source of truth
regarding font-synthesis into FontDescription (from RenderStyle).

Test: fast/css3-text/font-synthesis.html

* css/CSSPropertyNames.in: Mark font-synthesis as a FontProperty, which means that the relevant state is kept
inside FontDescription. This also means that the property must be marked as a "high priority" property, which
is enforced by it existing at the top of CSSPropertyNames.in. The move to FontDescription must be done so that
FontCacheIOS and FontCacheMac have access to the saved state.
* css/CSSComputedStyleDeclaration.cpp:
(WebCore::fontSynthesisFromStyle): Update to look for state in FontDescription instead of RenderStyle.
* css/CSSSegmentedFontFace.cpp:
(WebCore::CSSSegmentedFontFace::fontRanges): Guard the synthetic bold and synthic italic decision on the new
FontSynthesis state inside FontDescription.
* css/StyleBuilderConverter.h:
(WebCore::StyleBuilderConverter::convertFontSynthesis): Moved from
StyleBuilderCustom::applyValueFontSynthesis(). This is because of the updated options inside of
CSSPropertyNames.in
* css/StyleBuilderCustom.h:
(WebCore::StyleBuilderCustom::applyValueFontSynthesis): Deleted.
* platform/graphics/FontDescription.h:
(WebCore::FontDescription::FontDescription): Moved from RenderStyle.
(WebCore::FontDescription::fontSynthesis): Ditto.
(WebCore::FontDescription::setFontSynthesis): Ditto.
(WebCore::FontDescription::initialFontSynthesis): Ditto.
(WebCore::FontDescription::operator==): Update to include new state.
* platform/graphics/FontCache.h:
(WebCore::FontDescriptionFontDataCacheKey::makeFlagKey): Make FontDescription hashes sensitive to the new
state inside FontDescription.
* platform/graphics/ios/FontCacheIOS.mm:
(WebCore::FontCache::createFontPlatformData): Guard the synthetic bold and synthetic italic decision on the
new FontSynthesis state inside FontDescription.
* platform/graphics/mac/FontCacheMac.mm:
(WebCore::FontCache::createFontPlatformData): Ditto.
* platform/text/TextFlags.h: Move FontSynthesis type from RenderStyleConstants.
* rendering/style/RenderStyle.h:
(WebCore::RenderStyle::fontSynthesis): Moved to FontDescription.
(WebCore::RenderStyle::setFontSynthesis): Ditto.
(WebCore::RenderStyle::initialFontSynthesis): Ditto.
* rendering/style/RenderStyleConstants.h: Moved FontSynthesis type to TextFlags.
* rendering/style/StyleRareInheritedData.h: Remove state regarding font synthesis (moved to FontDescription)

LayoutTests:
Implement font-synthesis CSS property
https://bugs.webkit.org/show_bug.cgi?id=144305

Reviewed by Andreas Kling.

Test all combinations of font-weight, font-style, and font-synthesis.

* fast/css3-text/font-synthesis-expected.html: Added.
* fast/css3-text/font-synthesis.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183494 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win] Temporarily disable WebGL support
bfulgham@apple.com [Tue, 28 Apr 2015 18:04:11 +0000 (18:04 +0000)]
[Win] Temporarily disable WebGL support
https://bugs.webkit.org/show_bug.cgi?id=144291

* platform/win/TestExpectations:
* platform/win/js/dom/constructor-length-expected.txt: Added.
* platform/win/js/dom/global-constructors-attributes-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183493 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoVarargsForwardingPhase should only consider MovHints that have the candidate as a...
fpizlo@apple.com [Tue, 28 Apr 2015 18:04:10 +0000 (18:04 +0000)]
VarargsForwardingPhase should only consider MovHints that have the candidate as a child
https://bugs.webkit.org/show_bug.cgi?id=144340

Reviewed by Michael Saboff and Mark Lam.

Since we were considering all MovHints, we'd assume that the CreateDirectArguments or
CreateClosedArguments node was live so long as any MovHinted bytecode variable was alive.
Basically, we'd keep it alive until the end of the block. This maximized the chances of
there being an interfering operation, which would prevent elimination.

The fix is to only consider MovHints that have the arguments candidate as a child. We only
care to track the liveness of those bytecode locals that would need an arguments object
recovery on OSR exit.

This is a speed-up on V8Spider/raytrace and Octane/raytrace because it undoes the regression
introduced in http://trac.webkit.org/changeset/183406.

* dfg/DFGVarargsForwardingPhase.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183492 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Make all options actually options
mcatanzaro@igalia.com [Tue, 28 Apr 2015 18:01:57 +0000 (18:01 +0000)]
[GTK] Make all options actually options
https://bugs.webkit.org/show_bug.cgi?id=144106

Reviewed by Martin Robinson.

Use WEBKIT_OPTION_DEFINE to set ENABLE_GTKDOC, ENABLE_INTROSPECTION, ENABLE_X11_TARGET, and
ENABLE_WAYLAND_TARGET.

* Source/cmake/OptionsGTK.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183491 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSharedBuffer::copy is not computing the buffer size correctly when having m_dataArray
youenn.fablet@crf.canon.fr [Tue, 28 Apr 2015 17:45:27 +0000 (17:45 +0000)]
SharedBuffer::copy is not computing the buffer size correctly when having m_dataArray
https://bugs.webkit.org/show_bug.cgi?id=144321

Reviewed by Darin Adler.

Patch correctness covered by existing tests.

* platform/SharedBuffer.cpp:
(WebCore::SharedBuffer::copy): Direct appending to m_dataArray and not using append method as this method updates the SharedBuffer size.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183489 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDisable App Nap in WebKitTestRunner UI process
ap@apple.com [Tue, 28 Apr 2015 17:44:55 +0000 (17:44 +0000)]
Disable App Nap in WebKitTestRunner UI process
https://bugs.webkit.org/show_bug.cgi?id=144308

Reviewed by Daniel Bates.

* WebKitTestRunner/mac/main.mm:
(disableAppNapInUIProcess):
(main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183488 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed gardening; add results for new tests.
jer.noble@apple.com [Tue, 28 Apr 2015 17:38:45 +0000 (17:38 +0000)]
Unreviewed gardening; add results for new tests.

* media/audio-playback-restriction-autoplay-expected.txt: Added.
* media/audio-playback-restriction-play-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183487 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Update JHBuild Mesa download directory
commit-queue@webkit.org [Tue, 28 Apr 2015 17:36:28 +0000 (17:36 +0000)]
[GTK] Update JHBuild Mesa download directory
https://bugs.webkit.org/show_bug.cgi?id=144318

Update the download directory of Mesa module since the previous one
doesn't exist any more.

Patch by Marcos Chavarría Teijeiro <chavarria1991@gmail.com> on 2015-04-28
Reviewed by Žan Doberšek.

* gtk/jhbuild.modules:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183486 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Curl] Memory leak.
peavo@outlook.com [Tue, 28 Apr 2015 17:28:47 +0000 (17:28 +0000)]
[Curl] Memory leak.
https://bugs.webkit.org/show_bug.cgi?id=144327

Reviewed by Brent Fulgham.

When a data url is loaded, the ResourceHandle is not released.

* platform/network/curl/ResourceHandleManager.cpp:
(WebCore::ResourceHandleManager::startJob):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183485 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win] PHP scripts fail if HTTP_ORIGIN is not set.
bfulgham@apple.com [Tue, 28 Apr 2015 17:05:30 +0000 (17:05 +0000)]
[Win] PHP scripts fail if HTTP_ORIGIN is not set.
https://bugs.webkit.org/show_bug.cgi?id=144333

Rubberstamped by Brady Eidson.

* http/tests/security/resources/cors-post-redirect-target.php: Check
that HTTP_ORIGIN key exists before using it to avoid PHP error.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183483 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix the link noted 'contribute to features' in status.html
utatane.tea@gmail.com [Tue, 28 Apr 2015 17:05:00 +0000 (17:05 +0000)]
Fix the link noted 'contribute to features' in status.html
https://bugs.webkit.org/show_bug.cgi?id=144331

Reviewed by Csaba Osztrogonác.

Fix the href of the link to http://www.webkit.org/coding/contributing.html.

* status.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183482 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r183477.
zandobersek@gmail.com [Tue, 28 Apr 2015 16:41:29 +0000 (16:41 +0000)]
Unreviewed, rolling out r183477.
https://bugs.webkit.org/show_bug.cgi?id=144334

Roll back the Mesa dependency bump in fear of uncaught
regressions while the testing bots are no-op. (Requested by
zdobersek on #webkit).

Reverted changeset:

"[GTK][jhbuild] Mesa build failure with llvm 3.6"
https://bugs.webkit.org/show_bug.cgi?id=144250
http://trac.webkit.org/changeset/183477

Patch by Commit Queue <commit-queue@webkit.org> on 2015-04-28

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183481 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2][Mac] Update WebContent process' sandbox profile for AWD
cdumez@apple.com [Tue, 28 Apr 2015 16:39:18 +0000 (16:39 +0000)]
[WK2][Mac] Update WebContent process' sandbox profile for AWD
https://bugs.webkit.org/show_bug.cgi?id=144315
<rdar://problem/20719293>

Reviewed by Alexey Proskuryakov.

Update sandbox profile for AWD similarly to what was done for iOS in
r182278.

* WebProcess/com.apple.WebProcess.sb.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183480 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win] Deactivate WebGL until Windows tests work properly
bfulgham@apple.com [Tue, 28 Apr 2015 15:56:04 +0000 (15:56 +0000)]
[Win] Deactivate WebGL until Windows tests work properly
https://bugs.webkit.org/show_bug.cgi?id=144291
<rdar://problem/20707307>

Reviewed by Zalan Bujtas.

Source/WebKit:

* WebKit.vcxproj/WebKit/WebKitANGLE.props: Added.
* WebKit.vcxproj/WebKit/WebKitCommon.props: Remove ANGLE link
libraries from common property sheet.
* WebKit.vcxproj/WebKit/WebKitDebugWinCairo.props: Reference
the WebKitANGLE property sheet.
* WebKit.vcxproj/WebKit/WebKitReleaseWinCairo.props: Ditto.

Source/WTF:

* wtf/FeatureDefines.h: Force WebGL off for Windows builds.

WebKitLibraries:

* win/tools/vsprops/FeatureDefines.props: Turn off the WEB_GL
feature for the Apple Windows port.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183479 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix the Mavericks Debug build after r183467.
aestes@apple.com [Tue, 28 Apr 2015 15:52:39 +0000 (15:52 +0000)]
Fix the Mavericks Debug build after r183467.

* platform/network/ResourceRequestBase.h:
(WebCore::ResourceRequestBase::requester): Removed WEBCORE_EXPORT.
(WebCore::ResourceRequestBase::setRequester): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183478 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK][jhbuild] Mesa build failure with llvm 3.6
jdiggs@igalia.com [Tue, 28 Apr 2015 15:47:34 +0000 (15:47 +0000)]
[GTK][jhbuild] Mesa build failure with llvm 3.6
https://bugs.webkit.org/show_bug.cgi?id=144250

Reviewed by Žan Doberšek.

Bump the version of Mesa to 10.5.4

* gtk/jhbuild.modules:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183477 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove WinCE cruft from cmake build system
ossy@webkit.org [Tue, 28 Apr 2015 15:19:12 +0000 (15:19 +0000)]
Remove WinCE cruft from cmake build system
https://bugs.webkit.org/show_bug.cgi?id=144325

Reviewed by Gyuyoung Kim.

* CMakeLists.txt:
* create_jit_stubs: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183476 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove make-file-arrays.py
ossy@webkit.org [Tue, 28 Apr 2015 15:11:54 +0000 (15:11 +0000)]
Remove make-file-arrays.py
https://bugs.webkit.org/show_bug.cgi?id=144324

Reviewed by Gyuyoung Kim.

* make-file-arrays.py: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183475 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd expectations for tests added in r183428 and r183450.
aestes@apple.com [Tue, 28 Apr 2015 14:37:10 +0000 (14:37 +0000)]
Add expectations for tests added in r183428 and r183450.

* js/regress/typed-array-get-set-by-val-profiling-expected.txt: Added.
* media/audio-playback-restriction-autoplay-expected.txt: Added.
* media/audio-playback-restriction-play-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183473 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix.
antti@apple.com [Tue, 28 Apr 2015 14:19:54 +0000 (14:19 +0000)]
Build fix.

* platform/network/ios/ResourceRequestIOS.mm:
(WebCore::ResourceRequest::ResourceRequest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183472 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: add a separate overview for the Rendering Frames timeline
commit-queue@webkit.org [Tue, 28 Apr 2015 13:30:22 +0000 (13:30 +0000)]
Web Inspector: add a separate overview for the Rendering Frames timeline
https://bugs.webkit.org/show_bug.cgi?id=144245

Patch by Matt Baker <mattbaker@apple.com> on 2015-04-28
Reviewed by Timothy Hatcher.

The Timelines tab now supports two view modes, labeled "Timelines" and "Rendering Frames", with each mode
having a seperate timeline overview and selected range. The Timelines mode displays the original time-based
graphs (Network, Layout, and Script), and switching to the Rendering Frames mode displays a frames graph
with a timeline ruler which shows frame indices.

If the rendering frames timeline is supported by the backend, a navigation bar is added to the timelines
sidebar with controls for switching between view modes. In addition to user-triggered view mode changes, the
timelines sidebar will automatically change the current view mode in response to content view changes that
are triggered elsewhere in the UI (for example, when a timeline that does not belong to the current view
mode is selected via the navigation path).

* UserInterface/Controllers/TimelineManager.js:
(WebInspector.TimelineManager.prototype.get activeRecording): Deleted.
(WebInspector.TimelineManager.prototype.get recordings): Deleted.
Removed unused code.

* UserInterface/Main.html:
New files.

* UserInterface/Models/RenderingFrameTimelineRecord.js:
(WebInspector.RenderingFrameTimelineRecord):
(WebInspector.RenderingFrameTimelineRecord.resetFrameIndex):
(WebInspector.RenderingFrameTimelineRecord.prototype.get frameIndex):
(WebInspector.RenderingFrameTimelineRecord.prototype.get frameNumber):
* UserInterface/Models/TimelineRecording.js:
(WebInspector.TimelineRecording.prototype.reset):
Frame number is now derived from frame index.

* UserInterface/Views/LinearTimelineOverview.js:
(WebInspector.LinearTimelineOverview):
(WebInspector.LinearTimelineOverview.prototype.canShowTimeline):
New overview class for time-based timeline graphs.

* UserInterface/Views/RenderingFrameDetailsSidebarPanel.js:
(WebInspector.RenderingFrameDetailsSidebarPanel):
(WebInspector.RenderingFrameDetailsSidebarPanel.prototype.updateRangeSelection):
(WebInspector.RenderingFrameDetailsSidebarPanel.prototype._getSelectedRecords):
Changed type of selected range from time to frame indices.

* UserInterface/Views/RenderingFrameTimelineOverview.js:
(WebInspector.RenderingFrameTimelineOverview.this.timelineRuler.formatLabelCallback):
(WebInspector.RenderingFrameTimelineOverview):
(WebInspector.RenderingFrameTimelineOverview.prototype.canShowTimeline):
New overview class for frame-based timeline graphs.

* UserInterface/Views/RenderingFrameTimelineOverviewGraph.css:
(.timeline-overview > .graphs-container > .timeline-overview-graph.rendering-frame):
* UserInterface/Views/RenderingFrameTimelineOverviewGraph.js:
(WebInspector.RenderingFrameTimelineOverviewGraph):
(WebInspector.RenderingFrameTimelineOverviewGraph.prototype._updateDividers.createDividerAtPosition.get if):
Updated to use frame index ruler and removed frame combining logic.

* UserInterface/Views/RenderingFrameTimelineView.js:
(WebInspector.RenderingFrameTimelineView):
Increased column widths.

* UserInterface/Views/TimelineOverview.css:
(.timeline-overview > .graphs-container):
(body.mac-platform.legacy .timeline-overview > .graphs-container):
Updated for new ruler height.

* UserInterface/Views/TimelineOverview.js:
(WebInspector.TimelineOverview):
(WebInspector.TimelineOverview.prototype.get secondsPerPixel):
(WebInspector.TimelineOverview.prototype.set secondsPerPixel):
(WebInspector.TimelineOverview.prototype.get visibleDuration):
(WebInspector.TimelineOverview.prototype.set selectionDuration):
(WebInspector.TimelineOverview.prototype.updateLayout):
(WebInspector.TimelineOverview.prototype.get timelineRuler):
(WebInspector.TimelineOverview.prototype.canShowTimeline):
(WebInspector.TimelineOverview.prototype._handleScrollEvent):
(WebInspector.TimelineOverview.prototype._handleWheelEvent):
(WebInspector.TimelineOverview.prototype._timelineRemoved):
(WebInspector.TimelineOverview.prototype._timeRangeSelectionChanged):
(WebInspector.TimelineOverview.prototype.set currentTime): Deleted.
(WebInspector.TimelineOverview.prototype.updateLayoutIfNeeded): Deleted.
(WebInspector.TimelineOverview.prototype._needsLayout): Deleted.
Converted to base class. Derived classes provide settings and permitted timeline types.

* UserInterface/Views/TimelineRecordFrame.css:
(.timeline-record-frame):
(.timeline-record-frame > .frame):
(.timeline-record-frame > .frame > .duration):
(.timeline-record-frame > .frame > .duration.timeline-record-type-network):
(.timeline-record-frame > .frame > .duration.timeline-record-type-layout):
(.timeline-record-frame > .frame > .duration.timeline-record-type-script):
(.timeline-record-frame > .frame > .duration:last-child): Deleted.
New graph frames styles.

* UserInterface/Views/TimelineRecordFrame.js:
(WebInspector.TimelineRecordFrame):
(WebInspector.TimelineRecordFrame.prototype.get record):
(WebInspector.TimelineRecordFrame.prototype.set record):
(WebInspector.TimelineRecordFrame.prototype._updateChildElements.createDurationElement):
(WebInspector.TimelineRecordFrame.prototype.get element): Deleted.
Updated to use frame index ruler.
Removed dropped frames and frame combining logic.

* UserInterface/Views/TimelineRecordingContentView.js:
(WebInspector.TimelineRecordingContentView):
(WebInspector.TimelineRecordingContentView.prototype.shown):
(WebInspector.TimelineRecordingContentView.prototype.hidden):
(WebInspector.TimelineRecordingContentView.prototype.updateLayout):
(WebInspector.TimelineRecordingContentView.prototype._currentContentViewDidChange):
(WebInspector.TimelineRecordingContentView.prototype._updateTimes):
(WebInspector.TimelineRecordingContentView.prototype._updateTimelineOverviewHeight):
(WebInspector.TimelineRecordingContentView.prototype._getRenderingFrameTimeline):
(WebInspector.TimelineRecordingContentView.prototype._timelineRemoved):
(WebInspector.TimelineRecordingContentView.prototype._timelineCountChanged):
(WebInspector.TimelineRecordingContentView.prototype._recordingReset):
(WebInspector.TimelineRecordingContentView.prototype._timeRangeSelectionChanged):
(WebInspector.TimelineRecordingContentView.prototype.get currentTimelineView): Deleted.
(WebInspector.TimelineRecordingContentView.prototype.goForward): Deleted.
(WebInspector.TimelineRecordingContentView.prototype.matchTreeElementAgainstCustomFilters): Deleted.
(WebInspector.TimelineRecordingContentView.prototype._contentViewSupplementalRepresentedObjectsDidChange): Deleted.
(WebInspector.TimelineRecordingContentView.prototype._update): Deleted.
(WebInspector.TimelineRecordingContentView.prototype._debuggerPaused): Deleted.
(WebInspector.TimelineRecordingContentView.prototype._debuggerResumed): Deleted.
(WebInspector.TimelineRecordingContentView.prototype._recordingTimesUpdated): Deleted.
Maintains two timeline overviews, switching between them based on the current content view.
Updated tree element filtering to support both overview rulers.

* UserInterface/Views/TimelineRuler.css:
(.timeline-ruler > .header):
(body.mac-platform.legacy .timeline-ruler > .header):
(.timeline-ruler > .header > .divider > .label):
(.timeline-ruler > .selection-drag):
(body.mac-platform.legacy .timeline-ruler > .selection-drag):
(.timeline-ruler > .selection-handle):
(body.mac-platform.legacy .timeline-ruler > .selection-handle):
(.timeline-ruler > .header > .divider): Deleted.
(.timeline-ruler > .selection-drag:active): Deleted.
Updated for new ruler height.

* UserInterface/Views/TimelineRuler.js:
(WebInspector.TimelineRuler):
(WebInspector.TimelineRuler.prototype.set formatLabelCallback):
(WebInspector.TimelineRuler.prototype.updateLayout):
(WebInspector.TimelineRuler.prototype._formatDividerLabelText):
(WebInspector.TimelineRuler.prototype.set allowsClippedLabels): Deleted.
(WebInspector.TimelineRuler.prototype._updateSelection): Deleted.
Added ability to customize label formatting.

* UserInterface/Views/TimelineSidebarPanel.css:
(.panel.timeline .navigation-bar.timelines-view):
(.sidebar > .panel.navigation.timeline > .title-bar):
(body.mac-platform.legacy .sidebar > .panel.navigation.timeline > .title-bar):
(.sidebar > .panel.navigation.timeline > .title-bar.timelines):
(.sidebar > .panel.navigation.timeline > .navigation-bar-container):
(.sidebar > .panel.navigation.timeline > .navigation-bar-container > .navigation-bar):
(.sidebar > .panel.navigation.timeline > .title-bar.timeline-events):
(body.mac-platform.legacy .sidebar > .panel.navigation.timeline > .title-bar.timeline-events):
(.sidebar > .panel.navigation.timeline > .timelines-content):
(.sidebar > .panel.navigation.timeline > .empty-content-placeholder):
(body.mac-platform.legacy .sidebar > .panel.navigation.timeline > :matches(.content, .empty-content-placeholder)):
(.sidebar > .panel.navigation.timeline > .timelines-content li.item.selected + li.item): Deleted.
(body.window-inactive .sidebar > .panel.navigation.timeline > .timelines-content li.item.selected + li.item): Deleted.
Updated for new navigation bar and ruler height.

* UserInterface/Views/TimelineSidebarPanel.js:
(WebInspector.TimelineSidebarPanel):
(WebInspector.TimelineSidebarPanel.prototype.get viewMode):
(WebInspector.TimelineSidebarPanel.prototype.canShowTimeline):
(WebInspector.TimelineSidebarPanel.showTimelineViewForTimeline.this._timelineTreeElementMap.get select):
(WebInspector.TimelineSidebarPanel.get this):
(WebInspector.TimelineSidebarPanel._timelineAdded.set this):
Added support for two view modes: Timelines and Rendering Frames.

* WebInspectorUI.vcxproj/WebInspectorUI.vcxproj:
* WebInspectorUI.vcxproj/WebInspectorUI.vcxproj.filters:
New files.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183469 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Add new tab button to make it easy to bring a closed tab back
timothy@apple.com [Tue, 28 Apr 2015 13:04:38 +0000 (13:04 +0000)]
Web Inspector: Add new tab button to make it easy to bring a closed tab back
https://bugs.webkit.org/show_bug.cgi?id=144297

Reviewed by Joseph Pecoraro.

* Localizations/en.lproj/localizedStrings.js:
* UserInterface/Base/Main.js:
(WebInspector.contentLoaded):
(WebInspector.isTabTypeAllowed):
(WebInspector._tabContentViewForType):
(WebInspector._rememberOpenTabs):
(WebInspector._updateNewTabButtonState):
(WebInspector._newTabItemClicked):
(WebInspector.isNewTabWithTypeAllowed):
(WebInspector.createNewTab):
(WebInspector.activateExtraDomains):
(WebInspector._restoreCookieForOpenTabs):
(WebInspector._saveCookieForOpenTabs):
(WebInspector._isTabTypeAllowed): Deleted.
* UserInterface/Images/NewTab.svg: Added.
* UserInterface/Images/NewTabPlus.svg:
* UserInterface/Main.html:
* UserInterface/Views/NewTabContentView.css: Added.
(.new-tab.tab.content-view):
(.new-tab.tab.content-view > .tab-item):
(.new-tab.tab.content-view > .tab-item.disabled):
(.new-tab.tab.content-view > .tab-item > .box):
(.new-tab.tab.content-view > .tab-item:not(.disabled):active > .box):
(.new-tab.tab.content-view > .tab-item > .box > img):
(.new-tab.tab.content-view > .tab-item > label):
* UserInterface/Views/NewTabContentView.js: Added.
(WebInspector.NewTabContentView):
(WebInspector.NewTabContentView.prototype.get type):
(WebInspector.NewTabContentView.prototype._createNewTab):
* UserInterface/Views/TabBar.css:
(.tab-bar > .item:not(.disabled).selected):
(.tab-bar > .item.new-tab-button:not(.disabled):hover):
(.tab-bar:not(.animating) > .item:not(.selected, .disabled):hover + .item):
(.tab-bar > .item.disabled > .icon):
(.tab-bar > .item.new-tab-button:not(.disabled):hover > .icon):
(.tab-bar > .item.selected): Deleted.
(.tab-bar > .item.new-tab-button:hover): Deleted.
(.tab-bar:not(.animating) > .item:not(.selected):hover + .item): Deleted.
(.tab-bar > .item.new-tab-button:hover > .icon): Deleted.
* UserInterface/Views/TabBar.js:
(WebInspector.TabBar.prototype.set selectedTabBarItem):
(WebInspector.TabBar.prototype._handleMouseDown):
* UserInterface/Views/TabBarItem.js:
(WebInspector.TabBarItem.prototype.get disabled):
(WebInspector.TabBarItem.prototype.set disabled):
* UserInterface/Views/TabBrowser.js:
(WebInspector.TabBrowser):
(WebInspector.TabBrowser.prototype.addTabForContentView):
(WebInspector.TabBrowser.prototype.showTabForContentView):
(WebInspector.TabBrowser.prototype.closeTabForContentView):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183468 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoNetwork Cache: Disk cache getting filled by YouTube video data
antti@apple.com [Tue, 28 Apr 2015 12:49:30 +0000 (12:49 +0000)]
Network Cache: Disk cache getting filled by YouTube video data
https://bugs.webkit.org/show_bug.cgi?id=144259

Reviewed by Darin Adler.

Source/WebCore:

MSE media is loaded via XHR and tends to eventually fill the cache.

YouTube serves the media chunks cacheable, however they are rarely (if ever) reused.
We can reduce disk writes and keep more useful resources around by not caching them

Test: http/tests/cache/disk-cache/disk-cache-media.html

* loader/DocumentLoader.cpp:
(WebCore::DocumentLoader::startLoadingMainResource):

    Set the requester.

* loader/cache/CachedRawResource.cpp:
(WebCore::CachedRawResource::CachedRawResource):
* page/DiagnosticLoggingKeys.cpp:
(WebCore::DiagnosticLoggingKeys::streamingMedia):
* page/DiagnosticLoggingKeys.h:
* platform/network/ResourceRequestBase.cpp:
(WebCore::ResourceRequestBase::adopt):
(WebCore::ResourceRequestBase::copyData):
(WebCore::equalIgnoringHeaderFields):
* platform/network/ResourceRequestBase.h:
(WebCore::ResourceRequestBase::requester):
(WebCore::ResourceRequestBase::setRequester):

    Add requester type to the request object. Currently this is main resource, xhr or unspecified.

(WebCore::ResourceRequestBase::encodeWithoutPlatformData):
(WebCore::ResourceRequestBase::decodeWithoutPlatformData):
* platform/network/cf/ResourceRequest.h:
(WebCore::ResourceRequest::deprecatedSetMainResourceRequest): Deleted.
(WebCore::ResourceRequest::deprecatedIsMainResourceRequest): Deleted.

    Replace this iOS only field with shared mechanism.

* platform/network/ios/QuickLook.mm:
(WebCore::QuickLookHandle::create):
* xml/XMLHttpRequest.cpp:
(WebCore::XMLHttpRequest::createRequest):

    Set the requester.

Source/WebKit2:

MSE media is loaded via XHR and tends to eventually fill the cache.

YouTube serves the media chunks cacheable, however they are rarely (if ever) reused.
We can reduce disk writes and keep more useful resources around by not caching them

* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::didReceiveResponseAsync):
(WebKit::NetworkResourceLoader::didRetrieveCacheEntry):
* NetworkProcess/cache/NetworkCache.cpp:
(WebKit::NetworkCache::makeStoreDecision):

    Don't store if the requester is XHR and response has video or audio content type.

* NetworkProcess/cache/NetworkCache.h:
* NetworkProcess/cache/NetworkCacheStatistics.cpp:
(WebKit::NetworkCache::storeDecisionToDiagnosticKey):
* Shared/Network/NetworkResourceLoadParameters.cpp:
(WebKit::NetworkResourceLoadParameters::NetworkResourceLoadParameters):
(WebKit::NetworkResourceLoadParameters::encode):
(WebKit::NetworkResourceLoadParameters::decode):
* Shared/Network/NetworkResourceLoadParameters.h:

    The requester type is now part of the ResourceRequest, no need for separate parameter.

* Shared/mac/WebCoreArgumentCodersMac.mm:
(IPC::ArgumentCoder<ResourceRequest>::encodePlatformData):
(IPC::ArgumentCoder<ResourceRequest>::decodePlatformData):
* WebProcess/Network/WebResourceLoadScheduler.cpp:
(WebKit::WebResourceLoadScheduler::scheduleLoad):

LayoutTests:

* http/tests/cache/disk-cache/disk-cache-media-expected.txt: Added.
* http/tests/cache/disk-cache/disk-cache-media.html: Added.
* http/tests/cache/disk-cache/resources/cache-test.js:
(generateTestURL):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183467 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix windows build error in WebCore related to bulk build.
commit-queue@webkit.org [Tue, 28 Apr 2015 11:36:17 +0000 (11:36 +0000)]
Fix windows build error in WebCore related to bulk build.
https://bugs.webkit.org/show_bug.cgi?id=144313

Patch by Namhoon Kim <nakim@ea.com> on 2015-04-28
Reviewed by Csaba Osztrogonác.

No new tests because there is no behavior change.

* css/CSSAllInOne.cpp:
* dom/DOMAllInOne.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183466 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win] Enable ES6 template literals
peavo@outlook.com [Tue, 28 Apr 2015 10:45:55 +0000 (10:45 +0000)]
[Win] Enable ES6 template literals
https://bugs.webkit.org/show_bug.cgi?id=144316

Reviewed by Csaba Osztrogonác.

* win/tools/vsprops/FeatureDefines.props:
* win/tools/vsprops/FeatureDefinesCairo.props:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183465 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] API::UserContentURLPattern creation functions should return Ref<>
zandobersek@gmail.com [Tue, 28 Apr 2015 08:09:22 +0000 (08:09 +0000)]
[WK2] API::UserContentURLPattern creation functions should return Ref<>
https://bugs.webkit.org/show_bug.cgi?id=144225

Reviewed by Darin Adler.

Have API::UserContentURLPattern creation functions return Ref<>.
Call-sites are also updated, using and operating on the returned Ref<>
object where possible.

* Shared/API/APIUserContentURLPattern.h:
(API::UserContentURLPattern::create):
* Shared/API/c/WKUserContentURLPattern.cpp:
(WKUserContentURLPatternCreate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183463 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] API::URLRequest, API::URLResponse creation functions should return Ref<>
zandobersek@gmail.com [Tue, 28 Apr 2015 08:06:58 +0000 (08:06 +0000)]
[WK2] API::URLRequest, API::URLResponse creation functions should return Ref<>
https://bugs.webkit.org/show_bug.cgi?id=144224

Reviewed by Darin Adler.

Have API::URLRequest and API::URLResponse creation functions return Ref<>.
The call-sites are also updated, using and operating on the returned Ref<> object
where possible.

* Shared/API/APIURLRequest.h:
(API::URLRequest::create):
* Shared/API/APIURLResponse.h:
(API::URLResponse::create):
* Shared/API/c/WKURLRequest.cpp:
(WKURLRequestCreateWithWKURL):
* Shared/API/c/mac/WKURLRequestNS.mm:
(WKURLRequestCreateWithNSURLRequest):
* Shared/API/c/mac/WKURLResponseNS.mm:
(WKURLResponseCreateWithNSURLResponse):
* UIProcess/API/C/WKContext.cpp:
(WKContextSetDownloadClient):
* UIProcess/API/C/WKDownload.cpp:
(WKDownloadCopyRequest):
* UIProcess/API/C/WKNavigationDataRef.cpp:
(WKNavigationDataCopyOriginalRequest):
* UIProcess/API/C/WKPage.cpp:
(WKPageSetPagePolicyClient):
(WKPageSetPageUIClient):
* UIProcess/Cocoa/NavigationState.mm:
(WebKit::NavigationState::NavigationClient::decidePolicyForNavigationAction):
* UIProcess/Network/CustomProtocols/soup/WebSoupCustomProtocolRequestManagerClient.cpp:
(WebKit::WebSoupCustomProtocolRequestManagerClient::startLoading):
* WebProcess/InjectedBundle/API/gtk/WebKitWebPage.cpp:
(willSendRequestForFrame):
* WebProcess/InjectedBundle/InjectedBundlePagePolicyClient.cpp:
(WebKit::InjectedBundlePagePolicyClient::decidePolicyForNavigationAction):
(WebKit::InjectedBundlePagePolicyClient::decidePolicyForNewWindowAction):
(WebKit::InjectedBundlePagePolicyClient::decidePolicyForResponse):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183462 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoEliminate styleDidChange with StyleDifferenceEqual when updates are actually necessary
simon.fraser@apple.com [Tue, 28 Apr 2015 06:52:20 +0000 (06:52 +0000)]
Eliminate styleDidChange with StyleDifferenceEqual when updates are actually necessary
https://bugs.webkit.org/show_bug.cgi?id=144198

Followup: fix assertions seen in fullscreen and pseudo-element tests. These code
paths set the style to the existing pointer, but with a SyntheticStyleChange.
We have to avoid an early return in this case.

* rendering/RenderElement.cpp:
(WebCore::RenderElement::setStyle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183461 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] API::Point, API::Size, API::Rect creation functions should return Ref<>
zandobersek@gmail.com [Tue, 28 Apr 2015 06:35:38 +0000 (06:35 +0000)]
[WK2] API::Point, API::Size, API::Rect creation functions should return Ref<>
https://bugs.webkit.org/show_bug.cgi?id=144223

Reviewed by Darin Adler.

Have API::Point, API::Size and API::Rect creation functions return Ref<>.
Call-sites are also updated, using and operating on the returned Ref<> object
where possible.

* Shared/API/APIGeometry.h:
(API::Size::create):
(API::Point::create):
(API::Rect::create):
* Shared/API/c/WKGeometry.cpp:
(WKPointCreate):
(WKSizeCreate):
(WKRectCreate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183460 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] API::Error creation functions should return Ref<>
zandobersek@gmail.com [Tue, 28 Apr 2015 06:31:29 +0000 (06:31 +0000)]
[WK2] API::Error creation functions should return Ref<>
https://bugs.webkit.org/show_bug.cgi?id=144222

Reviewed by Darin Adler.

Have API::Error creation functions return Ref<>.
The call-sites are also updated, using and operating
on the returned Ref<> object where possible.

* Shared/API/APIError.h:
(API::Error::create):
* Shared/API/c/cf/WKErrorCF.cpp:
(WKErrorCreateWithCFError):
* UIProcess/API/C/WKOriginDataManager.cpp:
(WKOriginDataManagerDeleteEntriesForOrigin):
(WKOriginDataManagerDeleteEntriesModifiedBetweenDates):
(WKOriginDataManagerDeleteAllEntries):
* UIProcess/API/C/WKPage.cpp:
(toGenericCallbackFunction):
(WKPageForceRepaint):
(WKPageValidateCommand):
(WKPageComputePagesForPrinting):
* UIProcess/API/C/WKPluginSiteDataManager.cpp:
(WKPluginSiteDataManagerClearSiteData):
(WKPluginSiteDataManagerClearAllSiteData):
* UIProcess/GenericCallback.h:
(WebKit::toGenericCallbackFunction):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::printFinishedCallback):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183459 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRegExp matches arrays should use contiguous indexing.
akling@apple.com [Tue, 28 Apr 2015 05:54:02 +0000 (05:54 +0000)]
RegExp matches arrays should use contiguous indexing.
<https://webkit.org/b/144286>

Reviewed by Geoffrey Garen.

We had a custom Structure being used for RegExp matches arrays that would
put the arrays into SlowPutArrayStorageShape mode. This was just left
from when matches arrays were custom, lazily initialized objects.

This change removes that Structure and switches the matches arrays to
using the default ContiguousShape Structure. This allows the FTL JIT
to compile the inner loop of the Octane/regexp benchmark.

Also made a version of initializeIndex() [inline] that takes the indexing
type in an argument, allowing createRegExpMatchesArray() to initialize
the entire array without branching on the indexing type for each entry.

~3% progression on Octane/regexp.

* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
(JSC::JSGlobalObject::visitChildren):
* runtime/JSGlobalObject.h:
(JSC::JSGlobalObject::mapStructure):
(JSC::JSGlobalObject::regExpMatchesArrayStructure): Deleted.
* runtime/JSObject.h:
(JSC::JSObject::initializeIndex):
* runtime/RegExpMatchesArray.cpp:
(JSC::createRegExpMatchesArray):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183458 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Remove unnecessary semicolon from WebKitNotificationProvider.h.
commit-queue@webkit.org [Tue, 28 Apr 2015 05:33:12 +0000 (05:33 +0000)]
[GTK] Remove unnecessary semicolon from WebKitNotificationProvider.h.
https://bugs.webkit.org/show_bug.cgi?id=144312

Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-04-27
Reviewed by Martin Robinson.

No new tests, no behavior change.

* UIProcess/API/gtk/WebKitNotificationProvider.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183457 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, fix GTK build after r183452
mcatanzaro@igalia.com [Tue, 28 Apr 2015 05:24:10 +0000 (05:24 +0000)]
Unreviewed, fix GTK build after r183452

* Source/cmake/OptionsGTK.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183456 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Cannot scroll Timeline Overview horizontally, scrollbar cannot be...
commit-queue@webkit.org [Tue, 28 Apr 2015 05:13:45 +0000 (05:13 +0000)]
Web Inspector: Cannot scroll Timeline Overview horizontally, scrollbar cannot be interacted with
https://bugs.webkit.org/show_bug.cgi?id=144302

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-04-27
Reviewed by Timothy Hatcher.

* UserInterface/Views/TimelineOverview.css:
(.timeline-overview > .scroll-container):
Increase the z-index of the scroll container over anything in the
overview so that it can be interacted with.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183455 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoEliminate styleDidChange with StyleDifferenceEqual when updates are actually necessary
simon.fraser@apple.com [Tue, 28 Apr 2015 04:59:48 +0000 (04:59 +0000)]
Eliminate styleDidChange with StyleDifferenceEqual when updates are actually necessary
https://bugs.webkit.org/show_bug.cgi?id=144198

Reviewed by Darin Adler, Antti Koivisto.

Source/WebCore:

SyntheticStyleChange style recalcs are triggered for cases where behavior depends
on state which is outside of RenderStyle; this includes triggering compositing for
animations, for video and canvas, and for iframes with composited content.

In these cases, we'd run through RenderElement::setStyle() and its fan-out, but
with diff == StyleDifferenceEqual, and so be unable to determine if there
is actual work to be done.

This patch enforces the contract that the diff is never StyleDifferenceEqual if
compositing or other work has to happen from setStyle(). This is achieved by
passing in a 'hasSideEffects' flag, which causes the diff to become at least
StyleDifferenceRecompositeLayer.

RenderLayerCompositor::layerStyleChanged() can now safely early return
if the diff is equal. Future patches will reduce redundant work even more.

Test: compositing/animation/no-style-recalc-during-accelerated-animation.html

* page/animation/AnimationBase.h:
(WebCore::AnimationBase::animate): Returns a bool now if the state changed.
(WebCore::AnimationBase::state):
* page/animation/AnimationController.cpp:
(WebCore::AnimationController::updateAnimations): bool out param which indicates
whether any animations changed state.
* page/animation/AnimationController.h:
* page/animation/CompositeAnimation.cpp:
(WebCore::CompositeAnimation::animate): If any transitions or animations changed
state, set the animationStateChanged out param to true.
* page/animation/CompositeAnimation.h:
* page/animation/ImplicitAnimation.cpp:
(WebCore::ImplicitAnimation::animate): Return true if the state changed.
* page/animation/ImplicitAnimation.h:
* page/animation/KeyframeAnimation.cpp:
(WebCore::KeyframeAnimation::animate): Return true if the state changed.
* page/animation/KeyframeAnimation.h:
* rendering/RenderElement.cpp:
(WebCore::RenderElement::adjustStyleDifference): We may enter here now with diff
!= StyleDifferenceEqual, but still need to do the check to see if layers changed.
(WebCore::RenderElement::initializeStyle): When setting style for the first time,
don't use StyleDifferenceEqual.
(WebCore::RenderElement::setStyle): Additional flag to indicate whether this style
change involves side effects. If the diff is equal but the flag is set, change
the diff to StyleDifferenceRecompositeLayer (the "lowest" non-zero diff).
* rendering/RenderElement.h:
(WebCore::RenderElement::setAnimatableStyle): Pass true to setStyle() if hasSideEffects
is true, or if animation state changed.
* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::styleChanged): Pass the diff down.
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::layerStyleChanged): Return if the diff is equal.
* rendering/RenderLayerCompositor.h:
* rendering/style/RenderStyleConstants.h: StyleDifferenceNewStyle is used when
setting style for the first time.
* style/StyleResolveTree.cpp:
(WebCore::Style::createRendererIfNeeded): Provide animationsChanged bool (which is unused).
(WebCore::Style::resolveLocal): If the style change is synthetic, set the flag that
says there are side-effects.

LayoutTests:

New test that detects whether a "hardware" animation is firing the style recalc
timer on every frame, which happened during development of this patch.

* compositing/animation/no-style-recalc-during-accelerated-animation-expected.txt: Added.
* compositing/animation/no-style-recalc-during-accelerated-animation.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183454 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFTL failed to initialize arguments.callee on the slow path as well as the fast path
fpizlo@apple.com [Tue, 28 Apr 2015 04:44:39 +0000 (04:44 +0000)]
FTL failed to initialize arguments.callee on the slow path as well as the fast path
https://bugs.webkit.org/show_bug.cgi?id=144293

Reviewed by Mark Lam.

The slow path doesn't fully initialize DirectArguments - it leaves callee blank. So, we need
to initialize the callee on the common path after the fast and slow path.

* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::compileCreateDirectArguments):
* tests/stress/arguments-callee-uninitialized.js: Added.
(foo):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183453 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] Add one single option to control all OpenGL-related options
mcatanzaro@igalia.com [Tue, 28 Apr 2015 04:35:58 +0000 (04:35 +0000)]
[GTK] Add one single option to control all OpenGL-related options
https://bugs.webkit.org/show_bug.cgi?id=144105

Reviewed by Martin Robinson.

.:

Add public ENABLE_OPENGL option, which is mandatory for ENABLE_3D_TRANSFORMS,
ENABLE_ACCELERATED_2D_CANVAS, ENABLE_GLES2, ENABLE_THREADED_COMPOSITOR, ENABLE_WEBGL,
USE_REDIRECTED_XCOMPOSITE_WINDOW, and USE_GSTREAMER_GL. Make ENABLE_GLES2 a proper WebKit
option and move code around accordingly. Use WEBKIT_OPTION_DEPEND and rely on it. Add a
large comment to explain why default value of ENABLE_ACCELERATED_2D_CANVAS is based on the
presence of CairoGL.

* Source/cmake/OptionsGTK.cmake:

Source/WebCore:

Test WTF_USE_TEXTURE_MAPPER since USE_TEXTURE_MAPPER has been removed.

* PlatformGTK.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183452 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] ENABLE_SMOOTH_SCROLLING should be private
mcatanzaro@igalia.com [Tue, 28 Apr 2015 04:28:54 +0000 (04:28 +0000)]
[GTK] ENABLE_SMOOTH_SCROLLING should be private
https://bugs.webkit.org/show_bug.cgi?id=144306

Reviewed by Martin Robinson.

Make ENABLE_SMOOTH_SCROLLING private instead of public.

* Source/cmake/OptionsGTK.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183451 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[JSC] Add support for typed arrays to the Array profiling
benjamin@webkit.org [Tue, 28 Apr 2015 04:16:21 +0000 (04:16 +0000)]
[JSC] Add support for typed arrays to the Array profiling
https://bugs.webkit.org/show_bug.cgi?id=143913

Patch by Benjamin Poulain <bpoulain@apple.com> on 2015-04-27
Reviewed by Filip Pizlo.

Source/JavaScriptCore:

This patch adds ArrayModes for every typed arrays. Having that information
let us generate better GetByVal and PutByVal when the type speculation
are not good enough.

A typical case where this is useful is any basic block for which the type
of the object is always more restrictive than the speculation (for example,
a basic block gated by a branch only taken for on type).

* bytecode/ArrayProfile.cpp:
(JSC::dumpArrayModes):
* bytecode/ArrayProfile.h:
(JSC::arrayModeFromStructure):
* dfg/DFGArrayMode.cpp:
(JSC::DFG::ArrayMode::fromObserved):
(JSC::DFG::ArrayMode::refine):
Maintain the refine() semantic. We do not support OutOfBounds access
for GetByVal on typed array.

* runtime/IndexingType.h:
* tests/stress/typed-array-get-by-val-profiling.js: Added.
(testArray.testCode):
(testArray):
* tests/stress/typed-array-put-by-val-profiling.js: Added.
(testArray.testCode):
(testArray):

LayoutTests:

* js/regress/script-tests/typed-array-get-set-by-val-profiling.js: Added.
* js/regress/typed-array-get-set-by-val-profiling.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183450 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago<rdar://problem/20721931> REGRESSION (r178753): All legacy framework headers are...
mitz@apple.com [Tue, 28 Apr 2015 03:43:37 +0000 (03:43 +0000)]
<rdar://problem/20721931> REGRESSION (r178753): All legacy framework headers are missing availability annotations
https://bugs.webkit.org/show_bug.cgi?id=144311

Reviewed by Adele Peterson.

* postprocess-headers.sh: Changed “macsox” to “macosx”.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183449 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBreak up select text into multiple tests
commit-queue@webkit.org [Tue, 28 Apr 2015 03:05:26 +0000 (03:05 +0000)]
Break up select text into multiple tests
https://bugs.webkit.org/show_bug.cgi?id=144301

Patch by Doug Russell <d_russell@apple.com> on 2015-04-27
Reviewed by Alexey Proskuryakov.

* platform/mac/TestExpectations:
* platform/mac/accessibility/select-text-expected.txt: Removed.
* platform/mac/accessibility/select-text.html: Removed.
* platform/mac/accessibility/select-text/select-text-1-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-1.html: Added.
* platform/mac/accessibility/select-text/select-text-135546-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-135546.html: Added.
* platform/mac/accessibility/select-text/select-text-135575-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-135575.html: Added.
* platform/mac/accessibility/select-text/select-text-2-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-2.html: Added.
* platform/mac/accessibility/select-text/select-text-3-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-3.html: Added.
* platform/mac/accessibility/select-text/select-text-4-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-4.html: Added.
* platform/mac/accessibility/select-text/select-text-5-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-5.html: Added.
* platform/mac/accessibility/select-text/select-text-6-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-6.html: Added.
* platform/mac/accessibility/select-text/select-text-7-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-7.html: Added.
* platform/mac/accessibility/select-text/select-text-8-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-8.html: Added.
* platform/mac/accessibility/select-text/select-text-9-expected.txt: Added.
* platform/mac/accessibility/select-text/select-text-9.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183448 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoPDF action menu fixes
bfulgham@apple.com [Tue, 28 Apr 2015 02:43:40 +0000 (02:43 +0000)]
PDF action menu fixes
https://bugs.webkit.org/show_bug.cgi?id=144299
<rdar://problem/20702215>

Reviewed by Tim Horton.

Make two corrections to how PDFs are handled:
1. When calculating the view rect for the user's selection, make sure
that we get coordinates for the correct PDF page. The existing code assumed
that the current PDFLayerControler's current page was correct, but this will
not be true if you zoom the PDF out so that several pages are displayed at
once. Each selection keeps track of the page it is referenced against.

2. Revise the offsets calculated for the TextIndicator to take into account
the font descender (as well as the ascender), and to adjust by the scaled
amount of margin around the selected text.

* WebProcess/Plugins/PDF/PDFPlugin.mm:
(WebKit::PDFPlugin::viewRectForSelection): Use correct page for calculating
the coordinates.
* WebProcess/WebPage/mac/WebPageMac.mm:
(WebKit::WebPage::dictionaryPopupInfoForSelectionInPDFPlugin): Include font 'descendant'
and (scaled) margin when adjusting the hit target for the TextIndicator to draw.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183447 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, roll out r183438 "RegExp matches arrays should use contiguous indexing...
fpizlo@apple.com [Tue, 28 Apr 2015 02:14:26 +0000 (02:14 +0000)]
Unreviewed, roll out r183438 "RegExp matches arrays should use contiguous indexing". It
causes many debug test failures.

* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
(JSC::JSGlobalObject::visitChildren):
* runtime/JSGlobalObject.h:
(JSC::JSGlobalObject::regExpMatchesArrayStructure):
* runtime/JSObject.h:
(JSC::JSObject::initializeIndex):
* runtime/RegExpMatchesArray.cpp:
(JSC::createRegExpMatchesArray):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183446 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd '__init__.py' and fix known bug in benchmark runner.
commit-queue@webkit.org [Tue, 28 Apr 2015 01:06:14 +0000 (01:06 +0000)]
Add '__init__.py' and fix known bug in benchmark runner.
https://bugs.webkit.org/show_bug.cgi?id=144296

Patch by Dewei Zhu <dewei_zhu@apple.com> on 2015-04-27
Reviewed by Ryosuke Niwa

* Scripts/webkitpy/benchmark_runner/__init__.py: Make python to recognize this module.
* Scripts/webkitpy/benchmark_runner/benchmark_builder/__init__.py: Make python to recognize this module.
* Scripts/webkitpy/benchmark_runner/benchmark_runner.py:
(BenchmarkRunner.__init__):
(BenchmarkRunner.execute): Fix logic bug that '--output-file' option doesn't work, and empty output file bug.
* Scripts/webkitpy/benchmark_runner/browser_driver/__init__.py: Make python to recognize this module.
* Scripts/webkitpy/benchmark_runner/http_server_driver/__init__.py: Make python to recognize this module.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183443 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoConsolidate one-line flag-related header files into TextFlags.h
mmaxfield@apple.com [Tue, 28 Apr 2015 00:59:13 +0000 (00:59 +0000)]
Consolidate one-line flag-related header files into TextFlags.h
https://bugs.webkit.org/show_bug.cgi?id=144295

Reviewed by Tim Horton.

There were a collection of single-line header files throughout platform/ which contain
single-line type declaractions of flags related to text layout & rendering. This patch
consolidates all these single-line headers into TextFlags.h

No new tests because there is no behavior change.

* WebCore.vcxproj/WebCore.vcxproj:
* WebCore.vcxproj/WebCore.vcxproj.filters:
* WebCore.xcodeproj/project.pbxproj:
* css/CSSFontFace.h:
* css/CSSPrimitiveValueMappings.h:
* loader/cache/CachedFont.h:
* loader/cache/CachedTextTrack.h:
* page/Settings.h:
* platform/DragImage.h:
* platform/graphics/FontDescription.h:
* platform/graphics/FontOrientation.h: Removed.
* platform/graphics/FontPlatformData.h:
* platform/graphics/FontRenderingMode.h: Removed.
* platform/graphics/FontSmoothingMode.h: Removed.
* platform/graphics/FontTraitsMask.h: Removed.
* platform/graphics/FontWidthVariant.h: Removed.
* platform/graphics/TextRenderingMode.h: Removed.
* platform/graphics/cairo/FontCustomPlatformData.h:
* platform/graphics/freetype/FontPlatformData.h:
* platform/graphics/mac/FontCustomPlatformData.h:
* platform/graphics/win/FontCustomPlatformData.h:
* platform/text/NonCJKGlyphOrientation.h: Removed.
* platform/text/TextFlags.h:
* rendering/TextPainter.h:
* style/StyleFontSizeFunctions.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183442 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd naver to team.html
ryuan.choi@navercorp.com [Tue, 28 Apr 2015 00:46:21 +0000 (00:46 +0000)]
Add naver to team.html
https://bugs.webkit.org/show_bug.cgi?id=144298

Reviewed by Gyuyoung Kim.

* team.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183441 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoStatus page: fix the email links
benjamin@webkit.org [Tue, 28 Apr 2015 00:31:05 +0000 (00:31 +0000)]
Status page: fix the email links

Patch by Benjamin Poulain <bpoulain@apple.com> on 2015-04-27

* status.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183440 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMore build fix.
ap@apple.com [Tue, 28 Apr 2015 00:22:26 +0000 (00:22 +0000)]
More build fix.

* WebKitTestRunner/InjectedBundle/mac/AccessibilityNotificationHandler.mm:
(-[AccessibilityNotificationHandler _notificationReceived:]):
Use -objectForKey:, not array subscript.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183439 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRegExp matches arrays should use contiguous indexing.
akling@apple.com [Tue, 28 Apr 2015 00:14:47 +0000 (00:14 +0000)]
RegExp matches arrays should use contiguous indexing.
<https://webkit.org/b/144286>

Reviewed by Geoffrey Garen.

We had a custom Structure being used for RegExp matches arrays that would
put the arrays into SlowPutArrayStorageShape mode. This was just left
from when matches arrays were custom, lazily initialized objects.

This change removes that Structure and switches the matches arrays to
using the default ContiguousShape Structure. This allows the FTL JIT
to compile the inner loop of the Octane/regexp benchmark.

Also made a version of initializeIndex() [inline] that takes the indexing
type in an argument, allowing createRegExpMatchesArray() to initialize
the entire array without branching on the indexing type for each entry.

~3% progression on Octane/regexp.

* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
(JSC::JSGlobalObject::visitChildren):
* runtime/JSGlobalObject.h:
(JSC::JSGlobalObject::mapStructure):
(JSC::JSGlobalObject::regExpMatchesArrayStructure): Deleted.
* runtime/JSObject.h:
(JSC::JSObject::initializeIndex):
* runtime/RegExpMatchesArray.cpp:
(JSC::createRegExpMatchesArray):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183438 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoForm control may be associated with the wrong HTML Form element after form id change
dbates@webkit.org [Tue, 28 Apr 2015 00:03:55 +0000 (00:03 +0000)]
Form control may be associated with the wrong HTML Form element after form id change
https://bugs.webkit.org/show_bug.cgi?id=133456
<rdar://problem/17095055>

Reviewed by Andy Estes.

Source/WebCore:

Fixes an issue where a form control may be associated with the wrong HTML Form element
after the id of the HTML Form element associated with the form control is changed when
there is more than one HTML Form element with the same id in the document. Specifically,
a form control that has an HTML form attribute value X will always be associated with
some HTML Form element f where f.id = X regardless of whether f.id is subsequently
changed.

Tests: fast/forms/change-form-id-to-be-unique-then-submit-form.html
       fast/forms/change-form-id-to-be-unique.html

* dom/Element.cpp:
(WebCore::Element::attributeChanged): Notify observers when the id of an element changed.
(WebCore::Element::updateId): Added parameter NotifyObservers (defaults to NotifyObservers::Yes),
as to whether we should notify observers of the id change.
(WebCore::Element::updateIdForTreeScope): Ditto.
(WebCore::Element::willModifyAttribute): Do not notify observers of the id change immediately. As
indicated by the name of this method, we plan to modify the DOM attribute id of the element, but
we have not actually modified it when this method is called. Instead we will notify observers
in Element::attributeChanged(), which is called after the DOM attribute id is modified.
(WebCore::Element::cloneAttributesFromElement): Ditto.
* dom/Element.h: Defined enum class NotifyObservers.
* dom/TreeScope.cpp:
(WebCore::TreeScope::addElementById): Added boolean parameter notifyObservers (defaults to true)
as to whether we should dispatch a notification to all observers.
(WebCore::TreeScope::removeElementById): Ditto.
* dom/TreeScope.h:

LayoutTests:

Add tests to ensure that we associate the correct HTML Form element with a
<select> after changing the id of its associated HTML form element.

* fast/forms/change-form-id-to-be-unique-expected.txt: Added.
* fast/forms/change-form-id-to-be-unique-then-submit-form-expected.txt: Added.
* fast/forms/change-form-id-to-be-unique-then-submit-form.html: Added.
* fast/forms/change-form-id-to-be-unique.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183436 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoCatch unexpected exits in DumpRenderTree
ap@apple.com [Mon, 27 Apr 2015 23:34:06 +0000 (23:34 +0000)]
Catch unexpected exits in DumpRenderTree
https://bugs.webkit.org/show_bug.cgi?id=144288

Reviewed by Tim Horton.

* DumpRenderTree/mac/DumpRenderTree.mm:
(atexitFunction):
(DumpRenderTreeMain):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183435 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix
ap@apple.com [Mon, 27 Apr 2015 23:18:33 +0000 (23:18 +0000)]
Build fix

* DumpRenderTree/mac/AccessibilityNotificationHandler.mm:
(-[AccessibilityNotificationHandler _notificationReceived:]):
Use -objectForKey:, not array subscript.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183434 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoReduce allocations and memory usage when compiling content extensions.
achristensen@apple.com [Mon, 27 Apr 2015 22:34:47 +0000 (22:34 +0000)]
Reduce allocations and memory usage when compiling content extensions.
https://bugs.webkit.org/show_bug.cgi?id=144277

Reviewed by Benjamin Poulain.

Covered by existing tests.

* contentextensions/Term.h:
(WebCore::ContentExtensions::Term::CharacterSet::set):
(WebCore::ContentExtensions::Term::CharacterSet::get):
(WebCore::ContentExtensions::Term::CharacterSet::invert):
(WebCore::ContentExtensions::Term::CharacterSet::inverted):
(WebCore::ContentExtensions::Term::CharacterSet::bitCount):
(WebCore::ContentExtensions::Term::CharacterSet::operator==):
(WebCore::ContentExtensions::Term::CharacterSet::hash):
(WebCore::ContentExtensions::Term::Term):
(WebCore::ContentExtensions::Term::addCharacter):
(WebCore::ContentExtensions::Term::isEndOfLineAssertion):
(WebCore::ContentExtensions::Term::isUniversalTransition):
(WebCore::ContentExtensions::Term::generateSubgraphForAtom):
Use two uint64_t's instead of a BitVector with a capacity of 128 bits.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183433 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRename WTF_USE_3D_GRAPHICS to ENABLE_GRAPHICS_CONTEXT_3D
mcatanzaro@igalia.com [Mon, 27 Apr 2015 22:30:47 +0000 (22:30 +0000)]
Rename WTF_USE_3D_GRAPHICS to ENABLE_GRAPHICS_CONTEXT_3D
https://bugs.webkit.org/show_bug.cgi?id=144193

Reviewed by Darin Adler.

.:

* Source/cmake/OptionsEfl.cmake:
* Source/cmake/OptionsGTK.cmake:
* Source/cmake/OptionsWinCairo.cmake:
* Source/cmake/WebKitFeatures.cmake:

Source/WebCore:

* CMakeLists.txt:
* platform/graphics/ANGLEWebKitBridge.cpp:
* platform/graphics/FormatConverter.cpp:
* platform/graphics/FormatConverter.h:
* platform/graphics/GLContext.h:
* platform/graphics/GraphicsContext3D.cpp:
* platform/graphics/GraphicsContext3DPrivate.cpp:
* platform/graphics/OpenGLShims.cpp:
* platform/graphics/cairo/GraphicsContext3DCairo.cpp:
* platform/graphics/cg/GraphicsContext3DCG.cpp:
* platform/graphics/egl/GLContextEGL.cpp:
* platform/graphics/egl/GLContextEGL.h:
* platform/graphics/glx/GLContextGLX.cpp:
* platform/graphics/glx/GLContextGLX.h:
* platform/graphics/mac/GraphicsContext3DMac.mm:
* platform/graphics/opengl/Extensions3DOpenGL.cpp:
* platform/graphics/opengl/Extensions3DOpenGLCommon.cpp:
* platform/graphics/opengl/Extensions3DOpenGLES.cpp:
* platform/graphics/opengl/GLPlatformContext.cpp:
* platform/graphics/opengl/GLPlatformSurface.cpp:
* platform/graphics/opengl/GraphicsContext3DOpenGL.cpp:
* platform/graphics/opengl/GraphicsContext3DOpenGLCommon.cpp:
* platform/graphics/opengl/GraphicsContext3DOpenGLES.cpp:
* platform/graphics/opengl/TemporaryOpenGLSetting.cpp:
* platform/graphics/win/GraphicsContext3DWin.cpp:

Source/WebKit2:

* CMakeLists.txt:

Source/WTF:

* wtf/FeatureDefines.h:
* wtf/Platform.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183432 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] Video not centered in element on retina devices
jer.noble@apple.com [Mon, 27 Apr 2015 22:24:07 +0000 (22:24 +0000)]
[iOS] Video not centered in element on retina devices
https://bugs.webkit.org/show_bug.cgi?id=144274

Reviewed by Simon Fraser.

In r173702, a transform was added to the video layer (and a matching, inverse transform
in the UIProcess), but this transform affects the position property of the video layer
used to position the video content within the element bounds when their aspect ratios
do not match.

To work around this problem, pre-apply the transform during -setPosition:.

* platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(-[WebVideoContainerLayer setPosition:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183431 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd a setting & restriction which prevents non-interactivte playback of audible media...
jer.noble@apple.com [Mon, 27 Apr 2015 22:08:43 +0000 (22:08 +0000)]
Add a setting & restriction which prevents non-interactivte playback of audible media elements.
https://bugs.webkit.org/show_bug.cgi?id=143486

Reviewed by Eric Carlson.

Source/WebCore:

Tests: media/audio-playback-restriction-autoplay.html
       media/audio-playback-restriction-play.html

To allow clients who want to allow non-user-interactive video-only playback, but still
restrict playback of audible media elements, add a new setting and matching restriction
which disallows playback of media elements containing audible characteristics.

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::HTMLMediaElement): Set the RequireUserGestureForAudioRateChange
    restriction if the audioPlaybackRequiresUserGesture() setting is set.
(WebCore::HTMLMediaElement::parseAttribute): Drive-by fix. Move the opening brace
    out of the #if, as this confuses the heck out of diff and makes all subsequent
    changes appear to be within HTMLMediaElement::parseAttribute.
(WebCore::HTMLMediaElement::autoplay): Remove the restriction check from within autoplay().
    It is checked again immediately after every autoplay() call site.
(WebCore::HTMLMediaElement::pauseInternal): Remove the iOS-only #if.
(WebCore::HTMLMediaElement::mediaPlayerDidAddAudioTrack): Check whether playback
    is permitted, and if not, pause.
(WebCore::HTMLMediaElement::mediaPlayerCharacteristicChanged): Ditto.
(WebCore::HTMLMediaElement::removeBehaviorsRestrictionsAfterFirstUserGesture): Ditto.
* html/HTMLMediaSession.cpp:
(WebCore::restrictionName): Since BehaviorRestrictions is a bitfield, check each bit
    individually.
(WebCore::HTMLMediaSession::removeBehaviorRestriction): Handle RequireUserGestureForAudioRateChange.
(WebCore::HTMLMediaSession::playbackPermitted): Check whether the element has audio and
    audio playback is restricted, and return false if so.
* html/HTMLMediaSession.h:
* page/Settings.cpp:
* page/Settings.in:
* testing/Internals.cpp:
(WebCore::Internals::setMediaSessionRestrictions): Added.
(WebCore::Internals::setMediaElementRestrictions): Added.
* testing/Internals.h:
* testing/Internals.idl:

LayoutTests:

* media/audio-playback-restriction-autoplay.html: Added.
* media/audio-playback-restriction-play.html: Added.
* media/video-test.js:
(setCaptionDisplayMode):
(runWithKeyDown.eventName.window.testRunner.eventSender.keyDown.string_appeared_here.string_appeared_here.thunk):
(runWithKeyDown):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183428 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[CMake] Should be possible for an option to conflict with other options
mcatanzaro@igalia.com [Mon, 27 Apr 2015 22:04:04 +0000 (22:04 +0000)]
[CMake] Should be possible for an option to conflict with other options
https://bugs.webkit.org/show_bug.cgi?id=143956

Reviewed by Martin Robinson.

Add WEBKIT_OPTION_CONFLICT macro, which fails the build if conflicting options are on.

* Source/cmake/WebKitFeatures.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183427 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoContext menus are not reshown after an orientation change
jhoneycutt@apple.com [Mon, 27 Apr 2015 21:57:55 +0000 (21:57 +0000)]
Context menus are not reshown after an orientation change

<https://bugs.webkit.org/show_bug.cgi?id=144143>
<rdar://problem/20679931>

Reviewed by Darin Adler.

* UIProcess/ios/WKActionSheet.mm:
(-[WKActionSheet updateSheetPosition]):
Return early if we're still in a rotation, not if the rotation is
complete.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183426 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Attempt to workaround exception in TimelineRuler
timothy@apple.com [Mon, 27 Apr 2015 21:43:04 +0000 (21:43 +0000)]
Web Inspector: Attempt to workaround exception in TimelineRuler
https://bugs.webkit.org/show_bug.cgi?id=144228

Reviewed by Joseph Pecoraro.

* UserInterface/Views/TimelineRuler.js:
(WebInspector.TimelineRuler.prototype.updateLayout):
Use firstChild instead of a custom _labelElement property on the dividerElement.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183425 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WebAudio] AudioContext does not remove user-gesture restriction during resume()
jer.noble@apple.com [Mon, 27 Apr 2015 21:35:03 +0000 (21:35 +0000)]
[WebAudio] AudioContext does not remove user-gesture restriction during resume()
https://bugs.webkit.org/show_bug.cgi?id=144211

Reviewed by Eric Carlson.

Source/WebCore:

Tests: webaudio/audiocontext-restriction-audiobuffersourcenode-start.html
       webaudio/audiocontext-restriction.html

Before the introduction of resume(), suspend(), and stop(), AudioContexts which required
a user-gesture would start normally, but would effectively mute their outputs. Now that
the AudioContext's state property is exposed to JavaScript, the AudioContext should stay
in the "suspended" state until the user-gesture restriction is lifted.

Add a new method, willBeginPlayback() which checks and potentially clears the context's
behavior restrictions before checking with the MediaSession. Call this new willBeginPlayback()
method when the state would transition to "running".

Because they may be called before any nodes are created, make sure to call lazyInitialize()
from within the JS-exposed resumePlayback(), suspendPlayback(), and stopPlayback() methods.

Instead of clearing the behavior restrictions directly, scheduled AudioNodes should instead
call a new method nodeWillBeginPlayback(). Because existing sites will call AudioNode.start()
inside a user-gesture handler to clear the user-gesture restriction, call startRendering()
from nodeWillBeginPlayback(). But because we don't want AudioNode.start() to resume playback
unconditionally, only do so when the user-gesture restriction is set.

Now that an AudioContext will not transition to "running" state without a user-gesture (if
that restriction is set), there's no reason to check for that restriction from inside
AudioDestinationNode.

Add some internal methods to set and clear AudioContext BehaviorRestrictions for testing purposes.

* Modules/webaudio/AudioBufferSourceNode.cpp:
(WebCore::AudioBufferSourceNode::startPlaying):
* Modules/webaudio/AudioContext.cpp:
(WebCore::AudioContext::nodeWillBeginPlayback):
(WebCore::AudioContext::willBeginPlayback):
(WebCore::AudioContext::willPausePlayback):
(WebCore::AudioContext::startRendering):
(WebCore::AudioContext::suspendContext):
(WebCore::AudioContext::resumeContext):
(WebCore::AudioContext::closeContext):
(WebCore::AudioContext::suspendPlayback):
(WebCore::AudioContext::mayResumePlayback):
* Modules/webaudio/AudioContext.h:
(WebCore::AudioContext::behaviorRestrictions):
(WebCore::AudioContext::userGestureRequiredForAudioStart):
(WebCore::AudioContext::pageConsentRequiredForAudioStart):
* Modules/webaudio/AudioDestinationNode.cpp:
(WebCore::AudioDestinationNode::render):
* Modules/webaudio/AudioScheduledSourceNode.cpp:
(WebCore::AudioScheduledSourceNode::start):
* testing/Internals.cpp:
(WebCore::Internals::setAudioContextRestrictions):
* testing/Internals.h:
* testing/Internals.idl:

LayoutTests:

* webaudio/audiocontext-restriction-audiobuffersourcenode-start-expected.txt: Added.
* webaudio/audiocontext-restriction-audiobuffersourcenode-start.html: Added.
* webaudio/audiocontext-restriction-expected.txt: Added.
* webaudio/audiocontext-restriction.html: Added.
* webaudio/resources/audio-testing.js:
(runWithKeyDown):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183424 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Elements tab details sidebar becomes non-expandable after using Consol...
timothy@apple.com [Mon, 27 Apr 2015 21:30:58 +0000 (21:30 +0000)]
Web Inspector: Elements tab details sidebar becomes non-expandable after using Console tab
https://bugs.webkit.org/show_bug.cgi?id=144212

Show the DOM tree content view again on main resource change. This better matches what we did
in the pre tab user interface with the DOM being off the main resource tree element.

Reviewed by Joseph Pecoraro.

* UserInterface/Views/ElementsTabContentView.js:
(WebInspector.ElementsTabContentView): Listen for main resource change. Call _showDOMTreeContentView.
(WebInspector.ElementsTabContentView.prototype.closed): Fix a leak and exception.
The removeEventListener was wrong. Add a new removeEventListener for WebInspector.Frame.
(WebInspector.ElementsTabContentView.prototype._showDOMTreeContentView): Added.
(WebInspector.ElementsTabContentView.prototype._mainFrameDidChange): Call _showDOMTreeContentView.
(WebInspector.ElementsTabContentView.prototype._mainResourceDidChange): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183423 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: TabContentView closed() isn't being called when a tab is closed /...
timothy@apple.com [Mon, 27 Apr 2015 21:28:57 +0000 (21:28 +0000)]
Web Inspector: TabContentView closed() isn't being called when a tab is closed / removed
https://bugs.webkit.org/show_bug.cgi?id=144229

Reviewed by Joseph Pecoraro.

* UserInterface/Views/ContentViewContainer.js:
(WebInspector.ContentViewContainer.prototype._disassociateFromContentView):
TabContentViews don't have a representedObject, so returning early and not
calling closed() was not correct.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183422 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION (r183373): ASSERT failed in wtf/SHA1.h
rniwa@webkit.org [Mon, 27 Apr 2015 21:08:35 +0000 (21:08 +0000)]
REGRESSION (r183373): ASSERT failed in wtf/SHA1.h
https://bugs.webkit.org/show_bug.cgi?id=144257

Temporarily disable skip these tests.

* tests/stress/template-literal-line-terminators.js:
* tests/stress/template-literal-syntax.js:
* tests/stress/template-literal.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183421 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix.
ap@apple.com [Mon, 27 Apr 2015 21:02:31 +0000 (21:02 +0000)]
Build fix.

* accessibility/mac/AXObjectCacheMac.mm: (WebCore::AXObjectCache::postTextReplacementPlatformNotification):
Use -setObject:forKey:, not array subscript.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183420 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFunction allocations shouldn't sink through Put operations
commit-queue@webkit.org [Mon, 27 Apr 2015 20:42:28 +0000 (20:42 +0000)]
Function allocations shouldn't sink through Put operations
https://bugs.webkit.org/show_bug.cgi?id=144176

Patch by Basile Clement <basile_clement@apple.com> on 2015-04-27
Reviewed by Filip Pizlo.

By design, we don't support function allocation sinking through any
related operation ; however object allocation can sink through PutByOffset et
al.

Currently, the checks to prevent function allocation to sink through
these are misguided and do not prevent anything ; function allocation sinking
through these operations is prevented as a side effect of requiring an
AllocatePropertyStorage through which the function allocation is seen as
escaping.

This changes it so that ObjectAllocationSinkingPhase::handleNode()
checks properly that only object allocations sink through related write
operations.

* dfg/DFGObjectAllocationSinkingPhase.cpp:
(JSC::DFG::ObjectAllocationSinkingPhase::lowerNonReadingOperationsOnPhantomAllocations):
(JSC::DFG::ObjectAllocationSinkingPhase::handleNode):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183419 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agofast/frames/flattening/iframe-flattening-resize-event-count.html times out on Yosemit...
ap@apple.com [Mon, 27 Apr 2015 20:34:33 +0000 (20:34 +0000)]
fast/frames/flattening/iframe-flattening-resize-event-count.html times out on Yosemite WK2
https://bugs.webkit.org/show_bug.cgi?id=144155

Reviewed by Tim Horton.

* WebKitTestRunner/TestController.cpp: (WTR::TestController::resetPreferencesToConsistentValues):
Disable App Nap. I don't have positive evidence that it's the culprit, but it could be,
and we clearly don't want App Nap while testing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183418 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWirelessTargetPicker should not be visible unless a file is playable
eric.carlson@apple.com [Mon, 27 Apr 2015 20:33:15 +0000 (20:33 +0000)]
WirelessTargetPicker should not be visible unless a file is playable
https://bugs.webkit.org/show_bug.cgi?id=144271
<rdar://problem/20712003>

Reviewed by Jer Noble.

* Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.isPlayable): New.
(Controller.prototype.setStatusHidden): Call updateWirelessTargetAvailable.
(Controller.prototype.updateWirelessTargetAvailable): Don't show the button until the
    status message has been hidden.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183417 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[CMake] Autogenerate cmakeconfig.h.cmake
mrobinson@webkit.org [Mon, 27 Apr 2015 20:30:15 +0000 (20:30 +0000)]
[CMake] Autogenerate cmakeconfig.h.cmake
https://bugs.webkit.org/show_bug.cgi?id=143997

Reviewed by Csaba Osztrogonác.

.:

* CMakeLists.txt: Create the configuration header as the last part of the cmake run.
* Source/CMakeLists.txt: No longer use configure_file to create cmakeconfig.h.
* Source/cmake/OptionsCommon.cmake: Expose WTF_CPU_ARM64_CORTEXA53 to the build explicitly.
* Source/cmake/OptionsGTK.cmake: Expose variables to the build using SET_AND_EXPOSE_TO_BUILD.
Do some other miscellaneous related cleanup.
* Source/cmake/OptionsEFL.cmake: Expose HAVE_LLVM to the build.
* Source/cmake/WebKitFeatures.cmake: Expose all WebKit features to the build. Add support for
build exposed variables. Add a macro to generate a configuration from them.
* Source/cmakeconfig.h.cmake: Removed.

Source/WebCore:

* PlatformGTK.cmake: Use the WTF_USE style variables, because those are the same
as the ones exposed to the build.

Source/WTF:

* wtf/Platform.h: Move a GTK-specific setting from cmakeconfig.h to here.

Tools:

* DumpRenderTree/TestNetscapePlugIn/CMakeLists.txt: Add XP_UNIX to the command-line
definitions since TestNetscapePlugIn does not use cmakeconfig.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183416 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed build fix after r183405
beidson@apple.com [Mon, 27 Apr 2015 20:28:55 +0000 (20:28 +0000)]
Unreviewed build fix after r183405

* WebActionPropertyBag.cpp:
(WebActionPropertyBag::Read): Cast the enum class to a LONG.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183415 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agordar://problem/16678392 Page visibility tests are broken in Yosemite
ap@apple.com [Mon, 27 Apr 2015 20:23:08 +0000 (20:23 +0000)]
rdar://problem/16678392 Page visibility tests are broken in Yosemite

* platform/mac-wk2/TestExpectations: They don't seem to be broken any more, unskipping.
Maybe <http://trac.webkit.org/changeset/181866> fixed them?
Also updated the next line (media/video-background-tab-playback.html), as it didn't match what bots see now.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183414 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSimple line layout: Wrong text offsetting when range does not start from the first...
zalan@apple.com [Mon, 27 Apr 2015 20:19:13 +0000 (20:19 +0000)]
Simple line layout: Wrong text offsetting when range does not start from the first renderer.
https://bugs.webkit.org/show_bug.cgi?id=144167
rdar://problem/20639857

Reviewed by Simon Fraser.

This patch ensures that TextIterator returns the right text when the input range starts
from a sibling node.

TextIterator::m_previousTextLengthInFlow keeps track of the current node offset from the parent.
Source/WebCore:

It is required to map simple line layout runs to RenderText positions.
This patch sets the offset value when the iteration start with a sibling node.

Test: fast/text/range-text-with-simple-line-layout.html

* editing/TextIterator.cpp:
(WebCore::TextIterator::TextIterator):
(WebCore::TextIterator::handleTextNode):

LayoutTests:

it is required to map simple line layout runs to RenderText positions.
This patch sets the offset value when the iteration start with a sibling node.

* fast/text/range-text-with-simple-line-layout-expected.txt: Added.
* fast/text/range-text-with-simple-line-layout.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183413 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] API::Dictionary creation functions should return Ref<>
zandobersek@gmail.com [Mon, 27 Apr 2015 20:15:12 +0000 (20:15 +0000)]
[WK2] API::Dictionary creation functions should return Ref<>
https://bugs.webkit.org/show_bug.cgi?id=144221

Reviewed by Darin Adler.

Have API::Dictionary creation functions return Ref<>.
The call-sites are also updated, using and operating
on the returned Ref<> object where possible.

* Shared/API/APIDictionary.cpp:
(API::Dictionary::create):
* Shared/API/APIDictionary.h:
* Shared/API/Cocoa/WKRemoteObjectCoder.mm:
(createEncodedObject):
* Shared/API/Cocoa/_WKRemoteObjectRegistry.mm:
(-[_WKRemoteObjectRegistry _sendInvocation:interface:]):
* Shared/API/c/WKDictionary.cpp:
(WKDictionaryCreate):
* Shared/API/c/WKMutableDictionary.cpp:
(WKMutableDictionaryCreate):
* Shared/Plugins/Netscape/PluginInformation.cpp:
(WebKit::createPluginInformationDictionary):
* Shared/Plugins/Netscape/PluginInformation.h:
* UIProcess/API/C/WKPage.cpp:
(WKPageSetPageUIClient):
* UIProcess/API/gtk/WebKitWebContext.cpp:
(webkit_web_context_prefetch_dns):
* UIProcess/API/gtk/WebKitWebView.cpp:
(webkit_web_view_get_snapshot):
* UIProcess/Plugins/PlugInAutoStartProvider.cpp:
(WebKit::PlugInAutoStartProvider::autoStartOriginsTableCopy):
* UIProcess/Plugins/PlugInAutoStartProvider.h:
* UIProcess/StatisticsRequest.cpp:
(WebKit::createDictionaryFromHashMap):
(WebKit::StatisticsRequest::completedRequest):
* UIProcess/WebFormClient.cpp:
(WebKit::WebFormClient::willSubmitForm):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::didFailToInitializePlugin):
* WebProcess/InjectedBundle/API/gtk/WebKitWebPage.cpp:
(didInitiateLoadForResource):
(willSendRequestForFrame):
(didReceiveResponseForResource):
(didReceiveContentLengthForResource):
(didFinishLoadForResource):
(didFailLoadForResource):
(webkitWebPageDidReceiveMessage):
* WebProcess/InjectedBundle/InjectedBundlePageFormClient.cpp:
(WebKit::InjectedBundlePageFormClient::willSendSubmitEvent):
(WebKit::InjectedBundlePageFormClient::willSubmitForm):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183412 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] API::Data creation functions should return Ref<>
zandobersek@gmail.com [Mon, 27 Apr 2015 20:13:53 +0000 (20:13 +0000)]
[WK2] API::Data creation functions should return Ref<>
https://bugs.webkit.org/show_bug.cgi?id=144220

Reviewed by Darin Adler.

Have API::Data creation functions return Ref<>.
The call-sites are also updated, using and operating
on the returned Ref<> object where possible.

* Shared/API/APIData.h:
(API::Data::createWithoutCopying):
(API::Data::create):
* Shared/API/c/WKData.cpp:
(WKDataCreate):
* Shared/API/c/mac/WKWebArchive.cpp:
(WKWebArchiveCopyData):
* Shared/API/c/mac/WKWebArchiveResource.cpp:
(WKWebArchiveResourceCopyData):
* Shared/APIWebArchive.h:
* Shared/APIWebArchive.mm:
(API::WebArchive::data):
* Shared/APIWebArchiveResource.h:
* Shared/APIWebArchiveResource.mm:
(API::WebArchiveResource::data):
* Shared/Cocoa/APIDataCocoa.mm:
(API::Data::createWithoutCopying):
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView loadData:MIMEType:characterEncodingName:baseURL:]):
* UIProcess/API/gtk/WebKitURISchemeRequest.cpp:
(webkitURISchemeRequestReadCallback):
* UIProcess/API/gtk/WebKitWebView.cpp:
(webkit_web_view_load_bytes):
* UIProcess/Downloads/DownloadProxy.cpp:
(WebKit::createData):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::dataCallback):
(WebKit::WebPageProxy::savePDFToFileInDownloadsFolder):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183411 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] API::URL creation functions should return Ref<>
zandobersek@gmail.com [Mon, 27 Apr 2015 20:09:41 +0000 (20:09 +0000)]
[WK2] API::URL creation functions should return Ref<>
https://bugs.webkit.org/show_bug.cgi?id=144219

Reviewed by Darin Adler.

Have API::URL creation functions return Ref<>.
The call-sites are also updated, using and operating in the
returned Ref<> object, where possible.

* Shared/API/APIURL.h:
(API::URL::create):
* Shared/API/c/WKSharedAPICast.h:
(WebKit::toCopiedURLAPI):
* Shared/API/c/WKURL.cpp:
(WKURLCreateWithUTF8CString):
(WKURLCreateWithBaseURL):
* UIProcess/API/gtk/WebKitWebResource.cpp:
(webkit_web_resource_get_data):
* UIProcess/WebIconDatabase.cpp:
(WebKit::WebIconDatabase::didChangeIconForPageURL):
(WebKit::WebIconDatabase::notifyIconDataReadyForPageURL):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::userAgent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183410 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2] API::String creation functions should return Ref<>
zandobersek@gmail.com [Mon, 27 Apr 2015 20:05:14 +0000 (20:05 +0000)]
[WK2] API::String creation functions should return Ref<>
https://bugs.webkit.org/show_bug.cgi?id=144218

Reviewed by Darin Adler.

Have API::String creation functions return Ref<>.
The call-sites are also updated, using and operating on the returned
Ref<> object, where possible.

* Shared/API/APIString.h:
* Shared/API/c/WKPluginInformation.cpp:
(WKPluginInformationBundleIdentifierKey):
(WKPluginInformationBundleVersionKey):
(WKPluginInformationBundleShortVersionKey):
(WKPluginInformationPathKey):
(WKPluginInformationDisplayNameKey):
(WKPluginInformationDefaultLoadPolicyKey):
(WKPluginInformationUpdatePastLastBlockedVersionIsKnownAvailableKey):
(WKPluginInformationHasSandboxProfileKey):
(WKPluginInformationFrameURLKey):
(WKPluginInformationMIMETypeKey):
(WKPluginInformationPageURLKey):
(WKPluginInformationPluginspageAttributeURLKey):
(WKPluginInformationPluginURLKey):
(WKPlugInInformationReplacementObscuredKey):
* Shared/API/c/WKSharedAPICast.h:
(WebKit::ProxyingRefPtr::ProxyingRefPtr):
(WebKit::toCopiedAPI):
* Shared/API/c/WKString.cpp:
(WKStringCreateWithUTF8CString):
(WKStringCreateWithJSString):
* UIProcess/API/C/WKDatabaseManager.cpp:
(WKDatabaseManagerGetOriginKey):
(WKDatabaseManagerGetOriginQuotaKey):
(WKDatabaseManagerGetOriginUsageKey):
(WKDatabaseManagerGetDatabaseDetailsKey):
(WKDatabaseManagerGetDatabaseDetailsNameKey):
(WKDatabaseManagerGetDatabaseDetailsDisplayNameKey):
(WKDatabaseManagerGetDatabaseDetailsExpectedUsageKey):
(WKDatabaseManagerGetDatabaseDetailsCurrentUsageKey):
(WKDatabaseManagerGetDatabaseDetailsCreationTimeKey):
(WKDatabaseManagerGetDatabaseDetailsModificationTimeKey):
* UIProcess/API/C/WKKeyValueStorageManager.cpp:
(WKKeyValueStorageManagerGetOriginKey):
(WKKeyValueStorageManagerGetCreationTimeKey):
(WKKeyValueStorageManagerGetModificationTimeKey):
* UIProcess/API/C/WKPage.cpp:
(WKPageGetSessionHistoryURLValueType):
(WKPageGetSessionBackForwardListItemValueType):
(toGenericCallbackFunction):
(WKPageValidateCommand):
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::didReceiveInvalidMessage):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183409 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r183393.
commit-queue@webkit.org [Mon, 27 Apr 2015 19:36:25 +0000 (19:36 +0000)]
Unreviewed, rolling out r183393.
https://bugs.webkit.org/show_bug.cgi?id=144272

Caused memory corruption detected by GuardMalloc (Requested by
ap on #webkit).

Reverted changeset:

"Synchronous XMLHttpRequest should get access to AppCache
resources stored as flat files"
https://bugs.webkit.org/show_bug.cgi?id=143711
http://trac.webkit.org/changeset/183393

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183408 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Curl] Favicons loaded from disc cache are ignored.
peavo@outlook.com [Mon, 27 Apr 2015 18:57:13 +0000 (18:57 +0000)]
[Curl] Favicons loaded from disc cache are ignored.
https://bugs.webkit.org/show_bug.cgi?id=143953

Reviewed by Alex Christensen.

When a favicon is loaded from the Curl disc cache, the icon data is thrown away.
This happens because we give a 304 response, which makes the icon loader ignore
the response. We can solve this by responding with 200 OK.

* platform/network/ResourceHandleInternal.h:
(WebCore::ResourceHandleInternal::ResourceHandleInternal):
* platform/network/curl/CurlCacheManager.cpp:
(WebCore::CurlCacheManager::didReceiveResponse):
* platform/network/curl/ResourceHandleManager.cpp:
(WebCore::headerCallback):
(WebCore::ResourceHandleManager::initializeHandle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183407 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoVarargsForwardingPhase should use bytecode liveness in addition to other uses to...
fpizlo@apple.com [Mon, 27 Apr 2015 18:49:15 +0000 (18:49 +0000)]
VarargsForwardingPhase should use bytecode liveness in addition to other uses to determine the last point that a candidate is used
https://bugs.webkit.org/show_bug.cgi?id=143843

Reviewed by Geoffrey Garen.

It will soon come to pass that Phantom isn't available at the time that
VarargsForwardingPhase runs. So, it needs to use some other mechanism for discovering when
a value dies for OSR.

This is simplified by two things:

1) The bytecode kill analysis is now reusable. This patch makes it even more reusable than
   before by polishing the API.

2) This phase already operates on one node at a time and allows itself to do a full search
   of the enclosing basic block for that node. This is fine because CreateDirectArguments
   and friends is a rarely occurring node. The fact that it operates on one node at a time
   makes it even easier to reason about OSR liveness - we just track the list of locals in
   which it is live.

This change has no effect right now but it is a necessary prerequisite to implementing
https://bugs.webkit.org/show_bug.cgi?id=143736.

* dfg/DFGBasicBlock.h:
(JSC::DFG::BasicBlock::tryAt):
* dfg/DFGForAllKills.h:
(JSC::DFG::forAllKilledOperands):
* dfg/DFGPhantomInsertionPhase.cpp:
* dfg/DFGVarargsForwardingPhase.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183406 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMake 'enum NavigationType' be an enum class
beidson@apple.com [Mon, 27 Apr 2015 18:46:15 +0000 (18:46 +0000)]
Make 'enum NavigationType' be an enum class
https://bugs.webkit.org/show_bug.cgi?id=144270

Reviewed by Alex Christensen.

Source/WebCore:

No new tests (No change in behavior).

I’ll be adding a new NavigationType for bug 144269 so it makes sense to make this
an enum class first.

* loader/FrameLoader.cpp:
(WebCore::FrameLoader::load):
(WebCore::FrameLoader::reload):
(WebCore::FrameLoader::loadDifferentDocumentItem):
* loader/FrameLoaderTypes.h:

* loader/NavigationAction.cpp:
(WebCore::navigationType):
(WebCore::NavigationAction::NavigationAction):

* loader/PolicyChecker.cpp:
(WebCore::PolicyChecker::checkNavigationPolicy):

* page/PerformanceNavigation.cpp:
(WebCore::PerformanceNavigation::type):

Source/WebKit/mac:

* WebCoreSupport/WebFrameLoaderClient.mm:
(WebFrameLoaderClient::actionDictionary):

Source/WebKit2:

* Shared/API/c/WKSharedAPICast.h:
(WebKit::toAPI):

* Shared/NavigationActionData.cpp:
(WebKit::NavigationActionData::NavigationActionData):

* UIProcess/API/Cocoa/WKNavigationAction.mm:
(toWKNavigationType):

* UIProcess/API/gtk/WebKitPrivate.cpp:
(toWebKitNavigationType):

* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::dispatchDecidePolicyForNavigationAction):

* WebProcess/WebPage/WebInspector.cpp:
(WebKit::WebInspector::openInNewTab):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183405 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix viewport units in Media Queries
yoav@yoav.ws [Mon, 27 Apr 2015 18:42:29 +0000 (18:42 +0000)]
Fix viewport units in Media Queries
https://bugs.webkit.org/show_bug.cgi?id=144260

Reviewed by Darin Adler.

Source/WebCore:

This patch makes sure that viewport units are considered "length units"
in the context of Media Queries, by having MediaQueryExp use the unit logic
that is in CSSPrimitiveValue.
It does that by turning the relevant methods in CSSPrimitiveValue into static.

It also makes sure that the logic for "resolution units" is not maintained separately
in MediaQueryExp, to avoid similiar issues in the future with resolution units.

Test: fast/media/mq-viewport-units.html

* css/CSSPrimitiveValue.h:
(WebCore::CSSPrimitiveValue::isViewportPercentageLength): Added a static variant.
(WebCore::CSSPrimitiveValue::isLength): Added a static variant.
(WebCore::CSSPrimitiveValue::isResolution): Added a static variant.
* css/MediaQueryExp.cpp:
(WebCore::featureWithValidPositiveLenghtOrNumber): Call CSSPrimitiveValue's length unit logic.
(WebCore::featureWithValidDensity): Call CSSPrimitiveValue's resolution unit logic.

LayoutTests:

These tests make sure that viewport units are working as expected inside of Media Queries.

* fast/media/mq-viewport-units-expected.txt: Added.
* fast/media/mq-viewport-units.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183404 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION(r182879): Images and video can no longer be downloaded
bfulgham@apple.com [Mon, 27 Apr 2015 18:38:29 +0000 (18:38 +0000)]
REGRESSION(r182879): Images and video can no longer be downloaded
https://bugs.webkit.org/show_bug.cgi?id=144267
<rdar://problem/20702844>

Reviewed by Simon Fraser.

Source/WebKit/mac:

* WebView/WebActionMenuController.mm:
(-[WebActionMenuController _createActionMenuItemForTag:]): Check for
"!WebCore::protocolIs(..., "file")" for image and media downloads.

Source/WebKit2:

* UIProcess/mac/WKActionMenuController.mm:
(-[WKActionMenuController _createActionMenuItemForTag:]): Check for
"!WebCore::protocolIs(..., "file")" for image and media downloads.

Tools:

Revise tests to check for proper download menu state.
1. Downloading a loal image should be disabled.
2. Downloading a local media file should be disabled.
3. Downloading non-downloadable media should be disabled.

* TestWebKitAPI/Tests/WebKit2ObjC/ActionMenus.mm:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183403 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMap#entries and Map#keys error for non-Maps is swapped
commit-queue@webkit.org [Mon, 27 Apr 2015 18:36:14 +0000 (18:36 +0000)]
Map#entries and Map#keys error for non-Maps is swapped
https://bugs.webkit.org/show_bug.cgi?id=144253

Patch by Jordan Harband <ljharb@gmail.com> on 2015-04-27
Reviewed by Simon Fraser.

Correcting error messages on Set/Map methods when called on
incompatible objects.

* runtime/MapPrototype.cpp:
(JSC::mapProtoFuncEntries):
(JSC::mapProtoFuncKeys):
* runtime/SetPrototype.cpp:
(JSC::setProtoFuncEntries):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183402 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRationalize DFG DCE handling of nodes that perform checks that propagate through AI
fpizlo@apple.com [Mon, 27 Apr 2015 18:31:26 +0000 (18:31 +0000)]
Rationalize DFG DCE handling of nodes that perform checks that propagate through AI
https://bugs.webkit.org/show_bug.cgi?id=144186

Reviewed by Geoffrey Garen.

If I do ArithAdd(Int32Use, Int32Use, CheckOverflow) then AI will prove that this returns
Int32. We may later perform code simplifications based on the proof that this is Int32, and
we may kill all DFG users of this ArithAdd. Then we may prove that there is no exit site at
which the ArithAdd is live. This seems like it is sufficient to then kill the ArithAdd,
except that we still need the overflow check!

Previously we mishandled this:

- In places where we want the overflow check we need to use MustGenerate(@ArithAdd) as a hack
  to keep it alive. That's dirty and it's just indicative of a deeper issue.

- Our MovHint removal doesn't do Phantom canonicalization which essentially makes it
  powerless. This was sort of hiding the bug.

- Nodes that have checks that AI leverages should always be NodeMustGenerate. You can't kill
  something that you are relying on for subsequent simplifications.

This fixes MovHint removal to also canonicalize Phantoms. This also adds ModeMustGenerate to
nodes that may perform checks that are used by AI to guarantee the result type. As a result,
we no longer need the weird MustGenerate node.

* dfg/DFGAbstractInterpreterInlines.h:
(JSC::DFG::AbstractInterpreter<AbstractStateType>::executeEffects):
* dfg/DFGArgumentsEliminationPhase.cpp:
* dfg/DFGClobberize.h:
(JSC::DFG::clobberize):
* dfg/DFGDCEPhase.cpp:
(JSC::DFG::DCEPhase::run):
* dfg/DFGDoesGC.cpp:
(JSC::DFG::doesGC):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
(JSC::DFG::FixupPhase::tryToRelaxRepresentation):
* dfg/DFGIntegerCheckCombiningPhase.cpp:
(JSC::DFG::IntegerCheckCombiningPhase::handleBlock):
(JSC::DFG::IntegerCheckCombiningPhase::insertMustAdd): Deleted.
* dfg/DFGMayExit.cpp:
(JSC::DFG::mayExit):
* dfg/DFGNode.h:
(JSC::DFG::Node::willHaveCodeGenOrOSR):
* dfg/DFGNodeType.h:
* dfg/DFGObjectAllocationSinkingPhase.cpp:
(JSC::DFG::ObjectAllocationSinkingPhase::handleNode):
* dfg/DFGPhantomCanonicalizationPhase.cpp:
(JSC::DFG::PhantomCanonicalizationPhase::run):
* dfg/DFGPhantomRemovalPhase.cpp:
(JSC::DFG::PhantomRemovalPhase::run):
* dfg/DFGPlan.cpp:
(JSC::DFG::Plan::compileInThreadImpl):
* dfg/DFGPredictionPropagationPhase.cpp:
(JSC::DFG::PredictionPropagationPhase::propagate):
* dfg/DFGSafeToExecute.h:
(JSC::DFG::safeToExecute):
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGTypeCheckHoistingPhase.cpp:
(JSC::DFG::TypeCheckHoistingPhase::identifyRedundantStructureChecks):
(JSC::DFG::TypeCheckHoistingPhase::identifyRedundantArrayChecks):
* dfg/DFGVarargsForwardingPhase.cpp:
* ftl/FTLCapabilities.cpp:
(JSC::FTL::canCompile):
* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::compileNode):
* tests/stress/fold-based-on-int32-proof-mul-branch.js: Added.
(foo):
* tests/stress/fold-based-on-int32-proof-mul.js: Added.
(foo):
* tests/stress/fold-based-on-int32-proof-or-zero.js: Added.
(foo):
* tests/stress/fold-based-on-int32-proof.js: Added.
(foo):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183401 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMerge run-launcher into run-minibrowser
ossy@webkit.org [Mon, 27 Apr 2015 18:19:37 +0000 (18:19 +0000)]
Merge run-launcher into run-minibrowser
https://bugs.webkit.org/show_bug.cgi?id=141928

Reviewed by Darin Adler.

* Scripts/run-iexploder-tests:
(runSafariWithIExploder):
* Scripts/run-launcher: Removed.
* Scripts/run-minibrowser:
* Scripts/webkitdirs.pm:
(launcherPath):
* Scripts/webkitpy/port/efl.py:
(EflPort.show_results_html_file):
* Scripts/webkitpy/port/efl_unittest.py:
(EflPortTest.test_show_results_html_file):
* Scripts/webkitpy/port/gtk.py:
(GtkPort.print_leaks_summary):
(GtkPort.show_results_html_file):
(GtkPort): Deleted.
* Scripts/webkitpy/port/gtk_unittest.py:
(GtkPortTest.test_show_results_html_file):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183400 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[CSS Grid Layout] Support for the justify-self and justify-items in grid layout
jfernandez@igalia.com [Mon, 27 Apr 2015 17:24:44 +0000 (17:24 +0000)]
[CSS Grid Layout] Support for the justify-self and justify-items in grid layout
https://bugs.webkit.org/show_bug.cgi?id=133280

Source/WebCore:

Reviewed by David Hyatt.

Implementation of justify-self and justify-items properties for grid. It supports
different writing-modes and direction. Margins, borders and paddings are also
considered when computing the final position and stretched size.

This patch applies also a quite important refactoring of the alignment logic in
order to share code between the two alignment dimensions, row-axis and column-axis.

Overflow handling is still missing and will be added later in a follow-up bug.

Tests: fast/css-grid-layout/grid-align-justify-margin-border-padding-vertical-lr.html
       fast/css-grid-layout/grid-align-justify-margin-border-padding-vertical-rl.html
       fast/css-grid-layout/grid-align-justify-margin-border-padding.html
       fast/css-grid-layout/justify-self-cell.html

* rendering/RenderGrid.cpp:
(WebCore::RenderGrid::layoutGridItems):
(WebCore::RenderGrid::columnAxisPositionForChild):
(WebCore::RenderGrid::rowAxisPositionForChild):
(WebCore::RenderGrid::rowPositionForChild):
(WebCore::RenderGrid::columnPositionForChild):
(WebCore::RenderGrid::findChildLogicalPosition):
(WebCore::RenderGrid::allowedToStretchLogicalHeightForChild): Deleted.
(WebCore::RenderGrid::needToStretchChildLogicalHeight): Deleted.
(WebCore::RenderGrid::marginLogicalHeightForChild): Deleted.
(WebCore::RenderGrid::availableAlignmentSpaceForChildBeforeStretching): Deleted.
(WebCore::RenderGrid::applyStretchAlignmentToChildIfNeeded): Deleted.
* rendering/RenderGrid.h:
* rendering/style/RenderStyle.cpp:
(WebCore::RenderStyle::resolveJustification):
* rendering/style/RenderStyle.h:

LayoutTests:

Implementation of justify-self and justify-items properties for grid.
Added additional testing for ensuring margin, padding and border are considered when
computing grid item's position.

Reviewed by David Hyatt.

* fast/css-grid-layout/grid-align-justify-margin-border-padding-expected.txt: Added.
* fast/css-grid-layout/grid-align-justify-margin-border-padding-vertical-lr-expected.txt: Added.
* fast/css-grid-layout/grid-align-justify-margin-border-padding-vertical-lr.html: Added.
* fast/css-grid-layout/grid-align-justify-margin-border-padding-vertical-rl-expected.txt: Added.
* fast/css-grid-layout/grid-align-justify-margin-border-padding-vertical-rl.html: Added.
* fast/css-grid-layout/grid-align-justify-margin-border-padding.html: Added.
* fast/css-grid-layout/justify-self-cell-expected.txt: Added.
* fast/css-grid-layout/justify-self-cell.html: Added.
* fast/css-grid-layout/resources/grid.css:
(.directionLTR): Added. Useful for defining orthogonal flows.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183399 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoCrashes under IDBDatabase::closeConnection
darin@apple.com [Mon, 27 Apr 2015 17:07:20 +0000 (17:07 +0000)]
Crashes under IDBDatabase::closeConnection
https://bugs.webkit.org/show_bug.cgi?id=141745

Reviewed by Alexey Proskuryakov.

* Modules/indexeddb/IDBDatabase.cpp:
(WebCore::IDBDatabase::~IDBDatabase): Do the work of close/closeConnection without
actually calling those functions.
(WebCore::IDBDatabase::closeConnection): Protect the database so it's not destroyed
in the middle of this function's execution.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183398 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed.
calvaris@igalia.com [Mon, 27 Apr 2015 14:56:18 +0000 (14:56 +0000)]
Unreviewed.

Fixed timeout expectation of "closed should fulfill with
undefined" test. This doesn't time out anymore.

* streams/reference-implementation/readable-stream-templated.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183397 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed.
calvaris@igalia.com [Mon, 27 Apr 2015 14:54:56 +0000 (14:54 +0000)]
Unreviewed.

Fixed timeout expectation of "closed should fulfill with
undefined" test. This doesn't time out anymore.

* streams/reference-implementation/readable-stream-templated.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183396 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Streams API] ReadableStream constructor start function should be able to close the...
youenn.fablet@crf.canon.fr [Mon, 27 Apr 2015 13:31:59 +0000 (13:31 +0000)]
[Streams API] ReadableStream constructor start function should be able to close the stream
https://bugs.webkit.org/show_bug.cgi?id=143363

Reviewed by Benjamin Poulain.

Source/WebCore:

Implements https://streams.spec.whatwg.org/#close-readable-stream.
When the "close" JS function is called, the stream is getting closed.
The stream state is changed to close and if it has a reader, the reader gets closed as well:
The reader resolves the closed promise and releases the stream.

Enabled the possibility to resolve a promise with any JS value.
This is used to resolve closed promise with jsUndefined and will be used for read promises in
the future as well, though of course it is not restricted to Streams.

Covered by reference tests that are now passing.

* Modules/streams/ReadableStream.h:
* Modules/streams/ReadableStream.cpp:
(WebCore::ReadableStream::changeStateToClosed): Called by the JS function 'close'.
* Modules/streams/ReadableStreamReader.cpp:
(WebCore::ReadableStreamReader::ReadableStreamReader): Initialized stream.
(WebCore::ReadableStreamReader::initialize): Added to handle state change at constructor time
in particular closed/errored state.
(WebCore::ReadableStreamReader::releaseStream): Releases the stream.
(WebCore::ReadableStreamReader::closed): Stores the promise callbacks and invokes success
immediately if the stream is already closed.
(WebCore::ReadableStreamReader::changeStateToClosed): Changes the internal state to closed,
resolves the promise and releases the stream.
* Modules/streams/ReadableStreamReader.h:
(WebCore::ReadableStreamReader::State): Added.
* bindings/js/JSDOMPromise.h:
(WebCore::DeferredWrapper::resolve<JSC::JSValue>): Adds the ability to resolve a promise with
a custom JS value.
* bindings/js/JSReadableStreamControllerCustom.cpp:
(WebCore::JSReadableStreamController::close): Not "notImplemented" anymore. Now it closes the
stream.
* bindings/js/JSReadableStreamReaderCustom.cpp:
(WebCore::JSReadableStreamReader::closed): Resolves the promise with undefined.

LayoutTests:

Updated expectations with new passes.

* streams/reference-implementation/bad-underlying-sources-expected.txt:
* streams/reference-implementation/readable-stream-reader-expected.txt:
* streams/reference-implementation/readable-stream-templated-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183395 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix the !ENABLE(CSS_GRID_LAYOUT) build after r183370
ossy@webkit.org [Mon, 27 Apr 2015 11:10:49 +0000 (11:10 +0000)]
Fix the !ENABLE(CSS_GRID_LAYOUT) build after r183370
https://bugs.webkit.org/show_bug.cgi?id=144255

Reviewed by Carlos Garcia Campos.

* rendering/RenderBox.cpp:
(WebCore::RenderBox::computeLogicalHeight):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183394 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSynchronous XMLHttpRequest should get access to AppCache resources stored as flat...
youenn.fablet@crf.canon.fr [Mon, 27 Apr 2015 11:03:24 +0000 (11:03 +0000)]
Synchronous XMLHttpRequest should get access to AppCache resources stored as flat files
https://bugs.webkit.org/show_bug.cgi?id=143711

Reviewed by Darin Adler.

Source/WebCore:

This patch checks whether a substitute resource data is stored in memory or in file for synchronous loads.
If data is stored in file, it reads the data through SharedBuffer::createWithContentsOfFile.
This patch refactors some routines to replace Vector<char> by SharedBuffer to transmit response data.

Test: http/tests/appcache/simple-video-sync.html

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::parseAttribute):
* loader/DocumentThreadableLoader.cpp:
(WebCore::DocumentThreadableLoader::loadRequest):
* loader/FrameLoader.cpp:
(WebCore::FrameLoader::loadResourceSynchronously):
* loader/FrameLoader.h:
* loader/appcache/ApplicationCacheHost.cpp:
(WebCore::ApplicationCacheHost::maybeLoadResource):
(WebCore::ApplicationCacheHost::createFileURL):
(WebCore::ApplicationCacheHost::maybeLoadSynchronously):
(WebCore::ApplicationCacheHost::maybeLoadFallbackSynchronously):
* loader/appcache/ApplicationCacheHost.h:
* xml/XSLTProcessorLibxslt.cpp:
(WebCore::docLoaderFunc):
* xml/parser/XMLDocumentParserLibxml2.cpp:
(WebCore::openFunc):

LayoutTests:

* http/tests/appcache/resources/fake-video.mp4: Added.
* http/tests/appcache/resources/simple-video-sync.manifest: Added.
* http/tests/appcache/simple-video-sync-expected.txt: Added.
* http/tests/appcache/simple-video-sync.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183393 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAlso skip it in run-javascriptcore-tests.
rniwa@webkit.org [Mon, 27 Apr 2015 10:23:20 +0000 (10:23 +0000)]
Also skip it in run-javascriptcore-tests.

* js/script-tests/class-syntax-semicolon.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@183392 268f45cc-cd09-0410-ab3c-d52691b4dbfc